diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..7f4232f3fc2b899585592281597904f58cb2a239
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 2, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 4, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 6, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 8, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 10, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 12, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 14, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 16, "episode_return": 355.0000041872263, "episode_return_env": 355.0000041872263, "game_score": 75.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2886, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2886}
+{"episode_id": 18, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..cd4f1f9d7623a28ad31785bb2a9bbf3e96329a6d
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 46.000000566244125, "episode_return_env": 46.000000566244125, "game_score": 9.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 390, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 390}
+{"episode_id": 5, "episode_return": 199.9000023677945, "episode_return_env": 199.9000023677945, "game_score": 42.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1632, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1632}
+{"episode_id": 7, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 9, "episode_return": 157.9000018760562, "episode_return_env": 157.9000018760562, "game_score": 33.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1293, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1293}
+{"episode_id": 11, "episode_return": 438.1000051572919, "episode_return_env": 438.1000051572919, "game_score": 93.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3555, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3555}
+{"episode_id": 13, "episode_return": 115.90000138431787, "episode_return_env": 115.90000138431787, "game_score": 24.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 954, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 954}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 19, "episode_return": 256.10000302642584, "episode_return_env": 256.10000302642584, "game_score": 54.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2086, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2086}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..35f182f12316a45d5c621de94d63f9287845b818
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 270.0000031888485, "episode_return_env": 270.0000031888485, "game_score": 57.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2198, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2198}
+{"episode_id": 2, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 4, "episode_return": 242.80000287294388, "episode_return_env": 242.80000287294388, "game_score": 51.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1980, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1980}
+{"episode_id": 6, "episode_return": 400.4000047147274, "episode_return_env": 400.4000047147274, "game_score": 85.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3250, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3250}
+{"episode_id": 8, "episode_return": 118.80000141263008, "episode_return_env": 118.80000141263008, "game_score": 25.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 974, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 974}
+{"episode_id": 10, "episode_return": 129.90000154823065, "episode_return_env": 129.90000154823065, "game_score": 27.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1067, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1067}
+{"episode_id": 12, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 14, "episode_return": 151.3000017926097, "episode_return_env": 151.3000017926097, "game_score": 32.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1236, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1236}
+{"episode_id": 16, "episode_return": 186.10000220686197, "episode_return_env": 186.10000220686197, "game_score": 39.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1521, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1521}
+{"episode_id": 18, "episode_return": 143.90000171214342, "episode_return_env": 143.90000171214342, "game_score": 30.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1180, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1180}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..78427bf08a0d5d77e245c1837eca5daacf4b11a2
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 7, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 9, "episode_return": 275.4000032544136, "episode_return_env": 275.4000032544136, "game_score": 58.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2243, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2243}
+{"episode_id": 11, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 19, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..7d9aba29e1b7064ce0972740ecc1137a949661ae
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 2, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 4, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 6, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 8, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 10, "episode_return": 9.100000135600567, "episode_return_env": 9.100000135600567, "game_score": 1.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 93, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 93}
+{"episode_id": 12, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 14, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 59.90000072866678, "episode_return_env": 59.90000072866678, "game_score": 12.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 502, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 502}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..08a222349c7c92b8aac6c3886c22fa82a622fcb3
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 7, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 9, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 11, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 19, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..823672cd132017e6d712a83fe1f636eaa4415849
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 9.200000137090683, "episode_return_env": 9.200000137090683, "game_score": 1.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 94, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 94}
+{"episode_id": 5, "episode_return": 181.70000215619802, "episode_return_env": 181.70000215619802, "game_score": 38.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1486, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1486}
+{"episode_id": 7, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 9, "episode_return": 443.10000521689653, "episode_return_env": 443.10000521689653, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3596, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3596}
+{"episode_id": 11, "episode_return": 59.90000072866678, "episode_return_env": 59.90000072866678, "game_score": 12.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 502, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 502}
+{"episode_id": 13, "episode_return": 279.6000033020973, "episode_return_env": 279.6000033020973, "game_score": 59.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2276, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2276}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 19, "episode_return": 275.10000324994326, "episode_return_env": 275.10000324994326, "game_score": 58.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2240, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2240}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..f77d61d35a8dfe0382eabb5c3ff8285d5b7bf86b
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 2, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 4, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 6, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 8, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 10, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 12, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 14, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..f41b79fcfa14331842b2987bd8096df48629020e
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 2, "episode_return": 46.60000057518482, "episode_return_env": 46.60000057518482, "game_score": 9.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 396, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 396}
+{"episode_id": 4, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 6, "episode_return": 166.90000198036432, "episode_return_env": 166.90000198036432, "game_score": 35.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1365, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1365}
+{"episode_id": 8, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 10, "episode_return": 339.90000400692225, "episode_return_env": 339.90000400692225, "game_score": 72.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2762, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2762}
+{"episode_id": 12, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 14, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..5bfe0155aa86a39da52357297ddd7f9cb6537ce0
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 2, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 4, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 6, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 8, "episode_return": 386.9000045582652, "episode_return_env": 386.9000045582652, "game_score": 82.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3142, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3142}
+{"episode_id": 10, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 12, "episode_return": 190.90000226348639, "episode_return_env": 190.90000226348639, "game_score": 40.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1560, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1560}
+{"episode_id": 14, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..cf237387b1682e36504b0b675be9d5c30f1bd465
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 32.40000040829182, "episode_return_env": 32.40000040829182, "game_score": 6.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 281, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 281}
+{"episode_id": 2, "episode_return": 9.00000013411045, "episode_return_env": 9.00000013411045, "game_score": 1.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 92, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 92}
+{"episode_id": 4, "episode_return": 88.50000106543303, "episode_return_env": 88.50000106543303, "game_score": 18.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 734, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 734}
+{"episode_id": 6, "episode_return": 261.1000030860305, "episode_return_env": 261.1000030860305, "game_score": 55.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2127, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2127}
+{"episode_id": 8, "episode_return": 13.300000183284283, "episode_return_env": 13.300000183284283, "game_score": 2.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 126, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 126}
+{"episode_id": 10, "episode_return": 51.100000627338886, "episode_return_env": 51.100000627338886, "game_score": 10.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 432, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 432}
+{"episode_id": 12, "episode_return": 121.10000144690275, "episode_return_env": 121.10000144690275, "game_score": 25.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 997, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 997}
+{"episode_id": 14, "episode_return": 13.400000184774399, "episode_return_env": 13.400000184774399, "game_score": 2.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 127, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 127}
+{"episode_id": 16, "episode_return": 162.6000019311905, "episode_return_env": 162.6000019311905, "game_score": 34.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1331, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1331}
+{"episode_id": 18, "episode_return": 134.9000016078353, "episode_return_env": 134.9000016078353, "game_score": 28.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1108, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1108}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..30e71c2ac87735d04a1ca878c0b9402a66a0ee60
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 218.80000258982182, "episode_return_env": 218.80000258982182, "game_score": 46.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1785, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1785}
+{"episode_id": 5, "episode_return": 68.90000083297491, "episode_return_env": 68.90000083297491, "game_score": 14.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 574, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 574}
+{"episode_id": 7, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 9, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 11, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 13, "episode_return": 419.3000049367547, "episode_return_env": 419.3000049367547, "game_score": 89.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3403, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3403}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 36.800000458955765, "episode_return_env": 36.800000458955765, "game_score": 7.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 316, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 316}
+{"episode_id": 19, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..a6156a995f9d74231e89fd92ceb2fe247c6e7d16
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 25}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..a6156a995f9d74231e89fd92ceb2fe247c6e7d16
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 25}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..4c3e4504c76c259f47e4cbe5be12e16e8e40e345
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,25 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 367.9000043347478, "episode_return_env": 367.9000043347478, "game_score": 78.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2988, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2988}
+{"episode_id": 5, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 7, "episode_return": 116.20000138878822, "episode_return_env": 116.20000138878822, "game_score": 24.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 957, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 957}
+{"episode_id": 9, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 11, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 19, "episode_return": 17.900000236928463, "episode_return_env": 17.900000236928463, "game_score": 3.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 163, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 163}
+{"episode_id": 21, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 63, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 23, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 65, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 25, "episode_return": 160.8000019043684, "episode_return_env": 160.8000019043684, "game_score": 34.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 67, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1313, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1313}
+{"episode_id": 27, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 69, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 29, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 71, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 31, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 73, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 33, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 75, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 35, "episode_return": 353.900004170835, "episode_return_env": 353.900004170835, "game_score": 75.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 77, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2875, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2875}
+{"episode_id": 37, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 79, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 39, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 81, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 41, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 83, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 43, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 85, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 45, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 87, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 47, "episode_return": 251.40000297129154, "episode_return_env": 251.40000297129154, "game_score": 53.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 89, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2048, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2048}
+{"episode_id": 49, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 91, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/summary.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/summary.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..ce7dad56d7abf845dd57d6f02d76df86c6434c49
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_kv_memory_flex/summary.jsonl
@@ -0,0 +1,8 @@
+{"steps": 500}
+{"steps": 1000}
+{"steps": 1500}
+{"steps": 2000}
+{"steps": 2500}
+{"steps": 3000}
+{"steps": 3500}
+{"steps": 4000}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1000.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1000.json
new file mode 100644
index 0000000000000000000000000000000000000000..d92248aecc40d3aba78ef27bfe02d55c7f6d7c23
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1000.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 189.86000225022434,
+      "mean_length": 1550.75,
+      "std_reward": 127.10599033053973,
+      "std_length": 1024.5753693604,
+      "episode_rewards": [
+        71.20000085234642,
+        444.6000052243471,
+        167.50000198930502,
+        32.80000041425228,
+        66.70000080019236,
+        314.8000037074089,
+        93.60000112652779,
+        144.80000172555447,
+        233.40000276267529,
+        242.80000287294388,
+        107.60000129044056,
+        327.00000385940075,
+        223.80000264942646,
+        23.40000030398369,
+        237.90000281482935,
+        336.100003965199,
+        125.80000150203705,
+        97.8000011742115,
+        61.00000074505806,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        588,
+        3600,
+        1371,
+        285,
+        552,
+        2556,
+        776,
+        1189,
+        1904,
+        1980,
+        889,
+        2660,
+        1826,
+        209,
+        1940,
+        2733,
+        1035,
+        809,
+        513,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 189.86000225022434,
+    "mean_length": 1550.75,
+    "std_reward": 127.10599033053973,
+    "std_length": 1024.5753693604,
+    "task_count": 1,
+    "macro_mean_reward": 189.86000225022434,
+    "macro_mean_length": 1550.75,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1250.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1250.json
new file mode 100644
index 0000000000000000000000000000000000000000..bb75acb722c76f823578f86b2a5766b108fab6cd
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1250.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 273.75500322692096,
+      "mean_length": 2223.75,
+      "std_reward": 186.2417630023561,
+      "std_length": 1500.723254800831,
+      "episode_rewards": [
+        13.600000187754631,
+        291.3000034317374,
+        76.80000092089176,
+        6.800000101327896,
+        65.10000079125166,
+        444.6000052243471,
+        444.6000052243471,
+        230.80000272393227,
+        34.80000042915344,
+        252.00000298023224,
+        6.800000101327896,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        51.100000627338886,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        129,
+        2366,
+        635,
+        70,
+        545,
+        3600,
+        3600,
+        1878,
+        296,
+        2054,
+        70,
+        3600,
+        3600,
+        3600,
+        432,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 273.75500322692096,
+    "mean_length": 2223.75,
+    "std_reward": 186.2417630023561,
+    "std_length": 1500.723254800831,
+    "task_count": 1,
+    "macro_mean_reward": 273.75500322692096,
+    "macro_mean_length": 2223.75,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1500.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1500.json
new file mode 100644
index 0000000000000000000000000000000000000000..3a7f05acfe2573afb8e1378f2821ecc61b6886ea
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1500.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 332.3600039117038,
+      "mean_length": 2695.55,
+      "std_reward": 174.17145606979872,
+      "std_length": 1403.4020263274526,
+      "episode_rewards": [
+        18.700000248849392,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        169.80000200867653,
+        23.40000030398369,
+        444.6000052243471,
+        444.6000052243471,
+        108.80000129342079,
+        444.6000052243471,
+        444.6000052243471,
+        81.30000097304583,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        20.80000026524067,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        171,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        1385,
+        209,
+        3600,
+        3600,
+        892,
+        3600,
+        3600,
+        671,
+        3600,
+        3600,
+        3600,
+        3600,
+        183,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 332.3600039117038,
+    "mean_length": 2695.55,
+    "std_reward": 174.17145606979872,
+    "std_length": 1403.4020263274526,
+    "task_count": 1,
+    "macro_mean_reward": 332.3600039117038,
+    "macro_mean_length": 2695.55,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1750.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1750.json
new file mode 100644
index 0000000000000000000000000000000000000000..cb9647ca568962867f6f83777294940fbb3b6483
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_1750.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 262.9750031013042,
+      "mean_length": 2137.6,
+      "std_reward": 133.6748079966357,
+      "std_length": 1077.1758166613286,
+      "episode_rewards": [
+        261.40000309050083,
+        235.30000277608633,
+        444.6000052243471,
+        444.6000052243471,
+        230.80000272393227,
+        31.900000400841236,
+        158.70000188797712,
+        76.80000092089176,
+        146.80000174045563,
+        444.6000052243471,
+        286.80000337958336,
+        146.80000174045563,
+        444.6000052243471,
+        444.6000052243471,
+        146.80000174045563,
+        81.30000097304583,
+        356.8000041991472,
+        333.30000392347574,
+        228.2000027000904,
+        314.8000037074089
+      ],
+      "episode_lengths": [
+        2130,
+        1914,
+        3600,
+        3600,
+        1878,
+        276,
+        1301,
+        635,
+        1200,
+        3600,
+        2330,
+        1200,
+        3600,
+        3600,
+        1200,
+        671,
+        2895,
+        2705,
+        1861,
+        2556
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 262.9750031013042,
+    "mean_length": 2137.6,
+    "std_reward": 133.6748079966357,
+    "std_length": 1077.1758166613286,
+    "task_count": 1,
+    "macro_mean_reward": 262.9750031013042,
+    "macro_mean_length": 2137.6,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2000.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2000.json
new file mode 100644
index 0000000000000000000000000000000000000000..3f6421be1a9b5323db9d4099d99841fba40766e6
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2000.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 214.70500253476203,
+      "mean_length": 1747.45,
+      "std_reward": 161.25401726393875,
+      "std_length": 1301.0273431023654,
+      "episode_rewards": [
+        435.8000051230192,
+        67.30000080913305,
+        127.80000151693821,
+        444.6000052243471,
+        76.80000092089176,
+        34.80000042915344,
+        444.6000052243471,
+        328.8000038713217,
+        25.300000317394733,
+        421.80000495910645,
+        34.80000042915344,
+        67.30000080913305,
+        118.80000141263008,
+        393.8000046312809,
+        29.800000369548798,
+        221.30000261217356,
+        398.80000469088554,
+        305.3000035956502,
+        71.80000086128712,
+        244.80000288784504
+      ],
+      "episode_lengths": [
+        3532,
+        558,
+        1046,
+        3600,
+        635,
+        296,
+        3600,
+        2669,
+        219,
+        3419,
+        296,
+        558,
+        974,
+        3193,
+        255,
+        1801,
+        3234,
+        2479,
+        594,
+        1991
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 214.70500253476203,
+    "mean_length": 1747.45,
+    "std_reward": 161.25401726393875,
+    "std_length": 1301.0273431023654,
+    "task_count": 1,
+    "macro_mean_reward": 214.70500253476203,
+    "macro_mean_length": 1747.45,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2250.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2250.json
new file mode 100644
index 0000000000000000000000000000000000000000..0b57585a0d5dfa231c77b6df46677feefadc6486
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2250.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 354.3150041695684,
+      "mean_length": 2873.35,
+      "std_reward": 135.2535348989007,
+      "std_length": 1089.2268025989813,
+      "episode_rewards": [
+        444.6000052243471,
+        113.40000134706497,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        158.0000018775463,
+        116.00000138580799,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        172.00000204145908,
+        409.9000048264861,
+        143.90000171214342,
+        193.300002284348
+      ],
+      "episode_lengths": [
+        3600,
+        929,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        1294,
+        955,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        1407,
+        3327,
+        1180,
+        1575
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 354.3150041695684,
+    "mean_length": 2873.35,
+    "std_reward": 135.2535348989007,
+    "std_length": 1089.2268025989813,
+    "task_count": 1,
+    "macro_mean_reward": 354.3150041695684,
+    "macro_mean_length": 2873.35,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_250.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_250.json
new file mode 100644
index 0000000000000000000000000000000000000000..ab123ccf28f0df0bdfe5f0135fa48dd6470d63e2
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_250.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 22.400000289827585,
+      "mean_length": 199.45,
+      "std_reward": 25.600390922011698,
+      "std_length": 206.7983256702046,
+      "episode_rewards": [
+        12.900000177323818,
+        8.400000125169754,
+        4.700000084936619,
+        13.700000189244747,
+        26.90000034123659,
+        9.200000137090683,
+        18.700000248849392,
+        93.20000112056732,
+        17.900000236928463,
+        15.300000198185444,
+        18.800000250339508,
+        13.800000190734863,
+        18.700000248849392,
+        17.900000236928463,
+        9.200000137090683,
+        9.200000137090683,
+        8.400000125169754,
+        10.500000141561031,
+        18.700000248849392,
+        101.9000012204051
+      ],
+      "episode_lengths": [
+        122,
+        86,
+        58,
+        130,
+        235,
+        94,
+        171,
+        772,
+        163,
+        137,
+        172,
+        131,
+        171,
+        163,
+        94,
+        94,
+        86,
+        98,
+        171,
+        841
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 22.400000289827585,
+    "mean_length": 199.45,
+    "std_reward": 25.600390922011698,
+    "std_length": 206.7983256702046,
+    "task_count": 1,
+    "macro_mean_reward": 22.400000289827585,
+    "macro_mean_length": 199.45,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2500.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2500.json
new file mode 100644
index 0000000000000000000000000000000000000000..ab7ea8cc80d531592d6a139a5b56ad695d995c18
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2500.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 318.88500375412406,
+      "mean_length": 2587.3,
+      "std_reward": 154.78360002135435,
+      "std_length": 1247.67135496492,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        193.300002284348,
+        444.6000052243471,
+        90.80000108480453,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        149.50000178068876,
+        67.30000080913305,
+        444.6000052243471,
+        440.80000518262386,
+        272.8000032156706,
+        141.80000168085098,
+        398.80000469088554,
+        444.6000052243471,
+        76.80000092089176,
+        99.80000118911266
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3600,
+        3600,
+        1575,
+        3600,
+        748,
+        3600,
+        3600,
+        3600,
+        1227,
+        558,
+        3600,
+        3573,
+        2217,
+        1159,
+        3234,
+        3600,
+        635,
+        820
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 318.88500375412406,
+    "mean_length": 2587.3,
+    "std_reward": 154.78360002135435,
+    "std_length": 1247.67135496492,
+    "task_count": 1,
+    "macro_mean_reward": 318.88500375412406,
+    "macro_mean_length": 2587.3,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2750.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2750.json
new file mode 100644
index 0000000000000000000000000000000000000000..cfb1a8df5d3976664c345319118071543dc4dfc5
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_2750.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 256.680003028363,
+      "mean_length": 2086.95,
+      "std_reward": 167.62305705644414,
+      "std_length": 1350.2858762128856,
+      "episode_rewards": [
+        444.6000052243471,
+        20.80000026524067,
+        214.20000253617764,
+        88.10000105947256,
+        55.40000067651272,
+        200.20000237226486,
+        90.80000108480453,
+        32.00000040233135,
+        444.6000052243471,
+        116.40000139176846,
+        444.6000052243471,
+        444.6000052243471,
+        118.80000141263008,
+        444.6000052243471,
+        146.80000174045563,
+        444.6000052243471,
+        444.6000052243471,
+        144.400001719594,
+        348.9000041112304,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        183,
+        1748,
+        730,
+        466,
+        1635,
+        748,
+        277,
+        3600,
+        959,
+        3600,
+        3600,
+        974,
+        3600,
+        1200,
+        3600,
+        3600,
+        1185,
+        2834,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 256.680003028363,
+    "mean_length": 2086.95,
+    "std_reward": 167.62305705644414,
+    "std_length": 1350.2858762128856,
+    "task_count": 1,
+    "macro_mean_reward": 256.680003028363,
+    "macro_mean_length": 2086.95,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3000.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3000.json
new file mode 100644
index 0000000000000000000000000000000000000000..30ff927e1b423dfdf179d1bb391d38f40fde69a0
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3000.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 283.24500333927574,
+      "mean_length": 2301.4,
+      "std_reward": 161.9693615741844,
+      "std_length": 1305.6022135397902,
+      "episode_rewards": [
+        444.6000052243471,
+        359.40000423789024,
+        258.8000030517578,
+        398.80000469088554,
+        106.40000127255917,
+        31.900000400841236,
+        90.80000108480453,
+        331.4000039100647,
+        79.5000009611249,
+        288.4000034034252,
+        444.6000052243471,
+        444.6000052243471,
+        118.80000141263008,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        81.30000097304583,
+        386.40000455081463,
+        20.80000026524067
+      ],
+      "episode_lengths": [
+        3600,
+        2921,
+        2104,
+        3234,
+        877,
+        276,
+        748,
+        2695,
+        662,
+        2346,
+        3600,
+        3600,
+        974,
+        3600,
+        3600,
+        3600,
+        3600,
+        671,
+        3137,
+        183
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 283.24500333927574,
+    "mean_length": 2301.4,
+    "std_reward": 161.9693615741844,
+    "std_length": 1305.6022135397902,
+    "task_count": 1,
+    "macro_mean_reward": 283.24500333927574,
+    "macro_mean_length": 2301.4,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3250.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3250.json
new file mode 100644
index 0000000000000000000000000000000000000000..4bb1f160581b80160aede30dc6e06225fd81741e
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3250.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 376.70500443093476,
+      "mean_length": 3053.6,
+      "std_reward": 120.41256086491829,
+      "std_length": 969.9987835043918,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        249.3000029399991,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        9.200000137090683,
+        202.80000239610672,
+        169.80000200867653,
+        444.6000052243471,
+        433.10000509768724,
+        361.3000042513013,
+        444.6000052243471,
+        328.8000038713217,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        2027,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        94,
+        1652,
+        1385,
+        3600,
+        3514,
+        2931,
+        3600,
+        2669,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 376.70500443093476,
+    "mean_length": 3053.6,
+    "std_reward": 120.41256086491829,
+    "std_length": 969.9987835043918,
+    "task_count": 1,
+    "macro_mean_reward": 376.70500443093476,
+    "macro_mean_length": 3053.6,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3500.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3500.json
new file mode 100644
index 0000000000000000000000000000000000000000..ba1a8928d63c2ab9e93a4b151aab5817e56023b8
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3500.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 416.78500490002335,
+      "mean_length": 3376.7,
+      "std_reward": 63.41565555134738,
+      "std_length": 510.38819539640605,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        387.600004568696,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        258.100003041327,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        355.1000041887164,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        221.30000261217356,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3149,
+        3600,
+        3600,
+        3600,
+        2097,
+        3600,
+        3600,
+        3600,
+        3600,
+        2887,
+        3600,
+        3600,
+        3600,
+        1801,
+        3600,
+        3600,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 416.78500490002335,
+    "mean_length": 3376.7,
+    "std_reward": 63.41565555134738,
+    "std_length": 510.38819539640605,
+    "task_count": 1,
+    "macro_mean_reward": 416.78500490002335,
+    "macro_mean_length": 3376.7,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3750.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3750.json
new file mode 100644
index 0000000000000000000000000000000000000000..9562ae22bd54861acd52f4d80cda68da859030d5
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_3750.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 411.52000483796,
+      "mean_length": 3334.25,
+      "std_reward": 55.36169859795965,
+      "std_length": 445.4333704382733,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        393.8000046312809,
+        272.8000032156706,
+        328.8000038713217,
+        444.6000052243471,
+        444.6000052243471,
+        384.80000452697277,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        440.80000518262386,
+        444.6000052243471,
+        328.8000038713217,
+        300.80000354349613,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3193,
+        2217,
+        2669,
+        3600,
+        3600,
+        3121,
+        3600,
+        3600,
+        3600,
+        3600,
+        3573,
+        3600,
+        2669,
+        2443,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 411.52000483796,
+    "mean_length": 3334.25,
+    "std_reward": 55.36169859795965,
+    "std_length": 445.4333704382733,
+    "task_count": 1,
+    "macro_mean_reward": 411.52000483796,
+    "macro_mean_length": 3334.25,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_4000.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_4000.json
new file mode 100644
index 0000000000000000000000000000000000000000..81a41da5a17e577eed813a9999ca544aa385c068
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_4000.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 377.76000444367526,
+      "mean_length": 3062.25,
+      "std_reward": 126.19024436197301,
+      "std_length": 1016.0166767824237,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        431.30000507086515,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        328.8000038713217,
+        79.40000095963478,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        83.20000100135803,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        118.80000141263008,
+        289.30000341683626,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3496,
+        3600,
+        3600,
+        3600,
+        3600,
+        2669,
+        661,
+        3600,
+        3600,
+        3600,
+        3600,
+        690,
+        3600,
+        3600,
+        3600,
+        974,
+        2355,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 4000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 377.76000444367526,
+    "mean_length": 3062.25,
+    "std_reward": 126.19024436197301,
+    "std_length": 1016.0166767824237,
+    "task_count": 1,
+    "macro_mean_reward": 377.76000444367526,
+    "macro_mean_length": 3062.25,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_500.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_500.json
new file mode 100644
index 0000000000000000000000000000000000000000..b41fb04f9a2256af49ccfe0b505116ab520599e6
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_500.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 49.130000597983596,
+      "mean_length": 412.3,
+      "std_reward": 48.54585621156259,
+      "std_length": 392.63979676033864,
+      "episode_rewards": [
+        6.800000101327896,
+        11.30000015348196,
+        41.70000051707029,
+        186.2000022083521,
+        53.30000064522028,
+        6.800000101327896,
+        90.80000108480453,
+        6.800000101327896,
+        43.80000053346157,
+        15.600000202655792,
+        34.80000042915344,
+        8.900000132620335,
+        95.3000011369586,
+        20.80000026524067,
+        29.800000369548798,
+        6.800000101327896,
+        76.80000092089176,
+        81.20000097155571,
+        144.30000171810389,
+        20.80000026524067
+      ],
+      "episode_lengths": [
+        70,
+        106,
+        356,
+        1522,
+        445,
+        70,
+        748,
+        70,
+        368,
+        140,
+        296,
+        91,
+        784,
+        183,
+        255,
+        70,
+        635,
+        670,
+        1184,
+        183
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 49.130000597983596,
+    "mean_length": 412.3,
+    "std_reward": 48.54585621156259,
+    "std_length": 392.63979676033864,
+    "task_count": 1,
+    "macro_mean_reward": 49.130000597983596,
+    "macro_mean_length": 412.3,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_750.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_750.json
new file mode 100644
index 0000000000000000000000000000000000000000..2d4ba50620f32fcb68c8948a04744cafcb6b94e7
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/mid_train/step_750.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 267.25000315085055,
+      "mean_length": 2171.8,
+      "std_reward": 160.98253665124207,
+      "std_length": 1298.493419313321,
+      "episode_rewards": [
+        69.40000084042549,
+        410.10000482946634,
+        426.8000050187111,
+        444.6000052243471,
+        314.8000037074089,
+        444.6000052243471,
+        272.8000032156706,
+        110.90000132471323,
+        444.6000052243471,
+        11.30000015348196,
+        90.80000108480453,
+        272.8000032156706,
+        444.6000052243471,
+        444.6000052243471,
+        423.9000049903989,
+        328.8000038713217,
+        76.80000092089176,
+        66.70000080019236,
+        67.30000080913305,
+        178.80000211298466
+      ],
+      "episode_lengths": [
+        579,
+        3329,
+        3460,
+        3600,
+        2556,
+        3600,
+        2217,
+        913,
+        3600,
+        106,
+        748,
+        2217,
+        3600,
+        3600,
+        3440,
+        2669,
+        635,
+        552,
+        558,
+        1457
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 267.25000315085055,
+    "mean_length": 2171.8,
+    "std_reward": 160.98253665124207,
+    "std_length": 1298.493419313321,
+    "task_count": 1,
+    "macro_mean_reward": 267.25000315085055,
+    "macro_mean_length": 2171.8,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/post_train/step_4000.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/post_train/step_4000.json
new file mode 100644
index 0000000000000000000000000000000000000000..778246f039875e2221107fe7254699567a821aa9
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/eval/post_train/step_4000.json
@@ -0,0 +1,440 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 100,
+      "mean_reward": 352.05800414279105,
+      "mean_length": 2855.02,
+      "std_reward": 141.33372174823657,
+      "std_length": 1138.840058831792,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        431.30000507086515,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        328.8000038713217,
+        79.40000095963478,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        83.20000100135803,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        118.80000141263008,
+        289.30000341683626,
+        444.6000052243471,
+        219.10000259429216,
+        412.8000048547983,
+        34.80000042915344,
+        444.6000052243471,
+        65.40000079572201,
+        373.400004401803,
+        20.80000026524067,
+        185.90000220388174,
+        401.40000472962856,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        43.400000527501106,
+        118.80000141263008,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        188.80000223219395,
+        444.6000052243471,
+        272.8000032156706,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        179.30000212043524,
+        444.6000052243471,
+        216.8000025600195,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        155.80000184476376,
+        167.70000199228525,
+        444.6000052243471,
+        444.6000052243471,
+        104.80000124871731,
+        221.30000261217356,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        379.700004465878,
+        444.6000052243471,
+        286.80000337958336,
+        444.6000052243471,
+        34.80000042915344,
+        230.80000272393227,
+        444.6000052243471,
+        444.6000052243471,
+        20.80000026524067,
+        444.6000052243471,
+        319.30000375956297,
+        13.600000187754631,
+        163.50000194460154,
+        347.3000040873885,
+        444.6000052243471,
+        179.30000212043524,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        188.80000223219395,
+        444.6000052243471,
+        426.8000050187111,
+        444.6000052243471,
+        286.80000337958336,
+        431.30000507086515,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        443.500005222857,
+        444.6000052243471,
+        444.6000052243471,
+        62.80000075697899
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3496,
+        3600,
+        3600,
+        3600,
+        3600,
+        2669,
+        661,
+        3600,
+        3600,
+        3600,
+        3600,
+        690,
+        3600,
+        3600,
+        3600,
+        974,
+        2355,
+        3600,
+        1788,
+        3347,
+        296,
+        3600,
+        548,
+        3034,
+        183,
+        1519,
+        3260,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        364,
+        974,
+        3600,
+        3600,
+        3600,
+        1539,
+        3600,
+        2217,
+        3600,
+        3600,
+        3600,
+        3600,
+        1462,
+        3600,
+        1765,
+        3600,
+        3600,
+        3600,
+        3600,
+        1272,
+        1373,
+        3600,
+        3600,
+        861,
+        1801,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3079,
+        3600,
+        2330,
+        3600,
+        296,
+        1878,
+        3600,
+        3600,
+        183,
+        3600,
+        2592,
+        129,
+        1340,
+        2818,
+        3600,
+        1462,
+        3600,
+        3600,
+        3600,
+        1539,
+        3600,
+        3460,
+        3600,
+        2330,
+        3496,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        522
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19,
+        20,
+        21,
+        22,
+        23,
+        24,
+        25,
+        26,
+        27,
+        28,
+        29,
+        30,
+        31,
+        32,
+        33,
+        34,
+        35,
+        36,
+        37,
+        38,
+        39,
+        40,
+        41,
+        42,
+        43,
+        44,
+        45,
+        46,
+        47,
+        48,
+        49,
+        50,
+        51,
+        52,
+        53,
+        54,
+        55,
+        56,
+        57,
+        58,
+        59,
+        60,
+        61,
+        62,
+        63,
+        64,
+        65,
+        66,
+        67,
+        68,
+        69,
+        70,
+        71,
+        72,
+        73,
+        74,
+        75,
+        76,
+        77,
+        78,
+        79,
+        80,
+        81,
+        82,
+        83,
+        84,
+        85,
+        86,
+        87,
+        88,
+        89,
+        90,
+        91,
+        92,
+        93,
+        94,
+        95,
+        96,
+        97,
+        98,
+        99
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "post_train",
+    "step": 4000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 100,
+    "mean_reward": 352.05800414279105,
+    "mean_length": 2855.02,
+    "std_reward": 141.33372174823657,
+    "std_length": 1138.840058831792,
+    "task_count": 1,
+    "macro_mean_reward": 352.05800414279105,
+    "macro_mean_length": 2855.02,
+    "distributed_eval": false
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/hydra/.hydra/hydra.yaml b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/hydra/.hydra/hydra.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..08ed491eea315897d6afd05552575c875c162dc7
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/hydra/.hydra/hydra.yaml
@@ -0,0 +1,369 @@
+hydra:
+  run:
+    dir: ${run_root_dir}/${run_id}/hydra
+  sweep:
+    dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task:
+    - model=openvla
+    - env=flappy
+    - init=bridge
+    - mode=single
+    - ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++framework.qwenvl.attn_implementation=flash_attention_2
+    - ++framework.qwenvl.enable_gradient_checkpointing=true
+    - ++framework.action_model.state_dim=7
+    - ++framework.action_model.loss_type=discrete_ce
+    - ++framework.action_model.action_horizon=1
+    - ++framework.action_model.future_action_window_size=0
+    - ++framework.action_model.past_action_window_size=0
+    - ++framework.action_model.action_dim=7
+    - ++framework.action_model.action_env_dim=2
+    - ++framework.kv_memory.enabled=false
+    - ++framework.kv_memory.window=4
+    - ++framework.kv_memory.rollout_len=8
+    - ++framework.kv_memory.packed_train=false
+    - ++framework.kv_memory.rebased_sink=true
+    - ++framework.name=QwenOFT
+    - ++datasets.vla_data.dataset_py=lerobot_datasets
+    - ++datasets.vla_data.include_state=true
+    - ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games
+    - ++datasets.vla_data.data_mix=flappy_train
+    - ++datasets.vla_data.eval_data_mix=null
+    - ++datasets.vla_data.custom_mixtures_path=null
+    - ++datasets.vla_data.action_type=discrete
+    - ++datasets.vla_data.sequential_step_sampling=false
+    - ++datasets.vla_data.eval_sequential_step_sampling=null
+    - ++datasets.vla_data.num_workers=8
+    - ++datasets.vla_data.eval_num_workers=8
+    - ++datasets.vla_data.prefetch_factor=4
+    - ++datasets.vla_data.persistent_workers=true
+    - ++datasets.vla_data.pin_memory=true
+    - ++datasets.vla_data.shuffle=true
+    - ++datasets.vla_data.action_balance.enabled=false
+    - ++datasets.vla_data.action_balance.strategy=balanced_epoch
+    - ++datasets.vla_data.action_balance.action_key=action_id
+    - ++datasets.vla_data.action_balance.target_flap_fraction=0.3
+    - ++datasets.vla_data.action_balance.noop_id=0
+    - ++datasets.vla_data.action_balance.flap_id=1
+    - ++datasets.vla_data.latency_curriculum.enabled=false
+    - ++datasets.vla_data.latency_curriculum.strategy=exclusive
+    - ++datasets.vla_data.latency_curriculum.latencies=null
+    - ++datasets.vla_data.latency_curriculum.phase_steps=null
+    - ++datasets.vla_data.latency_curriculum.phase_distributions=null
+    - ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0
+    - ++datasets.vla_data.latency_curriculum.replay_passes=0.25
+    - ++datasets.vla_data.latency_curriculum.target_total_passes=2.0
+    - ++datasets.vla_data.latency_curriculum.final_equalization=true
+    - ++datasets.vla_data.latency_curriculum.step_budget_mode=auto
+    - ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false
+    - ++datasets.vla_data.latency_curriculum.save_at_phase_end=false
+    - ++datasets.vla_data.latency_curriculum.computed_plan=null
+    - ++datasets.vla_data.per_device_batch_size=8
+    - ++datasets.vla_data.load_all_data_for_training=true
+    - ++datasets.vla_data.num_obs_frames=4
+    - ++datasets.vla_data.image_mode=multiframe
+    - ++datasets.vla_data.prompt_mode=raw
+    - ++datasets.vla_data.stitch_grid=[2,2]
+    - ++datasets.vla_data.obs_image_size=null
+    - ++datasets.vla_data.video_backend=torchvision_av
+    - ++dataset.source_hf=
+    - ++dataset.config_name=null
+    - ++dataset.source_subdir=null
+    - ++dataset.converted_name=flappy_train
+    - ++dataset.single_source_hf=
+    - ++dataset.mixed_source_hf=
+    - ++dataset.single_converted_name=flappy_train
+    - ++dataset.mixed_converted_name=flappy_mixed_latency_train
+    - ++dataset.single_latency_filter=null
+    - ++dataset.mixed_latency_filter=null
+    - ++dataset.force_download=false
+    - ++dataset.setup_force=false
+    - ++dataset.skip_verification=false
+    - ++dataset.target_latency_unit=raw_frames
+    - ++dataset.verify_rows=200
+    - ++dataset.max_episodes=null
+    - ++dataset.episodes_per_latency=null
+    - ++dataset.latency_filter=null
+    - ++dataset.debug_subset.enabled=false
+    - ++dataset.debug_subset.max_episodes=5
+    - ++dataset.debug_subset.suffix=debug
+    - ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct
+    - ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1
+    - ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1
+    - ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt
+    - ++trainer.max_train_steps=4000
+    - ++trainer.num_warmup_steps=100
+    - ++trainer.save_interval=500
+    - ++trainer.eval_interval=250
+    - ++trainer.eval_num_batches=50
+    - ++trainer.per_latency_eval_num_batches=null
+    - ++trainer.eval_action_classification=false
+    - ++trainer.eval_action_classification_interval=null
+    - ++trainer.cc_f1_tolerance=1
+    - ++trainer.learning_rate.base=2e-05
+    - ++trainer.learning_rate.qwen_vl_interface=1e-05
+    - ++trainer.learning_rate.action_model=0.0001
+    - ++trainer.lr_scheduler_type=cosine_with_min_lr
+    - ++trainer.scheduler_specific_kwargs.min_lr=1e-06
+    - ++trainer.freeze_modules=
+    - ++trainer.freeze_vit=false
+    - ++trainer.freeze_tied_embedding=false
+    - ++trainer.freeze_llm_layers=[]
+    - ++trainer.loss_scale.vla=1.0
+    - ++trainer.loss_scale.vlm=0.1
+    - ++trainer.max_grad_norm=1.0
+    - ++trainer.weight_decay=0.0
+    - ++trainer.logging_frequency=1
+    - ++trainer.profile_timing.enabled=false
+    - ++trainer.profile_timing.log_interval=10
+    - ++trainer.gradient_clipping=1.0
+    - ++trainer.gradient_accumulation_steps=16
+    - ++trainer.distributed_backend=deepspeed
+    - ++trainer.is_resume=false
+    - ++trainer.pretrained_checkpoint=null
+    - ++trainer.resume_step=0
+    - ++trainer.reload_modules=null
+    - ++trainer.optimizer.name=AdamW
+    - ++trainer.optimizer.betas=[0.9,0.95]
+    - ++trainer.optimizer.eps=1e-08
+    - ++trainer.optimizer.weight_decay=1e-08
+    - ++trainer.optimizer.fused=true
+    - ++trainer.save_format=pt
+    - ++workspace_dir=WORKSPACE_DIR
+    - ++run_root_dir=results/Checkpoints
+    - ++seed=42
+    - ++auth.env_file=null
+    - ++auth.hf_token_env=HF_TOKEN
+    - ++auth.wandb_api_key_env=WANDB_API_KEY
+    - ++paths.run_root_dir=results/Checkpoints
+    - ++paths.dataset_local_dir=data/memory/flappy_fix_latency_3_200ep_7k2steps
+    - ++paths.dataset_cache_dir=null
+    - ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml
+    - ++rl_games.model_alias=openvla
+    - ++rl_games.env_eval.image_size=224
+    - ++rl_games.env_eval.frameskip=1
+    - ++rl_games.env_eval.image_transform=raw_rgb
+    - ++rl_games.env_eval.prompt_mode=raw
+    - ++rl_games.env_eval.ghost_trail.history_frames=5
+    - ++rl_games.env_eval.ghost_trail.gamma=1.3
+    - ++rl_games.env_eval.ghost_trail.min_alpha=35
+    - ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0
+    - ++rl_games.env_eval.ghost_trail.ground_fraction=0.22
+    - ++rl_games.env_eval.seed=42
+    - ++rl_games.env_eval.fixed_episode_seeds=true
+    - ++rl_games.env_eval.latency_seed_stride=0
+    - ++rl_games.env_eval.task_seed_stride=0
+    - ++rl_games.env_eval.task_description=
+    - ++rl_games.env_eval.eval_parallel_envs=5
+    - ++rl_games.env_eval.action_chunk_execution.enabled=false
+    - ++rl_games.env_eval.action_chunk_execution.chunk_size=null
+    - ++rl_games.env_eval.enabled=true
+    - ++rl_games.env_eval.eval_backend=latency_bench
+    - ++rl_games.env_eval.distributed_mode=rank_sharded
+    - ++rl_games.env_eval.vectorized.enabled=false
+    - ++rl_games.env_eval.vectorized.batch_size=1
+    - ++rl_games.env_eval.latency.prompt_map_path=null
+    - ++rl_games.env_eval.latency.mode=single
+    - ++rl_games.env_eval.latency.values=[0]
+    - ++rl_games.env_eval.mid_train.enabled=true
+    - ++rl_games.env_eval.mid_train.interval_steps=250
+    - ++rl_games.env_eval.mid_train.latencies=[3]
+    - ++rl_games.env_eval.mid_train.num_episodes=20
+    - ++rl_games.env_eval.mid_train.max_steps_per_episode=3600
+    - ++rl_games.env_eval.post_train.enabled=true
+    - ++rl_games.env_eval.post_train.latencies=[3]
+    - ++rl_games.env_eval.post_train.num_episodes=50
+    - ++rl_games.env_eval.post_train.max_steps_per_episode=3600
+    - ++rl_games.task=flappy
+    - ++rl_games.initialization_mode=bridge
+    - ++rl_games.action_carrier=bridge
+    - ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action
+    - ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action
+    - ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++checkpoint.load=auto
+    - ++checkpoint.hf_repo_id=null
+    - ++checkpoint.save_best_model=false
+    - ++checkpoint.save_final_model=true
+    - ++checkpoint.save_pt_file=false
+    - ++checkpoint.save_training_state=true
+    - ++checkpoint.save_safetensors_file=true
+    - ++checkpoint.local.keep_last_n=1
+    - ++checkpoint.sync.enabled=false
+    - ++checkpoint.sync.repo_id=null
+    - ++checkpoint.sync.keep_last_n=0
+    - ++checkpoint.sync.sync_every_n_checkpoints=1
+    - ++checkpoint.sync.resume_policy=local_latest
+    - ++run_id=flappy_fix_latency_3_200ep_7k2steps_plain_multi
+    - ++output_dir=null
+    - ++config_yaml=null
+    - ++is_debug=false
+    - ++version_id=0.21
+    - ++run_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints
+    - ++trainer.is_resume=true
+    - ++trainer.pretrained_checkpoint=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state
+    - ++trainer.resume_step=4000
+    - ++datasets.vla_data.data_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps
+    - ++datasets.vla_data.data_mix=flappy_train__bridge
+    - ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val
+    - ++framework.qwenvl.base_vlm=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++rl_games.env_eval.latency.prompt_map_path=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json
+  job:
+    name: train_starvla_hydra
+    chdir: false
+    override_dirname: ++auth.env_file=null,++auth.hf_token_env=HF_TOKEN,++auth.wandb_api_key_env=WANDB_API_KEY,++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action,++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action,++checkpoint.hf_repo_id=null,++checkpoint.load=auto,++checkpoint.local.keep_last_n=1,++checkpoint.save_best_model=false,++checkpoint.save_final_model=true,++checkpoint.save_pt_file=false,++checkpoint.save_safetensors_file=true,++checkpoint.save_training_state=true,++checkpoint.sync.enabled=false,++checkpoint.sync.keep_last_n=0,++checkpoint.sync.repo_id=null,++checkpoint.sync.resume_policy=local_latest,++checkpoint.sync.sync_every_n_checkpoints=1,++config_yaml=null,++dataset.config_name=null,++dataset.converted_name=flappy_train,++dataset.debug_subset.enabled=false,++dataset.debug_subset.max_episodes=5,++dataset.debug_subset.suffix=debug,++dataset.episodes_per_latency=null,++dataset.force_download=false,++dataset.latency_filter=null,++dataset.max_episodes=null,++dataset.mixed_converted_name=flappy_mixed_latency_train,++dataset.mixed_latency_filter=null,++dataset.mixed_source_hf=,++dataset.setup_force=false,++dataset.single_converted_name=flappy_train,++dataset.single_latency_filter=null,++dataset.single_source_hf=,++dataset.skip_verification=false,++dataset.source_hf=,++dataset.source_subdir=null,++dataset.target_latency_unit=raw_frames,++dataset.verify_rows=200,++datasets.vla_data.action_balance.action_key=action_id,++datasets.vla_data.action_balance.enabled=false,++datasets.vla_data.action_balance.flap_id=1,++datasets.vla_data.action_balance.noop_id=0,++datasets.vla_data.action_balance.strategy=balanced_epoch,++datasets.vla_data.action_balance.target_flap_fraction=0.3,++datasets.vla_data.action_type=discrete,++datasets.vla_data.custom_mixtures_path=null,++datasets.vla_data.data_mix=flappy_train,++datasets.vla_data.data_mix=flappy_train__bridge,++datasets.vla_data.data_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps,++datasets.vla_data.data_root_dir=playground/Datasets/rl_games,++datasets.vla_data.dataset_py=lerobot_datasets,++datasets.vla_data.eval_data_mix=flappy_train__bridge__val,++datasets.vla_data.eval_data_mix=null,++datasets.vla_data.eval_num_workers=8,++datasets.vla_data.eval_sequential_step_sampling=null,++datasets.vla_data.image_mode=multiframe,++datasets.vla_data.include_state=true,++datasets.vla_data.latency_curriculum.computed_plan=null,++datasets.vla_data.latency_curriculum.enabled=false,++datasets.vla_data.latency_curriculum.eval_at_phase_end=false,++datasets.vla_data.latency_curriculum.final_equalization=true,++datasets.vla_data.latency_curriculum.latencies=null,++datasets.vla_data.latency_curriculum.new_latency_passes=1.0,++datasets.vla_data.latency_curriculum.phase_distributions=null,++datasets.vla_data.latency_curriculum.phase_steps=null,++datasets.vla_data.latency_curriculum.replay_passes=0.25,++datasets.vla_data.latency_curriculum.save_at_phase_end=false,++datasets.vla_data.latency_curriculum.step_budget_mode=auto,++datasets.vla_data.latency_curriculum.strategy=exclusive,++datasets.vla_data.latency_curriculum.target_total_passes=2.0,++datasets.vla_data.load_all_data_for_training=true,++datasets.vla_data.num_obs_frames=4,++datasets.vla_data.num_workers=8,++datasets.vla_data.obs_image_size=null,++datasets.vla_data.per_device_batch_size=8,++datasets.vla_data.persistent_workers=true,++datasets.vla_data.pin_memory=true,++datasets.vla_data.prefetch_factor=4,++datasets.vla_data.prompt_mode=raw,++datasets.vla_data.sequential_step_sampling=false,++datasets.vla_data.shuffle=true,++datasets.vla_data.stitch_grid=[2,2],++datasets.vla_data.video_backend=torchvision_av,++framework.action_model.action_dim=7,++framework.action_model.action_env_dim=2,++framework.action_model.action_horizon=1,++framework.action_model.future_action_window_size=0,++framework.action_model.loss_type=discrete_ce,++framework.action_model.past_action_window_size=0,++framework.action_model.state_dim=7,++framework.kv_memory.enabled=false,++framework.kv_memory.packed_train=false,++framework.kv_memory.rebased_sink=true,++framework.kv_memory.rollout_len=8,++framework.kv_memory.window=4,++framework.name=QwenOFT,++framework.qwenvl.attn_implementation=flash_attention_2,++framework.qwenvl.base_vlm=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.enable_gradient_checkpointing=true,++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt,++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1,++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1,++is_debug=false,++output_dir=null,++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml,++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++paths.dataset_cache_dir=null,++paths.dataset_local_dir=data/memory/flappy_fix_latency_3_200ep_7k2steps,++paths.run_root_dir=results/Checkpoints,++rl_games.action_carrier=bridge,++rl_games.env_eval.action_chunk_execution.chunk_size=null,++rl_games.env_eval.action_chunk_execution.enabled=false,++rl_games.env_eval.distributed_mode=rank_sharded,++rl_games.env_eval.enabled=true,++rl_games.env_eval.eval_backend=latency_bench,++rl_games.env_eval.eval_parallel_envs=5,++rl_games.env_eval.fixed_episode_seeds=true,++rl_games.env_eval.frameskip=1,++rl_games.env_eval.ghost_trail.gamma=1.3,++rl_games.env_eval.ghost_trail.ground_fraction=0.22,++rl_games.env_eval.ghost_trail.history_frames=5,++rl_games.env_eval.ghost_trail.min_alpha=35,++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0,++rl_games.env_eval.image_size=224,++rl_games.env_eval.image_transform=raw_rgb,++rl_games.env_eval.latency.mode=single,++rl_games.env_eval.latency.prompt_map_path=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json,++rl_games.env_eval.latency.prompt_map_path=null,++rl_games.env_eval.latency.values=[0],++rl_games.env_eval.latency_seed_stride=0,++rl_games.env_eval.mid_train.enabled=true,++rl_games.env_eval.mid_train.interval_steps=250,++rl_games.env_eval.mid_train.latencies=[3],++rl_games.env_eval.mid_train.max_steps_per_episode=3600,++rl_games.env_eval.mid_train.num_episodes=20,++rl_games.env_eval.post_train.enabled=true,++rl_games.env_eval.post_train.latencies=[3],++rl_games.env_eval.post_train.max_steps_per_episode=3600,++rl_games.env_eval.post_train.num_episodes=50,++rl_games.env_eval.prompt_mode=raw,++rl_games.env_eval.seed=42,++rl_games.env_eval.task_description=,++rl_games.env_eval.task_seed_stride=0,++rl_games.env_eval.vectorized.batch_size=1,++rl_games.env_eval.vectorized.enabled=false,++rl_games.initialization_mode=bridge,++rl_games.model_alias=openvla,++rl_games.task=flappy,++run_id=flappy_fix_latency_3_200ep_7k2steps_plain_multi,++run_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints,++run_root_dir=results/Checkpoints,++seed=42,++trainer.cc_f1_tolerance=1,++trainer.distributed_backend=deepspeed,++trainer.eval_action_classification=false,++trainer.eval_action_classification_interval=null,++trainer.eval_interval=250,++trainer.eval_num_batches=50,++trainer.freeze_llm_layers=[],++trainer.freeze_modules=,++trainer.freeze_tied_embedding=false,++trainer.freeze_vit=false,++trainer.gradient_accumulation_steps=16,++trainer.gradient_clipping=1.0,++trainer.is_resume=false,++trainer.is_resume=true,++trainer.learning_rate.action_model=0.0001,++trainer.learning_rate.base=2e-05,++trainer.learning_rate.qwen_vl_interface=1e-05,++trainer.logging_frequency=1,++trainer.loss_scale.vla=1.0,++trainer.loss_scale.vlm=0.1,++trainer.lr_scheduler_type=cosine_with_min_lr,++trainer.max_grad_norm=1.0,++trainer.max_train_steps=4000,++trainer.num_warmup_steps=100,++trainer.optimizer.betas=[0.9,0.95],++trainer.optimizer.eps=1e-08,++trainer.optimizer.fused=true,++trainer.optimizer.name=AdamW,++trainer.optimizer.weight_decay=1e-08,++trainer.per_latency_eval_num_batches=null,++trainer.pretrained_checkpoint=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state,++trainer.pretrained_checkpoint=null,++trainer.profile_timing.enabled=false,++trainer.profile_timing.log_interval=10,++trainer.reload_modules=null,++trainer.resume_step=0,++trainer.resume_step=4000,++trainer.save_format=pt,++trainer.save_interval=500,++trainer.scheduler_specific_kwargs.min_lr=1e-06,++trainer.weight_decay=0.0,++version_id=0.21,++workspace_dir=WORKSPACE_DIR,env=flappy,init=bridge,mode=single,model=openvla
+    id: ???
+    num: ???
+    config_name: train
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.3.4
+    version_base: '1.1'
+    cwd: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/examples/rl_games/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/hydra
+    choices:
+      cross_task_setup: null
+      checkpoint: default
+      mode: single
+      init: bridge
+      env: flappy
+      model: openvla
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/hydra/.hydra/overrides.yaml b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/hydra/.hydra/overrides.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..d51fa2a7a26f17bf9474c89c6f768f514b3ba9e3
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/hydra/.hydra/overrides.yaml
@@ -0,0 +1,209 @@
+- model=openvla
+- env=flappy
+- init=bridge
+- mode=single
+- ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++framework.qwenvl.attn_implementation=flash_attention_2
+- ++framework.qwenvl.enable_gradient_checkpointing=true
+- ++framework.action_model.state_dim=7
+- ++framework.action_model.loss_type=discrete_ce
+- ++framework.action_model.action_horizon=1
+- ++framework.action_model.future_action_window_size=0
+- ++framework.action_model.past_action_window_size=0
+- ++framework.action_model.action_dim=7
+- ++framework.action_model.action_env_dim=2
+- ++framework.kv_memory.enabled=false
+- ++framework.kv_memory.window=4
+- ++framework.kv_memory.rollout_len=8
+- ++framework.kv_memory.packed_train=false
+- ++framework.kv_memory.rebased_sink=true
+- ++framework.name=QwenOFT
+- ++datasets.vla_data.dataset_py=lerobot_datasets
+- ++datasets.vla_data.include_state=true
+- ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games
+- ++datasets.vla_data.data_mix=flappy_train
+- ++datasets.vla_data.eval_data_mix=null
+- ++datasets.vla_data.custom_mixtures_path=null
+- ++datasets.vla_data.action_type=discrete
+- ++datasets.vla_data.sequential_step_sampling=false
+- ++datasets.vla_data.eval_sequential_step_sampling=null
+- ++datasets.vla_data.num_workers=8
+- ++datasets.vla_data.eval_num_workers=8
+- ++datasets.vla_data.prefetch_factor=4
+- ++datasets.vla_data.persistent_workers=true
+- ++datasets.vla_data.pin_memory=true
+- ++datasets.vla_data.shuffle=true
+- ++datasets.vla_data.action_balance.enabled=false
+- ++datasets.vla_data.action_balance.strategy=balanced_epoch
+- ++datasets.vla_data.action_balance.action_key=action_id
+- ++datasets.vla_data.action_balance.target_flap_fraction=0.3
+- ++datasets.vla_data.action_balance.noop_id=0
+- ++datasets.vla_data.action_balance.flap_id=1
+- ++datasets.vla_data.latency_curriculum.enabled=false
+- ++datasets.vla_data.latency_curriculum.strategy=exclusive
+- ++datasets.vla_data.latency_curriculum.latencies=null
+- ++datasets.vla_data.latency_curriculum.phase_steps=null
+- ++datasets.vla_data.latency_curriculum.phase_distributions=null
+- ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0
+- ++datasets.vla_data.latency_curriculum.replay_passes=0.25
+- ++datasets.vla_data.latency_curriculum.target_total_passes=2.0
+- ++datasets.vla_data.latency_curriculum.final_equalization=true
+- ++datasets.vla_data.latency_curriculum.step_budget_mode=auto
+- ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false
+- ++datasets.vla_data.latency_curriculum.save_at_phase_end=false
+- ++datasets.vla_data.latency_curriculum.computed_plan=null
+- ++datasets.vla_data.per_device_batch_size=8
+- ++datasets.vla_data.load_all_data_for_training=true
+- ++datasets.vla_data.num_obs_frames=4
+- ++datasets.vla_data.image_mode=multiframe
+- ++datasets.vla_data.prompt_mode=raw
+- ++datasets.vla_data.stitch_grid=[2,2]
+- ++datasets.vla_data.obs_image_size=null
+- ++datasets.vla_data.video_backend=torchvision_av
+- ++dataset.source_hf=
+- ++dataset.config_name=null
+- ++dataset.source_subdir=null
+- ++dataset.converted_name=flappy_train
+- ++dataset.single_source_hf=
+- ++dataset.mixed_source_hf=
+- ++dataset.single_converted_name=flappy_train
+- ++dataset.mixed_converted_name=flappy_mixed_latency_train
+- ++dataset.single_latency_filter=null
+- ++dataset.mixed_latency_filter=null
+- ++dataset.force_download=false
+- ++dataset.setup_force=false
+- ++dataset.skip_verification=false
+- ++dataset.target_latency_unit=raw_frames
+- ++dataset.verify_rows=200
+- ++dataset.max_episodes=null
+- ++dataset.episodes_per_latency=null
+- ++dataset.latency_filter=null
+- ++dataset.debug_subset.enabled=false
+- ++dataset.debug_subset.max_episodes=5
+- ++dataset.debug_subset.suffix=debug
+- ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct
+- ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1
+- ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1
+- ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt
+- ++trainer.max_train_steps=4000
+- ++trainer.num_warmup_steps=100
+- ++trainer.save_interval=500
+- ++trainer.eval_interval=250
+- ++trainer.eval_num_batches=50
+- ++trainer.per_latency_eval_num_batches=null
+- ++trainer.eval_action_classification=false
+- ++trainer.eval_action_classification_interval=null
+- ++trainer.cc_f1_tolerance=1
+- ++trainer.learning_rate.base=2e-05
+- ++trainer.learning_rate.qwen_vl_interface=1e-05
+- ++trainer.learning_rate.action_model=0.0001
+- ++trainer.lr_scheduler_type=cosine_with_min_lr
+- ++trainer.scheduler_specific_kwargs.min_lr=1e-06
+- ++trainer.freeze_modules=
+- ++trainer.freeze_vit=false
+- ++trainer.freeze_tied_embedding=false
+- ++trainer.freeze_llm_layers=[]
+- ++trainer.loss_scale.vla=1.0
+- ++trainer.loss_scale.vlm=0.1
+- ++trainer.max_grad_norm=1.0
+- ++trainer.weight_decay=0.0
+- ++trainer.logging_frequency=1
+- ++trainer.profile_timing.enabled=false
+- ++trainer.profile_timing.log_interval=10
+- ++trainer.gradient_clipping=1.0
+- ++trainer.gradient_accumulation_steps=16
+- ++trainer.distributed_backend=deepspeed
+- ++trainer.is_resume=false
+- ++trainer.pretrained_checkpoint=null
+- ++trainer.resume_step=0
+- ++trainer.reload_modules=null
+- ++trainer.optimizer.name=AdamW
+- ++trainer.optimizer.betas=[0.9,0.95]
+- ++trainer.optimizer.eps=1e-08
+- ++trainer.optimizer.weight_decay=1e-08
+- ++trainer.optimizer.fused=true
+- ++trainer.save_format=pt
+- ++workspace_dir=WORKSPACE_DIR
+- ++run_root_dir=results/Checkpoints
+- ++seed=42
+- ++auth.env_file=null
+- ++auth.hf_token_env=HF_TOKEN
+- ++auth.wandb_api_key_env=WANDB_API_KEY
+- ++paths.run_root_dir=results/Checkpoints
+- ++paths.dataset_local_dir=data/memory/flappy_fix_latency_3_200ep_7k2steps
+- ++paths.dataset_cache_dir=null
+- ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml
+- ++rl_games.model_alias=openvla
+- ++rl_games.env_eval.image_size=224
+- ++rl_games.env_eval.frameskip=1
+- ++rl_games.env_eval.image_transform=raw_rgb
+- ++rl_games.env_eval.prompt_mode=raw
+- ++rl_games.env_eval.ghost_trail.history_frames=5
+- ++rl_games.env_eval.ghost_trail.gamma=1.3
+- ++rl_games.env_eval.ghost_trail.min_alpha=35
+- ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0
+- ++rl_games.env_eval.ghost_trail.ground_fraction=0.22
+- ++rl_games.env_eval.seed=42
+- ++rl_games.env_eval.fixed_episode_seeds=true
+- ++rl_games.env_eval.latency_seed_stride=0
+- ++rl_games.env_eval.task_seed_stride=0
+- ++rl_games.env_eval.task_description=
+- ++rl_games.env_eval.eval_parallel_envs=5
+- ++rl_games.env_eval.action_chunk_execution.enabled=false
+- ++rl_games.env_eval.action_chunk_execution.chunk_size=null
+- ++rl_games.env_eval.enabled=true
+- ++rl_games.env_eval.eval_backend=latency_bench
+- ++rl_games.env_eval.distributed_mode=rank_sharded
+- ++rl_games.env_eval.vectorized.enabled=false
+- ++rl_games.env_eval.vectorized.batch_size=1
+- ++rl_games.env_eval.latency.prompt_map_path=null
+- ++rl_games.env_eval.latency.mode=single
+- ++rl_games.env_eval.latency.values=[0]
+- ++rl_games.env_eval.mid_train.enabled=true
+- ++rl_games.env_eval.mid_train.interval_steps=250
+- ++rl_games.env_eval.mid_train.latencies=[3]
+- ++rl_games.env_eval.mid_train.num_episodes=20
+- ++rl_games.env_eval.mid_train.max_steps_per_episode=3600
+- ++rl_games.env_eval.post_train.enabled=true
+- ++rl_games.env_eval.post_train.latencies=[3]
+- ++rl_games.env_eval.post_train.num_episodes=50
+- ++rl_games.env_eval.post_train.max_steps_per_episode=3600
+- ++rl_games.task=flappy
+- ++rl_games.initialization_mode=bridge
+- ++rl_games.action_carrier=bridge
+- ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action
+- ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action
+- ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++checkpoint.load=auto
+- ++checkpoint.hf_repo_id=null
+- ++checkpoint.save_best_model=false
+- ++checkpoint.save_final_model=true
+- ++checkpoint.save_pt_file=false
+- ++checkpoint.save_training_state=true
+- ++checkpoint.save_safetensors_file=true
+- ++checkpoint.local.keep_last_n=1
+- ++checkpoint.sync.enabled=false
+- ++checkpoint.sync.repo_id=null
+- ++checkpoint.sync.keep_last_n=0
+- ++checkpoint.sync.sync_every_n_checkpoints=1
+- ++checkpoint.sync.resume_policy=local_latest
+- ++run_id=flappy_fix_latency_3_200ep_7k2steps_plain_multi
+- ++output_dir=null
+- ++config_yaml=null
+- ++is_debug=false
+- ++version_id=0.21
+- ++run_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints
+- ++trainer.is_resume=true
+- ++trainer.pretrained_checkpoint=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state
+- ++trainer.resume_step=4000
+- ++datasets.vla_data.data_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps
+- ++datasets.vla_data.data_mix=flappy_train__bridge
+- ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val
+- ++framework.qwenvl.base_vlm=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++rl_games.env_eval.latency.prompt_map_path=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/hydra/train_starvla_hydra.log b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/hydra/train_starvla_hydra.log
new file mode 100644
index 0000000000000000000000000000000000000000..30788c32c9eba4c570cec72c9da1c2028856c408
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/hydra/train_starvla_hydra.log
@@ -0,0 +1,5292 @@
+[2026-07-28 05:49:56,659][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 05:49:56,663][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21'
+[2026-07-28 05:49:56,663][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 05:49:56,880][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 05:49:56,880][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 05:49:56,880][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 05:49:56,880][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 05:49:57,606][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 05:49:57,621][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 05:49:57,696][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 05:49:57,711][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 05:50:26,574][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt
+[2026-07-28 05:50:26,575][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt
+[2026-07-28 05:50:26,586][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 05:50:26,586][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 05:50:27,582][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 05:50:27,583][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 05:50:27,583][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 05:50:27,583][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 05:50:27,584][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 05:50:27,584][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 05:50:27,584][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 05:50:27,585][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 05:50:27,585][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 05:50:27,585][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 05:50:27,586][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 05:50:27,586][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 05:50:27,586][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 05:50:27,587][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 05:50:27,587][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 05:50:27,587][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 05:50:27,587][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 05:50:27,587][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 05:50:27,587][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 05:50:27,588][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 05:50:27,588][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 05:50:27,588][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 05:50:27,589][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 05:50:27,589][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 05:50:27,589][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 05:50:27,590][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 05:50:27,590][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 05:50:27,590][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 05:50:27,590][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 05:50:27,590][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 05:50:27,590][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 05:50:27,754][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 05:50:27,755][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 05:50:27,755][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 05:50:27,755][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 05:50:27,756][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 05:50:27,756][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 05:50:27,756][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 05:50:27,756][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 05:50:27,757][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 05:50:27,757][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 05:50:27,757][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 05:50:27,757][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 05:50:27,757][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 05:50:27,758][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 05:50:27,758][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 05:50:27,758][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 05:50:27,759][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 05:50:27,759][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 05:50:27,759][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 05:50:27,760][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 05:50:27,760][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 05:50:27,760][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 05:50:27,761][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 05:50:27,761][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 05:50:27,761][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 05:50:27,761][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 05:50:27,761][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 05:50:27,761][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 05:50:27,762][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 05:50:27,762][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 05:50:27,762][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 05:50:29,075][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 05:50:29,079][starVLA.dataloader][INFO] - Created train VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 05:50:29,082][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 05:50:29,102][starVLA.dataloader][INFO] - Created eval VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 05:50:29,107][starVLA.training.train_starvla][INFO] - LR Group qwen_vl_interface: lr=1e-05, num_params=713
+[2026-07-28 05:50:29,107][starVLA.training.train_starvla][INFO] - LR Group action_model: lr=0.0001, num_params=16
+[2026-07-28 05:50:30,411][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch.
+[2026-07-28 05:50:30,420][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/config.full.yaml
+[2026-07-28 05:50:30,431][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/config.yaml
+[2026-07-28 05:50:30,432][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch.
+[2026-07-28 05:50:30,436][accelerate.accelerator][INFO] - Since you passed both train and evaluation dataloader, `is_train_batch_min` (here True will decide the `train_batch_size` (8).
+[2026-07-28 05:50:41,780][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-07-28 05:50:41,780][starVLA.training.train_starvla][INFO] -   Total optimization steps = 4000
+[2026-07-28 05:50:41,781][starVLA.training.train_starvla][INFO] -   Per device batch size = 8
+[2026-07-28 05:50:41,781][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-07-28 05:50:41,781][starVLA.training.train_starvla][INFO] -   Total batch size = 256
+[2026-07-28 05:52:00,660][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.7024374008178711, 'train/grad_norm_pre_clip': 33.67003631591797, 'timing/data': 0.000317954458296299, 'timing/model': 0.8336081691086292, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0})
+[2026-07-28 05:52:08,411][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.7015745639801025, 'train/grad_norm_pre_clip': 31.274425506591797, 'timing/data': 0.0003387844190001488, 'timing/model': 0.645293845795095, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0})
+[2026-07-28 05:52:15,752][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.701816201210022, 'train/grad_norm_pre_clip': 32.03770446777344, 'timing/data': 0.0003205621615052223, 'timing/model': 0.531942936591804, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0})
+[2026-07-28 05:52:23,060][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.7002396583557129, 'train/grad_norm_pre_clip': 33.905879974365234, 'timing/data': 0.0005586165934801102, 'timing/model': 0.5347318286076188, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0})
+[2026-07-28 05:52:30,374][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.6934170722961426, 'train/grad_norm_pre_clip': 32.17133331298828, 'timing/data': 0.00031821150332689285, 'timing/model': 0.5324147744104266, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0})
+[2026-07-28 05:52:37,673][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6720224618911743, 'train/grad_norm_pre_clip': 31.189193725585938, 'timing/data': 0.0003390554338693619, 'timing/model': 0.5400613453239202, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0})
+[2026-07-28 05:52:44,980][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6600353121757507, 'train/grad_norm_pre_clip': 34.54467010498047, 'timing/data': 0.0002984730526804924, 'timing/model': 0.5289625767618418, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0})
+[2026-07-28 05:52:52,296][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.5734919309616089, 'train/grad_norm_pre_clip': 29.376617431640625, 'timing/data': 0.00036383606493473053, 'timing/model': 0.5278053795918822, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0})
+[2026-07-28 05:52:59,617][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.5594220161437988, 'train/grad_norm_pre_clip': 25.30658531188965, 'timing/data': 0.00032441504299640656, 'timing/model': 0.5307797929272056, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0})
+[2026-07-28 05:53:06,929][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.452495276927948, 'train/grad_norm_pre_clip': 8.168952941894531, 'timing/data': 0.0003380170091986656, 'timing/model': 0.5349666243419051, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0})
+[2026-07-28 05:53:14,260][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.45226627588272095, 'train/grad_norm_pre_clip': 4.6503214836120605, 'timing/data': 0.00042280927300453186, 'timing/model': 0.5305145820602775, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0})
+[2026-07-28 05:53:21,606][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.4584832787513733, 'train/grad_norm_pre_clip': 0.9239698052406311, 'timing/data': 0.0003361338749527931, 'timing/model': 0.5306807039305568, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0})
+[2026-07-28 05:53:28,895][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.45377182960510254, 'train/grad_norm_pre_clip': 2.951399803161621, 'timing/data': 0.0003343923017382622, 'timing/model': 0.5338439960032701, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.0})
+[2026-07-28 05:53:36,164][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.4996251165866852, 'train/grad_norm_pre_clip': 6.223056793212891, 'timing/data': 0.0003779437392950058, 'timing/model': 0.5276427324861288, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.0})
+[2026-07-28 05:53:43,441][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.4491029977798462, 'train/grad_norm_pre_clip': 3.357671022415161, 'timing/data': 0.0004337728023529053, 'timing/model': 0.5300613343715668, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.0})
+[2026-07-28 05:53:50,756][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.4549543261528015, 'train/grad_norm_pre_clip': 2.6662840843200684, 'timing/data': 0.00032493099570274353, 'timing/model': 0.5326464185491204, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.0})
+[2026-07-28 05:53:58,017][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.43470990657806396, 'train/grad_norm_pre_clip': 1.9686044454574585, 'timing/data': 0.0004044407978653908, 'timing/model': 0.5307163586840034, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.0})
+[2026-07-28 05:54:05,275][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.5205451250076294, 'train/grad_norm_pre_clip': 6.296026706695557, 'timing/data': 0.0003743506968021393, 'timing/model': 0.5273581380024552, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.0})
+[2026-07-28 05:54:12,529][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.4234807789325714, 'train/grad_norm_pre_clip': 4.019733905792236, 'timing/data': 0.0003252122551202774, 'timing/model': 0.5294093182310462, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.0})
+[2026-07-28 05:54:19,800][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.5386861562728882, 'train/grad_norm_pre_clip': 12.483633995056152, 'timing/data': 0.00034087058156728745, 'timing/model': 0.5324580976739526, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0})
+[2026-07-28 05:54:27,250][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.40044134855270386, 'train/grad_norm_pre_clip': 3.8599164485931396, 'timing/data': 0.0004141293466091156, 'timing/model': 0.5633037202060223, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.0})
+[2026-07-28 05:54:34,645][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.45412689447402954, 'train/grad_norm_pre_clip': 9.067339897155762, 'timing/data': 0.00043095555156469345, 'timing/model': 0.5489839967340231, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.0})
+[2026-07-28 05:54:42,218][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.45713645219802856, 'train/grad_norm_pre_clip': 7.317044734954834, 'timing/data': 0.00034017208963632584, 'timing/model': 0.5489029809832573, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.0})
+[2026-07-28 05:54:49,555][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.3971293568611145, 'train/grad_norm_pre_clip': 5.081205368041992, 'timing/data': 0.0004272330552339554, 'timing/model': 0.5404059337452054, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.0})
+[2026-07-28 05:54:56,872][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.3721742033958435, 'train/grad_norm_pre_clip': 6.225526809692383, 'timing/data': 0.00034969672560691833, 'timing/model': 0.5370620600879192, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.0})
+[2026-07-28 05:55:04,152][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.34609657526016235, 'train/grad_norm_pre_clip': 8.534976959228516, 'timing/data': 0.00036977604031562805, 'timing/model': 0.5278726862743497, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.0})
+[2026-07-28 05:55:11,423][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.3472138047218323, 'train/grad_norm_pre_clip': 9.789068222045898, 'timing/data': 0.00040294602513313293, 'timing/model': 0.5418718596920371, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.0})
+[2026-07-28 05:55:18,681][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.3766789436340332, 'train/grad_norm_pre_clip': 15.946943283081055, 'timing/data': 0.0004161437973380089, 'timing/model': 0.5332411965355277, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01})
+[2026-07-28 05:55:25,909][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.28566962480545044, 'train/grad_norm_pre_clip': 6.09666109085083, 'timing/data': 0.0003415672108530998, 'timing/model': 0.5304734613746405, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01})
+[2026-07-28 05:55:33,929][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.34513550996780396, 'train/grad_norm_pre_clip': 22.230836868286133, 'timing/data': 0.00042413827031850815, 'timing/model': 0.5422706035897136, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01})
+[2026-07-28 05:55:41,157][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.34353190660476685, 'train/grad_norm_pre_clip': 14.810571670532227, 'timing/data': 0.0008559189736843109, 'timing/model': 0.5297310361638665, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01})
+[2026-07-28 05:55:48,388][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.36126232147216797, 'train/grad_norm_pre_clip': 25.55312728881836, 'timing/data': 0.00032913871109485626, 'timing/model': 0.5290772467851639, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01})
+[2026-07-28 05:55:55,630][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.392955482006073, 'train/grad_norm_pre_clip': 20.650814056396484, 'timing/data': 0.00043559540063142776, 'timing/model': 0.5310938414186239, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01})
+[2026-07-28 05:56:02,872][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.315117210149765, 'train/grad_norm_pre_clip': 23.05707359313965, 'timing/data': 0.00048027560114860535, 'timing/model': 0.5305130779743195, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01})
+[2026-07-28 05:56:10,110][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.3412806987762451, 'train/grad_norm_pre_clip': 7.037160396575928, 'timing/data': 0.00033659208565950394, 'timing/model': 0.530001881532371, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01})
+[2026-07-28 05:56:17,337][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.3735227882862091, 'train/grad_norm_pre_clip': 27.646343231201172, 'timing/data': 0.00033117085695266724, 'timing/model': 0.5313838962465525, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01})
+[2026-07-28 05:56:24,594][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.3374643325805664, 'train/grad_norm_pre_clip': 14.082565307617188, 'timing/data': 0.0006152288988232613, 'timing/model': 0.5289789047092199, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01})
+[2026-07-28 05:56:31,878][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.27306073904037476, 'train/grad_norm_pre_clip': 6.141484260559082, 'timing/data': 0.00037474557757377625, 'timing/model': 0.5287113664671779, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.01})
+[2026-07-28 05:56:39,152][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.3299521207809448, 'train/grad_norm_pre_clip': 14.28785228729248, 'timing/data': 0.0003654761239886284, 'timing/model': 0.5308855259791017, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.01})
+[2026-07-28 05:56:46,497][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.19982391595840454, 'train/grad_norm_pre_clip': 4.653558254241943, 'timing/data': 0.0003812601789832115, 'timing/model': 0.5620584562420845, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.01})
+[2026-07-28 05:56:53,790][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.2576076090335846, 'train/grad_norm_pre_clip': 10.258792877197266, 'timing/data': 0.00033186841756105423, 'timing/model': 0.5285145165398717, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.01})
+[2026-07-28 05:57:01,062][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.260918527841568, 'train/grad_norm_pre_clip': 10.33397102355957, 'timing/data': 0.0004457375034689903, 'timing/model': 0.5295219048857689, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.01})
+[2026-07-28 05:57:08,308][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.16413646936416626, 'train/grad_norm_pre_clip': 6.321646690368652, 'timing/data': 0.0003321561962366104, 'timing/model': 0.5296097500249743, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.01})
+[2026-07-28 05:57:15,559][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.16536979377269745, 'train/grad_norm_pre_clip': 8.660696029663086, 'timing/data': 0.0004154238849878311, 'timing/model': 0.5298750475049019, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.01})
+[2026-07-28 05:57:22,792][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.16813403367996216, 'train/grad_norm_pre_clip': 10.6522216796875, 'timing/data': 0.00038321129977703094, 'timing/model': 0.531306667253375, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.01})
+[2026-07-28 05:57:30,016][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.2032897025346756, 'train/grad_norm_pre_clip': 11.97035026550293, 'timing/data': 0.0003134729340672493, 'timing/model': 0.5278204334899783, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.01})
+[2026-07-28 05:57:37,238][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.19411250948905945, 'train/grad_norm_pre_clip': 20.360671997070312, 'timing/data': 0.0003346409648656845, 'timing/model': 0.5285686925053596, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.01})
+[2026-07-28 05:57:44,508][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.2243797779083252, 'train/grad_norm_pre_clip': 34.44025802612305, 'timing/data': 0.00038591399788856506, 'timing/model': 0.5331661114469171, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.01})
+[2026-07-28 05:57:51,826][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.17144986987113953, 'train/grad_norm_pre_clip': 11.005759239196777, 'timing/data': 0.0003921985626220703, 'timing/model': 0.5321990139782429, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.01})
+[2026-07-28 05:57:59,088][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.20619329810142517, 'train/grad_norm_pre_clip': 21.500656127929688, 'timing/data': 0.00043893419206142426, 'timing/model': 0.5284759569913149, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.01})
+[2026-07-28 05:58:06,365][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.2211373746395111, 'train/grad_norm_pre_clip': 16.59986114501953, 'timing/data': 0.00032640714198350906, 'timing/model': 0.5337627194821835, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.01})
+[2026-07-28 05:58:13,615][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.23520635068416595, 'train/grad_norm_pre_clip': 32.2686882019043, 'timing/data': 0.00034017395228147507, 'timing/model': 0.528851774521172, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.01})
+[2026-07-28 05:58:20,884][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.23611825704574585, 'train/grad_norm_pre_clip': 33.55447769165039, 'timing/data': 0.0003761323168873787, 'timing/model': 0.5281317038461566, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.01})
+[2026-07-28 05:58:28,121][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.19454747438430786, 'train/grad_norm_pre_clip': 17.19769859313965, 'timing/data': 0.0003812294453382492, 'timing/model': 0.5279563460499048, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.01})
+[2026-07-28 05:58:35,353][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.16669441759586334, 'train/grad_norm_pre_clip': 16.33798599243164, 'timing/data': 0.0003382870927453041, 'timing/model': 0.5285967662930489, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.01})
+[2026-07-28 05:58:42,611][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.20637938380241394, 'train/grad_norm_pre_clip': 21.898160934448242, 'timing/data': 0.0003300374373793602, 'timing/model': 0.5275574894621968, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.01})
+[2026-07-28 05:58:49,898][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.25953930616378784, 'train/grad_norm_pre_clip': 28.96543312072754, 'timing/data': 0.0003783665597438812, 'timing/model': 0.5260228700935841, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.01})
+[2026-07-28 05:58:57,159][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.2701089382171631, 'train/grad_norm_pre_clip': 20.38739585876465, 'timing/data': 0.0006847837939858437, 'timing/model': 0.5279702693223953, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.01})
+[2026-07-28 05:59:04,424][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.20804637670516968, 'train/grad_norm_pre_clip': 15.239349365234375, 'timing/data': 0.0004025781527161598, 'timing/model': 0.5296359704807401, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.01})
+[2026-07-28 05:59:11,660][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.24393726885318756, 'train/grad_norm_pre_clip': 24.958065032958984, 'timing/data': 0.00034232158213853836, 'timing/model': 0.5260472362861037, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.01})
+[2026-07-28 05:59:19,010][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.16969475150108337, 'train/grad_norm_pre_clip': 11.700553894042969, 'timing/data': 0.000400785356760025, 'timing/model': 0.5277437930926681, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.01})
+[2026-07-28 05:59:26,296][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.2286628633737564, 'train/grad_norm_pre_clip': 21.220685958862305, 'timing/data': 0.0004035048186779022, 'timing/model': 0.5280703734606504, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.01})
+[2026-07-28 05:59:33,613][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.1537810117006302, 'train/grad_norm_pre_clip': 11.040543556213379, 'timing/data': 0.00041218381375074387, 'timing/model': 0.529879747889936, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.01})
+[2026-07-28 05:59:40,858][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.22576531767845154, 'train/grad_norm_pre_clip': 24.24481773376465, 'timing/data': 0.00031076185405254364, 'timing/model': 0.5284704202786088, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.01})
+[2026-07-28 05:59:48,177][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.16184988617897034, 'train/grad_norm_pre_clip': 13.572067260742188, 'timing/data': 0.0003381427377462387, 'timing/model': 0.5284498939290643, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.01})
+[2026-07-28 05:59:55,444][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.18137413263320923, 'train/grad_norm_pre_clip': 15.467320442199707, 'timing/data': 0.00033784378319978714, 'timing/model': 0.5288012977689505, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.01})
+[2026-07-28 06:00:02,745][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.15259206295013428, 'train/grad_norm_pre_clip': 11.577957153320312, 'timing/data': 0.00030697695910930634, 'timing/model': 0.528897562995553, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.01})
+[2026-07-28 06:00:10,350][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.20464301109313965, 'train/grad_norm_pre_clip': 10.054410934448242, 'timing/data': 0.0003149528056383133, 'timing/model': 0.5280129490420222, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.01})
+[2026-07-28 06:00:17,872][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.1877986490726471, 'train/grad_norm_pre_clip': 7.745436668395996, 'timing/data': 0.00031718984246253967, 'timing/model': 0.526360179297626, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.01})
+[2026-07-28 06:00:25,145][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.16244208812713623, 'train/grad_norm_pre_clip': 5.607160568237305, 'timing/data': 0.00031282566487789154, 'timing/model': 0.529093774035573, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.01})
+[2026-07-28 06:00:32,443][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.1777474284172058, 'train/grad_norm_pre_clip': 7.872598648071289, 'timing/data': 0.0003710957244038582, 'timing/model': 0.5309805348515511, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.01})
+[2026-07-28 06:00:39,688][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.23860321938991547, 'train/grad_norm_pre_clip': 12.9111909866333, 'timing/data': 0.000321987085044384, 'timing/model': 0.5266072833910584, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.01})
+[2026-07-28 06:00:46,897][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.2454296052455902, 'train/grad_norm_pre_clip': 9.971592903137207, 'timing/data': 0.0003170361742377281, 'timing/model': 0.5270251017063856, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.01})
+[2026-07-28 06:00:54,130][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.1846042275428772, 'train/grad_norm_pre_clip': 4.78640604019165, 'timing/data': 0.0004063853994011879, 'timing/model': 0.5285357562825084, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.01})
+[2026-07-28 06:01:01,371][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.1305568814277649, 'train/grad_norm_pre_clip': 10.133856773376465, 'timing/data': 0.0003253193572163582, 'timing/model': 0.5279019707813859, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.01})
+[2026-07-28 06:01:08,616][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.20220963656902313, 'train/grad_norm_pre_clip': 18.045351028442383, 'timing/data': 0.00034166593104600906, 'timing/model': 0.528505302965641, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.01})
+[2026-07-28 06:01:15,840][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.22093604505062103, 'train/grad_norm_pre_clip': 17.544696807861328, 'timing/data': 0.0003051813691854477, 'timing/model': 0.5300053972750902, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.01})
+[2026-07-28 06:01:23,050][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.2026362121105194, 'train/grad_norm_pre_clip': 11.510557174682617, 'timing/data': 0.0003478238359093666, 'timing/model': 0.5270820241421461, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.01})
+[2026-07-28 06:01:30,402][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.19959717988967896, 'train/grad_norm_pre_clip': 10.739416122436523, 'timing/data': 0.000325152650475502, 'timing/model': 0.5278497450053692, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.01})
+[2026-07-28 06:01:37,649][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.1482941210269928, 'train/grad_norm_pre_clip': 7.036162376403809, 'timing/data': 0.000322192907333374, 'timing/model': 0.5268719056621194, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.01})
+[2026-07-28 06:01:44,867][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.18191875517368317, 'train/grad_norm_pre_clip': 12.245190620422363, 'timing/data': 0.0003319326788187027, 'timing/model': 0.5271220896393061, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.01})
+[2026-07-28 06:01:52,097][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.17000780999660492, 'train/grad_norm_pre_clip': 6.417335033416748, 'timing/data': 0.0003093201667070389, 'timing/model': 0.5278318310156465, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.02})
+[2026-07-28 06:01:59,320][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.2078617513179779, 'train/grad_norm_pre_clip': 10.096346855163574, 'timing/data': 0.00032893382012844086, 'timing/model': 0.5281479004770517, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.02})
+[2026-07-28 06:02:06,584][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.15356729924678802, 'train/grad_norm_pre_clip': 7.350108623504639, 'timing/data': 0.0003335680812597275, 'timing/model': 0.5300746960565448, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.02})
+[2026-07-28 06:02:13,820][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.2941374182701111, 'train/grad_norm_pre_clip': 25.743576049804688, 'timing/data': 0.0003959173336625099, 'timing/model': 0.5316407596692443, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.02})
+[2026-07-28 06:02:21,087][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.20507800579071045, 'train/grad_norm_pre_clip': 14.30822467803955, 'timing/data': 0.00039158575236797333, 'timing/model': 0.5378076024353504, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.02})
+[2026-07-28 06:02:28,319][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.18905037641525269, 'train/grad_norm_pre_clip': 9.001924514770508, 'timing/data': 0.0003802673891186714, 'timing/model': 0.5296671725809574, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.02})
+[2026-07-28 06:02:35,534][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.1712026447057724, 'train/grad_norm_pre_clip': 7.995051860809326, 'timing/data': 0.0003303084522485733, 'timing/model': 0.52811926510185, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.02})
+[2026-07-28 06:02:42,789][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.16453583538532257, 'train/grad_norm_pre_clip': 3.8328816890716553, 'timing/data': 0.000304175540804863, 'timing/model': 0.5276371752843261, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.02})
+[2026-07-28 06:02:50,003][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.2116396725177765, 'train/grad_norm_pre_clip': 10.819324493408203, 'timing/data': 0.00030271243304014206, 'timing/model': 0.5266015147790313, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.02})
+[2026-07-28 06:02:57,236][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.16174958646297455, 'train/grad_norm_pre_clip': 6.738084316253662, 'timing/data': 0.0003308253362774849, 'timing/model': 0.5286465613171458, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.02})
+[2026-07-28 06:03:04,479][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.2232222855091095, 'train/grad_norm_pre_clip': 13.986787796020508, 'timing/data': 0.0003322949633002281, 'timing/model': 0.5290192812681198, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.02})
+[2026-07-28 06:03:11,717][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.19306586682796478, 'train/grad_norm_pre_clip': 8.262458801269531, 'timing/data': 0.00037497375160455704, 'timing/model': 0.5310038728639483, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.02})
+[2026-07-28 06:03:19,050][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.15577691793441772, 'train/grad_norm_pre_clip': 11.994475364685059, 'timing/data': 0.0003460003063082695, 'timing/model': 0.5302670327946544, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.02})
+[2026-07-28 06:03:26,310][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.2037504017353058, 'train/grad_norm_pre_clip': 9.933059692382812, 'timing/data': 0.00032011885195970535, 'timing/model': 0.5272972946986556, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.02})
+[2026-07-28 06:03:33,569][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.15942257642745972, 'train/grad_norm_pre_clip': 5.312713623046875, 'timing/data': 0.0003744969144463539, 'timing/model': 0.5281398482620716, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.02})
+[2026-07-28 06:03:40,895][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.17092055082321167, 'train/grad_norm_pre_clip': 4.043694496154785, 'timing/data': 0.0003444282338023186, 'timing/model': 0.529584962874651, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.02})
+[2026-07-28 06:03:48,142][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.23939263820648193, 'train/grad_norm_pre_clip': 18.359146118164062, 'timing/data': 0.00032227393239736557, 'timing/model': 0.5274921534582973, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.02})
+[2026-07-28 06:03:55,439][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.188429057598114, 'train/grad_norm_pre_clip': 7.642422199249268, 'timing/data': 0.0003155479207634926, 'timing/model': 0.5295940861105919, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.02})
+[2026-07-28 06:04:02,708][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.11098355799913406, 'train/grad_norm_pre_clip': 2.9162557125091553, 'timing/data': 0.00032521970570087433, 'timing/model': 0.5298359142616391, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.02})
+[2026-07-28 06:04:10,053][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.13632217049598694, 'train/grad_norm_pre_clip': 5.929408073425293, 'timing/data': 0.0003316560760140419, 'timing/model': 0.5308993514627218, 'learning_rate/qwen_vl_interface': 9.999998458888285e-06, 'learning_rate/action_model': 9.999998458888285e-05, 'epoch': 0.02})
+[2026-07-28 06:04:17,327][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.20234352350234985, 'train/grad_norm_pre_clip': 9.119263648986816, 'timing/data': 0.0003248807042837143, 'timing/model': 0.5282922480255365, 'learning_rate/qwen_vl_interface': 9.999993835554141e-06, 'learning_rate/action_model': 9.999993835554141e-05, 'epoch': 0.02})
+[2026-07-28 06:04:24,583][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.20980113744735718, 'train/grad_norm_pre_clip': 12.897795677185059, 'timing/data': 0.00033473502844572067, 'timing/model': 0.5280585074797273, 'learning_rate/qwen_vl_interface': 9.999986130000566e-06, 'learning_rate/action_model': 9.999986130000566e-05, 'epoch': 0.02})
+[2026-07-28 06:04:31,850][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.12315820157527924, 'train/grad_norm_pre_clip': 5.087514400482178, 'timing/data': 0.00038054492324590683, 'timing/model': 0.5326638324186206, 'learning_rate/qwen_vl_interface': 9.99997534223256e-06, 'learning_rate/action_model': 9.99997534223256e-05, 'epoch': 0.02})
+[2026-07-28 06:04:39,376][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.1459025740623474, 'train/grad_norm_pre_clip': 6.613966941833496, 'timing/data': 0.000327317975461483, 'timing/model': 0.5266889818012714, 'learning_rate/qwen_vl_interface': 9.999961472257127e-06, 'learning_rate/action_model': 9.999961472257127e-05, 'epoch': 0.02})
+[2026-07-28 06:04:46,589][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.1792462170124054, 'train/grad_norm_pre_clip': 5.092395305633545, 'timing/data': 0.00036945659667253494, 'timing/model': 0.5286022834479809, 'learning_rate/qwen_vl_interface': 9.999944520083262e-06, 'learning_rate/action_model': 9.999944520083261e-05, 'epoch': 0.02})
+[2026-07-28 06:04:53,893][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.13508038222789764, 'train/grad_norm_pre_clip': 5.689759731292725, 'timing/data': 0.00031103286892175674, 'timing/model': 0.5277079921215773, 'learning_rate/qwen_vl_interface': 9.999924485721968e-06, 'learning_rate/action_model': 9.999924485721968e-05, 'epoch': 0.02})
+[2026-07-28 06:05:01,302][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.16522806882858276, 'train/grad_norm_pre_clip': 5.486186504364014, 'timing/data': 0.0003273477777838707, 'timing/model': 0.5278655784204602, 'learning_rate/qwen_vl_interface': 9.999901369186244e-06, 'learning_rate/action_model': 9.999901369186244e-05, 'epoch': 0.02})
+[2026-07-28 06:05:08,583][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.21184566617012024, 'train/grad_norm_pre_clip': 6.6547465324401855, 'timing/data': 0.0003092195838689804, 'timing/model': 0.5261358162388206, 'learning_rate/qwen_vl_interface': 9.999875170491092e-06, 'learning_rate/action_model': 9.999875170491091e-05, 'epoch': 0.02})
+[2026-07-28 06:05:16,133][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.17604222893714905, 'train/grad_norm_pre_clip': 4.514275550842285, 'timing/data': 0.000309760682284832, 'timing/model': 0.5320217655971646, 'learning_rate/qwen_vl_interface': 9.999845889653509e-06, 'learning_rate/action_model': 9.999845889653508e-05, 'epoch': 0.02})
+[2026-07-28 06:05:23,394][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.1362641453742981, 'train/grad_norm_pre_clip': 4.521178722381592, 'timing/data': 0.0003528594970703125, 'timing/model': 0.5365365659818053, 'learning_rate/qwen_vl_interface': 9.999813526692496e-06, 'learning_rate/action_model': 9.999813526692495e-05, 'epoch': 0.02})
+[2026-07-28 06:05:30,627][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.18687818944454193, 'train/grad_norm_pre_clip': 7.215977191925049, 'timing/data': 0.0003663981333374977, 'timing/model': 0.5339179523289204, 'learning_rate/qwen_vl_interface': 9.999778081629055e-06, 'learning_rate/action_model': 9.999778081629054e-05, 'epoch': 0.02})
+[2026-07-28 06:05:37,928][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.13263589143753052, 'train/grad_norm_pre_clip': 4.991947650909424, 'timing/data': 0.0003748331218957901, 'timing/model': 0.5434991773217916, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.02})
+[2026-07-28 06:05:45,417][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.22486382722854614, 'train/grad_norm_pre_clip': 12.229019165039062, 'timing/data': 0.00031493697315454483, 'timing/model': 0.528710138052702, 'learning_rate/qwen_vl_interface': 9.999697945288881e-06, 'learning_rate/action_model': 9.999697945288881e-05, 'epoch': 0.02})
+[2026-07-28 06:05:52,632][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.2231580913066864, 'train/grad_norm_pre_clip': 14.726171493530273, 'timing/data': 0.0003299480304121971, 'timing/model': 0.5283460104838014, 'learning_rate/qwen_vl_interface': 9.999653254064153e-06, 'learning_rate/action_model': 9.999653254064152e-05, 'epoch': 0.02})
+[2026-07-28 06:05:59,835][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.2550440728664398, 'train/grad_norm_pre_clip': 10.96418571472168, 'timing/data': 0.0003858190029859543, 'timing/model': 0.5278234481811523, 'learning_rate/qwen_vl_interface': 9.99960548084099e-06, 'learning_rate/action_model': 9.99960548084099e-05, 'epoch': 0.02})
+[2026-07-28 06:06:07,049][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.24249231815338135, 'train/grad_norm_pre_clip': 17.872982025146484, 'timing/data': 0.00033533666282892227, 'timing/model': 0.5307837212458253, 'learning_rate/qwen_vl_interface': 9.999554625650398e-06, 'learning_rate/action_model': 9.999554625650399e-05, 'epoch': 0.02})
+[2026-07-28 06:06:14,284][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.13138744235038757, 'train/grad_norm_pre_clip': 5.946702003479004, 'timing/data': 0.0003702649846673012, 'timing/model': 0.5282464483752847, 'learning_rate/qwen_vl_interface': 9.999500688525376e-06, 'learning_rate/action_model': 9.999500688525375e-05, 'epoch': 0.02})
+[2026-07-28 06:06:21,513][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.17697249352931976, 'train/grad_norm_pre_clip': 8.346096992492676, 'timing/data': 0.00034664105623960495, 'timing/model': 0.5295926136896014, 'learning_rate/qwen_vl_interface': 9.99944366950092e-06, 'learning_rate/action_model': 9.999443669500919e-05, 'epoch': 0.02})
+[2026-07-28 06:06:28,726][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.13971611857414246, 'train/grad_norm_pre_clip': 3.538818836212158, 'timing/data': 0.00032395869493484497, 'timing/model': 0.5308911334723234, 'learning_rate/qwen_vl_interface': 9.99938356861403e-06, 'learning_rate/action_model': 9.99938356861403e-05, 'epoch': 0.02})
+[2026-07-28 06:06:35,942][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.18113023042678833, 'train/grad_norm_pre_clip': 5.260613918304443, 'timing/data': 0.0003354167565703392, 'timing/model': 0.5301416264846921, 'learning_rate/qwen_vl_interface': 9.999320385903709e-06, 'learning_rate/action_model': 9.999320385903708e-05, 'epoch': 0.02})
+[2026-07-28 06:06:43,202][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.14663061499595642, 'train/grad_norm_pre_clip': 5.9588942527771, 'timing/data': 0.00033683981746435165, 'timing/model': 0.526775368489325, 'learning_rate/qwen_vl_interface': 9.999254121410948e-06, 'learning_rate/action_model': 9.999254121410949e-05, 'epoch': 0.02})
+[2026-07-28 06:06:50,485][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.16964766383171082, 'train/grad_norm_pre_clip': 3.8426589965820312, 'timing/data': 0.0003442838788032532, 'timing/model': 0.5284167435020208, 'learning_rate/qwen_vl_interface': 9.999184775178751e-06, 'learning_rate/action_model': 9.999184775178751e-05, 'epoch': 0.02})
+[2026-07-28 06:06:57,704][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.14922207593917847, 'train/grad_norm_pre_clip': 8.442703247070312, 'timing/data': 0.00032946188002824783, 'timing/model': 0.527799746952951, 'learning_rate/qwen_vl_interface': 9.999112347252117e-06, 'learning_rate/action_model': 9.999112347252116e-05, 'epoch': 0.02})
+[2026-07-28 06:07:04,906][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.14634130895137787, 'train/grad_norm_pre_clip': 4.526506423950195, 'timing/data': 0.00038969796150922775, 'timing/model': 0.527720469981432, 'learning_rate/qwen_vl_interface': 9.99903683767804e-06, 'learning_rate/action_model': 9.99903683767804e-05, 'epoch': 0.02})
+[2026-07-28 06:07:12,101][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.14461475610733032, 'train/grad_norm_pre_clip': 3.2483267784118652, 'timing/data': 0.0003142794594168663, 'timing/model': 0.5276281721889973, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.02})
+[2026-07-28 06:07:19,308][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.14153437316417694, 'train/grad_norm_pre_clip': 5.782490253448486, 'timing/data': 0.00034546665847301483, 'timing/model': 0.5271940035745502, 'learning_rate/qwen_vl_interface': 9.998876573785547e-06, 'learning_rate/action_model': 9.998876573785547e-05, 'epoch': 0.02})
+[2026-07-28 06:07:26,518][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.15871703624725342, 'train/grad_norm_pre_clip': 4.667313575744629, 'timing/data': 0.00036635622382164, 'timing/model': 0.5274158734828234, 'learning_rate/qwen_vl_interface': 9.998791819571126e-06, 'learning_rate/action_model': 9.998791819571127e-05, 'epoch': 0.02})
+[2026-07-28 06:07:33,785][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.1335936188697815, 'train/grad_norm_pre_clip': 3.6396591663360596, 'timing/data': 0.00041187088936567307, 'timing/model': 0.5301539162173867, 'learning_rate/qwen_vl_interface': 9.998703983917252e-06, 'learning_rate/action_model': 9.998703983917252e-05, 'epoch': 0.02})
+[2026-07-28 06:07:41,021][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.13919764757156372, 'train/grad_norm_pre_clip': 2.7873241901397705, 'timing/data': 0.00042419787496328354, 'timing/model': 0.5275179613381624, 'learning_rate/qwen_vl_interface': 9.998613066880919e-06, 'learning_rate/action_model': 9.998613066880918e-05, 'epoch': 0.02})
+[2026-07-28 06:07:48,268][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.11681362241506577, 'train/grad_norm_pre_clip': 4.102765083312988, 'timing/data': 0.0003410950303077698, 'timing/model': 0.5268466779962182, 'learning_rate/qwen_vl_interface': 9.99851906852112e-06, 'learning_rate/action_model': 9.998519068521119e-05, 'epoch': 0.02})
+[2026-07-28 06:07:55,506][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.1558775007724762, 'train/grad_norm_pre_clip': 3.341837167739868, 'timing/data': 0.0003634924069046974, 'timing/model': 0.5353432828560472, 'learning_rate/qwen_vl_interface': 9.998421988898851e-06, 'learning_rate/action_model': 9.998421988898851e-05, 'epoch': 0.02})
+[2026-07-28 06:08:02,748][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.10148751735687256, 'train/grad_norm_pre_clip': 2.97904109954834, 'timing/data': 0.00033918116241693497, 'timing/model': 0.5303549971431494, 'learning_rate/qwen_vl_interface': 9.998321828077107e-06, 'learning_rate/action_model': 9.998321828077106e-05, 'epoch': 0.02})
+[2026-07-28 06:08:09,975][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.0898047387599945, 'train/grad_norm_pre_clip': 3.0631279945373535, 'timing/data': 0.000319531187415123, 'timing/model': 0.5300365891307592, 'learning_rate/qwen_vl_interface': 9.99821858612088e-06, 'learning_rate/action_model': 9.99821858612088e-05, 'epoch': 0.02})
+[2026-07-28 06:08:17,217][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.1854950189590454, 'train/grad_norm_pre_clip': 16.287948608398438, 'timing/data': 0.0003228476271033287, 'timing/model': 0.5288814157247543, 'learning_rate/qwen_vl_interface': 9.998112263097164e-06, 'learning_rate/action_model': 9.998112263097165e-05, 'epoch': 0.02})
+[2026-07-28 06:08:24,456][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.15147382020950317, 'train/grad_norm_pre_clip': 7.392197132110596, 'timing/data': 0.00033208727836608887, 'timing/model': 0.5299431309103966, 'learning_rate/qwen_vl_interface': 9.99800285907495e-06, 'learning_rate/action_model': 9.99800285907495e-05, 'epoch': 0.02})
+[2026-07-28 06:08:31,677][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.1550397276878357, 'train/grad_norm_pre_clip': 7.2956156730651855, 'timing/data': 0.00032737385481595993, 'timing/model': 0.5303119635209441, 'learning_rate/qwen_vl_interface': 9.997890374125228e-06, 'learning_rate/action_model': 9.997890374125227e-05, 'epoch': 0.03})
+[2026-07-28 06:08:38,906][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.19465357065200806, 'train/grad_norm_pre_clip': 7.820765495300293, 'timing/data': 0.0003293994814157486, 'timing/model': 0.5292357690632343, 'learning_rate/qwen_vl_interface': 9.997774808320992e-06, 'learning_rate/action_model': 9.997774808320992e-05, 'epoch': 0.03})
+[2026-07-28 06:08:46,152][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.12507672607898712, 'train/grad_norm_pre_clip': 3.5808396339416504, 'timing/data': 0.0007910355925559998, 'timing/model': 0.5310270516201854, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.03})
+[2026-07-28 06:08:53,388][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.12546157836914062, 'train/grad_norm_pre_clip': 7.001732349395752, 'timing/data': 0.0003406573086977005, 'timing/model': 0.5308411503210664, 'learning_rate/qwen_vl_interface': 9.997534434450922e-06, 'learning_rate/action_model': 9.997534434450921e-05, 'epoch': 0.03})
+[2026-07-28 06:09:00,627][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.15981166064739227, 'train/grad_norm_pre_clip': 4.624497890472412, 'timing/data': 0.0003346716985106468, 'timing/model': 0.5308432737365365, 'learning_rate/qwen_vl_interface': 9.997409626541065e-06, 'learning_rate/action_model': 9.997409626541066e-05, 'epoch': 0.03})
+[2026-07-28 06:09:07,850][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.15100863575935364, 'train/grad_norm_pre_clip': 3.884943962097168, 'timing/data': 0.000318366102874279, 'timing/model': 0.5278699528425932, 'learning_rate/qwen_vl_interface': 9.997281738088647e-06, 'learning_rate/action_model': 9.997281738088646e-05, 'epoch': 0.03})
+[2026-07-28 06:09:15,362][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.1185467392206192, 'train/grad_norm_pre_clip': 5.625893592834473, 'timing/data': 0.00030443258583545685, 'timing/model': 0.8180910916998982, 'learning_rate/qwen_vl_interface': 9.997150769176648e-06, 'learning_rate/action_model': 9.997150769176647e-05, 'epoch': 0.03})
+[2026-07-28 06:09:22,636][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.13359197974205017, 'train/grad_norm_pre_clip': 4.636960983276367, 'timing/data': 0.00039832666516304016, 'timing/model': 0.5349621986970305, 'learning_rate/qwen_vl_interface': 9.997016719890052e-06, 'learning_rate/action_model': 9.997016719890052e-05, 'epoch': 0.03})
+[2026-07-28 06:09:29,903][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.14438748359680176, 'train/grad_norm_pre_clip': 5.439245700836182, 'timing/data': 0.000329592265188694, 'timing/model': 0.5271635800600052, 'learning_rate/qwen_vl_interface': 9.996879590315847e-06, 'learning_rate/action_model': 9.996879590315846e-05, 'epoch': 0.03})
+[2026-07-28 06:09:37,121][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.14265479147434235, 'train/grad_norm_pre_clip': 4.755518913269043, 'timing/data': 0.0003965906798839569, 'timing/model': 0.5293146818876266, 'learning_rate/qwen_vl_interface': 9.99673938054301e-06, 'learning_rate/action_model': 9.996739380543009e-05, 'epoch': 0.03})
+[2026-07-28 06:09:44,428][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.18577346205711365, 'train/grad_norm_pre_clip': 9.584854125976562, 'timing/data': 0.0004408247768878937, 'timing/model': 0.529100339859724, 'learning_rate/qwen_vl_interface': 9.996596090662525e-06, 'learning_rate/action_model': 9.996596090662525e-05, 'epoch': 0.03})
+[2026-07-28 06:09:51,738][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.20499274134635925, 'train/grad_norm_pre_clip': 13.720232963562012, 'timing/data': 0.00032557640224695206, 'timing/model': 0.5335827665403485, 'learning_rate/qwen_vl_interface': 9.996449720767369e-06, 'learning_rate/action_model': 9.996449720767368e-05, 'epoch': 0.03})
+[2026-07-28 06:09:59,030][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.14021068811416626, 'train/grad_norm_pre_clip': 7.182021617889404, 'timing/data': 0.0003360006958246231, 'timing/model': 0.5305525865405798, 'learning_rate/qwen_vl_interface': 9.99630027095252e-06, 'learning_rate/action_model': 9.99630027095252e-05, 'epoch': 0.03})
+[2026-07-28 06:10:06,552][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.13817748427391052, 'train/grad_norm_pre_clip': 6.5296854972839355, 'timing/data': 0.0003135884180665016, 'timing/model': 0.5272405240684748, 'learning_rate/qwen_vl_interface': 9.996147741314956e-06, 'learning_rate/action_model': 9.996147741314957e-05, 'epoch': 0.03})
+[2026-07-28 06:10:13,774][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.13337945938110352, 'train/grad_norm_pre_clip': 4.261722564697266, 'timing/data': 0.00031949859112501144, 'timing/model': 0.5272345431149006, 'learning_rate/qwen_vl_interface': 9.99599213195365e-06, 'learning_rate/action_model': 9.99599213195365e-05, 'epoch': 0.03})
+[2026-07-28 06:10:20,981][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.23292572796344757, 'train/grad_norm_pre_clip': 18.70435905456543, 'timing/data': 0.0003697909414768219, 'timing/model': 0.5285492101684213, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.03})
+[2026-07-28 06:10:28,201][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.15585952997207642, 'train/grad_norm_pre_clip': 5.469541072845459, 'timing/data': 0.0003107041120529175, 'timing/model': 0.5249716844409704, 'learning_rate/qwen_vl_interface': 9.995671674465709e-06, 'learning_rate/action_model': 9.995671674465709e-05, 'epoch': 0.03})
+[2026-07-28 06:10:35,409][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.15683293342590332, 'train/grad_norm_pre_clip': 3.379070997238159, 'timing/data': 0.0003373194485902786, 'timing/model': 0.5287332003936172, 'learning_rate/qwen_vl_interface': 9.99550682654701e-06, 'learning_rate/action_model': 9.995506826547011e-05, 'epoch': 0.03})
+[2026-07-28 06:10:42,614][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.19073109328746796, 'train/grad_norm_pre_clip': 5.896751880645752, 'timing/data': 0.0003137495368719101, 'timing/model': 0.5270266868174076, 'learning_rate/qwen_vl_interface': 9.995338899320457e-06, 'learning_rate/action_model': 9.995338899320456e-05, 'epoch': 0.03})
+[2026-07-28 06:10:49,838][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.12327257543802261, 'train/grad_norm_pre_clip': 3.0056543350219727, 'timing/data': 0.00038807839155197144, 'timing/model': 0.5283610811457038, 'learning_rate/qwen_vl_interface': 9.995167892895011e-06, 'learning_rate/action_model': 9.99516789289501e-05, 'epoch': 0.03})
+[2026-07-28 06:10:57,069][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.16172392666339874, 'train/grad_norm_pre_clip': 8.44103717803955, 'timing/data': 0.0003243768587708473, 'timing/model': 0.532128949649632, 'learning_rate/qwen_vl_interface': 9.994993807381635e-06, 'learning_rate/action_model': 9.994993807381635e-05, 'epoch': 0.03})
+[2026-07-28 06:11:04,292][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.2043972909450531, 'train/grad_norm_pre_clip': 12.721603393554688, 'timing/data': 0.0003428291529417038, 'timing/model': 0.5277803726494312, 'learning_rate/qwen_vl_interface': 9.994816642893294e-06, 'learning_rate/action_model': 9.994816642893293e-05, 'epoch': 0.03})
+[2026-07-28 06:11:11,521][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.10601113736629486, 'train/grad_norm_pre_clip': 3.7319185733795166, 'timing/data': 0.0003220345824956894, 'timing/model': 0.5354724060744047, 'learning_rate/qwen_vl_interface': 9.994636399544948e-06, 'learning_rate/action_model': 9.994636399544947e-05, 'epoch': 0.03})
+[2026-07-28 06:11:18,771][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.13853758573532104, 'train/grad_norm_pre_clip': 2.923736810684204, 'timing/data': 0.00032145529985427856, 'timing/model': 0.5299096815288067, 'learning_rate/qwen_vl_interface': 9.994453077453552e-06, 'learning_rate/action_model': 9.994453077453552e-05, 'epoch': 0.03})
+[2026-07-28 06:11:26,018][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.11862340569496155, 'train/grad_norm_pre_clip': 7.948518753051758, 'timing/data': 0.00038586556911468506, 'timing/model': 0.529513768851757, 'learning_rate/qwen_vl_interface': 9.994266676738065e-06, 'learning_rate/action_model': 9.994266676738065e-05, 'epoch': 0.03})
+[2026-07-28 06:11:33,247][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.1596313863992691, 'train/grad_norm_pre_clip': 3.64727520942688, 'timing/data': 0.0003353012725710869, 'timing/model': 0.5302637601271272, 'learning_rate/qwen_vl_interface': 9.99407719751944e-06, 'learning_rate/action_model': 9.994077197519439e-05, 'epoch': 0.03})
+[2026-07-28 06:11:40,491][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.12332287430763245, 'train/grad_norm_pre_clip': 2.619598865509033, 'timing/data': 0.00038647837936878204, 'timing/model': 0.5365725941956043, 'learning_rate/qwen_vl_interface': 9.993884639920626e-06, 'learning_rate/action_model': 9.993884639920626e-05, 'epoch': 0.03})
+[2026-07-28 06:11:47,732][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.09509450197219849, 'train/grad_norm_pre_clip': 2.990053653717041, 'timing/data': 0.000375429168343544, 'timing/model': 0.5313675962388515, 'learning_rate/qwen_vl_interface': 9.993689004066573e-06, 'learning_rate/action_model': 9.993689004066573e-05, 'epoch': 0.03})
+[2026-07-28 06:11:54,994][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.11343428492546082, 'train/grad_norm_pre_clip': 2.260322093963623, 'timing/data': 0.0003236718475818634, 'timing/model': 0.5377821018919349, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.03})
+[2026-07-28 06:12:02,340][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.11047220975160599, 'train/grad_norm_pre_clip': 1.992570161819458, 'timing/data': 0.000564970076084137, 'timing/model': 0.5322091514244676, 'learning_rate/qwen_vl_interface': 9.993288498102531e-06, 'learning_rate/action_model': 9.993288498102532e-05, 'epoch': 0.03})
+[2026-07-28 06:12:09,561][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.11670458316802979, 'train/grad_norm_pre_clip': 2.905242443084717, 'timing/data': 0.00030931178480386734, 'timing/model': 0.5276976097375154, 'learning_rate/qwen_vl_interface': 9.993083628252424e-06, 'learning_rate/action_model': 9.993083628252424e-05, 'epoch': 0.03})
+[2026-07-28 06:12:16,783][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.10935313999652863, 'train/grad_norm_pre_clip': 3.376394271850586, 'timing/data': 0.0003379387781023979, 'timing/model': 0.5293327076360583, 'learning_rate/qwen_vl_interface': 9.992875680666849e-06, 'learning_rate/action_model': 9.992875680666849e-05, 'epoch': 0.03})
+[2026-07-28 06:12:24,001][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.17106975615024567, 'train/grad_norm_pre_clip': 7.867595672607422, 'timing/data': 0.0003371797502040863, 'timing/model': 0.5282859373837709, 'learning_rate/qwen_vl_interface': 9.992664655480736e-06, 'learning_rate/action_model': 9.992664655480735e-05, 'epoch': 0.03})
+[2026-07-28 06:12:31,242][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.28355929255485535, 'train/grad_norm_pre_clip': 15.815386772155762, 'timing/data': 0.0003182925283908844, 'timing/model': 0.5296223629266024, 'learning_rate/qwen_vl_interface': 9.992450552831018e-06, 'learning_rate/action_model': 9.992450552831017e-05, 'epoch': 0.03})
+[2026-07-28 06:12:38,497][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.2639783024787903, 'train/grad_norm_pre_clip': 16.785005569458008, 'timing/data': 0.00032909680157899857, 'timing/model': 0.5284697413444519, 'learning_rate/qwen_vl_interface': 9.992233372856623e-06, 'learning_rate/action_model': 9.992233372856624e-05, 'epoch': 0.03})
+[2026-07-28 06:12:45,800][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.19742587208747864, 'train/grad_norm_pre_clip': 10.550714492797852, 'timing/data': 0.0004586493596434593, 'timing/model': 0.5397589402273297, 'learning_rate/qwen_vl_interface': 9.99201311569848e-06, 'learning_rate/action_model': 9.99201311569848e-05, 'epoch': 0.03})
+[2026-07-28 06:12:53,062][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.15448105335235596, 'train/grad_norm_pre_clip': 8.410223960876465, 'timing/data': 0.00035115238279104233, 'timing/model': 0.5381721695885062, 'learning_rate/qwen_vl_interface': 9.99178978149951e-06, 'learning_rate/action_model': 9.99178978149951e-05, 'epoch': 0.03})
+[2026-07-28 06:13:00,316][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.08570289611816406, 'train/grad_norm_pre_clip': 3.711392402648926, 'timing/data': 0.0004405910149216652, 'timing/model': 0.5381392119452357, 'learning_rate/qwen_vl_interface': 9.99156337040463e-06, 'learning_rate/action_model': 9.991563370404629e-05, 'epoch': 0.03})
+[2026-07-28 06:13:07,562][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.1266532987356186, 'train/grad_norm_pre_clip': 3.4490513801574707, 'timing/data': 0.00032497476786375046, 'timing/model': 0.5288946311920881, 'learning_rate/qwen_vl_interface': 9.991333882560759e-06, 'learning_rate/action_model': 9.991333882560759e-05, 'epoch': 0.03})
+[2026-07-28 06:13:14,803][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.22651848196983337, 'train/grad_norm_pre_clip': 9.981379508972168, 'timing/data': 0.0003365064039826393, 'timing/model': 0.5379310511052608, 'learning_rate/qwen_vl_interface': 9.991101318116806e-06, 'learning_rate/action_model': 9.991101318116806e-05, 'epoch': 0.03})
+[2026-07-28 06:13:22,108][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.20727242529392242, 'train/grad_norm_pre_clip': 4.374927997589111, 'timing/data': 0.00031695421785116196, 'timing/model': 0.5320084756240249, 'learning_rate/qwen_vl_interface': 9.990865677223684e-06, 'learning_rate/action_model': 9.990865677223684e-05, 'epoch': 0.03})
+[2026-07-28 06:13:29,330][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.1660493016242981, 'train/grad_norm_pre_clip': 6.455294132232666, 'timing/data': 0.00032034236937761307, 'timing/model': 0.5266233989968896, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.03})
+[2026-07-28 06:13:36,558][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.12930597364902496, 'train/grad_norm_pre_clip': 3.152656316757202, 'timing/data': 0.0003481311723589897, 'timing/model': 0.5321280192583799, 'learning_rate/qwen_vl_interface': 9.990385166703534e-06, 'learning_rate/action_model': 9.990385166703533e-05, 'epoch': 0.03})
+[2026-07-28 06:13:43,787][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.158851757645607, 'train/grad_norm_pre_clip': 7.934603214263916, 'timing/data': 0.0003716368228197098, 'timing/model': 0.5313716093078256, 'learning_rate/qwen_vl_interface': 9.990140297388306e-06, 'learning_rate/action_model': 9.990140297388307e-05, 'epoch': 0.03})
+[2026-07-28 06:13:51,037][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.13044674694538116, 'train/grad_norm_pre_clip': 4.051201820373535, 'timing/data': 0.000500931404531002, 'timing/model': 0.5294111501425505, 'learning_rate/qwen_vl_interface': 9.989892352247505e-06, 'learning_rate/action_model': 9.989892352247504e-05, 'epoch': 0.03})
+[2026-07-28 06:13:58,550][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.137849360704422, 'train/grad_norm_pre_clip': 6.159013748168945, 'timing/data': 0.0004305141046643257, 'timing/model': 0.530967821367085, 'learning_rate/qwen_vl_interface': 9.989641331442014e-06, 'learning_rate/action_model': 9.989641331442013e-05, 'epoch': 0.03})
+[2026-07-28 06:14:05,832][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.18421289324760437, 'train/grad_norm_pre_clip': 7.916823863983154, 'timing/data': 0.0003207605332136154, 'timing/model': 0.5276332870125771, 'learning_rate/qwen_vl_interface': 9.989387235134722e-06, 'learning_rate/action_model': 9.989387235134722e-05, 'epoch': 0.03})
+[2026-07-28 06:14:13,077][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.0947112962603569, 'train/grad_norm_pre_clip': 7.881208896636963, 'timing/data': 0.0003273468464612961, 'timing/model': 0.5287829572334886, 'learning_rate/qwen_vl_interface': 9.98913006349051e-06, 'learning_rate/action_model': 9.989130063490508e-05, 'epoch': 0.03})
+[2026-07-28 06:14:20,290][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.25729799270629883, 'train/grad_norm_pre_clip': 5.167147636413574, 'timing/data': 0.00032744649797677994, 'timing/model': 0.5269268210977316, 'learning_rate/qwen_vl_interface': 9.988869816676247e-06, 'learning_rate/action_model': 9.988869816676246e-05, 'epoch': 0.03})
+[2026-07-28 06:14:27,510][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.10216207802295685, 'train/grad_norm_pre_clip': 3.916288375854492, 'timing/data': 0.00032686255872249603, 'timing/model': 0.5264883674681187, 'learning_rate/qwen_vl_interface': 9.988606494860811e-06, 'learning_rate/action_model': 9.988606494860812e-05, 'epoch': 0.03})
+[2026-07-28 06:14:35,031][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.17553356289863586, 'train/grad_norm_pre_clip': 4.555591106414795, 'timing/data': 0.00033696088939905167, 'timing/model': 0.527384877204895, 'learning_rate/qwen_vl_interface': 9.98834009821507e-06, 'learning_rate/action_model': 9.988340098215069e-05, 'epoch': 0.03})
+[2026-07-28 06:14:42,286][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.1311572790145874, 'train/grad_norm_pre_clip': 2.7423384189605713, 'timing/data': 0.00035298336297273636, 'timing/model': 0.5313519351184368, 'learning_rate/qwen_vl_interface': 9.98807062691188e-06, 'learning_rate/action_model': 9.988070626911879e-05, 'epoch': 0.03})
+[2026-07-28 06:14:49,870][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.1341029703617096, 'train/grad_norm_pre_clip': 3.7190871238708496, 'timing/data': 0.00048195384442806244, 'timing/model': 0.5314857270568609, 'learning_rate/qwen_vl_interface': 9.987798081126099e-06, 'learning_rate/action_model': 9.987798081126099e-05, 'epoch': 0.03})
+[2026-07-28 06:14:57,146][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.1479499787092209, 'train/grad_norm_pre_clip': 3.7550277709960938, 'timing/data': 0.0004001446068286896, 'timing/model': 0.5356140378862619, 'learning_rate/qwen_vl_interface': 9.987522461034584e-06, 'learning_rate/action_model': 9.987522461034584e-05, 'epoch': 0.03})
+[2026-07-28 06:15:04,422][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.10831049084663391, 'train/grad_norm_pre_clip': 3.9372360706329346, 'timing/data': 0.00034009572118520737, 'timing/model': 0.5331763327121735, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.04})
+[2026-07-28 06:15:11,682][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.09796766936779022, 'train/grad_norm_pre_clip': 2.691194534301758, 'timing/data': 0.00034323520958423615, 'timing/model': 0.531083763577044, 'learning_rate/qwen_vl_interface': 9.98696199865172e-06, 'learning_rate/action_model': 9.98696199865172e-05, 'epoch': 0.04})
+[2026-07-28 06:15:18,907][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.10202053189277649, 'train/grad_norm_pre_clip': 4.09560489654541, 'timing/data': 0.00040726829320192337, 'timing/model': 0.5296178832650185, 'learning_rate/qwen_vl_interface': 9.986677156724054e-06, 'learning_rate/action_model': 9.986677156724054e-05, 'epoch': 0.04})
+[2026-07-28 06:15:26,125][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.12793397903442383, 'train/grad_norm_pre_clip': 4.943708896636963, 'timing/data': 0.0003514392301440239, 'timing/model': 0.5267849620431662, 'learning_rate/qwen_vl_interface': 9.986389241218003e-06, 'learning_rate/action_model': 9.986389241218003e-05, 'epoch': 0.04})
+[2026-07-28 06:15:33,380][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.24088197946548462, 'train/grad_norm_pre_clip': 8.141899108886719, 'timing/data': 0.00040075182914733887, 'timing/model': 0.5283605270087719, 'learning_rate/qwen_vl_interface': 9.986098252320397e-06, 'learning_rate/action_model': 9.986098252320397e-05, 'epoch': 0.04})
+[2026-07-28 06:15:40,881][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.17631645500659943, 'train/grad_norm_pre_clip': 6.537606716156006, 'timing/data': 0.0003335867077112198, 'timing/model': 0.5271812528371811, 'learning_rate/qwen_vl_interface': 9.985804190220054e-06, 'learning_rate/action_model': 9.985804190220053e-05, 'epoch': 0.04})
+[2026-07-28 06:15:48,092][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.13380084931850433, 'train/grad_norm_pre_clip': 6.065043926239014, 'timing/data': 0.00033144373446702957, 'timing/model': 0.5296756513416767, 'learning_rate/qwen_vl_interface': 9.985507055107788e-06, 'learning_rate/action_model': 9.985507055107788e-05, 'epoch': 0.04})
+[2026-07-28 06:15:55,312][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.1325712949037552, 'train/grad_norm_pre_clip': 4.520786285400391, 'timing/data': 0.00038306042551994324, 'timing/model': 0.5293144453316927, 'learning_rate/qwen_vl_interface': 9.985206847176406e-06, 'learning_rate/action_model': 9.985206847176406e-05, 'epoch': 0.04})
+[2026-07-28 06:16:02,546][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.15099671483039856, 'train/grad_norm_pre_clip': 7.167413711547852, 'timing/data': 0.0003814930096268654, 'timing/model': 0.5268913777545094, 'learning_rate/qwen_vl_interface': 9.984903566620711e-06, 'learning_rate/action_model': 9.984903566620712e-05, 'epoch': 0.04})
+[2026-07-28 06:16:09,802][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.12215916067361832, 'train/grad_norm_pre_clip': 4.499102592468262, 'timing/data': 0.0003881072625517845, 'timing/model': 0.5297204647213221, 'learning_rate/qwen_vl_interface': 9.984597213637497e-06, 'learning_rate/action_model': 9.984597213637497e-05, 'epoch': 0.04})
+[2026-07-28 06:16:17,042][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.15767265856266022, 'train/grad_norm_pre_clip': 5.095218658447266, 'timing/data': 0.0003768056631088257, 'timing/model': 0.5289239529520273, 'learning_rate/qwen_vl_interface': 9.984287788425555e-06, 'learning_rate/action_model': 9.984287788425555e-05, 'epoch': 0.04})
+[2026-07-28 06:16:24,293][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.11987672001123428, 'train/grad_norm_pre_clip': 2.9343791007995605, 'timing/data': 0.00039413198828697205, 'timing/model': 0.5305717634037137, 'learning_rate/qwen_vl_interface': 9.983975291185666e-06, 'learning_rate/action_model': 9.983975291185667e-05, 'epoch': 0.04})
+[2026-07-28 06:16:31,567][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.11662493646144867, 'train/grad_norm_pre_clip': 4.059971809387207, 'timing/data': 0.0004081735387444496, 'timing/model': 0.5270601380616426, 'learning_rate/qwen_vl_interface': 9.983659722120607e-06, 'learning_rate/action_model': 9.983659722120608e-05, 'epoch': 0.04})
+[2026-07-28 06:16:38,856][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.11147705465555191, 'train/grad_norm_pre_clip': 5.41676664352417, 'timing/data': 0.0005024485290050507, 'timing/model': 0.5288549782708287, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.04})
+[2026-07-28 06:16:46,115][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.09758509695529938, 'train/grad_norm_pre_clip': 0.8200662136077881, 'timing/data': 0.00040570273995399475, 'timing/model': 0.5307513317093253, 'learning_rate/qwen_vl_interface': 9.983019369336048e-06, 'learning_rate/action_model': 9.983019369336048e-05, 'epoch': 0.04})
+[2026-07-28 06:16:53,385][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.1103268712759018, 'train/grad_norm_pre_clip': 2.7761080265045166, 'timing/data': 0.0003615235909819603, 'timing/model': 0.5275609921664, 'learning_rate/qwen_vl_interface': 9.982694586032066e-06, 'learning_rate/action_model': 9.982694586032065e-05, 'epoch': 0.04})
+[2026-07-28 06:17:00,682][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.1536971926689148, 'train/grad_norm_pre_clip': 6.314487934112549, 'timing/data': 0.000370267778635025, 'timing/model': 0.5336803700774908, 'learning_rate/qwen_vl_interface': 9.982366731733948e-06, 'learning_rate/action_model': 9.982366731733947e-05, 'epoch': 0.04})
+[2026-07-28 06:17:07,974][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.10781221091747284, 'train/grad_norm_pre_clip': 2.99953293800354, 'timing/data': 0.00035534240305423737, 'timing/model': 0.5303126079961658, 'learning_rate/qwen_vl_interface': 9.982035806654437e-06, 'learning_rate/action_model': 9.982035806654437e-05, 'epoch': 0.04})
+[2026-07-28 06:17:15,259][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.21852752566337585, 'train/grad_norm_pre_clip': 10.683979034423828, 'timing/data': 0.0004216255620121956, 'timing/model': 0.5304998308420181, 'learning_rate/qwen_vl_interface': 9.981701811008267e-06, 'learning_rate/action_model': 9.981701811008267e-05, 'epoch': 0.04})
+[2026-07-28 06:17:22,655][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.13268522918224335, 'train/grad_norm_pre_clip': 4.343482494354248, 'timing/data': 0.00045457202941179276, 'timing/model': 0.5329676400870085, 'learning_rate/qwen_vl_interface': 9.981364745012163e-06, 'learning_rate/action_model': 9.981364745012163e-05, 'epoch': 0.04})
+[2026-07-28 06:17:30,007][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.0901007279753685, 'train/grad_norm_pre_clip': 4.800471305847168, 'timing/data': 0.0003497833386063576, 'timing/model': 0.5301669901236892, 'learning_rate/qwen_vl_interface': 9.981024608884842e-06, 'learning_rate/action_model': 9.981024608884843e-05, 'epoch': 0.04})
+[2026-07-28 06:17:37,429][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.08719430863857269, 'train/grad_norm_pre_clip': 2.595734119415283, 'timing/data': 0.00047949422150850296, 'timing/model': 0.527527910657227, 'learning_rate/qwen_vl_interface': 9.980681402847017e-06, 'learning_rate/action_model': 9.980681402847016e-05, 'epoch': 0.04})
+[2026-07-28 06:17:44,742][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.11956561356782913, 'train/grad_norm_pre_clip': 2.6676015853881836, 'timing/data': 0.00047394726425409317, 'timing/model': 0.5294999331235886, 'learning_rate/qwen_vl_interface': 9.980335127121391e-06, 'learning_rate/action_model': 9.98033512712139e-05, 'epoch': 0.04})
+[2026-07-28 06:17:51,968][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.08512972295284271, 'train/grad_norm_pre_clip': 3.0081849098205566, 'timing/data': 0.0003475705161690712, 'timing/model': 0.5286699254065752, 'learning_rate/qwen_vl_interface': 9.979985781932658e-06, 'learning_rate/action_model': 9.979985781932657e-05, 'epoch': 0.04})
+[2026-07-28 06:17:59,234][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.13411004841327667, 'train/grad_norm_pre_clip': 5.446550369262695, 'timing/data': 0.00040199607610702515, 'timing/model': 0.5267318785190582, 'learning_rate/qwen_vl_interface': 9.979633367507501e-06, 'learning_rate/action_model': 9.979633367507502e-05, 'epoch': 0.04})
+[2026-07-28 06:18:06,464][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.13759681582450867, 'train/grad_norm_pre_clip': 3.247908353805542, 'timing/data': 0.00034529250115156174, 'timing/model': 0.5321405259892344, 'learning_rate/qwen_vl_interface': 9.979277884074604e-06, 'learning_rate/action_model': 9.979277884074604e-05, 'epoch': 0.04})
+[2026-07-28 06:18:13,732][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.11400862783193588, 'train/grad_norm_pre_clip': 5.047634601593018, 'timing/data': 0.00035401247441768646, 'timing/model': 0.5295414291322231, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.04})
+[2026-07-28 06:18:20,973][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.12042145431041718, 'train/grad_norm_pre_clip': 3.358743190765381, 'timing/data': 0.0003495998680591583, 'timing/model': 0.5332823116332293, 'learning_rate/qwen_vl_interface': 9.978557711110245e-06, 'learning_rate/action_model': 9.978557711110244e-05, 'epoch': 0.04})
+[2026-07-28 06:18:28,215][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.10693474113941193, 'train/grad_norm_pre_clip': 1.783373236656189, 'timing/data': 0.00037457793951034546, 'timing/model': 0.5309893265366554, 'learning_rate/qwen_vl_interface': 9.978193022046097e-06, 'learning_rate/action_model': 9.978193022046097e-05, 'epoch': 0.04})
+[2026-07-28 06:18:35,711][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.16730555891990662, 'train/grad_norm_pre_clip': 5.965268611907959, 'timing/data': 0.0006272848695516586, 'timing/model': 0.5290470896288753, 'learning_rate/qwen_vl_interface': 9.977825264908832e-06, 'learning_rate/action_model': 9.977825264908832e-05, 'epoch': 0.04})
+[2026-07-28 06:18:42,955][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.13943515717983246, 'train/grad_norm_pre_clip': 2.7522525787353516, 'timing/data': 0.0003545619547367096, 'timing/model': 0.5326204877346754, 'learning_rate/qwen_vl_interface': 9.97745443993708e-06, 'learning_rate/action_model': 9.97745443993708e-05, 'epoch': 0.04})
+[2026-07-28 06:18:50,197][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.07186955958604813, 'train/grad_norm_pre_clip': 1.9683141708374023, 'timing/data': 0.0003391122445464134, 'timing/model': 0.5294299451634288, 'learning_rate/qwen_vl_interface': 9.977080547371468e-06, 'learning_rate/action_model': 9.977080547371468e-05, 'epoch': 0.04})
+[2026-07-28 06:18:57,531][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.08813357353210449, 'train/grad_norm_pre_clip': 2.9235899448394775, 'timing/data': 0.00035125110298395157, 'timing/model': 0.529530955478549, 'learning_rate/qwen_vl_interface': 9.97670358745461e-06, 'learning_rate/action_model': 9.97670358745461e-05, 'epoch': 0.04})
+[2026-07-28 06:19:04,828][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.2277432084083557, 'train/grad_norm_pre_clip': 8.796868324279785, 'timing/data': 0.0003720177337527275, 'timing/model': 0.531658542342484, 'learning_rate/qwen_vl_interface': 9.976323560431112e-06, 'learning_rate/action_model': 9.976323560431111e-05, 'epoch': 0.04})
+[2026-07-28 06:19:12,087][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.12555500864982605, 'train/grad_norm_pre_clip': 8.238978385925293, 'timing/data': 0.00034733396023511887, 'timing/model': 0.5292682917788625, 'learning_rate/qwen_vl_interface': 9.975940466547567e-06, 'learning_rate/action_model': 9.975940466547567e-05, 'epoch': 0.04})
+[2026-07-28 06:19:19,344][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.13787546753883362, 'train/grad_norm_pre_clip': 3.338386058807373, 'timing/data': 0.0004213927313685417, 'timing/model': 0.5311114192008972, 'learning_rate/qwen_vl_interface': 9.975554306052563e-06, 'learning_rate/action_model': 9.975554306052562e-05, 'epoch': 0.04})
+[2026-07-28 06:19:26,647][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.12734080851078033, 'train/grad_norm_pre_clip': 2.498225688934326, 'timing/data': 0.00040380842983722687, 'timing/model': 0.5345209510996938, 'learning_rate/qwen_vl_interface': 9.975165079196674e-06, 'learning_rate/action_model': 9.975165079196673e-05, 'epoch': 0.04})
+[2026-07-28 06:19:33,872][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.10873385518789291, 'train/grad_norm_pre_clip': 4.097443580627441, 'timing/data': 0.00037035346031188965, 'timing/model': 0.5279475962743163, 'learning_rate/qwen_vl_interface': 9.974772786232466e-06, 'learning_rate/action_model': 9.974772786232466e-05, 'epoch': 0.04})
+[2026-07-28 06:19:41,450][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.1704210638999939, 'train/grad_norm_pre_clip': 4.296196460723877, 'timing/data': 0.0003376677632331848, 'timing/model': 0.529884459450841, 'learning_rate/qwen_vl_interface': 9.974377427414493e-06, 'learning_rate/action_model': 9.974377427414493e-05, 'epoch': 0.04})
+[2026-07-28 06:19:48,873][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.12173204869031906, 'train/grad_norm_pre_clip': 3.693159580230713, 'timing/data': 0.0004161028191447258, 'timing/model': 0.5292784459888935, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.04})
+[2026-07-28 06:19:56,114][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.18972386419773102, 'train/grad_norm_pre_clip': 4.5710978507995605, 'timing/data': 0.0003464650362730026, 'timing/model': 0.529731061309576, 'learning_rate/qwen_vl_interface': 9.973577513245417e-06, 'learning_rate/action_model': 9.973577513245417e-05, 'epoch': 0.04})
+[2026-07-28 06:20:03,351][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.20404499769210815, 'train/grad_norm_pre_clip': 4.910457611083984, 'timing/data': 0.0005070902407169342, 'timing/model': 0.5355806853622198, 'learning_rate/qwen_vl_interface': 9.973172958413372e-06, 'learning_rate/action_model': 9.97317295841337e-05, 'epoch': 0.04})
+[2026-07-28 06:20:10,598][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.13948488235473633, 'train/grad_norm_pre_clip': 3.33880352973938, 'timing/data': 0.00045733898878097534, 'timing/model': 0.5284558590501547, 'learning_rate/qwen_vl_interface': 9.97276533876567e-06, 'learning_rate/action_model': 9.97276533876567e-05, 'epoch': 0.04})
+[2026-07-28 06:20:17,826][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.1664963811635971, 'train/grad_norm_pre_clip': 3.8402259349823, 'timing/data': 0.00040750205516815186, 'timing/model': 0.5278587341308594, 'learning_rate/qwen_vl_interface': 9.972354654566816e-06, 'learning_rate/action_model': 9.972354654566816e-05, 'epoch': 0.04})
+[2026-07-28 06:20:25,070][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.1253703087568283, 'train/grad_norm_pre_clip': 4.869285583496094, 'timing/data': 0.0003196327015757561, 'timing/model': 0.529589637182653, 'learning_rate/qwen_vl_interface': 9.971940906083295e-06, 'learning_rate/action_model': 9.971940906083295e-05, 'epoch': 0.04})
+[2026-07-28 06:20:32,305][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.18204811215400696, 'train/grad_norm_pre_clip': 6.129208564758301, 'timing/data': 0.00032936781644821167, 'timing/model': 0.5319134751334786, 'learning_rate/qwen_vl_interface': 9.971524093583586e-06, 'learning_rate/action_model': 9.971524093583586e-05, 'epoch': 0.04})
+[2026-07-28 06:20:39,548][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.15155260264873505, 'train/grad_norm_pre_clip': 5.014395236968994, 'timing/data': 0.00039432384073734283, 'timing/model': 0.5288336519151926, 'learning_rate/qwen_vl_interface': 9.971104217338154e-06, 'learning_rate/action_model': 9.971104217338153e-05, 'epoch': 0.04})
+[2026-07-28 06:20:46,786][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.1367245316505432, 'train/grad_norm_pre_clip': 3.017174005508423, 'timing/data': 0.0003604758530855179, 'timing/model': 0.527588133700192, 'learning_rate/qwen_vl_interface': 9.970681277619452e-06, 'learning_rate/action_model': 9.970681277619452e-05, 'epoch': 0.04})
+[2026-07-28 06:20:54,036][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.13912490010261536, 'train/grad_norm_pre_clip': 5.830143928527832, 'timing/data': 0.0003297077491879463, 'timing/model': 0.5300341360270977, 'learning_rate/qwen_vl_interface': 9.97025527470192e-06, 'learning_rate/action_model': 9.970255274701919e-05, 'epoch': 0.04})
+[2026-07-28 06:21:01,285][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.15341943502426147, 'train/grad_norm_pre_clip': 5.189215183258057, 'timing/data': 0.0003131190314888954, 'timing/model': 0.5308801485225558, 'learning_rate/qwen_vl_interface': 9.969826208861987e-06, 'learning_rate/action_model': 9.969826208861987e-05, 'epoch': 0.04})
+[2026-07-28 06:21:08,517][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.1568007469177246, 'train/grad_norm_pre_clip': 3.5843286514282227, 'timing/data': 0.0003820061683654785, 'timing/model': 0.5288966353982687, 'learning_rate/qwen_vl_interface': 9.96939408037807e-06, 'learning_rate/action_model': 9.96939408037807e-05, 'epoch': 0.04})
+[2026-07-28 06:21:15,763][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.08638143539428711, 'train/grad_norm_pre_clip': 3.1383371353149414, 'timing/data': 0.00034605246037244797, 'timing/model': 0.530309627763927, 'learning_rate/qwen_vl_interface': 9.968958889530573e-06, 'learning_rate/action_model': 9.968958889530573e-05, 'epoch': 0.04})
+[2026-07-28 06:21:23,010][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.12749406695365906, 'train/grad_norm_pre_clip': 3.7439091205596924, 'timing/data': 0.00039076898247003555, 'timing/model': 0.5303457640111446, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.04})
+[2026-07-28 06:21:30,239][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.10747506469488144, 'train/grad_norm_pre_clip': 2.6259944438934326, 'timing/data': 0.0003611985594034195, 'timing/model': 0.531059785746038, 'learning_rate/qwen_vl_interface': 9.968079321876382e-06, 'learning_rate/action_model': 9.968079321876382e-05, 'epoch': 0.04})
+[2026-07-28 06:21:37,488][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.18066264688968658, 'train/grad_norm_pre_clip': 5.810922622680664, 'timing/data': 0.00044760387390851974, 'timing/model': 0.5330400839447975, 'learning_rate/qwen_vl_interface': 9.967634945640432e-06, 'learning_rate/action_model': 9.967634945640431e-05, 'epoch': 0.05})
+[2026-07-28 06:21:44,773][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.12218044698238373, 'train/grad_norm_pre_clip': 2.0520267486572266, 'timing/data': 0.0003465069457888603, 'timing/model': 0.5368317160755396, 'learning_rate/qwen_vl_interface': 9.967187508182385e-06, 'learning_rate/action_model': 9.967187508182385e-05, 'epoch': 0.05})
+[2026-07-28 06:21:52,017][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.09583638608455658, 'train/grad_norm_pre_clip': 2.082139492034912, 'timing/data': 0.0004764823243021965, 'timing/model': 0.5422630133107305, 'learning_rate/qwen_vl_interface': 9.966737009792577e-06, 'learning_rate/action_model': 9.966737009792577e-05, 'epoch': 0.05})
+[2026-07-28 06:21:59,263][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.10395832359790802, 'train/grad_norm_pre_clip': 2.3319380283355713, 'timing/data': 0.00034805014729499817, 'timing/model': 0.5321664391085505, 'learning_rate/qwen_vl_interface': 9.966283450763333e-06, 'learning_rate/action_model': 9.966283450763333e-05, 'epoch': 0.05})
+[2026-07-28 06:22:06,509][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.20012547075748444, 'train/grad_norm_pre_clip': 6.604739665985107, 'timing/data': 0.0003347797319293022, 'timing/model': 0.5280238073319197, 'learning_rate/qwen_vl_interface': 9.965826831388962e-06, 'learning_rate/action_model': 9.965826831388961e-05, 'epoch': 0.05})
+[2026-07-28 06:22:13,724][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 250
+[2026-07-28 06:23:29,242][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 250
+[2026-07-28 06:25:03,926][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 250
+[2026-07-28 06:25:03,927][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.09295956790447235, 'train/grad_norm_pre_clip': 1.1155766248703003, 'eval/action_loss/seconds': 75.51415343768895, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.10881078988313675, 'eval/latency_3/loss': 0.10881078988313675, 'eval/flappy/loss': 0.10881078988313675, 'eval/flappy/latency_3/loss': 0.10881078988313675, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 22.400000289827585, 'rl_games_eval/mid_train/mean_length': 199.45, 'rl_games_eval/mid_train/std_reward': 25.600390922011698, 'rl_games_eval/mid_train/std_length': 206.7983256702046, 'rl_games_eval/mid_train/macro_mean_reward': 22.400000289827585, 'rl_games_eval/mid_train/macro_mean_length': 199.45, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 22.400000289827585, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 199.45, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 25.600390922011698, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 206.7983256702046, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 22.400000289827585, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003173351287841797, 'timing/model': 0.5270479079335928, 'learning_rate/qwen_vl_interface': 9.965367151965757e-06, 'learning_rate/action_model': 9.965367151965756e-05, 'epoch': 0.05})
+[2026-07-28 06:25:13,781][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.11318802833557129, 'train/grad_norm_pre_clip': 2.585028648376465, 'timing/data': 0.0003658728674054146, 'timing/model': 0.8893365310505033, 'learning_rate/qwen_vl_interface': 9.964904412792002e-06, 'learning_rate/action_model': 9.964904412792002e-05, 'epoch': 0.05})
+[2026-07-28 06:25:22,994][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.09685856848955154, 'train/grad_norm_pre_clip': 1.8883252143859863, 'timing/data': 0.00031986646354198456, 'timing/model': 0.6175991995260119, 'learning_rate/qwen_vl_interface': 9.964438614167964e-06, 'learning_rate/action_model': 9.964438614167964e-05, 'epoch': 0.05})
+[2026-07-28 06:25:32,246][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.09220102429389954, 'train/grad_norm_pre_clip': 2.996077537536621, 'timing/data': 0.0003228709101676941, 'timing/model': 0.6189725324511528, 'learning_rate/qwen_vl_interface': 9.963969756395892e-06, 'learning_rate/action_model': 9.96396975639589e-05, 'epoch': 0.05})
+[2026-07-28 06:25:41,586][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.13333089649677277, 'train/grad_norm_pre_clip': 3.863117218017578, 'timing/data': 0.0004002731293439865, 'timing/model': 0.785951005294919, 'learning_rate/qwen_vl_interface': 9.963497839780025e-06, 'learning_rate/action_model': 9.963497839780024e-05, 'epoch': 0.05})
+[2026-07-28 06:25:50,979][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.12722861766815186, 'train/grad_norm_pre_clip': 5.8970770835876465, 'timing/data': 0.0003796154633164406, 'timing/model': 0.6708623645827174, 'learning_rate/qwen_vl_interface': 9.963022864626582e-06, 'learning_rate/action_model': 9.963022864626581e-05, 'epoch': 0.05})
+[2026-07-28 06:26:00,298][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.12238430976867676, 'train/grad_norm_pre_clip': 4.106616497039795, 'timing/data': 0.0003902539610862732, 'timing/model': 0.6221617069095373, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.05})
+[2026-07-28 06:26:09,567][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.15690867602825165, 'train/grad_norm_pre_clip': 4.974050998687744, 'timing/data': 0.000492975115776062, 'timing/model': 0.6204840149730444, 'learning_rate/qwen_vl_interface': 9.962063739941783e-06, 'learning_rate/action_model': 9.962063739941783e-05, 'epoch': 0.05})
+[2026-07-28 06:26:18,991][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.11139895021915436, 'train/grad_norm_pre_clip': 2.8017897605895996, 'timing/data': 0.00034657958894968033, 'timing/model': 0.8601525351405144, 'learning_rate/qwen_vl_interface': 9.961579591032791e-06, 'learning_rate/action_model': 9.96157959103279e-05, 'epoch': 0.05})
+[2026-07-28 06:26:28,394][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.10925298929214478, 'train/grad_norm_pre_clip': 3.7436625957489014, 'timing/data': 0.0003865780308842659, 'timing/model': 0.892358997836709, 'learning_rate/qwen_vl_interface': 9.961092384830956e-06, 'learning_rate/action_model': 9.961092384830956e-05, 'epoch': 0.05})
+[2026-07-28 06:26:37,626][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.20707765221595764, 'train/grad_norm_pre_clip': 9.099160194396973, 'timing/data': 0.0003445744514465332, 'timing/model': 0.6763885365799069, 'learning_rate/qwen_vl_interface': 9.96060212165242e-06, 'learning_rate/action_model': 9.96060212165242e-05, 'epoch': 0.05})
+[2026-07-28 06:26:46,950][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.10268674790859222, 'train/grad_norm_pre_clip': 2.6053848266601562, 'timing/data': 0.0005075307562947273, 'timing/model': 0.6243654740974307, 'learning_rate/qwen_vl_interface': 9.96010880181531e-06, 'learning_rate/action_model': 9.960108801815309e-05, 'epoch': 0.05})
+[2026-07-28 06:26:56,398][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.11980558931827545, 'train/grad_norm_pre_clip': 3.3141098022460938, 'timing/data': 0.0003249906003475189, 'timing/model': 0.6215428868308663, 'learning_rate/qwen_vl_interface': 9.959612425639736e-06, 'learning_rate/action_model': 9.959612425639735e-05, 'epoch': 0.05})
+[2026-07-28 06:27:05,553][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.1070452630519867, 'train/grad_norm_pre_clip': 2.8784849643707275, 'timing/data': 0.00032640527933835983, 'timing/model': 0.619336630217731, 'learning_rate/qwen_vl_interface': 9.95911299344779e-06, 'learning_rate/action_model': 9.959112993447789e-05, 'epoch': 0.05})
+[2026-07-28 06:27:14,782][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.14363574981689453, 'train/grad_norm_pre_clip': 4.8964457511901855, 'timing/data': 0.000322168692946434, 'timing/model': 0.618428461253643, 'learning_rate/qwen_vl_interface': 9.958610505563546e-06, 'learning_rate/action_model': 9.958610505563546e-05, 'epoch': 0.05})
+[2026-07-28 06:27:24,422][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.135748490691185, 'train/grad_norm_pre_clip': 3.921971321105957, 'timing/data': 0.00039398297667503357, 'timing/model': 0.8791050622239709, 'learning_rate/qwen_vl_interface': 9.95810496231307e-06, 'learning_rate/action_model': 9.95810496231307e-05, 'epoch': 0.05})
+[2026-07-28 06:27:33,613][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.10178886353969574, 'train/grad_norm_pre_clip': 2.3632309436798096, 'timing/data': 0.0003278963267803192, 'timing/model': 0.6224180599674582, 'learning_rate/qwen_vl_interface': 9.957596364024396e-06, 'learning_rate/action_model': 9.957596364024396e-05, 'epoch': 0.05})
+[2026-07-28 06:27:42,585][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.10756540298461914, 'train/grad_norm_pre_clip': 2.9066152572631836, 'timing/data': 0.0003426000475883484, 'timing/model': 0.6200162414461374, 'learning_rate/qwen_vl_interface': 9.957084711027553e-06, 'learning_rate/action_model': 9.957084711027552e-05, 'epoch': 0.05})
+[2026-07-28 06:27:51,786][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.093720942735672, 'train/grad_norm_pre_clip': 2.0774567127227783, 'timing/data': 0.00031693466007709503, 'timing/model': 0.6240423740819097, 'learning_rate/qwen_vl_interface': 9.956570003654544e-06, 'learning_rate/action_model': 9.956570003654544e-05, 'epoch': 0.05})
+[2026-07-28 06:28:01,414][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.10769860446453094, 'train/grad_norm_pre_clip': 2.0766847133636475, 'timing/data': 0.00032347068190574646, 'timing/model': 0.619046539068222, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.05})
+[2026-07-28 06:28:10,784][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.129516139626503, 'train/grad_norm_pre_clip': 2.2774312496185303, 'timing/data': 0.00031247176229953766, 'timing/model': 0.6182662453502417, 'learning_rate/qwen_vl_interface': 9.955531427117964e-06, 'learning_rate/action_model': 9.955531427117964e-05, 'epoch': 0.05})
+[2026-07-28 06:28:19,989][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.05942050367593765, 'train/grad_norm_pre_clip': 3.2056000232696533, 'timing/data': 0.0003099050372838974, 'timing/model': 0.6187196616083384, 'learning_rate/qwen_vl_interface': 9.955007558628316e-06, 'learning_rate/action_model': 9.955007558628316e-05, 'epoch': 0.05})
+[2026-07-28 06:28:29,165][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.10049837827682495, 'train/grad_norm_pre_clip': 3.0059990882873535, 'timing/data': 0.00042298901826143265, 'timing/model': 0.6206113304942846, 'learning_rate/qwen_vl_interface': 9.954480637110346e-06, 'learning_rate/action_model': 9.954480637110344e-05, 'epoch': 0.05})
+[2026-07-28 06:28:38,639][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.13360652327537537, 'train/grad_norm_pre_clip': 4.825632572174072, 'timing/data': 0.0003232760354876518, 'timing/model': 0.6187643464654684, 'learning_rate/qwen_vl_interface': 9.953950662905963e-06, 'learning_rate/action_model': 9.953950662905964e-05, 'epoch': 0.05})
+[2026-07-28 06:28:48,093][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.08449471741914749, 'train/grad_norm_pre_clip': 2.400965452194214, 'timing/data': 0.000438460148870945, 'timing/model': 0.6232583122327924, 'learning_rate/qwen_vl_interface': 9.953417636359067e-06, 'learning_rate/action_model': 9.953417636359067e-05, 'epoch': 0.05})
+[2026-07-28 06:28:57,433][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.10140477120876312, 'train/grad_norm_pre_clip': 2.6277365684509277, 'timing/data': 0.00039958395063877106, 'timing/model': 0.620474667288363, 'learning_rate/qwen_vl_interface': 9.95288155781553e-06, 'learning_rate/action_model': 9.95288155781553e-05, 'epoch': 0.05})
+[2026-07-28 06:29:06,844][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.14515967667102814, 'train/grad_norm_pre_clip': 2.2661523818969727, 'timing/data': 0.0003321059048175812, 'timing/model': 0.8475636122748256, 'learning_rate/qwen_vl_interface': 9.952342427623209e-06, 'learning_rate/action_model': 9.952342427623208e-05, 'epoch': 0.05})
+[2026-07-28 06:29:16,229][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.1416437327861786, 'train/grad_norm_pre_clip': 3.7490835189819336, 'timing/data': 0.0003980277106165886, 'timing/model': 0.8394511556252837, 'learning_rate/qwen_vl_interface': 9.95180024613194e-06, 'learning_rate/action_model': 9.951800246131939e-05, 'epoch': 0.05})
+[2026-07-28 06:29:25,387][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.16764840483665466, 'train/grad_norm_pre_clip': 3.476836919784546, 'timing/data': 0.00041336193680763245, 'timing/model': 0.6194485267624259, 'learning_rate/qwen_vl_interface': 9.951255013693537e-06, 'learning_rate/action_model': 9.951255013693536e-05, 'epoch': 0.05})
+[2026-07-28 06:29:34,524][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.09204400330781937, 'train/grad_norm_pre_clip': 3.5552570819854736, 'timing/data': 0.0004178723320364952, 'timing/model': 0.618346493691206, 'learning_rate/qwen_vl_interface': 9.950706730661797e-06, 'learning_rate/action_model': 9.950706730661797e-05, 'epoch': 0.05})
+[2026-07-28 06:29:43,698][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.052059952169656754, 'train/grad_norm_pre_clip': 1.7733436822891235, 'timing/data': 0.00042063649743795395, 'timing/model': 0.6192488595843315, 'learning_rate/qwen_vl_interface': 9.950155397392494e-06, 'learning_rate/action_model': 9.950155397392493e-05, 'epoch': 0.05})
+[2026-07-28 06:29:53,091][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.0910482406616211, 'train/grad_norm_pre_clip': 1.940176010131836, 'timing/data': 0.00036488473415374756, 'timing/model': 0.8489197576418519, 'learning_rate/qwen_vl_interface': 9.949601014243382e-06, 'learning_rate/action_model': 9.949601014243382e-05, 'epoch': 0.05})
+[2026-07-28 06:30:02,470][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.09968240559101105, 'train/grad_norm_pre_clip': 2.1825308799743652, 'timing/data': 0.000329664908349514, 'timing/model': 0.6221524262800813, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.05})
+[2026-07-28 06:30:11,549][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.09757110476493835, 'train/grad_norm_pre_clip': 3.7113544940948486, 'timing/data': 0.00038818269968032837, 'timing/model': 0.6172887347638607, 'learning_rate/qwen_vl_interface': 9.948483099746645e-06, 'learning_rate/action_model': 9.948483099746645e-05, 'epoch': 0.05})
+[2026-07-28 06:30:20,511][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.14181001484394073, 'train/grad_norm_pre_clip': 2.7919909954071045, 'timing/data': 0.0003378307446837425, 'timing/model': 0.6243097484111786, 'learning_rate/qwen_vl_interface': 9.947919569124423e-06, 'learning_rate/action_model': 9.947919569124422e-05, 'epoch': 0.05})
+[2026-07-28 06:30:29,514][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.10867384076118469, 'train/grad_norm_pre_clip': 2.1528732776641846, 'timing/data': 0.0003288816660642624, 'timing/model': 0.616052514873445, 'learning_rate/qwen_vl_interface': 9.947352990073196e-06, 'learning_rate/action_model': 9.947352990073195e-05, 'epoch': 0.05})
+[2026-07-28 06:30:38,621][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.12870506942272186, 'train/grad_norm_pre_clip': 2.3650386333465576, 'timing/data': 0.0003389492630958557, 'timing/model': 0.6179328504949808, 'learning_rate/qwen_vl_interface': 9.946783362960612e-06, 'learning_rate/action_model': 9.946783362960612e-05, 'epoch': 0.05})
+[2026-07-28 06:30:47,770][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.08496841788291931, 'train/grad_norm_pre_clip': 1.5308631658554077, 'timing/data': 0.00036308541893959045, 'timing/model': 0.8498533191159368, 'learning_rate/qwen_vl_interface': 9.946210688156297e-06, 'learning_rate/action_model': 9.946210688156297e-05, 'epoch': 0.05})
+[2026-07-28 06:30:57,127][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.11548812687397003, 'train/grad_norm_pre_clip': 3.787951946258545, 'timing/data': 0.00034037698060274124, 'timing/model': 0.6218594238162041, 'learning_rate/qwen_vl_interface': 9.945634966031852e-06, 'learning_rate/action_model': 9.945634966031851e-05, 'epoch': 0.05})
+[2026-07-28 06:31:06,270][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.1293395459651947, 'train/grad_norm_pre_clip': 3.389244556427002, 'timing/data': 0.00032547395676374435, 'timing/model': 0.6179874297231436, 'learning_rate/qwen_vl_interface': 9.945056196960858e-06, 'learning_rate/action_model': 9.945056196960858e-05, 'epoch': 0.05})
+[2026-07-28 06:31:15,570][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.07821187376976013, 'train/grad_norm_pre_clip': 1.8261507749557495, 'timing/data': 0.0003187805414199829, 'timing/model': 0.6640717219561338, 'learning_rate/qwen_vl_interface': 9.944474381318871e-06, 'learning_rate/action_model': 9.94447438131887e-05, 'epoch': 0.05})
+[2026-07-28 06:31:24,872][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.12871068716049194, 'train/grad_norm_pre_clip': 4.452404499053955, 'timing/data': 0.0004048151895403862, 'timing/model': 0.6169681316241622, 'learning_rate/qwen_vl_interface': 9.943889519483424e-06, 'learning_rate/action_model': 9.943889519483424e-05, 'epoch': 0.05})
+[2026-07-28 06:31:34,201][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.1651979684829712, 'train/grad_norm_pre_clip': 4.404116630554199, 'timing/data': 0.0003252839669585228, 'timing/model': 0.6184850307181478, 'learning_rate/qwen_vl_interface': 9.94330161183403e-06, 'learning_rate/action_model': 9.94330161183403e-05, 'epoch': 0.05})
+[2026-07-28 06:31:43,560][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.15934894979000092, 'train/grad_norm_pre_clip': 4.556929111480713, 'timing/data': 0.00039126165211200714, 'timing/model': 0.6217589657753706, 'learning_rate/qwen_vl_interface': 9.942710658752175e-06, 'learning_rate/action_model': 9.942710658752174e-05, 'epoch': 0.05})
+[2026-07-28 06:31:52,695][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.13545091450214386, 'train/grad_norm_pre_clip': 5.736920356750488, 'timing/data': 0.0003486955538392067, 'timing/model': 0.6156474379822612, 'learning_rate/qwen_vl_interface': 9.942116660621319e-06, 'learning_rate/action_model': 9.94211666062132e-05, 'epoch': 0.05})
+[2026-07-28 06:32:02,129][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.09086611866950989, 'train/grad_norm_pre_clip': 3.3866841793060303, 'timing/data': 0.0004189368337392807, 'timing/model': 0.6191247506067157, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.05})
+[2026-07-28 06:32:11,474][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.10108549892902374, 'train/grad_norm_pre_clip': 3.1993966102600098, 'timing/data': 0.0004032449796795845, 'timing/model': 0.7888459321111441, 'learning_rate/qwen_vl_interface': 9.940919530756347e-06, 'learning_rate/action_model': 9.940919530756347e-05, 'epoch': 0.05})
+[2026-07-28 06:32:20,833][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.10411763191223145, 'train/grad_norm_pre_clip': 3.6464786529541016, 'timing/data': 0.0003321925178170204, 'timing/model': 0.6193286590278149, 'learning_rate/qwen_vl_interface': 9.940316399799032e-06, 'learning_rate/action_model': 9.940316399799033e-05, 'epoch': 0.05})
+[2026-07-28 06:32:30,025][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.06570202857255936, 'train/grad_norm_pre_clip': 1.1861821413040161, 'timing/data': 0.00036506354808807373, 'timing/model': 0.6181469177827239, 'learning_rate/qwen_vl_interface': 9.939710225346329e-06, 'learning_rate/action_model': 9.939710225346329e-05, 'epoch': 0.05})
+[2026-07-28 06:32:39,381][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.13486018776893616, 'train/grad_norm_pre_clip': 3.202505588531494, 'timing/data': 0.0004013329744338989, 'timing/model': 0.6166635900735855, 'learning_rate/qwen_vl_interface': 9.939101007791575e-06, 'learning_rate/action_model': 9.939101007791575e-05, 'epoch': 0.05})
+[2026-07-28 06:32:48,771][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.12118560075759888, 'train/grad_norm_pre_clip': 3.7807185649871826, 'timing/data': 0.00033447518944740295, 'timing/model': 0.8531698929145932, 'learning_rate/qwen_vl_interface': 9.938488747530087e-06, 'learning_rate/action_model': 9.938488747530086e-05, 'epoch': 0.06})
+[2026-07-28 06:32:57,936][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.09094621241092682, 'train/grad_norm_pre_clip': 4.691878318786621, 'timing/data': 0.00032481737434864044, 'timing/model': 0.618374046869576, 'learning_rate/qwen_vl_interface': 9.937873444959151e-06, 'learning_rate/action_model': 9.937873444959151e-05, 'epoch': 0.06})
+[2026-07-28 06:33:07,063][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.10650205612182617, 'train/grad_norm_pre_clip': 2.791707754135132, 'timing/data': 0.00032757874578237534, 'timing/model': 0.6175354020670056, 'learning_rate/qwen_vl_interface': 9.937255100478036e-06, 'learning_rate/action_model': 9.937255100478036e-05, 'epoch': 0.06})
+[2026-07-28 06:33:16,449][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.127506822347641, 'train/grad_norm_pre_clip': 6.044828414916992, 'timing/data': 0.0003975452855229378, 'timing/model': 0.6252677580341697, 'learning_rate/qwen_vl_interface': 9.93663371448797e-06, 'learning_rate/action_model': 9.93663371448797e-05, 'epoch': 0.06})
+[2026-07-28 06:33:25,897][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.12282954901456833, 'train/grad_norm_pre_clip': 3.1989850997924805, 'timing/data': 0.0003788033500313759, 'timing/model': 0.6205992316827178, 'learning_rate/qwen_vl_interface': 9.936009287392173e-06, 'learning_rate/action_model': 9.936009287392171e-05, 'epoch': 0.06})
+[2026-07-28 06:33:34,954][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.12288527190685272, 'train/grad_norm_pre_clip': 4.26549768447876, 'timing/data': 0.00033242255449295044, 'timing/model': 0.6195288002490997, 'learning_rate/qwen_vl_interface': 9.935381819595822e-06, 'learning_rate/action_model': 9.935381819595822e-05, 'epoch': 0.06})
+[2026-07-28 06:33:44,150][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.09322814643383026, 'train/grad_norm_pre_clip': 2.9768481254577637, 'timing/data': 0.00032625719904899597, 'timing/model': 0.6188742937520146, 'learning_rate/qwen_vl_interface': 9.934751311506078e-06, 'learning_rate/action_model': 9.934751311506077e-05, 'epoch': 0.06})
+[2026-07-28 06:33:53,327][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.09089749306440353, 'train/grad_norm_pre_clip': 2.6953210830688477, 'timing/data': 0.0003240341320633888, 'timing/model': 0.6239184401929379, 'learning_rate/qwen_vl_interface': 9.93411776353207e-06, 'learning_rate/action_model': 9.934117763532069e-05, 'epoch': 0.06})
+[2026-07-28 06:34:02,960][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.08183309435844421, 'train/grad_norm_pre_clip': 1.5476959943771362, 'timing/data': 0.0003315722569823265, 'timing/model': 0.839749913662672, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.06})
+[2026-07-28 06:34:12,197][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.10952600091695786, 'train/grad_norm_pre_clip': 5.297059059143066, 'timing/data': 0.00033265911042690277, 'timing/model': 0.6199129754677415, 'learning_rate/qwen_vl_interface': 9.932841549577642e-06, 'learning_rate/action_model': 9.932841549577641e-05, 'epoch': 0.06})
+[2026-07-28 06:34:21,392][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.13313552737236023, 'train/grad_norm_pre_clip': 4.082218647003174, 'timing/data': 0.00032476335763931274, 'timing/model': 0.6172927049919963, 'learning_rate/qwen_vl_interface': 9.932198884425343e-06, 'learning_rate/action_model': 9.932198884425342e-05, 'epoch': 0.06})
+[2026-07-28 06:34:30,820][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.11799199879169464, 'train/grad_norm_pre_clip': 2.497957468032837, 'timing/data': 0.00033425353467464447, 'timing/model': 0.8532649772241712, 'learning_rate/qwen_vl_interface': 9.931553181045022e-06, 'learning_rate/action_model': 9.931553181045022e-05, 'epoch': 0.06})
+[2026-07-28 06:34:40,014][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.07904735207557678, 'train/grad_norm_pre_clip': 3.710848808288574, 'timing/data': 0.00032375752925872803, 'timing/model': 0.6178702702745795, 'learning_rate/qwen_vl_interface': 9.930904439855667e-06, 'learning_rate/action_model': 9.930904439855667e-05, 'epoch': 0.06})
+[2026-07-28 06:34:49,359][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.09663932770490646, 'train/grad_norm_pre_clip': 1.7085469961166382, 'timing/data': 0.00033541396260261536, 'timing/model': 0.6204035868868232, 'learning_rate/qwen_vl_interface': 9.930252661278245e-06, 'learning_rate/action_model': 9.930252661278244e-05, 'epoch': 0.06})
+[2026-07-28 06:34:58,467][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.11489799618721008, 'train/grad_norm_pre_clip': 3.8688547611236572, 'timing/data': 0.00034928135573863983, 'timing/model': 0.62106744479388, 'learning_rate/qwen_vl_interface': 9.92959784573568e-06, 'learning_rate/action_model': 9.92959784573568e-05, 'epoch': 0.06})
+[2026-07-28 06:35:07,636][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.08099201321601868, 'train/grad_norm_pre_clip': 1.8349277973175049, 'timing/data': 0.0003650849685072899, 'timing/model': 0.6187826637178659, 'learning_rate/qwen_vl_interface': 9.92893999365288e-06, 'learning_rate/action_model': 9.92893999365288e-05, 'epoch': 0.06})
+[2026-07-28 06:35:17,238][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.08446323871612549, 'train/grad_norm_pre_clip': 2.6917030811309814, 'timing/data': 0.0004035094752907753, 'timing/model': 0.6198488250374794, 'learning_rate/qwen_vl_interface': 9.928279105456717e-06, 'learning_rate/action_model': 9.928279105456718e-05, 'epoch': 0.06})
+[2026-07-28 06:35:26,163][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.07105772197246552, 'train/grad_norm_pre_clip': 1.4943766593933105, 'timing/data': 0.00033335015177726746, 'timing/model': 0.6202372573316097, 'learning_rate/qwen_vl_interface': 9.927615181576034e-06, 'learning_rate/action_model': 9.927615181576034e-05, 'epoch': 0.06})
+[2026-07-28 06:35:35,293][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.0800158828496933, 'train/grad_norm_pre_clip': 2.527442216873169, 'timing/data': 0.00032543670386075974, 'timing/model': 0.6189855895936489, 'learning_rate/qwen_vl_interface': 9.926948222441644e-06, 'learning_rate/action_model': 9.926948222441643e-05, 'epoch': 0.06})
+[2026-07-28 06:35:44,412][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.06561645865440369, 'train/grad_norm_pre_clip': 2.4544806480407715, 'timing/data': 0.00032041966915130615, 'timing/model': 0.6198429754003882, 'learning_rate/qwen_vl_interface': 9.926278228486328e-06, 'learning_rate/action_model': 9.926278228486327e-05, 'epoch': 0.06})
+[2026-07-28 06:35:54,000][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.11676307022571564, 'train/grad_norm_pre_clip': 3.4063916206359863, 'timing/data': 0.0003402894362807274, 'timing/model': 0.8488129293546081, 'learning_rate/qwen_vl_interface': 9.92560520014484e-06, 'learning_rate/action_model': 9.92560520014484e-05, 'epoch': 0.06})
+[2026-07-28 06:36:03,155][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.08331260085105896, 'train/grad_norm_pre_clip': 3.7203145027160645, 'timing/data': 0.000338665209710598, 'timing/model': 0.6179511211812496, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.06})
+[2026-07-28 06:36:12,347][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.08791134506464005, 'train/grad_norm_pre_clip': 3.291445255279541, 'timing/data': 0.00042837299406528473, 'timing/model': 0.6170928422361612, 'learning_rate/qwen_vl_interface': 9.924250042052197e-06, 'learning_rate/action_model': 9.924250042052197e-05, 'epoch': 0.06})
+[2026-07-28 06:36:21,790][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.11892183870077133, 'train/grad_norm_pre_clip': 3.769294023513794, 'timing/data': 0.0003383653238415718, 'timing/model': 0.8744550682604313, 'learning_rate/qwen_vl_interface': 9.92356791318039e-06, 'learning_rate/action_model': 9.92356791318039e-05, 'epoch': 0.06})
+[2026-07-28 06:36:30,947][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.14113913476467133, 'train/grad_norm_pre_clip': 4.62691593170166, 'timing/data': 0.0003225291147828102, 'timing/model': 0.616298571228981, 'learning_rate/qwen_vl_interface': 9.922882751681104e-06, 'learning_rate/action_model': 9.922882751681104e-05, 'epoch': 0.06})
+[2026-07-28 06:36:40,113][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.09773192554712296, 'train/grad_norm_pre_clip': 3.5118699073791504, 'timing/data': 0.00031955912709236145, 'timing/model': 0.6183846900239587, 'learning_rate/qwen_vl_interface': 9.922194557998932e-06, 'learning_rate/action_model': 9.922194557998933e-05, 'epoch': 0.06})
+[2026-07-28 06:36:49,483][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.10584264993667603, 'train/grad_norm_pre_clip': 3.9100077152252197, 'timing/data': 0.0003191158175468445, 'timing/model': 0.6174662988632917, 'learning_rate/qwen_vl_interface': 9.921503332580439e-06, 'learning_rate/action_model': 9.921503332580439e-05, 'epoch': 0.06})
+[2026-07-28 06:36:58,610][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.134621262550354, 'train/grad_norm_pre_clip': 3.0764880180358887, 'timing/data': 0.00031882524490356445, 'timing/model': 0.6174382856115699, 'learning_rate/qwen_vl_interface': 9.92080907587415e-06, 'learning_rate/action_model': 9.92080907587415e-05, 'epoch': 0.06})
+[2026-07-28 06:37:07,991][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.1468663215637207, 'train/grad_norm_pre_clip': 4.7640228271484375, 'timing/data': 0.0004292065277695656, 'timing/model': 0.615982417948544, 'learning_rate/qwen_vl_interface': 9.92011178833056e-06, 'learning_rate/action_model': 9.92011178833056e-05, 'epoch': 0.06})
+[2026-07-28 06:37:17,337][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.1128939688205719, 'train/grad_norm_pre_clip': 3.285381555557251, 'timing/data': 0.00034672487527132034, 'timing/model': 0.6179469637572765, 'learning_rate/qwen_vl_interface': 9.919411470402136e-06, 'learning_rate/action_model': 9.919411470402136e-05, 'epoch': 0.06})
+[2026-07-28 06:37:26,448][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.07659196853637695, 'train/grad_norm_pre_clip': 3.2660417556762695, 'timing/data': 0.00034351181238889694, 'timing/model': 0.6162027921527624, 'learning_rate/qwen_vl_interface': 9.918708122543305e-06, 'learning_rate/action_model': 9.918708122543305e-05, 'epoch': 0.06})
+[2026-07-28 06:37:35,740][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.09565470367670059, 'train/grad_norm_pre_clip': 3.2660746574401855, 'timing/data': 0.00033894553780555725, 'timing/model': 0.6172537337988615, 'learning_rate/qwen_vl_interface': 9.918001745210458e-06, 'learning_rate/action_model': 9.918001745210456e-05, 'epoch': 0.06})
+[2026-07-28 06:37:45,414][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.17252129316329956, 'train/grad_norm_pre_clip': 3.1244640350341797, 'timing/data': 0.00032841041684150696, 'timing/model': 0.786289619281888, 'learning_rate/qwen_vl_interface': 9.917292338861958e-06, 'learning_rate/action_model': 9.917292338861957e-05, 'epoch': 0.06})
+[2026-07-28 06:37:54,651][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.09567517787218094, 'train/grad_norm_pre_clip': 3.755276679992676, 'timing/data': 0.00034220609813928604, 'timing/model': 0.6200411431491375, 'learning_rate/qwen_vl_interface': 9.916579903958134e-06, 'learning_rate/action_model': 9.916579903958133e-05, 'epoch': 0.06})
+[2026-07-28 06:38:03,862][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.05389697849750519, 'train/grad_norm_pre_clip': 1.5293952226638794, 'timing/data': 0.0003102375194430351, 'timing/model': 0.6181454425677657, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.06})
+[2026-07-28 06:38:13,012][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.09808972477912903, 'train/grad_norm_pre_clip': 2.989741802215576, 'timing/data': 0.000377618707716465, 'timing/model': 0.6186277959495783, 'learning_rate/qwen_vl_interface': 9.915145950335631e-06, 'learning_rate/action_model': 9.915145950335632e-05, 'epoch': 0.06})
+[2026-07-28 06:38:22,401][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.0641249567270279, 'train/grad_norm_pre_clip': 3.647412061691284, 'timing/data': 0.0003226865082979202, 'timing/model': 0.6197758009657264, 'learning_rate/qwen_vl_interface': 9.914424432547434e-06, 'learning_rate/action_model': 9.914424432547433e-05, 'epoch': 0.06})
+[2026-07-28 06:38:31,756][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.12074355036020279, 'train/grad_norm_pre_clip': 1.6713826656341553, 'timing/data': 0.00032585952430963516, 'timing/model': 0.6229084767401218, 'learning_rate/qwen_vl_interface': 9.913699888064862e-06, 'learning_rate/action_model': 9.913699888064861e-05, 'epoch': 0.06})
+[2026-07-28 06:38:40,938][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.12371261417865753, 'train/grad_norm_pre_clip': 4.335604190826416, 'timing/data': 0.0003329869359731674, 'timing/model': 0.6189291225746274, 'learning_rate/qwen_vl_interface': 9.912972317358066e-06, 'learning_rate/action_model': 9.912972317358066e-05, 'epoch': 0.06})
+[2026-07-28 06:38:50,290][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.10469333827495575, 'train/grad_norm_pre_clip': 2.2192611694335938, 'timing/data': 0.0003611501306295395, 'timing/model': 0.8508143294602633, 'learning_rate/qwen_vl_interface': 9.912241720899157e-06, 'learning_rate/action_model': 9.912241720899157e-05, 'epoch': 0.06})
+[2026-07-28 06:38:59,424][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.08761443942785263, 'train/grad_norm_pre_clip': 6.119393825531006, 'timing/data': 0.00044927094131708145, 'timing/model': 0.6175096044316888, 'learning_rate/qwen_vl_interface': 9.911508099162215e-06, 'learning_rate/action_model': 9.911508099162214e-05, 'epoch': 0.06})
+[2026-07-28 06:39:08,646][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.10523981600999832, 'train/grad_norm_pre_clip': 4.414035797119141, 'timing/data': 0.0003451639786362648, 'timing/model': 0.6195758562535048, 'learning_rate/qwen_vl_interface': 9.910771452623276e-06, 'learning_rate/action_model': 9.910771452623275e-05, 'epoch': 0.06})
+[2026-07-28 06:39:18,087][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.15590372681617737, 'train/grad_norm_pre_clip': 6.037247180938721, 'timing/data': 0.0003228522837162018, 'timing/model': 0.6180712003260851, 'learning_rate/qwen_vl_interface': 9.910031781760342e-06, 'learning_rate/action_model': 9.910031781760342e-05, 'epoch': 0.06})
+[2026-07-28 06:39:27,218][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.05802544951438904, 'train/grad_norm_pre_clip': 2.007880449295044, 'timing/data': 0.0003376733511686325, 'timing/model': 0.6185126574710011, 'learning_rate/qwen_vl_interface': 9.909289087053378e-06, 'learning_rate/action_model': 9.909289087053378e-05, 'epoch': 0.06})
+[2026-07-28 06:39:36,567][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.09970816969871521, 'train/grad_norm_pre_clip': 2.8614068031311035, 'timing/data': 0.00033523794263601303, 'timing/model': 0.6195634044706821, 'learning_rate/qwen_vl_interface': 9.90854336898431e-06, 'learning_rate/action_model': 9.90854336898431e-05, 'epoch': 0.06})
+[2026-07-28 06:39:45,902][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.09876228123903275, 'train/grad_norm_pre_clip': 3.022531032562256, 'timing/data': 0.0003158748149871826, 'timing/model': 0.8406145870685577, 'learning_rate/qwen_vl_interface': 9.907794628037028e-06, 'learning_rate/action_model': 9.907794628037028e-05, 'epoch': 0.06})
+[2026-07-28 06:39:55,005][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.08089016377925873, 'train/grad_norm_pre_clip': 3.065486431121826, 'timing/data': 0.00040884315967559814, 'timing/model': 0.6182351280003786, 'learning_rate/qwen_vl_interface': 9.90704286469738e-06, 'learning_rate/action_model': 9.90704286469738e-05, 'epoch': 0.06})
+[2026-07-28 06:40:04,161][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.12857778370380402, 'train/grad_norm_pre_clip': 4.881289482116699, 'timing/data': 0.0003356505185365677, 'timing/model': 0.6178508624434471, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.06})
+[2026-07-28 06:40:13,266][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.11729893088340759, 'train/grad_norm_pre_clip': 5.266537189483643, 'timing/data': 0.0003197239711880684, 'timing/model': 0.6160815684124827, 'learning_rate/qwen_vl_interface': 9.905530272794195e-06, 'learning_rate/action_model': 9.905530272794195e-05, 'epoch': 0.06})
+[2026-07-28 06:40:22,617][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.09422077238559723, 'train/grad_norm_pre_clip': 2.778574228286743, 'timing/data': 0.00032894592732191086, 'timing/model': 0.8578677922487259, 'learning_rate/qwen_vl_interface': 9.904769445212163e-06, 'learning_rate/action_model': 9.904769445212162e-05, 'epoch': 0.06})
+[2026-07-28 06:40:32,051][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.06734202802181244, 'train/grad_norm_pre_clip': 0.9502338171005249, 'timing/data': 0.000340338796377182, 'timing/model': 0.8499154942110181, 'learning_rate/qwen_vl_interface': 9.904005597200772e-06, 'learning_rate/action_model': 9.904005597200772e-05, 'epoch': 0.06})
+[2026-07-28 06:40:41,180][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.05464722961187363, 'train/grad_norm_pre_clip': 2.3977296352386475, 'timing/data': 0.0003835642710328102, 'timing/model': 0.6208983054384589, 'learning_rate/qwen_vl_interface': 9.903238729255678e-06, 'learning_rate/action_model': 9.903238729255677e-05, 'epoch': 0.06})
+[2026-07-28 06:40:50,442][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.12060883641242981, 'train/grad_norm_pre_clip': 4.2365241050720215, 'timing/data': 0.0003271196037530899, 'timing/model': 0.6163710672408342, 'learning_rate/qwen_vl_interface': 9.902468841874492e-06, 'learning_rate/action_model': 9.902468841874492e-05, 'epoch': 0.06})
+[2026-07-28 06:40:59,638][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.12461435049772263, 'train/grad_norm_pre_clip': 3.6171298027038574, 'timing/data': 0.0003427397459745407, 'timing/model': 0.6169526055455208, 'learning_rate/qwen_vl_interface': 9.901695935556784e-06, 'learning_rate/action_model': 9.901695935556783e-05, 'epoch': 0.06})
+[2026-07-28 06:41:09,019][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.10355212539434433, 'train/grad_norm_pre_clip': 5.120621681213379, 'timing/data': 0.00032228417694568634, 'timing/model': 0.6171734239906073, 'learning_rate/qwen_vl_interface': 9.900920010804088e-06, 'learning_rate/action_model': 9.900920010804088e-05, 'epoch': 0.07})
+[2026-07-28 06:41:18,127][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.12031346559524536, 'train/grad_norm_pre_clip': 1.9343681335449219, 'timing/data': 0.0004703979939222336, 'timing/model': 0.6194855608046055, 'learning_rate/qwen_vl_interface': 9.900141068119891e-06, 'learning_rate/action_model': 9.900141068119891e-05, 'epoch': 0.07})
+[2026-07-28 06:41:27,304][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.06804250180721283, 'train/grad_norm_pre_clip': 2.5611867904663086, 'timing/data': 0.0003264229744672775, 'timing/model': 0.6176761304959655, 'learning_rate/qwen_vl_interface': 9.89935910800964e-06, 'learning_rate/action_model': 9.89935910800964e-05, 'epoch': 0.07})
+[2026-07-28 06:41:36,683][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.09182494878768921, 'train/grad_norm_pre_clip': 2.452773094177246, 'timing/data': 0.0004084799438714981, 'timing/model': 0.6237944969907403, 'learning_rate/qwen_vl_interface': 9.898574130980743e-06, 'learning_rate/action_model': 9.898574130980742e-05, 'epoch': 0.07})
+[2026-07-28 06:41:45,867][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.11073081195354462, 'train/grad_norm_pre_clip': 7.039622783660889, 'timing/data': 0.00035465601831674576, 'timing/model': 0.6187793081626296, 'learning_rate/qwen_vl_interface': 9.89778613754256e-06, 'learning_rate/action_model': 9.89778613754256e-05, 'epoch': 0.07})
+[2026-07-28 06:41:54,991][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.09355408698320389, 'train/grad_norm_pre_clip': 4.732100963592529, 'timing/data': 0.0004078429192304611, 'timing/model': 0.6170732695609331, 'learning_rate/qwen_vl_interface': 9.896995128206412e-06, 'learning_rate/action_model': 9.896995128206411e-05, 'epoch': 0.07})
+[2026-07-28 06:42:04,151][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.1423640251159668, 'train/grad_norm_pre_clip': 3.600982904434204, 'timing/data': 0.00038408953696489334, 'timing/model': 0.6187391579151154, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.07})
+[2026-07-28 06:42:13,554][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.10757964104413986, 'train/grad_norm_pre_clip': 4.149837017059326, 'timing/data': 0.00033155549317598343, 'timing/model': 0.6167661547660828, 'learning_rate/qwen_vl_interface': 9.89540406389529e-06, 'learning_rate/action_model': 9.89540406389529e-05, 'epoch': 0.07})
+[2026-07-28 06:42:22,694][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.1811772584915161, 'train/grad_norm_pre_clip': 4.988002777099609, 'timing/data': 0.00040012504905462265, 'timing/model': 0.6158627234399319, 'learning_rate/qwen_vl_interface': 9.894604009952742e-06, 'learning_rate/action_model': 9.89460400995274e-05, 'epoch': 0.07})
+[2026-07-28 06:42:32,245][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.12189775705337524, 'train/grad_norm_pre_clip': 5.1756978034973145, 'timing/data': 0.0003458075225353241, 'timing/model': 0.8433974990621209, 'learning_rate/qwen_vl_interface': 9.893800942177073e-06, 'learning_rate/action_model': 9.893800942177073e-05, 'epoch': 0.07})
+[2026-07-28 06:42:41,412][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.14045456051826477, 'train/grad_norm_pre_clip': 5.396177291870117, 'timing/data': 0.0003517419099807739, 'timing/model': 0.6204517390578985, 'learning_rate/qwen_vl_interface': 9.892994861089393e-06, 'learning_rate/action_model': 9.892994861089393e-05, 'epoch': 0.07})
+[2026-07-28 06:42:50,563][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.133035346865654, 'train/grad_norm_pre_clip': 5.42151403427124, 'timing/data': 0.0003277435898780823, 'timing/model': 0.6244560359045863, 'learning_rate/qwen_vl_interface': 9.892185767212756e-06, 'learning_rate/action_model': 9.892185767212756e-05, 'epoch': 0.07})
+[2026-07-28 06:42:59,495][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.10921235382556915, 'train/grad_norm_pre_clip': 3.333772897720337, 'timing/data': 0.0003215717151761055, 'timing/model': 0.6194732030853629, 'learning_rate/qwen_vl_interface': 9.891373661072172e-06, 'learning_rate/action_model': 9.891373661072173e-05, 'epoch': 0.07})
+[2026-07-28 06:43:08,620][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.0806288942694664, 'train/grad_norm_pre_clip': 2.02282977104187, 'timing/data': 0.0003365585580468178, 'timing/model': 0.6172733660787344, 'learning_rate/qwen_vl_interface': 9.89055854319461e-06, 'learning_rate/action_model': 9.890558543194611e-05, 'epoch': 0.07})
+[2026-07-28 06:43:18,001][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.09749063849449158, 'train/grad_norm_pre_clip': 4.142368316650391, 'timing/data': 0.00035773031413555145, 'timing/model': 0.6163098644465208, 'learning_rate/qwen_vl_interface': 9.889740414108994e-06, 'learning_rate/action_model': 9.889740414108995e-05, 'epoch': 0.07})
+[2026-07-28 06:43:27,337][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.1129227876663208, 'train/grad_norm_pre_clip': 3.512244939804077, 'timing/data': 0.0003315890207886696, 'timing/model': 0.8435666710138321, 'learning_rate/qwen_vl_interface': 9.888919274346196e-06, 'learning_rate/action_model': 9.888919274346195e-05, 'epoch': 0.07})
+[2026-07-28 06:43:36,608][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.09477672725915909, 'train/grad_norm_pre_clip': 1.4007712602615356, 'timing/data': 0.0004132455214858055, 'timing/model': 0.6162334606051445, 'learning_rate/qwen_vl_interface': 9.888095124439043e-06, 'learning_rate/action_model': 9.888095124439043e-05, 'epoch': 0.07})
+[2026-07-28 06:43:45,978][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.10379491746425629, 'train/grad_norm_pre_clip': 4.003896713256836, 'timing/data': 0.0003311038017272949, 'timing/model': 0.6181555157527328, 'learning_rate/qwen_vl_interface': 9.88726796492232e-06, 'learning_rate/action_model': 9.88726796492232e-05, 'epoch': 0.07})
+[2026-07-28 06:43:55,462][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.07541335374116898, 'train/grad_norm_pre_clip': 1.4274722337722778, 'timing/data': 0.0003891279920935631, 'timing/model': 0.6171384481713176, 'learning_rate/qwen_vl_interface': 9.886437796332759e-06, 'learning_rate/action_model': 9.88643779633276e-05, 'epoch': 0.07})
+[2026-07-28 06:44:04,579][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.08702217042446136, 'train/grad_norm_pre_clip': 2.4370360374450684, 'timing/data': 0.0003908742219209671, 'timing/model': 0.6185691952705383, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.07})
+[2026-07-28 06:44:13,704][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.12456735968589783, 'train/grad_norm_pre_clip': 2.1456315517425537, 'timing/data': 0.0003188112750649452, 'timing/model': 0.6189302345737815, 'learning_rate/qwen_vl_interface': 9.884768434091832e-06, 'learning_rate/action_model': 9.884768434091832e-05, 'epoch': 0.07})
+[2026-07-28 06:44:23,272][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.05578691512346268, 'train/grad_norm_pre_clip': 1.2860161066055298, 'timing/data': 0.0004074210301041603, 'timing/model': 0.8321526357904077, 'learning_rate/qwen_vl_interface': 9.883929241523698e-06, 'learning_rate/action_model': 9.883929241523697e-05, 'epoch': 0.07})
+[2026-07-28 06:44:32,219][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.1134495735168457, 'train/grad_norm_pre_clip': 4.598765850067139, 'timing/data': 0.00040425825864076614, 'timing/model': 0.6167643424123526, 'learning_rate/qwen_vl_interface': 9.883087042049185e-06, 'learning_rate/action_model': 9.883087042049186e-05, 'epoch': 0.07})
+[2026-07-28 06:44:41,340][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.10844762623310089, 'train/grad_norm_pre_clip': 2.494805335998535, 'timing/data': 0.00042234640568494797, 'timing/model': 0.617674233391881, 'learning_rate/qwen_vl_interface': 9.882241836214794e-06, 'learning_rate/action_model': 9.882241836214794e-05, 'epoch': 0.07})
+[2026-07-28 06:44:50,457][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.1233481764793396, 'train/grad_norm_pre_clip': 3.208548069000244, 'timing/data': 0.0004107421264052391, 'timing/model': 0.6191193759441376, 'learning_rate/qwen_vl_interface': 9.881393624568965e-06, 'learning_rate/action_model': 9.881393624568965e-05, 'epoch': 0.07})
+[2026-07-28 06:44:59,574][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.11858142912387848, 'train/grad_norm_pre_clip': 3.7236177921295166, 'timing/data': 0.0003204662352800369, 'timing/model': 0.6155305868014693, 'learning_rate/qwen_vl_interface': 9.880542407662098e-06, 'learning_rate/action_model': 9.880542407662097e-05, 'epoch': 0.07})
+[2026-07-28 06:45:08,985][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.09544240683317184, 'train/grad_norm_pre_clip': 2.5047378540039062, 'timing/data': 0.00034758448600769043, 'timing/model': 0.8690170319750905, 'learning_rate/qwen_vl_interface': 9.879688186046535e-06, 'learning_rate/action_model': 9.879688186046534e-05, 'epoch': 0.07})
+[2026-07-28 06:45:18,072][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.12132330983877182, 'train/grad_norm_pre_clip': 2.5227015018463135, 'timing/data': 0.0003262460231781006, 'timing/model': 0.6154781421646476, 'learning_rate/qwen_vl_interface': 9.87883096027657e-06, 'learning_rate/action_model': 9.87883096027657e-05, 'epoch': 0.07})
+[2026-07-28 06:45:27,139][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.13076628744602203, 'train/grad_norm_pre_clip': 2.567178249359131, 'timing/data': 0.0003478648141026497, 'timing/model': 0.6191479759290814, 'learning_rate/qwen_vl_interface': 9.87797073090845e-06, 'learning_rate/action_model': 9.87797073090845e-05, 'epoch': 0.07})
+[2026-07-28 06:45:36,257][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.09146448224782944, 'train/grad_norm_pre_clip': 0.9622531533241272, 'timing/data': 0.0003314148634672165, 'timing/model': 0.620550730265677, 'learning_rate/qwen_vl_interface': 9.877107498500368e-06, 'learning_rate/action_model': 9.877107498500367e-05, 'epoch': 0.07})
+[2026-07-28 06:45:45,390][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.0935264304280281, 'train/grad_norm_pre_clip': 2.6239871978759766, 'timing/data': 0.0003228159621357918, 'timing/model': 0.6165081979706883, 'learning_rate/qwen_vl_interface': 9.876241263612467e-06, 'learning_rate/action_model': 9.876241263612465e-05, 'epoch': 0.07})
+[2026-07-28 06:45:54,760][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.0899757444858551, 'train/grad_norm_pre_clip': 1.4210704565048218, 'timing/data': 0.00032520759850740433, 'timing/model': 0.6195093216374516, 'learning_rate/qwen_vl_interface': 9.875372026806835e-06, 'learning_rate/action_model': 9.875372026806835e-05, 'epoch': 0.07})
+[2026-07-28 06:46:03,897][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.08915852755308151, 'train/grad_norm_pre_clip': 1.7407667636871338, 'timing/data': 0.0003245510160923004, 'timing/model': 0.6181893460452557, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.07})
+[2026-07-28 06:46:13,249][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.08624692261219025, 'train/grad_norm_pre_clip': 4.638806343078613, 'timing/data': 0.00036140717566013336, 'timing/model': 0.6210496723651886, 'learning_rate/qwen_vl_interface': 9.873624549700485e-06, 'learning_rate/action_model': 9.873624549700484e-05, 'epoch': 0.07})
+[2026-07-28 06:46:22,647][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.13392332196235657, 'train/grad_norm_pre_clip': 3.6476006507873535, 'timing/data': 0.00034529995173215866, 'timing/model': 0.6211410481482744, 'learning_rate/qwen_vl_interface': 9.872746310533683e-06, 'learning_rate/action_model': 9.872746310533684e-05, 'epoch': 0.07})
+[2026-07-28 06:46:31,608][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.10517233610153198, 'train/grad_norm_pre_clip': 3.6744189262390137, 'timing/data': 0.0006473362445831299, 'timing/model': 0.61911291629076, 'learning_rate/qwen_vl_interface': 9.871865071716991e-06, 'learning_rate/action_model': 9.87186507171699e-05, 'epoch': 0.07})
+[2026-07-28 06:46:40,523][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.0648718923330307, 'train/grad_norm_pre_clip': 4.157019138336182, 'timing/data': 0.0003301817923784256, 'timing/model': 0.6202497556805611, 'learning_rate/qwen_vl_interface': 9.870980833822231e-06, 'learning_rate/action_model': 9.87098083382223e-05, 'epoch': 0.07})
+[2026-07-28 06:46:49,691][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.11185826361179352, 'train/grad_norm_pre_clip': 3.547088146209717, 'timing/data': 0.000528351403772831, 'timing/model': 0.6281089400872588, 'learning_rate/qwen_vl_interface': 9.87009359742318e-06, 'learning_rate/action_model': 9.87009359742318e-05, 'epoch': 0.07})
+[2026-07-28 06:46:59,296][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.12717542052268982, 'train/grad_norm_pre_clip': 4.641373634338379, 'timing/data': 0.00033247098326683044, 'timing/model': 0.617865345440805, 'learning_rate/qwen_vl_interface': 9.86920336309555e-06, 'learning_rate/action_model': 9.86920336309555e-05, 'epoch': 0.07})
+[2026-07-28 06:47:08,692][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.0853474885225296, 'train/grad_norm_pre_clip': 2.641248941421509, 'timing/data': 0.0003391178324818611, 'timing/model': 0.619657538831234, 'learning_rate/qwen_vl_interface': 9.86831013141701e-06, 'learning_rate/action_model': 9.86831013141701e-05, 'epoch': 0.07})
+[2026-07-28 06:47:17,827][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.06991159915924072, 'train/grad_norm_pre_clip': 6.969385147094727, 'timing/data': 0.00034117139875888824, 'timing/model': 0.6172534953802824, 'learning_rate/qwen_vl_interface': 9.867413902967166e-06, 'learning_rate/action_model': 9.867413902967166e-05, 'epoch': 0.07})
+[2026-07-28 06:47:26,995][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.1354426145553589, 'train/grad_norm_pre_clip': 4.945489883422852, 'timing/data': 0.00037401728332042694, 'timing/model': 0.6189040886238217, 'learning_rate/qwen_vl_interface': 9.86651467832757e-06, 'learning_rate/action_model': 9.86651467832757e-05, 'epoch': 0.07})
+[2026-07-28 06:47:36,449][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.06907418370246887, 'train/grad_norm_pre_clip': 3.1974101066589355, 'timing/data': 0.0004205629229545593, 'timing/model': 0.6176733132451773, 'learning_rate/qwen_vl_interface': 9.865612458081718e-06, 'learning_rate/action_model': 9.865612458081719e-05, 'epoch': 0.07})
+[2026-07-28 06:47:45,587][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.11134733259677887, 'train/grad_norm_pre_clip': 4.836210250854492, 'timing/data': 0.0004090918228030205, 'timing/model': 0.6187049336731434, 'learning_rate/qwen_vl_interface': 9.864707242815057e-06, 'learning_rate/action_model': 9.864707242815056e-05, 'epoch': 0.07})
+[2026-07-28 06:47:55,068][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.07352033257484436, 'train/grad_norm_pre_clip': 1.7449383735656738, 'timing/data': 0.00035253074020147324, 'timing/model': 0.6183474669232965, 'learning_rate/qwen_vl_interface': 9.863799033114966e-06, 'learning_rate/action_model': 9.863799033114966e-05, 'epoch': 0.07})
+[2026-07-28 06:48:04,259][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.054458118975162506, 'train/grad_norm_pre_clip': 1.9695253372192383, 'timing/data': 0.00032881367951631546, 'timing/model': 0.6223634509369731, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.07})
+[2026-07-28 06:48:13,619][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.12734022736549377, 'train/grad_norm_pre_clip': 4.351394176483154, 'timing/data': 0.00032086577266454697, 'timing/model': 0.851090288721025, 'learning_rate/qwen_vl_interface': 9.861973632773748e-06, 'learning_rate/action_model': 9.861973632773748e-05, 'epoch': 0.07})
+[2026-07-28 06:48:22,982][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.07327345013618469, 'train/grad_norm_pre_clip': 3.436774253845215, 'timing/data': 0.00034054089337587357, 'timing/model': 0.6513474145904183, 'learning_rate/qwen_vl_interface': 9.861056443317105e-06, 'learning_rate/action_model': 9.861056443317105e-05, 'epoch': 0.07})
+[2026-07-28 06:48:32,442][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.18355271220207214, 'train/grad_norm_pre_clip': 4.174715518951416, 'timing/data': 0.0003168080002069473, 'timing/model': 0.6192129803821445, 'learning_rate/qwen_vl_interface': 9.860136261796e-06, 'learning_rate/action_model': 9.860136261795998e-05, 'epoch': 0.07})
+[2026-07-28 06:48:41,565][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.11760564893484116, 'train/grad_norm_pre_clip': 3.6859302520751953, 'timing/data': 0.0003516245633363724, 'timing/model': 0.617275538854301, 'learning_rate/qwen_vl_interface': 9.859213088807521e-06, 'learning_rate/action_model': 9.85921308880752e-05, 'epoch': 0.07})
+[2026-07-28 06:48:50,695][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.1650145947933197, 'train/grad_norm_pre_clip': 5.347385883331299, 'timing/data': 0.0003272229805588722, 'timing/model': 0.6178630562499166, 'learning_rate/qwen_vl_interface': 9.858286924950714e-06, 'learning_rate/action_model': 9.858286924950713e-05, 'epoch': 0.07})
+[2026-07-28 06:48:59,787][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.07673071324825287, 'train/grad_norm_pre_clip': 2.1144280433654785, 'timing/data': 0.0003549838438630104, 'timing/model': 0.7896236358210444, 'learning_rate/qwen_vl_interface': 9.85735777082655e-06, 'learning_rate/action_model': 9.857357770826551e-05, 'epoch': 0.07})
+[2026-07-28 06:49:09,175][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.06517767906188965, 'train/grad_norm_pre_clip': 1.317785620689392, 'timing/data': 0.0003393888473510742, 'timing/model': 0.6200844319537282, 'learning_rate/qwen_vl_interface': 9.856425627037951e-06, 'learning_rate/action_model': 9.856425627037952e-05, 'epoch': 0.07})
+[2026-07-28 06:49:18,315][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.08052036166191101, 'train/grad_norm_pre_clip': 3.777376651763916, 'timing/data': 0.0004030987620353699, 'timing/model': 0.6177119668573141, 'learning_rate/qwen_vl_interface': 9.855490494189777e-06, 'learning_rate/action_model': 9.855490494189776e-05, 'epoch': 0.07})
+[2026-07-28 06:49:27,568][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.12978024780750275, 'train/grad_norm_pre_clip': 6.333254337310791, 'timing/data': 0.00046058837324380875, 'timing/model': 0.6189263155683875, 'learning_rate/qwen_vl_interface': 9.854552372888819e-06, 'learning_rate/action_model': 9.854552372888819e-05, 'epoch': 0.07})
+[2026-07-28 06:49:36,979][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.1686621904373169, 'train/grad_norm_pre_clip': 3.5234501361846924, 'timing/data': 0.00036194082349538803, 'timing/model': 0.8545110737904906, 'learning_rate/qwen_vl_interface': 9.853611263743817e-06, 'learning_rate/action_model': 9.853611263743817e-05, 'epoch': 0.08})
+[2026-07-28 06:49:46,332][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.11108752340078354, 'train/grad_norm_pre_clip': 3.5844764709472656, 'timing/data': 0.0003381175920367241, 'timing/model': 0.8481315393000841, 'learning_rate/qwen_vl_interface': 9.85266716736545e-06, 'learning_rate/action_model': 9.852667167365449e-05, 'epoch': 0.08})
+[2026-07-28 06:49:55,481][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.09581182897090912, 'train/grad_norm_pre_clip': 2.7301881313323975, 'timing/data': 0.00039483513683080673, 'timing/model': 0.6175739001482725, 'learning_rate/qwen_vl_interface': 9.851720084366324e-06, 'learning_rate/action_model': 9.851720084366324e-05, 'epoch': 0.08})
+[2026-07-28 06:50:04,651][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.17222929000854492, 'train/grad_norm_pre_clip': 5.321543216705322, 'timing/data': 0.0003317510709166527, 'timing/model': 0.617449326440692, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.08})
+[2026-07-28 06:50:14,081][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.12501439452171326, 'train/grad_norm_pre_clip': 5.1512532234191895, 'timing/data': 0.0003428906202316284, 'timing/model': 0.6195493750274181, 'learning_rate/qwen_vl_interface': 9.849816960965959e-06, 'learning_rate/action_model': 9.849816960965959e-05, 'epoch': 0.08})
+[2026-07-28 06:50:23,458][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.0507960319519043, 'train/grad_norm_pre_clip': 1.9309580326080322, 'timing/data': 0.0003266371786594391, 'timing/model': 0.6164455292746425, 'learning_rate/qwen_vl_interface': 9.848860921799633e-06, 'learning_rate/action_model': 9.848860921799632e-05, 'epoch': 0.08})
+[2026-07-28 06:50:32,553][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.10277169197797775, 'train/grad_norm_pre_clip': 3.3047168254852295, 'timing/data': 0.0003634607419371605, 'timing/model': 0.6170855071395636, 'learning_rate/qwen_vl_interface': 9.847901898482385e-06, 'learning_rate/action_model': 9.847901898482385e-05, 'epoch': 0.08})
+[2026-07-28 06:50:41,941][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.1729731261730194, 'train/grad_norm_pre_clip': 2.516477346420288, 'timing/data': 0.00035973265767097473, 'timing/model': 0.8459511315450072, 'learning_rate/qwen_vl_interface': 9.846939891636512e-06, 'learning_rate/action_model': 9.846939891636512e-05, 'epoch': 0.08})
+[2026-07-28 06:50:51,131][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.09931780397891998, 'train/grad_norm_pre_clip': 1.3605594635009766, 'timing/data': 0.0003281785175204277, 'timing/model': 0.6286359829828143, 'learning_rate/qwen_vl_interface': 9.845974901886253e-06, 'learning_rate/action_model': 9.845974901886252e-05, 'epoch': 0.08})
+[2026-07-28 06:51:00,577][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.11083844304084778, 'train/grad_norm_pre_clip': 2.0755887031555176, 'timing/data': 0.00034087058156728745, 'timing/model': 0.9007321847602725, 'learning_rate/qwen_vl_interface': 9.845006929857777e-06, 'learning_rate/action_model': 9.845006929857776e-05, 'epoch': 0.08})
+[2026-07-28 06:51:09,720][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.09905141592025757, 'train/grad_norm_pre_clip': 1.9775234460830688, 'timing/data': 0.0003331173211336136, 'timing/model': 0.617164134979248, 'learning_rate/qwen_vl_interface': 9.844035976179193e-06, 'learning_rate/action_model': 9.844035976179192e-05, 'epoch': 0.08})
+[2026-07-28 06:51:19,113][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.07564260065555573, 'train/grad_norm_pre_clip': 1.2109423875808716, 'timing/data': 0.000315885990858078, 'timing/model': 0.6167610427364707, 'learning_rate/qwen_vl_interface': 9.843062041480538e-06, 'learning_rate/action_model': 9.843062041480539e-05, 'epoch': 0.08})
+[2026-07-28 06:51:28,450][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.10584083199501038, 'train/grad_norm_pre_clip': 1.3453935384750366, 'timing/data': 0.0003281710669398308, 'timing/model': 0.6167186517268419, 'learning_rate/qwen_vl_interface': 9.842085126393793e-06, 'learning_rate/action_model': 9.842085126393792e-05, 'epoch': 0.08})
+[2026-07-28 06:51:37,568][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.07071797549724579, 'train/grad_norm_pre_clip': 3.1850852966308594, 'timing/data': 0.0003342600539326668, 'timing/model': 0.6297145755961537, 'learning_rate/qwen_vl_interface': 9.841105231552863e-06, 'learning_rate/action_model': 9.841105231552864e-05, 'epoch': 0.08})
+[2026-07-28 06:51:46,719][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.10767056047916412, 'train/grad_norm_pre_clip': 4.134379863739014, 'timing/data': 0.00032422784715890884, 'timing/model': 0.6217260584235191, 'learning_rate/qwen_vl_interface': 9.840122357593594e-06, 'learning_rate/action_model': 9.840122357593594e-05, 'epoch': 0.08})
+[2026-07-28 06:51:56,115][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.06981874257326126, 'train/grad_norm_pre_clip': 2.831594228744507, 'timing/data': 0.0003927517682313919, 'timing/model': 0.6201592935249209, 'learning_rate/qwen_vl_interface': 9.839136505153762e-06, 'learning_rate/action_model': 9.839136505153762e-05, 'epoch': 0.08})
+[2026-07-28 06:52:05,292][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.0756082609295845, 'train/grad_norm_pre_clip': 1.8316426277160645, 'timing/data': 0.0004125218838453293, 'timing/model': 0.6303465431556106, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.08})
+[2026-07-28 06:52:14,438][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.15812624990940094, 'train/grad_norm_pre_clip': 5.159767150878906, 'timing/data': 0.00041237007826566696, 'timing/model': 0.6187455868348479, 'learning_rate/qwen_vl_interface': 9.837155867393176e-06, 'learning_rate/action_model': 9.837155867393175e-05, 'epoch': 0.08})
+[2026-07-28 06:52:23,845][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.15398484468460083, 'train/grad_norm_pre_clip': 4.954563140869141, 'timing/data': 0.00046393461525440216, 'timing/model': 0.8762062061578035, 'learning_rate/qwen_vl_interface': 9.836161083357635e-06, 'learning_rate/action_model': 9.836161083357635e-05, 'epoch': 0.08})
+[2026-07-28 06:52:33,230][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.15795356035232544, 'train/grad_norm_pre_clip': 5.839984893798828, 'timing/data': 0.0003894539549946785, 'timing/model': 0.8493107818067074, 'learning_rate/qwen_vl_interface': 9.835163323411957e-06, 'learning_rate/action_model': 9.835163323411958e-05, 'epoch': 0.08})
+[2026-07-28 06:52:42,336][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.16667649149894714, 'train/grad_norm_pre_clip': 3.425771713256836, 'timing/data': 0.0003889799118041992, 'timing/model': 0.6162013215944171, 'learning_rate/qwen_vl_interface': 9.834162588203581e-06, 'learning_rate/action_model': 9.83416258820358e-05, 'epoch': 0.08})
+[2026-07-28 06:52:51,432][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.12136310338973999, 'train/grad_norm_pre_clip': 7.047240257263184, 'timing/data': 0.00032363925129175186, 'timing/model': 0.616435844451189, 'learning_rate/qwen_vl_interface': 9.833158878381872e-06, 'learning_rate/action_model': 9.83315887838187e-05, 'epoch': 0.08})
+[2026-07-28 06:53:00,557][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.11192931234836578, 'train/grad_norm_pre_clip': 1.9725555181503296, 'timing/data': 0.00033700279891490936, 'timing/model': 0.616959085687995, 'learning_rate/qwen_vl_interface': 9.832152194598123e-06, 'learning_rate/action_model': 9.832152194598122e-05, 'epoch': 0.08})
+[2026-07-28 06:53:09,950][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.1089719608426094, 'train/grad_norm_pre_clip': 1.8616523742675781, 'timing/data': 0.00033007655292749405, 'timing/model': 0.8498050514608622, 'learning_rate/qwen_vl_interface': 9.831142537505563e-06, 'learning_rate/action_model': 9.831142537505562e-05, 'epoch': 0.08})
+[2026-07-28 06:53:19,072][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.12170256674289703, 'train/grad_norm_pre_clip': 4.71510124206543, 'timing/data': 0.00039592012763023376, 'timing/model': 0.6184620084241033, 'learning_rate/qwen_vl_interface': 9.830129907759347e-06, 'learning_rate/action_model': 9.830129907759347e-05, 'epoch': 0.08})
+[2026-07-28 06:53:28,435][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.09459969401359558, 'train/grad_norm_pre_clip': 2.139333724975586, 'timing/data': 0.0003379303961992264, 'timing/model': 0.6175589896738529, 'learning_rate/qwen_vl_interface': 9.82911430601656e-06, 'learning_rate/action_model': 9.82911430601656e-05, 'epoch': 0.08})
+[2026-07-28 06:53:37,578][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.0964559018611908, 'train/grad_norm_pre_clip': 2.96596622467041, 'timing/data': 0.00034783314913511276, 'timing/model': 0.6185837835073471, 'learning_rate/qwen_vl_interface': 9.828095732936216e-06, 'learning_rate/action_model': 9.828095732936216e-05, 'epoch': 0.08})
+[2026-07-28 06:53:46,721][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.10102568566799164, 'train/grad_norm_pre_clip': 2.3115925788879395, 'timing/data': 0.0004825545474886894, 'timing/model': 0.6171943927183747, 'learning_rate/qwen_vl_interface': 9.827074189179254e-06, 'learning_rate/action_model': 9.827074189179253e-05, 'epoch': 0.08})
+[2026-07-28 06:53:55,921][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.1261344850063324, 'train/grad_norm_pre_clip': 4.009852409362793, 'timing/data': 0.0004114219918847084, 'timing/model': 0.6724485652521253, 'learning_rate/qwen_vl_interface': 9.82604967540854e-06, 'learning_rate/action_model': 9.82604967540854e-05, 'epoch': 0.08})
+[2026-07-28 06:54:05,185][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.11160656064748764, 'train/grad_norm_pre_clip': 2.1823346614837646, 'timing/data': 0.0003428282216191292, 'timing/model': 0.6183514762669802, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.08})
+[2026-07-28 06:54:14,554][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.07458727061748505, 'train/grad_norm_pre_clip': 2.1306488513946533, 'timing/data': 0.00039299391210079193, 'timing/model': 0.6225958010181785, 'learning_rate/qwen_vl_interface': 9.823991740486982e-06, 'learning_rate/action_model': 9.823991740486982e-05, 'epoch': 0.08})
+[2026-07-28 06:54:23,698][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.07409171760082245, 'train/grad_norm_pre_clip': 1.5662845373153687, 'timing/data': 0.00033421628177165985, 'timing/model': 0.6186237158253789, 'learning_rate/qwen_vl_interface': 9.822958320671506e-06, 'learning_rate/action_model': 9.822958320671506e-05, 'epoch': 0.08})
+[2026-07-28 06:54:32,835][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.10499522089958191, 'train/grad_norm_pre_clip': 3.0862317085266113, 'timing/data': 0.0003259405493736267, 'timing/model': 0.6174173103645444, 'learning_rate/qwen_vl_interface': 9.821921933513023e-06, 'learning_rate/action_model': 9.821921933513023e-05, 'epoch': 0.08})
+[2026-07-28 06:54:42,102][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.09614819288253784, 'train/grad_norm_pre_clip': 4.177680969238281, 'timing/data': 0.00038897711783647537, 'timing/model': 0.6172824716195464, 'learning_rate/qwen_vl_interface': 9.820882579684037e-06, 'learning_rate/action_model': 9.820882579684037e-05, 'epoch': 0.08})
+[2026-07-28 06:54:51,254][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.10957939922809601, 'train/grad_norm_pre_clip': 5.036886692047119, 'timing/data': 0.00038677919656038284, 'timing/model': 0.6184151107445359, 'learning_rate/qwen_vl_interface': 9.819840259858968e-06, 'learning_rate/action_model': 9.819840259858968e-05, 'epoch': 0.08})
+[2026-07-28 06:55:00,895][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.09684744477272034, 'train/grad_norm_pre_clip': 3.6602325439453125, 'timing/data': 0.0003498885780572891, 'timing/model': 0.8621243014931679, 'learning_rate/qwen_vl_interface': 9.818794974714167e-06, 'learning_rate/action_model': 9.818794974714168e-05, 'epoch': 0.08})
+[2026-07-28 06:55:10,114][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.06362242996692657, 'train/grad_norm_pre_clip': 2.6645660400390625, 'timing/data': 0.0003536716103553772, 'timing/model': 0.6186556089669466, 'learning_rate/qwen_vl_interface': 9.817746724927912e-06, 'learning_rate/action_model': 9.817746724927912e-05, 'epoch': 0.08})
+[2026-07-28 06:55:19,256][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.16298073530197144, 'train/grad_norm_pre_clip': 4.209120273590088, 'timing/data': 0.0005028583109378815, 'timing/model': 0.6180664310231805, 'learning_rate/qwen_vl_interface': 9.8166955111804e-06, 'learning_rate/action_model': 9.816695511180399e-05, 'epoch': 0.08})
+[2026-07-28 06:55:28,643][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.13516756892204285, 'train/grad_norm_pre_clip': 3.0979528427124023, 'timing/data': 0.00034230947494506836, 'timing/model': 0.8133200872689486, 'learning_rate/qwen_vl_interface': 9.815641334153746e-06, 'learning_rate/action_model': 9.815641334153746e-05, 'epoch': 0.08})
+[2026-07-28 06:55:37,818][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.06690145283937454, 'train/grad_norm_pre_clip': 2.0294618606567383, 'timing/data': 0.00034203287214040756, 'timing/model': 0.6176781868562102, 'learning_rate/qwen_vl_interface': 9.814584194532001e-06, 'learning_rate/action_model': 9.814584194532001e-05, 'epoch': 0.08})
+[2026-07-28 06:55:46,974][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.10489365458488464, 'train/grad_norm_pre_clip': 4.119851112365723, 'timing/data': 0.0003933170810341835, 'timing/model': 0.6163959549739957, 'learning_rate/qwen_vl_interface': 9.813524093001128e-06, 'learning_rate/action_model': 9.813524093001129e-05, 'epoch': 0.08})
+[2026-07-28 06:55:56,334][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.08368805795907974, 'train/grad_norm_pre_clip': 1.6513622999191284, 'timing/data': 0.0003233030438423157, 'timing/model': 0.6208652276545763, 'learning_rate/qwen_vl_interface': 9.81246103024902e-06, 'learning_rate/action_model': 9.81246103024902e-05, 'epoch': 0.08})
+[2026-07-28 06:56:05,683][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.08596557378768921, 'train/grad_norm_pre_clip': 1.9475398063659668, 'timing/data': 0.00032469164580106735, 'timing/model': 0.8526442442089319, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.08})
+[2026-07-28 06:56:14,847][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.08957227319478989, 'train/grad_norm_pre_clip': 4.450567722320557, 'timing/data': 0.00038217008113861084, 'timing/model': 0.6180781358852983, 'learning_rate/qwen_vl_interface': 9.810326023842245e-06, 'learning_rate/action_model': 9.810326023842245e-05, 'epoch': 0.08})
+[2026-07-28 06:56:24,030][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.08676813542842865, 'train/grad_norm_pre_clip': 4.809839248657227, 'timing/data': 0.0003347843885421753, 'timing/model': 0.6216266425326467, 'learning_rate/qwen_vl_interface': 9.809254081572963e-06, 'learning_rate/action_model': 9.809254081572961e-05, 'epoch': 0.08})
+[2026-07-28 06:56:33,468][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.10104377567768097, 'train/grad_norm_pre_clip': 2.0702691078186035, 'timing/data': 0.0003444589674472809, 'timing/model': 0.6192617835476995, 'learning_rate/qwen_vl_interface': 9.808179180853204e-06, 'learning_rate/action_model': 9.808179180853204e-05, 'epoch': 0.08})
+[2026-07-28 06:56:42,896][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.09367959201335907, 'train/grad_norm_pre_clip': 2.2807323932647705, 'timing/data': 0.0004101637750864029, 'timing/model': 0.6215739781036973, 'learning_rate/qwen_vl_interface': 9.807101322380464e-06, 'learning_rate/action_model': 9.807101322380464e-05, 'epoch': 0.08})
+[2026-07-28 06:56:52,085][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.09497344493865967, 'train/grad_norm_pre_clip': 1.4389948844909668, 'timing/data': 0.00034809764474630356, 'timing/model': 0.6175235938280821, 'learning_rate/qwen_vl_interface': 9.806020506854152e-06, 'learning_rate/action_model': 9.806020506854153e-05, 'epoch': 0.08})
+[2026-07-28 06:57:01,478][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.1157461404800415, 'train/grad_norm_pre_clip': 3.0891427993774414, 'timing/data': 0.0003282129764556885, 'timing/model': 0.8856914397329092, 'learning_rate/qwen_vl_interface': 9.804936734975596e-06, 'learning_rate/action_model': 9.804936734975596e-05, 'epoch': 0.08})
+[2026-07-28 06:57:10,759][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.08375076949596405, 'train/grad_norm_pre_clip': 2.6167426109313965, 'timing/data': 0.0003341455012559891, 'timing/model': 0.6161544341593981, 'learning_rate/qwen_vl_interface': 9.803850007448046e-06, 'learning_rate/action_model': 9.803850007448045e-05, 'epoch': 0.08})
+[2026-07-28 06:57:20,202][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.10408319532871246, 'train/grad_norm_pre_clip': 3.8045475482940674, 'timing/data': 0.0003260606899857521, 'timing/model': 0.6195539347827435, 'learning_rate/qwen_vl_interface': 9.802760324976666e-06, 'learning_rate/action_model': 9.802760324976667e-05, 'epoch': 0.08})
+[2026-07-28 06:57:29,345][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.0749727338552475, 'train/grad_norm_pre_clip': 2.9796693325042725, 'timing/data': 0.0003260485827922821, 'timing/model': 0.6181256575509906, 'learning_rate/qwen_vl_interface': 9.80166768826854e-06, 'learning_rate/action_model': 9.80166768826854e-05, 'epoch': 0.08})
+[2026-07-28 06:57:38,439][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.06578726321458817, 'train/grad_norm_pre_clip': 2.7307918071746826, 'timing/data': 0.00032631121575832367, 'timing/model': 0.6169948913156986, 'learning_rate/qwen_vl_interface': 9.800572098032668e-06, 'learning_rate/action_model': 9.800572098032668e-05, 'epoch': 0.08})
+[2026-07-28 06:57:47,845][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.08125246316194534, 'train/grad_norm_pre_clip': 5.943574905395508, 'timing/data': 0.00033961888402700424, 'timing/model': 0.6186990616843104, 'learning_rate/qwen_vl_interface': 9.799473554979966e-06, 'learning_rate/action_model': 9.799473554979965e-05, 'epoch': 0.08})
+[2026-07-28 06:57:56,918][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.08311678469181061, 'train/grad_norm_pre_clip': 2.626112461090088, 'timing/data': 0.00033395085483789444, 'timing/model': 0.6164112575352192, 'learning_rate/qwen_vl_interface': 9.798372059823268e-06, 'learning_rate/action_model': 9.798372059823268e-05, 'epoch': 0.09})
+[2026-07-28 06:58:06,006][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.09505381435155869, 'train/grad_norm_pre_clip': 1.4762195348739624, 'timing/data': 0.0003108903765678406, 'timing/model': 0.6210097977891564, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.09})
+[2026-07-28 06:58:15,178][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.12851785123348236, 'train/grad_norm_pre_clip': 3.181342124938965, 'timing/data': 0.0003437921404838562, 'timing/model': 0.6196312103420496, 'learning_rate/qwen_vl_interface': 9.79616021605879e-06, 'learning_rate/action_model': 9.796160216058788e-05, 'epoch': 0.09})
+[2026-07-28 06:58:24,504][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.07070654630661011, 'train/grad_norm_pre_clip': 1.760617733001709, 'timing/data': 0.0003658430650830269, 'timing/model': 0.8528912495821714, 'learning_rate/qwen_vl_interface': 9.795049868886247e-06, 'learning_rate/action_model': 9.795049868886247e-05, 'epoch': 0.09})
+[2026-07-28 06:58:33,869][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.060101475566625595, 'train/grad_norm_pre_clip': 1.4955724477767944, 'timing/data': 0.000425056554377079, 'timing/model': 0.8545733178034425, 'learning_rate/qwen_vl_interface': 9.793936572480192e-06, 'learning_rate/action_model': 9.793936572480192e-05, 'epoch': 0.09})
+[2026-07-28 06:58:43,089][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.06227649003267288, 'train/grad_norm_pre_clip': 1.82904052734375, 'timing/data': 0.0003281906247138977, 'timing/model': 0.6202553613111377, 'learning_rate/qwen_vl_interface': 9.792820327563028e-06, 'learning_rate/action_model': 9.792820327563028e-05, 'epoch': 0.09})
+[2026-07-28 06:58:52,280][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.09896839410066605, 'train/grad_norm_pre_clip': 2.2973525524139404, 'timing/data': 0.00032394565641880035, 'timing/model': 0.6166868824511766, 'learning_rate/qwen_vl_interface': 9.791701134859076e-06, 'learning_rate/action_model': 9.791701134859075e-05, 'epoch': 0.09})
+[2026-07-28 06:59:01,398][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.07031169533729553, 'train/grad_norm_pre_clip': 1.5802617073059082, 'timing/data': 0.0003787800669670105, 'timing/model': 0.6175611056387424, 'learning_rate/qwen_vl_interface': 9.790578995094565e-06, 'learning_rate/action_model': 9.790578995094563e-05, 'epoch': 0.09})
+[2026-07-28 06:59:10,873][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.10029950737953186, 'train/grad_norm_pre_clip': 3.4512245655059814, 'timing/data': 0.00039497390389442444, 'timing/model': 0.8436873452737927, 'learning_rate/qwen_vl_interface': 9.789453908997639e-06, 'learning_rate/action_model': 9.789453908997638e-05, 'epoch': 0.09})
+[2026-07-28 06:59:20,189][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.10673265904188156, 'train/grad_norm_pre_clip': 3.024090051651001, 'timing/data': 0.00031663477420806885, 'timing/model': 0.6173429759219289, 'learning_rate/qwen_vl_interface': 9.788325877298355e-06, 'learning_rate/action_model': 9.788325877298355e-05, 'epoch': 0.09})
+[2026-07-28 06:59:29,331][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.04641781002283096, 'train/grad_norm_pre_clip': 1.1370155811309814, 'timing/data': 0.0003265589475631714, 'timing/model': 0.6176800550892949, 'learning_rate/qwen_vl_interface': 9.787194900728682e-06, 'learning_rate/action_model': 9.787194900728683e-05, 'epoch': 0.09})
+[2026-07-28 06:59:38,451][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.06794197112321854, 'train/grad_norm_pre_clip': 1.5708386898040771, 'timing/data': 0.00033390801399946213, 'timing/model': 0.6176924081519246, 'learning_rate/qwen_vl_interface': 9.786060980022499e-06, 'learning_rate/action_model': 9.786060980022498e-05, 'epoch': 0.09})
+[2026-07-28 06:59:47,796][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.08456985652446747, 'train/grad_norm_pre_clip': 2.265622854232788, 'timing/data': 0.0003864932805299759, 'timing/model': 0.6216372465714812, 'learning_rate/qwen_vl_interface': 9.784924115915592e-06, 'learning_rate/action_model': 9.784924115915592e-05, 'epoch': 0.09})
+[2026-07-28 06:59:56,930][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.061094239354133606, 'train/grad_norm_pre_clip': 2.7539868354797363, 'timing/data': 0.00040305033326148987, 'timing/model': 0.6169852977618575, 'learning_rate/qwen_vl_interface': 9.783784309145663e-06, 'learning_rate/action_model': 9.783784309145661e-05, 'epoch': 0.09})
+[2026-07-28 07:00:06,067][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.06441643834114075, 'train/grad_norm_pre_clip': 1.612561583518982, 'timing/data': 0.0003526564687490463, 'timing/model': 0.6180733218789101, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.09})
+[2026-07-28 07:00:15,435][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.024322673678398132, 'train/grad_norm_pre_clip': 1.7807788848876953, 'timing/data': 0.0003489479422569275, 'timing/model': 0.6204255474731326, 'learning_rate/qwen_vl_interface': 9.78149587057707e-06, 'learning_rate/action_model': 9.78149587057707e-05, 'epoch': 0.09})
+[2026-07-28 07:00:24,858][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.14742036163806915, 'train/grad_norm_pre_clip': 4.450340270996094, 'timing/data': 0.000334358774125576, 'timing/model': 0.6189444763585925, 'learning_rate/qwen_vl_interface': 9.780347240263354e-06, 'learning_rate/action_model': 9.780347240263353e-05, 'epoch': 0.09})
+[2026-07-28 07:00:34,151][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.15602780878543854, 'train/grad_norm_pre_clip': 3.4093141555786133, 'timing/data': 0.0003751572221517563, 'timing/model': 0.616945618763566, 'learning_rate/qwen_vl_interface': 9.779195670256498e-06, 'learning_rate/action_model': 9.779195670256497e-05, 'epoch': 0.09})
+[2026-07-28 07:00:43,248][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.09465522319078445, 'train/grad_norm_pre_clip': 3.317106008529663, 'timing/data': 0.0003846157342195511, 'timing/model': 0.6179190101101995, 'learning_rate/qwen_vl_interface': 9.778041161303746e-06, 'learning_rate/action_model': 9.778041161303745e-05, 'epoch': 0.09})
+[2026-07-28 07:00:52,367][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.09885512292385101, 'train/grad_norm_pre_clip': 3.1021201610565186, 'timing/data': 0.00039803050458431244, 'timing/model': 0.6174589367583394, 'learning_rate/qwen_vl_interface': 9.776883714154242e-06, 'learning_rate/action_model': 9.776883714154241e-05, 'epoch': 0.09})
+[2026-07-28 07:01:01,721][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.08770573139190674, 'train/grad_norm_pre_clip': 2.9020652770996094, 'timing/data': 0.0003919573500752449, 'timing/model': 0.6155438022688031, 'learning_rate/qwen_vl_interface': 9.775723329559044e-06, 'learning_rate/action_model': 9.775723329559044e-05, 'epoch': 0.09})
+[2026-07-28 07:01:10,632][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.06760963797569275, 'train/grad_norm_pre_clip': 2.436781406402588, 'timing/data': 0.00039930734783411026, 'timing/model': 0.6175091788172722, 'learning_rate/qwen_vl_interface': 9.774560008271114e-06, 'learning_rate/action_model': 9.774560008271114e-05, 'epoch': 0.09})
+[2026-07-28 07:01:19,760][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.07148858159780502, 'train/grad_norm_pre_clip': 2.755711078643799, 'timing/data': 0.00035444460809230804, 'timing/model': 0.6198154184967279, 'learning_rate/qwen_vl_interface': 9.773393751045315e-06, 'learning_rate/action_model': 9.773393751045315e-05, 'epoch': 0.09})
+[2026-07-28 07:01:29,015][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.08151152729988098, 'train/grad_norm_pre_clip': 2.4428367614746094, 'timing/data': 0.00038575008511543274, 'timing/model': 0.7871496984735131, 'learning_rate/qwen_vl_interface': 9.772224558638422e-06, 'learning_rate/action_model': 9.772224558638422e-05, 'epoch': 0.09})
+[2026-07-28 07:01:38,357][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.10730122029781342, 'train/grad_norm_pre_clip': 2.7954351902008057, 'timing/data': 0.00035032257437705994, 'timing/model': 0.8464911244809628, 'learning_rate/qwen_vl_interface': 9.77105243180911e-06, 'learning_rate/action_model': 9.771052431809109e-05, 'epoch': 0.09})
+[2026-07-28 07:01:47,458][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.08362403512001038, 'train/grad_norm_pre_clip': 3.4164533615112305, 'timing/data': 0.00048084743320941925, 'timing/model': 0.6182494712993503, 'learning_rate/qwen_vl_interface': 9.76987737131796e-06, 'learning_rate/action_model': 9.76987737131796e-05, 'epoch': 0.09})
+[2026-07-28 07:01:56,602][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.07019583880901337, 'train/grad_norm_pre_clip': 1.1866097450256348, 'timing/data': 0.00033796578645706177, 'timing/model': 0.6176646575331688, 'learning_rate/qwen_vl_interface': 9.768699377927454e-06, 'learning_rate/action_model': 9.768699377927453e-05, 'epoch': 0.09})
+[2026-07-28 07:02:06,243][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.0558219775557518, 'train/grad_norm_pre_clip': 1.495641827583313, 'timing/data': 0.00042422767728567123, 'timing/model': 0.9606834147125483, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.09})
+[2026-07-28 07:02:15,416][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.08960072696208954, 'train/grad_norm_pre_clip': 2.186255693435669, 'timing/data': 0.0003936486318707466, 'timing/model': 0.616885875351727, 'learning_rate/qwen_vl_interface': 9.76633459550783e-06, 'learning_rate/action_model': 9.76633459550783e-05, 'epoch': 0.09})
+[2026-07-28 07:02:24,516][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.07241997867822647, 'train/grad_norm_pre_clip': 1.987596869468689, 'timing/data': 0.0003974679857492447, 'timing/model': 0.61795399710536, 'learning_rate/qwen_vl_interface': 9.765147808013193e-06, 'learning_rate/action_model': 9.765147808013192e-05, 'epoch': 0.09})
+[2026-07-28 07:02:33,868][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.16987130045890808, 'train/grad_norm_pre_clip': 5.111798286437988, 'timing/data': 0.00039079878479242325, 'timing/model': 0.6180773042142391, 'learning_rate/qwen_vl_interface': 9.763958090688165e-06, 'learning_rate/action_model': 9.763958090688164e-05, 'epoch': 0.09})
+[2026-07-28 07:02:43,249][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.1341462880373001, 'train/grad_norm_pre_clip': 2.043635129928589, 'timing/data': 0.00033598579466342926, 'timing/model': 0.6215874524787068, 'learning_rate/qwen_vl_interface': 9.762765444304737e-06, 'learning_rate/action_model': 9.762765444304737e-05, 'epoch': 0.09})
+[2026-07-28 07:02:52,372][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.07406836748123169, 'train/grad_norm_pre_clip': 2.5156643390655518, 'timing/data': 0.00033245980739593506, 'timing/model': 0.6193684292957187, 'learning_rate/qwen_vl_interface': 9.761569869636808e-06, 'learning_rate/action_model': 9.761569869636808e-05, 'epoch': 0.09})
+[2026-07-28 07:03:01,728][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.1723519265651703, 'train/grad_norm_pre_clip': 3.5461134910583496, 'timing/data': 0.00031893420964479446, 'timing/model': 0.6171395611017942, 'learning_rate/qwen_vl_interface': 9.760371367460172e-06, 'learning_rate/action_model': 9.760371367460171e-05, 'epoch': 0.09})
+[2026-07-28 07:03:10,807][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.07976457476615906, 'train/grad_norm_pre_clip': 1.808201551437378, 'timing/data': 0.000318269245326519, 'timing/model': 0.6174223944544792, 'learning_rate/qwen_vl_interface': 9.759169938552524e-06, 'learning_rate/action_model': 9.759169938552524e-05, 'epoch': 0.09})
+[2026-07-28 07:03:20,174][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.11056913435459137, 'train/grad_norm_pre_clip': 2.363633394241333, 'timing/data': 0.0003445819020271301, 'timing/model': 0.6187347834929824, 'learning_rate/qwen_vl_interface': 9.757965583693458e-06, 'learning_rate/action_model': 9.757965583693457e-05, 'epoch': 0.09})
+[2026-07-28 07:03:29,471][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.1109730452299118, 'train/grad_norm_pre_clip': 2.6621997356414795, 'timing/data': 0.0004142727702856064, 'timing/model': 0.7848503449931741, 'learning_rate/qwen_vl_interface': 9.756758303664467e-06, 'learning_rate/action_model': 9.756758303664468e-05, 'epoch': 0.09})
+[2026-07-28 07:03:38,587][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 500
+[2026-07-28 07:03:48,849][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 500
+[2026-07-28 07:06:37,721][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500
+[2026-07-28 07:06:37,722][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.10679058730602264, 'train/grad_norm_pre_clip': 4.517428398132324, 'eval/action_loss/seconds': 10.25863724667579, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.09126710891723633, 'eval/latency_3/loss': 0.09126710891723633, 'eval/flappy/loss': 0.09126710891723633, 'eval/flappy/latency_3/loss': 0.09126710891723633, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 49.130000597983596, 'rl_games_eval/mid_train/mean_length': 412.3, 'rl_games_eval/mid_train/std_reward': 48.54585621156259, 'rl_games_eval/mid_train/std_length': 392.63979676033864, 'rl_games_eval/mid_train/macro_mean_reward': 49.130000597983596, 'rl_games_eval/mid_train/macro_mean_length': 412.3, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 49.130000597983596, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 412.3, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 48.54585621156259, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 392.63979676033864, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 49.130000597983596, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00034506525844335556, 'timing/model': 0.6173278409987688, 'learning_rate/qwen_vl_interface': 9.755548099248942e-06, 'learning_rate/action_model': 9.75554809924894e-05, 'epoch': 0.09})
+[2026-07-28 07:06:37,725][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_500_state
+[2026-07-28 07:06:37,725][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 07:07:15,863][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_500_state/pytorch_model
+[2026-07-28 07:07:15,863][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_500_state/sampler.bin
+[2026-07-28 07:07:15,863][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_500_state/sampler_1.bin
+[2026-07-28 07:07:15,866][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_500_state/random_states_0.pkl
+[2026-07-28 07:07:24,105][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500
+[2026-07-28 07:07:24,105][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 07:07:24,126][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 07:07:33,497][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.10070520639419556, 'train/grad_norm_pre_clip': 2.4009361267089844, 'timing/data': 0.00041549280285835266, 'timing/model': 0.6169857466593385, 'learning_rate/qwen_vl_interface': 9.75433497123217e-06, 'learning_rate/action_model': 9.754334971232171e-05, 'epoch': 0.09})
+[2026-07-28 07:07:42,842][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.09397944062948227, 'train/grad_norm_pre_clip': 2.13088059425354, 'timing/data': 0.00035730190575122833, 'timing/model': 0.8497581779956818, 'learning_rate/qwen_vl_interface': 9.753118920401341e-06, 'learning_rate/action_model': 9.75311892040134e-05, 'epoch': 0.09})
+[2026-07-28 07:07:51,999][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.0469755083322525, 'train/grad_norm_pre_clip': 0.8849805593490601, 'timing/data': 0.0004042433574795723, 'timing/model': 0.6210446069017053, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.09})
+[2026-07-28 07:08:01,186][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.08489353209733963, 'train/grad_norm_pre_clip': 4.977962017059326, 'timing/data': 0.0003303978592157364, 'timing/model': 0.6180255142971873, 'learning_rate/qwen_vl_interface': 9.750678053455725e-06, 'learning_rate/action_model': 9.750678053455725e-05, 'epoch': 0.09})
+[2026-07-28 07:08:10,552][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.05500912666320801, 'train/grad_norm_pre_clip': 2.111128330230713, 'timing/data': 0.0003213007003068924, 'timing/model': 0.6178037030622363, 'learning_rate/qwen_vl_interface': 9.749453238924793e-06, 'learning_rate/action_model': 9.749453238924792e-05, 'epoch': 0.09})
+[2026-07-28 07:08:19,878][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.10765537619590759, 'train/grad_norm_pre_clip': 3.11604905128479, 'timing/data': 0.00033726543188095093, 'timing/model': 0.6158074093982577, 'learning_rate/qwen_vl_interface': 9.7482255047475e-06, 'learning_rate/action_model': 9.7482255047475e-05, 'epoch': 0.09})
+[2026-07-28 07:08:28,981][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.05592465400695801, 'train/grad_norm_pre_clip': 2.542872905731201, 'timing/data': 0.0003229128196835518, 'timing/model': 0.6165309650823474, 'learning_rate/qwen_vl_interface': 9.746994851720518e-06, 'learning_rate/action_model': 9.746994851720517e-05, 'epoch': 0.09})
+[2026-07-28 07:08:38,136][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.055459827184677124, 'train/grad_norm_pre_clip': 2.16127347946167, 'timing/data': 0.0003311010077595711, 'timing/model': 0.6186355231329799, 'learning_rate/qwen_vl_interface': 9.745761280642397e-06, 'learning_rate/action_model': 9.745761280642397e-05, 'epoch': 0.09})
+[2026-07-28 07:08:47,515][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.03305836394429207, 'train/grad_norm_pre_clip': 1.8302432298660278, 'timing/data': 0.00032938458025455475, 'timing/model': 0.8671603547409177, 'learning_rate/qwen_vl_interface': 9.744524792313591e-06, 'learning_rate/action_model': 9.74452479231359e-05, 'epoch': 0.09})
+[2026-07-28 07:08:56,654][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.1346059888601303, 'train/grad_norm_pre_clip': 3.564377784729004, 'timing/data': 0.00034787319600582123, 'timing/model': 0.6169534316286445, 'learning_rate/qwen_vl_interface': 9.743285387536445e-06, 'learning_rate/action_model': 9.743285387536444e-05, 'epoch': 0.09})
+[2026-07-28 07:09:05,767][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.09734000265598297, 'train/grad_norm_pre_clip': 2.81522536277771, 'timing/data': 0.0003476254642009735, 'timing/model': 0.6191059313714504, 'learning_rate/qwen_vl_interface': 9.742043067115193e-06, 'learning_rate/action_model': 9.742043067115192e-05, 'epoch': 0.09})
+[2026-07-28 07:09:14,987][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.07401323318481445, 'train/grad_norm_pre_clip': 2.442408561706543, 'timing/data': 0.0003240639343857765, 'timing/model': 0.6808247826993465, 'learning_rate/qwen_vl_interface': 9.740797831855965e-06, 'learning_rate/action_model': 9.740797831855965e-05, 'epoch': 0.09})
+[2026-07-28 07:09:24,297][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.08924373984336853, 'train/grad_norm_pre_clip': 2.348010778427124, 'timing/data': 0.00033507682383060455, 'timing/model': 0.617989213205874, 'learning_rate/qwen_vl_interface': 9.73954968256678e-06, 'learning_rate/action_model': 9.73954968256678e-05, 'epoch': 0.09})
+[2026-07-28 07:09:33,680][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.05748419463634491, 'train/grad_norm_pre_clip': 2.6148300170898438, 'timing/data': 0.00031970907002687454, 'timing/model': 0.6175587922334671, 'learning_rate/qwen_vl_interface': 9.738298620057547e-06, 'learning_rate/action_model': 9.738298620057546e-05, 'epoch': 0.09})
+[2026-07-28 07:09:42,808][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.10912767052650452, 'train/grad_norm_pre_clip': 2.115283727645874, 'timing/data': 0.00034892279654741287, 'timing/model': 0.6165197249501944, 'learning_rate/qwen_vl_interface': 9.73704464514007e-06, 'learning_rate/action_model': 9.737044645140068e-05, 'epoch': 0.09})
+[2026-07-28 07:09:51,895][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.08788622915744781, 'train/grad_norm_pre_clip': 2.638066053390503, 'timing/data': 0.00033336319029331207, 'timing/model': 0.6176826916635036, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.09})
+[2026-07-28 07:10:01,029][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.10911858826875687, 'train/grad_norm_pre_clip': 3.011096477508545, 'timing/data': 0.00034269876778125763, 'timing/model': 0.6189125450327992, 'learning_rate/qwen_vl_interface': 9.734527961337028e-06, 'learning_rate/action_model': 9.734527961337028e-05, 'epoch': 0.09})
+[2026-07-28 07:10:10,388][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.07393258064985275, 'train/grad_norm_pre_clip': 1.1926777362823486, 'timing/data': 0.0003375271335244179, 'timing/model': 0.8619040949270129, 'learning_rate/qwen_vl_interface': 9.733265254084514e-06, 'learning_rate/action_model': 9.733265254084513e-05, 'epoch': 0.1})
+[2026-07-28 07:10:19,752][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.07555459439754486, 'train/grad_norm_pre_clip': 1.1598052978515625, 'timing/data': 0.0004068193957209587, 'timing/model': 0.859236303716898, 'learning_rate/qwen_vl_interface': 9.73199963768985e-06, 'learning_rate/action_model': 9.731999637689849e-05, 'epoch': 0.1})
+[2026-07-28 07:10:28,873][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.08849867433309555, 'train/grad_norm_pre_clip': 1.2676137685775757, 'timing/data': 0.0004256702959537506, 'timing/model': 0.6179032670333982, 'learning_rate/qwen_vl_interface': 9.73073111297428e-06, 'learning_rate/action_model': 9.73073111297428e-05, 'epoch': 0.1})
+[2026-07-28 07:10:38,021][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.09112244844436646, 'train/grad_norm_pre_clip': 1.6627323627471924, 'timing/data': 0.00032437872141599655, 'timing/model': 0.6187460925430059, 'learning_rate/qwen_vl_interface': 9.72945968076094e-06, 'learning_rate/action_model': 9.729459680760939e-05, 'epoch': 0.1})
+[2026-07-28 07:10:47,145][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.04982713982462883, 'train/grad_norm_pre_clip': 1.1328628063201904, 'timing/data': 0.0004097083583474159, 'timing/model': 0.6161378752440214, 'learning_rate/qwen_vl_interface': 9.728185341874844e-06, 'learning_rate/action_model': 9.728185341874844e-05, 'epoch': 0.1})
+[2026-07-28 07:10:56,525][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.052366144955158234, 'train/grad_norm_pre_clip': 2.174684762954712, 'timing/data': 0.00033375155180692673, 'timing/model': 0.621767258271575, 'learning_rate/qwen_vl_interface': 9.7269080971429e-06, 'learning_rate/action_model': 9.726908097142899e-05, 'epoch': 0.1})
+[2026-07-28 07:11:05,841][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.14736127853393555, 'train/grad_norm_pre_clip': 3.4709177017211914, 'timing/data': 0.0004003634676337242, 'timing/model': 0.6181962648406625, 'learning_rate/qwen_vl_interface': 9.725627947393896e-06, 'learning_rate/action_model': 9.725627947393896e-05, 'epoch': 0.1})
+[2026-07-28 07:11:14,738][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.055858246982097626, 'train/grad_norm_pre_clip': 2.491468667984009, 'timing/data': 0.00033223628997802734, 'timing/model': 0.6175421653315425, 'learning_rate/qwen_vl_interface': 9.72434489345851e-06, 'learning_rate/action_model': 9.72434489345851e-05, 'epoch': 0.1})
+[2026-07-28 07:11:23,829][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.09498560428619385, 'train/grad_norm_pre_clip': 3.6234309673309326, 'timing/data': 0.00039935018867254257, 'timing/model': 0.615753335878253, 'learning_rate/qwen_vl_interface': 9.723058936169298e-06, 'learning_rate/action_model': 9.723058936169297e-05, 'epoch': 0.1})
+[2026-07-28 07:11:33,201][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.1283053755760193, 'train/grad_norm_pre_clip': 6.474320411682129, 'timing/data': 0.00040777865797281265, 'timing/model': 0.6201475765556097, 'learning_rate/qwen_vl_interface': 9.721770076360707e-06, 'learning_rate/action_model': 9.721770076360706e-05, 'epoch': 0.1})
+[2026-07-28 07:11:42,547][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.057905253022909164, 'train/grad_norm_pre_clip': 2.835355281829834, 'timing/data': 0.00035141967236995697, 'timing/model': 0.6181696113198996, 'learning_rate/qwen_vl_interface': 9.720478314869061e-06, 'learning_rate/action_model': 9.72047831486906e-05, 'epoch': 0.1})
+[2026-07-28 07:11:51,678][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.11994920670986176, 'train/grad_norm_pre_clip': 4.268604278564453, 'timing/data': 0.0003574499860405922, 'timing/model': 0.6179526904597878, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.1})
+[2026-07-28 07:12:00,815][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.1242360770702362, 'train/grad_norm_pre_clip': 2.4936070442199707, 'timing/data': 0.0004183342680335045, 'timing/model': 0.6189095173031092, 'learning_rate/qwen_vl_interface': 9.717886090191334e-06, 'learning_rate/action_model': 9.717886090191333e-05, 'epoch': 0.1})
+[2026-07-28 07:12:10,366][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.05638944357633591, 'train/grad_norm_pre_clip': 2.0336530208587646, 'timing/data': 0.00031980499625205994, 'timing/model': 0.8169490098953247, 'learning_rate/qwen_vl_interface': 9.716585628687317e-06, 'learning_rate/action_model': 9.716585628687316e-05, 'epoch': 0.1})
+[2026-07-28 07:12:19,499][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.08544397354125977, 'train/grad_norm_pre_clip': 2.86924147605896, 'timing/data': 0.0004469584673643112, 'timing/model': 0.6190064530819654, 'learning_rate/qwen_vl_interface': 9.715282268864379e-06, 'learning_rate/action_model': 9.715282268864378e-05, 'epoch': 0.1})
+[2026-07-28 07:12:28,588][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.1005048155784607, 'train/grad_norm_pre_clip': 2.0249972343444824, 'timing/data': 0.0004092017188668251, 'timing/model': 0.616532820276916, 'learning_rate/qwen_vl_interface': 9.713976011568256e-06, 'learning_rate/action_model': 9.713976011568256e-05, 'epoch': 0.1})
+[2026-07-28 07:12:37,715][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.10418856889009476, 'train/grad_norm_pre_clip': 2.160827159881592, 'timing/data': 0.00040596816688776016, 'timing/model': 0.6210055723786354, 'learning_rate/qwen_vl_interface': 9.712666857646565e-06, 'learning_rate/action_model': 9.712666857646564e-05, 'epoch': 0.1})
+[2026-07-28 07:12:47,273][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.08525919914245605, 'train/grad_norm_pre_clip': 2.036017656326294, 'timing/data': 0.0003197677433490753, 'timing/model': 0.8586385268718004, 'learning_rate/qwen_vl_interface': 9.711354807948798e-06, 'learning_rate/action_model': 9.711354807948798e-05, 'epoch': 0.1})
+[2026-07-28 07:12:56,375][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.08146802335977554, 'train/grad_norm_pre_clip': 2.9394328594207764, 'timing/data': 0.0003187675029039383, 'timing/model': 0.6173818008974195, 'learning_rate/qwen_vl_interface': 9.710039863326334e-06, 'learning_rate/action_model': 9.710039863326334e-05, 'epoch': 0.1})
+[2026-07-28 07:13:05,468][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.0866440087556839, 'train/grad_norm_pre_clip': 1.6086984872817993, 'timing/data': 0.0003286590799689293, 'timing/model': 0.6181263392791152, 'learning_rate/qwen_vl_interface': 9.708722024632425e-06, 'learning_rate/action_model': 9.708722024632424e-05, 'epoch': 0.1})
+[2026-07-28 07:13:14,557][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.058626044541597366, 'train/grad_norm_pre_clip': 1.4104061126708984, 'timing/data': 0.000403732992708683, 'timing/model': 0.6175480931997299, 'learning_rate/qwen_vl_interface': 9.7074012927222e-06, 'learning_rate/action_model': 9.7074012927222e-05, 'epoch': 0.1})
+[2026-07-28 07:13:23,937][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.07670759409666061, 'train/grad_norm_pre_clip': 3.4159293174743652, 'timing/data': 0.00039675459265708923, 'timing/model': 0.8400334697216749, 'learning_rate/qwen_vl_interface': 9.706077668452671e-06, 'learning_rate/action_model': 9.706077668452671e-05, 'epoch': 0.1})
+[2026-07-28 07:13:33,063][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.09230190515518188, 'train/grad_norm_pre_clip': 2.499108076095581, 'timing/data': 0.00037417467683553696, 'timing/model': 0.6177091915160418, 'learning_rate/qwen_vl_interface': 9.704751152682722e-06, 'learning_rate/action_model': 9.704751152682721e-05, 'epoch': 0.1})
+[2026-07-28 07:13:42,216][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.06275458633899689, 'train/grad_norm_pre_clip': 1.5660611391067505, 'timing/data': 0.0004007536917924881, 'timing/model': 0.6184618454426527, 'learning_rate/qwen_vl_interface': 9.703421746273115e-06, 'learning_rate/action_model': 9.703421746273114e-05, 'epoch': 0.1})
+[2026-07-28 07:13:51,378][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.10708288848400116, 'train/grad_norm_pre_clip': 2.6402792930603027, 'timing/data': 0.00034815724939107895, 'timing/model': 0.616041055880487, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.1})
+[2026-07-28 07:14:00,797][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.08865533769130707, 'train/grad_norm_pre_clip': 2.7456586360931396, 'timing/data': 0.00039481744170188904, 'timing/model': 0.6207418637350202, 'learning_rate/qwen_vl_interface': 9.700754264987347e-06, 'learning_rate/action_model': 9.700754264987347e-05, 'epoch': 0.1})
+[2026-07-28 07:14:09,925][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.11681988835334778, 'train/grad_norm_pre_clip': 3.153268814086914, 'timing/data': 0.0006252974271774292, 'timing/model': 0.6169450124725699, 'learning_rate/qwen_vl_interface': 9.69941619184209e-06, 'learning_rate/action_model': 9.699416191842089e-05, 'epoch': 0.1})
+[2026-07-28 07:14:19,307][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.10366815328598022, 'train/grad_norm_pre_clip': 2.2117457389831543, 'timing/data': 0.0003287447616457939, 'timing/model': 0.6203225012868643, 'learning_rate/qwen_vl_interface': 9.69807523151897e-06, 'learning_rate/action_model': 9.698075231518969e-05, 'epoch': 0.1})
+[2026-07-28 07:14:28,586][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.1015411764383316, 'train/grad_norm_pre_clip': 2.5307981967926025, 'timing/data': 0.0003298511728644371, 'timing/model': 0.7852995870634913, 'learning_rate/qwen_vl_interface': 9.696731384888125e-06, 'learning_rate/action_model': 9.696731384888125e-05, 'epoch': 0.1})
+[2026-07-28 07:14:37,734][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.08615527302026749, 'train/grad_norm_pre_clip': 2.162978410720825, 'timing/data': 0.00032539572566747665, 'timing/model': 0.617219801992178, 'learning_rate/qwen_vl_interface': 9.695384652821562e-06, 'learning_rate/action_model': 9.695384652821561e-05, 'epoch': 0.1})
+[2026-07-28 07:14:46,835][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.10717397183179855, 'train/grad_norm_pre_clip': 1.9299534559249878, 'timing/data': 0.00032265111804008484, 'timing/model': 0.6189124993979931, 'learning_rate/qwen_vl_interface': 9.694035036193162e-06, 'learning_rate/action_model': 9.694035036193161e-05, 'epoch': 0.1})
+[2026-07-28 07:14:56,180][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.06688357889652252, 'train/grad_norm_pre_clip': 2.143333673477173, 'timing/data': 0.00032540131360292435, 'timing/model': 0.8471972346305847, 'learning_rate/qwen_vl_interface': 9.692682535878674e-06, 'learning_rate/action_model': 9.692682535878673e-05, 'epoch': 0.1})
+[2026-07-28 07:15:05,554][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.1052098274230957, 'train/grad_norm_pre_clip': 5.0496826171875, 'timing/data': 0.0003439858555793762, 'timing/model': 0.8537857681512833, 'learning_rate/qwen_vl_interface': 9.691327152755721e-06, 'learning_rate/action_model': 9.691327152755722e-05, 'epoch': 0.1})
+[2026-07-28 07:15:14,685][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.09519851207733154, 'train/grad_norm_pre_clip': 1.6971158981323242, 'timing/data': 0.00034099724143743515, 'timing/model': 0.6201038779690862, 'learning_rate/qwen_vl_interface': 9.6899688877038e-06, 'learning_rate/action_model': 9.689968887703798e-05, 'epoch': 0.1})
+[2026-07-28 07:15:23,813][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.061603009700775146, 'train/grad_norm_pre_clip': 1.6387715339660645, 'timing/data': 0.00042602699249982834, 'timing/model': 0.6181851541623473, 'learning_rate/qwen_vl_interface': 9.688607741604269e-06, 'learning_rate/action_model': 9.688607741604269e-05, 'epoch': 0.1})
+[2026-07-28 07:15:33,148][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.08730901032686234, 'train/grad_norm_pre_clip': 1.755948543548584, 'timing/data': 0.0003931913524866104, 'timing/model': 0.6164814522489905, 'learning_rate/qwen_vl_interface': 9.687243715340365e-06, 'learning_rate/action_model': 9.687243715340364e-05, 'epoch': 0.1})
+[2026-07-28 07:15:42,560][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.07516464591026306, 'train/grad_norm_pre_clip': 1.7599661350250244, 'timing/data': 0.0003733942285180092, 'timing/model': 0.6211005691438913, 'learning_rate/qwen_vl_interface': 9.685876809797188e-06, 'learning_rate/action_model': 9.685876809797188e-05, 'epoch': 0.1})
+[2026-07-28 07:15:51,687][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.07118101418018341, 'train/grad_norm_pre_clip': 1.2106069326400757, 'timing/data': 0.0003313012421131134, 'timing/model': 0.6172289047390223, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.1})
+[2026-07-28 07:16:01,050][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.07199391722679138, 'train/grad_norm_pre_clip': 1.356750726699829, 'timing/data': 0.0003276318311691284, 'timing/model': 0.8484984999522567, 'learning_rate/qwen_vl_interface': 9.683134364422766e-06, 'learning_rate/action_model': 9.683134364422765e-05, 'epoch': 0.1})
+[2026-07-28 07:16:10,167][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.09677808731794357, 'train/grad_norm_pre_clip': 1.457133412361145, 'timing/data': 0.0003323471173644066, 'timing/model': 0.6173330191522837, 'learning_rate/qwen_vl_interface': 9.681758826371064e-06, 'learning_rate/action_model': 9.681758826371063e-05, 'epoch': 0.1})
+[2026-07-28 07:16:19,511][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.0990164577960968, 'train/grad_norm_pre_clip': 5.112559795379639, 'timing/data': 0.0003243470564484596, 'timing/model': 0.6175615834072232, 'learning_rate/qwen_vl_interface': 9.680380412599174e-06, 'learning_rate/action_model': 9.680380412599173e-05, 'epoch': 0.1})
+[2026-07-28 07:16:28,623][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.06954070925712585, 'train/grad_norm_pre_clip': 2.6405186653137207, 'timing/data': 0.00033422932028770447, 'timing/model': 0.6172296851873398, 'learning_rate/qwen_vl_interface': 9.678999124001533e-06, 'learning_rate/action_model': 9.678999124001533e-05, 'epoch': 0.1})
+[2026-07-28 07:16:37,740][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.09548035264015198, 'train/grad_norm_pre_clip': 1.506081223487854, 'timing/data': 0.0003910595551133156, 'timing/model': 0.6154273236170411, 'learning_rate/qwen_vl_interface': 9.67761496147445e-06, 'learning_rate/action_model': 9.677614961474449e-05, 'epoch': 0.1})
+[2026-07-28 07:16:47,105][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.11664901673793793, 'train/grad_norm_pre_clip': 1.8563657999038696, 'timing/data': 0.00036039575934410095, 'timing/model': 0.6217066245153546, 'learning_rate/qwen_vl_interface': 9.676227925916082e-06, 'learning_rate/action_model': 9.676227925916082e-05, 'epoch': 0.1})
+[2026-07-28 07:16:56,467][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.14328275620937347, 'train/grad_norm_pre_clip': 3.1249163150787354, 'timing/data': 0.000352972187101841, 'timing/model': 0.8560701366513968, 'learning_rate/qwen_vl_interface': 9.674838018226472e-06, 'learning_rate/action_model': 9.674838018226471e-05, 'epoch': 0.1})
+[2026-07-28 07:17:05,593][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.07796433568000793, 'train/grad_norm_pre_clip': 1.9556169509887695, 'timing/data': 0.0004214821383357048, 'timing/model': 0.6211188640445471, 'learning_rate/qwen_vl_interface': 9.67344523930751e-06, 'learning_rate/action_model': 9.673445239307511e-05, 'epoch': 0.1})
+[2026-07-28 07:17:14,777][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.06742221117019653, 'train/grad_norm_pre_clip': 1.2322945594787598, 'timing/data': 0.0003822939470410347, 'timing/model': 0.6187882712110877, 'learning_rate/qwen_vl_interface': 9.672049590062959e-06, 'learning_rate/action_model': 9.672049590062958e-05, 'epoch': 0.1})
+[2026-07-28 07:17:23,932][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.07927212119102478, 'train/grad_norm_pre_clip': 1.5229358673095703, 'timing/data': 0.0003960048779845238, 'timing/model': 0.617768332362175, 'learning_rate/qwen_vl_interface': 9.670651071398436e-06, 'learning_rate/action_model': 9.670651071398436e-05, 'epoch': 0.1})
+[2026-07-28 07:17:33,283][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.14161311089992523, 'train/grad_norm_pre_clip': 3.0943429470062256, 'timing/data': 0.0003269612789154053, 'timing/model': 0.842514798976481, 'learning_rate/qwen_vl_interface': 9.669249684221427e-06, 'learning_rate/action_model': 9.669249684221428e-05, 'epoch': 0.1})
+[2026-07-28 07:17:42,395][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.05650348588824272, 'train/grad_norm_pre_clip': 1.9052122831344604, 'timing/data': 0.00033371616154909134, 'timing/model': 0.6181419203057885, 'learning_rate/qwen_vl_interface': 9.66784542944128e-06, 'learning_rate/action_model': 9.66784542944128e-05, 'epoch': 0.1})
+[2026-07-28 07:17:51,743][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.10029464215040207, 'train/grad_norm_pre_clip': 0.9242041110992432, 'timing/data': 0.0003215465694665909, 'timing/model': 0.6188419051468372, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.1})
+[2026-07-28 07:18:00,868][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.10142989456653595, 'train/grad_norm_pre_clip': 1.1539098024368286, 'timing/data': 0.00032950565218925476, 'timing/model': 0.6192108700051904, 'learning_rate/qwen_vl_interface': 9.66502832071824e-06, 'learning_rate/action_model': 9.66502832071824e-05, 'epoch': 0.1})
+[2026-07-28 07:18:09,985][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.10134267807006836, 'train/grad_norm_pre_clip': 1.2372490167617798, 'timing/data': 0.000322527252137661, 'timing/model': 0.6158950505778193, 'learning_rate/qwen_vl_interface': 9.663615468603344e-06, 'learning_rate/action_model': 9.663615468603343e-05, 'epoch': 0.1})
+[2026-07-28 07:18:19,336][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.06826919317245483, 'train/grad_norm_pre_clip': 0.9342677593231201, 'timing/data': 0.0005209818482398987, 'timing/model': 0.6185373943299055, 'learning_rate/qwen_vl_interface': 9.662199752541287e-06, 'learning_rate/action_model': 9.662199752541287e-05, 'epoch': 0.1})
+[2026-07-28 07:18:28,451][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.07487320899963379, 'train/grad_norm_pre_clip': 2.166762590408325, 'timing/data': 0.0004392443224787712, 'timing/model': 0.6190753784030676, 'learning_rate/qwen_vl_interface': 9.660781173450711e-06, 'learning_rate/action_model': 9.660781173450712e-05, 'epoch': 0.11})
+[2026-07-28 07:18:37,577][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.07836014032363892, 'train/grad_norm_pre_clip': 1.7056694030761719, 'timing/data': 0.0003778841346502304, 'timing/model': 0.6188783152028918, 'learning_rate/qwen_vl_interface': 9.659359732252121e-06, 'learning_rate/action_model': 9.659359732252121e-05, 'epoch': 0.11})
+[2026-07-28 07:18:46,977][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.08314766734838486, 'train/grad_norm_pre_clip': 1.9436784982681274, 'timing/data': 0.000360758975148201, 'timing/model': 0.6203061379492283, 'learning_rate/qwen_vl_interface': 9.657935429867872e-06, 'learning_rate/action_model': 9.657935429867873e-05, 'epoch': 0.11})
+[2026-07-28 07:18:56,131][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.11038811504840851, 'train/grad_norm_pre_clip': 3.1498260498046875, 'timing/data': 0.00032476335763931274, 'timing/model': 0.6184120494872332, 'learning_rate/qwen_vl_interface': 9.65650826722218e-06, 'learning_rate/action_model': 9.656508267222179e-05, 'epoch': 0.11})
+[2026-07-28 07:19:05,252][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.12378518283367157, 'train/grad_norm_pre_clip': 2.714695692062378, 'timing/data': 0.00033811386674642563, 'timing/model': 0.616167507134378, 'learning_rate/qwen_vl_interface': 9.655078245241114e-06, 'learning_rate/action_model': 9.655078245241112e-05, 'epoch': 0.11})
+[2026-07-28 07:19:14,385][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.11456477642059326, 'train/grad_norm_pre_clip': 2.254725217819214, 'timing/data': 0.0003320220857858658, 'timing/model': 0.6180572090670466, 'learning_rate/qwen_vl_interface': 9.653645364852598e-06, 'learning_rate/action_model': 9.653645364852597e-05, 'epoch': 0.11})
+[2026-07-28 07:19:23,741][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.11545728147029877, 'train/grad_norm_pre_clip': 1.7559674978256226, 'timing/data': 0.00032434798777103424, 'timing/model': 0.617776757106185, 'learning_rate/qwen_vl_interface': 9.652209626986414e-06, 'learning_rate/action_model': 9.652209626986413e-05, 'epoch': 0.11})
+[2026-07-28 07:19:33,089][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.07378112524747849, 'train/grad_norm_pre_clip': 2.281769275665283, 'timing/data': 0.00032318104058504105, 'timing/model': 0.6190744042396545, 'learning_rate/qwen_vl_interface': 9.6507710325742e-06, 'learning_rate/action_model': 9.6507710325742e-05, 'epoch': 0.11})
+[2026-07-28 07:19:42,215][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.08397963643074036, 'train/grad_norm_pre_clip': 2.85795259475708, 'timing/data': 0.00031823012977838516, 'timing/model': 0.6190208289772272, 'learning_rate/qwen_vl_interface': 9.649329582549439e-06, 'learning_rate/action_model': 9.649329582549439e-05, 'epoch': 0.11})
+[2026-07-28 07:19:51,344][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.11065201461315155, 'train/grad_norm_pre_clip': 2.86555814743042, 'timing/data': 0.0003305841237306595, 'timing/model': 0.6223665103316307, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.11})
+[2026-07-28 07:20:00,684][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.13276848196983337, 'train/grad_norm_pre_clip': 2.817087411880493, 'timing/data': 0.0004651341587305069, 'timing/model': 0.6230622911825776, 'learning_rate/qwen_vl_interface': 9.646438119405503e-06, 'learning_rate/action_model': 9.646438119405503e-05, 'epoch': 0.11})
+[2026-07-28 07:20:10,089][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.07433371245861053, 'train/grad_norm_pre_clip': 2.3920488357543945, 'timing/data': 0.0003442317247390747, 'timing/model': 0.6194452531635761, 'learning_rate/qwen_vl_interface': 9.644988108162566e-06, 'learning_rate/action_model': 9.644988108162567e-05, 'epoch': 0.11})
+[2026-07-28 07:20:19,219][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.10076067596673965, 'train/grad_norm_pre_clip': 2.4567298889160156, 'timing/data': 0.0003960737958550453, 'timing/model': 0.61757074855268, 'learning_rate/qwen_vl_interface': 9.643535245059562e-06, 'learning_rate/action_model': 9.643535245059561e-05, 'epoch': 0.11})
+[2026-07-28 07:20:28,381][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.06874597072601318, 'train/grad_norm_pre_clip': 2.3389298915863037, 'timing/data': 0.00032717734575271606, 'timing/model': 0.6237643947824836, 'learning_rate/qwen_vl_interface': 9.642079531039235e-06, 'learning_rate/action_model': 9.642079531039235e-05, 'epoch': 0.11})
+[2026-07-28 07:20:37,890][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.07902561128139496, 'train/grad_norm_pre_clip': 1.6048599481582642, 'timing/data': 0.0003477642312645912, 'timing/model': 1.0047756480053067, 'learning_rate/qwen_vl_interface': 9.640620967046186e-06, 'learning_rate/action_model': 9.640620967046186e-05, 'epoch': 0.11})
+[2026-07-28 07:20:47,020][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.06686481088399887, 'train/grad_norm_pre_clip': 2.4900577068328857, 'timing/data': 0.0003485642373561859, 'timing/model': 0.6175137581303716, 'learning_rate/qwen_vl_interface': 9.63915955402686e-06, 'learning_rate/action_model': 9.63915955402686e-05, 'epoch': 0.11})
+[2026-07-28 07:20:56,151][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.057056158781051636, 'train/grad_norm_pre_clip': 1.5691949129104614, 'timing/data': 0.0003277398645877838, 'timing/model': 0.6175128659233451, 'learning_rate/qwen_vl_interface': 9.637695292929552e-06, 'learning_rate/action_model': 9.637695292929552e-05, 'epoch': 0.11})
+[2026-07-28 07:21:05,308][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.08372501283884048, 'train/grad_norm_pre_clip': 2.0807089805603027, 'timing/data': 0.0003276001662015915, 'timing/model': 0.6183302300050855, 'learning_rate/qwen_vl_interface': 9.636228184704405e-06, 'learning_rate/action_model': 9.636228184704405e-05, 'epoch': 0.11})
+[2026-07-28 07:21:14,690][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.10507754981517792, 'train/grad_norm_pre_clip': 3.56355881690979, 'timing/data': 0.00032306183129549026, 'timing/model': 0.859961936250329, 'learning_rate/qwen_vl_interface': 9.634758230303408e-06, 'learning_rate/action_model': 9.634758230303408e-05, 'epoch': 0.11})
+[2026-07-28 07:21:23,829][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.06504898518323898, 'train/grad_norm_pre_clip': 2.602844476699829, 'timing/data': 0.0003378409892320633, 'timing/model': 0.6181695768609643, 'learning_rate/qwen_vl_interface': 9.633285430680401e-06, 'learning_rate/action_model': 9.633285430680401e-05, 'epoch': 0.11})
+[2026-07-28 07:21:33,269][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.11232957243919373, 'train/grad_norm_pre_clip': 1.6641356945037842, 'timing/data': 0.0003320425748825073, 'timing/model': 0.6179991811513901, 'learning_rate/qwen_vl_interface': 9.631809786791066e-06, 'learning_rate/action_model': 9.631809786791066e-05, 'epoch': 0.11})
+[2026-07-28 07:21:42,421][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.05702720582485199, 'train/grad_norm_pre_clip': 1.3424601554870605, 'timing/data': 0.000345032662153244, 'timing/model': 0.6191016426309943, 'learning_rate/qwen_vl_interface': 9.630331299592932e-06, 'learning_rate/action_model': 9.630331299592933e-05, 'epoch': 0.11})
+[2026-07-28 07:21:51,813][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.07574225962162018, 'train/grad_norm_pre_clip': 2.1443121433258057, 'timing/data': 0.0003661205992102623, 'timing/model': 0.6246728347614408, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.11})
+[2026-07-28 07:22:01,229][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.10904461145401001, 'train/grad_norm_pre_clip': 3.6141364574432373, 'timing/data': 0.0003386670723557472, 'timing/model': 0.6178806265816092, 'learning_rate/qwen_vl_interface': 9.627365799109617e-06, 'learning_rate/action_model': 9.627365799109617e-05, 'epoch': 0.11})
+[2026-07-28 07:22:10,335][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.0889052003622055, 'train/grad_norm_pre_clip': 1.6709588766098022, 'timing/data': 0.00031948741525411606, 'timing/model': 0.6183779677376151, 'learning_rate/qwen_vl_interface': 9.625878787748712e-06, 'learning_rate/action_model': 9.625878787748713e-05, 'epoch': 0.11})
+[2026-07-28 07:22:19,475][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.0785704106092453, 'train/grad_norm_pre_clip': 1.9118287563323975, 'timing/data': 0.00033219531178474426, 'timing/model': 0.6175527684390545, 'learning_rate/qwen_vl_interface': 9.624388936927572e-06, 'learning_rate/action_model': 9.624388936927571e-05, 'epoch': 0.11})
+[2026-07-28 07:22:29,023][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.06581899523735046, 'train/grad_norm_pre_clip': 1.1508615016937256, 'timing/data': 0.00044073816388845444, 'timing/model': 0.7845286978408694, 'learning_rate/qwen_vl_interface': 9.622896247612943e-06, 'learning_rate/action_model': 9.622896247612942e-05, 'epoch': 0.11})
+[2026-07-28 07:22:38,150][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.11533987522125244, 'train/grad_norm_pre_clip': 1.9239150285720825, 'timing/data': 0.00032207928597927094, 'timing/model': 0.6223154105246067, 'learning_rate/qwen_vl_interface': 9.621400720773414e-06, 'learning_rate/action_model': 9.621400720773414e-05, 'epoch': 0.11})
+[2026-07-28 07:22:47,263][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.06806506216526031, 'train/grad_norm_pre_clip': 1.0095163583755493, 'timing/data': 0.0003439122810959816, 'timing/model': 0.6190161183476448, 'learning_rate/qwen_vl_interface': 9.619902357379416e-06, 'learning_rate/action_model': 9.619902357379415e-05, 'epoch': 0.11})
+[2026-07-28 07:22:56,634][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.07502129673957825, 'train/grad_norm_pre_clip': 1.4097028970718384, 'timing/data': 0.00032200850546360016, 'timing/model': 0.8561105635017157, 'learning_rate/qwen_vl_interface': 9.618401158403222e-06, 'learning_rate/action_model': 9.618401158403222e-05, 'epoch': 0.11})
+[2026-07-28 07:23:06,002][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.06620271503925323, 'train/grad_norm_pre_clip': 1.2269372940063477, 'timing/data': 0.0003306316211819649, 'timing/model': 0.6213963152840734, 'learning_rate/qwen_vl_interface': 9.616897124818944e-06, 'learning_rate/action_model': 9.616897124818945e-05, 'epoch': 0.11})
+[2026-07-28 07:23:15,150][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.0865924209356308, 'train/grad_norm_pre_clip': 4.811131954193115, 'timing/data': 0.0003309166058897972, 'timing/model': 0.6198007166385651, 'learning_rate/qwen_vl_interface': 9.615390257602532e-06, 'learning_rate/action_model': 9.615390257602531e-05, 'epoch': 0.11})
+[2026-07-28 07:23:24,060][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.05598292499780655, 'train/grad_norm_pre_clip': 3.059263229370117, 'timing/data': 0.0003657964989542961, 'timing/model': 0.6183844525367022, 'learning_rate/qwen_vl_interface': 9.613880557731776e-06, 'learning_rate/action_model': 9.613880557731776e-05, 'epoch': 0.11})
+[2026-07-28 07:23:33,208][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.08677564561367035, 'train/grad_norm_pre_clip': 2.225287437438965, 'timing/data': 0.00035093631595373154, 'timing/model': 0.8484382554888725, 'learning_rate/qwen_vl_interface': 9.612368026186304e-06, 'learning_rate/action_model': 9.612368026186304e-05, 'epoch': 0.11})
+[2026-07-28 07:23:42,693][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.089609295129776, 'train/grad_norm_pre_clip': 4.854145050048828, 'timing/data': 0.0003364700824022293, 'timing/model': 0.6162706967443228, 'learning_rate/qwen_vl_interface': 9.610852663947582e-06, 'learning_rate/action_model': 9.610852663947583e-05, 'epoch': 0.11})
+[2026-07-28 07:23:51,809][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.12048758566379547, 'train/grad_norm_pre_clip': 4.453705787658691, 'timing/data': 0.0003956947475671768, 'timing/model': 0.6185226002708077, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.11})
+[2026-07-28 07:24:00,914][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.07368924468755722, 'train/grad_norm_pre_clip': 1.847838044166565, 'timing/data': 0.0003306083381175995, 'timing/model': 0.6164962705224752, 'learning_rate/qwen_vl_interface': 9.607813451325431e-06, 'learning_rate/action_model': 9.60781345132543e-05, 'epoch': 0.11})
+[2026-07-28 07:24:10,022][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.1166260838508606, 'train/grad_norm_pre_clip': 2.9901397228240967, 'timing/data': 0.00033375993371009827, 'timing/model': 0.6167739033699036, 'learning_rate/qwen_vl_interface': 9.606289602914113e-06, 'learning_rate/action_model': 9.606289602914112e-05, 'epoch': 0.11})
+[2026-07-28 07:24:19,397][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.0776091143488884, 'train/grad_norm_pre_clip': 2.029170036315918, 'timing/data': 0.00031923968344926834, 'timing/model': 0.6190218739211559, 'learning_rate/qwen_vl_interface': 9.604762927753768e-06, 'learning_rate/action_model': 9.604762927753768e-05, 'epoch': 0.11})
+[2026-07-28 07:24:28,508][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.08502241969108582, 'train/grad_norm_pre_clip': 1.390992283821106, 'timing/data': 0.00031918007880449295, 'timing/model': 0.6166370548307896, 'learning_rate/qwen_vl_interface': 9.603233426835037e-06, 'learning_rate/action_model': 9.603233426835036e-05, 'epoch': 0.11})
+[2026-07-28 07:24:37,620][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.07445596903562546, 'train/grad_norm_pre_clip': 0.864939272403717, 'timing/data': 0.000318882055580616, 'timing/model': 0.6181086171418428, 'learning_rate/qwen_vl_interface': 9.601701101150399e-06, 'learning_rate/action_model': 9.601701101150398e-05, 'epoch': 0.11})
+[2026-07-28 07:24:46,911][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.0757596343755722, 'train/grad_norm_pre_clip': 1.43790864944458, 'timing/data': 0.0003253854811191559, 'timing/model': 0.7885202178731561, 'learning_rate/qwen_vl_interface': 9.600165951694162e-06, 'learning_rate/action_model': 9.60016595169416e-05, 'epoch': 0.11})
+[2026-07-28 07:24:56,039][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.0758577436208725, 'train/grad_norm_pre_clip': 1.1997277736663818, 'timing/data': 0.0003579268231987953, 'timing/model': 0.618788069114089, 'learning_rate/qwen_vl_interface': 9.598627979462466e-06, 'learning_rate/action_model': 9.598627979462467e-05, 'epoch': 0.11})
+[2026-07-28 07:25:05,380][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.05057405307888985, 'train/grad_norm_pre_clip': 2.4596164226531982, 'timing/data': 0.000361843965947628, 'timing/model': 0.849451513029635, 'learning_rate/qwen_vl_interface': 9.59708718545329e-06, 'learning_rate/action_model': 9.597087185453288e-05, 'epoch': 0.11})
+[2026-07-28 07:25:14,524][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.07113815099000931, 'train/grad_norm_pre_clip': 3.098543167114258, 'timing/data': 0.00033116061240434647, 'timing/model': 0.6188320871442556, 'learning_rate/qwen_vl_interface': 9.595543570666433e-06, 'learning_rate/action_model': 9.595543570666432e-05, 'epoch': 0.11})
+[2026-07-28 07:25:23,899][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.07904355227947235, 'train/grad_norm_pre_clip': 2.9719791412353516, 'timing/data': 0.00033031776547431946, 'timing/model': 0.8958406522870064, 'learning_rate/qwen_vl_interface': 9.593997136103533e-06, 'learning_rate/action_model': 9.593997136103533e-05, 'epoch': 0.11})
+[2026-07-28 07:25:33,036][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.08273305743932724, 'train/grad_norm_pre_clip': 1.6637088060379028, 'timing/data': 0.00040664151310920715, 'timing/model': 0.6203013584017754, 'learning_rate/qwen_vl_interface': 9.592447882768053e-06, 'learning_rate/action_model': 9.592447882768052e-05, 'epoch': 0.11})
+[2026-07-28 07:25:42,159][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.09063847362995148, 'train/grad_norm_pre_clip': 1.7749096155166626, 'timing/data': 0.00032336171716451645, 'timing/model': 0.6181013397872448, 'learning_rate/qwen_vl_interface': 9.590895811665284e-06, 'learning_rate/action_model': 9.590895811665285e-05, 'epoch': 0.11})
+[2026-07-28 07:25:51,510][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.07562541216611862, 'train/grad_norm_pre_clip': 4.403503894805908, 'timing/data': 0.00038115307688713074, 'timing/model': 0.6173069626092911, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.11})
+[2026-07-28 07:26:00,870][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.06465493142604828, 'train/grad_norm_pre_clip': 2.322551965713501, 'timing/data': 0.00040494464337825775, 'timing/model': 0.6160480938851833, 'learning_rate/qwen_vl_interface': 9.587783220188212e-06, 'learning_rate/action_model': 9.587783220188212e-05, 'epoch': 0.11})
+[2026-07-28 07:26:10,063][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.09845587611198425, 'train/grad_norm_pre_clip': 2.220304250717163, 'timing/data': 0.00032538361847400665, 'timing/model': 0.6230485811829567, 'learning_rate/qwen_vl_interface': 9.586222701833631e-06, 'learning_rate/action_model': 9.58622270183363e-05, 'epoch': 0.11})
+[2026-07-28 07:26:19,435][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.12244214862585068, 'train/grad_norm_pre_clip': 3.872617483139038, 'timing/data': 0.00038157869130373, 'timing/model': 0.6163521399721503, 'learning_rate/qwen_vl_interface': 9.584659369751216e-06, 'learning_rate/action_model': 9.584659369751215e-05, 'epoch': 0.11})
+[2026-07-28 07:26:28,810][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.09422584623098373, 'train/grad_norm_pre_clip': 3.4182868003845215, 'timing/data': 0.000336705707013607, 'timing/model': 0.8578814016655087, 'learning_rate/qwen_vl_interface': 9.583093224955397e-06, 'learning_rate/action_model': 9.583093224955397e-05, 'epoch': 0.11})
+[2026-07-28 07:26:37,925][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.07677848637104034, 'train/grad_norm_pre_clip': 3.2643558979034424, 'timing/data': 0.0005251672118902206, 'timing/model': 0.6177863711491227, 'learning_rate/qwen_vl_interface': 9.58152426846243e-06, 'learning_rate/action_model': 9.581524268462429e-05, 'epoch': 0.11})
+[2026-07-28 07:26:47,091][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.06024010479450226, 'train/grad_norm_pre_clip': 0.7538566589355469, 'timing/data': 0.0004022819921374321, 'timing/model': 0.6240069409832358, 'learning_rate/qwen_vl_interface': 9.579952501290388e-06, 'learning_rate/action_model': 9.579952501290387e-05, 'epoch': 0.11})
+[2026-07-28 07:26:56,515][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.105955570936203, 'train/grad_norm_pre_clip': 4.004485130310059, 'timing/data': 0.0003392193466424942, 'timing/model': 0.6180389383807778, 'learning_rate/qwen_vl_interface': 9.57837792445918e-06, 'learning_rate/action_model': 9.578377924459181e-05, 'epoch': 0.12})
+[2026-07-28 07:27:05,941][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.07265029102563858, 'train/grad_norm_pre_clip': 1.7150455713272095, 'timing/data': 0.00042604561895132065, 'timing/model': 0.6205237712711096, 'learning_rate/qwen_vl_interface': 9.57680053899053e-06, 'learning_rate/action_model': 9.576800538990529e-05, 'epoch': 0.12})
+[2026-07-28 07:27:15,055][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.10845375061035156, 'train/grad_norm_pre_clip': 2.467489242553711, 'timing/data': 0.0003245994448661804, 'timing/model': 0.6168638970702887, 'learning_rate/qwen_vl_interface': 9.575220345907984e-06, 'learning_rate/action_model': 9.575220345907984e-05, 'epoch': 0.12})
+[2026-07-28 07:27:24,186][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.07891973853111267, 'train/grad_norm_pre_clip': 2.170588254928589, 'timing/data': 0.0003884425386786461, 'timing/model': 0.6190976593643427, 'learning_rate/qwen_vl_interface': 9.573637346236915e-06, 'learning_rate/action_model': 9.573637346236915e-05, 'epoch': 0.12})
+[2026-07-28 07:27:33,516][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.09823648631572723, 'train/grad_norm_pre_clip': 4.883731365203857, 'timing/data': 0.00039803143590688705, 'timing/model': 0.6189046129584312, 'learning_rate/qwen_vl_interface': 9.572051541004513e-06, 'learning_rate/action_model': 9.572051541004513e-05, 'epoch': 0.12})
+[2026-07-28 07:27:42,869][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.06281676143407822, 'train/grad_norm_pre_clip': 2.4965224266052246, 'timing/data': 0.00032686907798051834, 'timing/model': 0.6169914789497852, 'learning_rate/qwen_vl_interface': 9.57046293123979e-06, 'learning_rate/action_model': 9.570462931239789e-05, 'epoch': 0.12})
+[2026-07-28 07:27:51,989][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.05196754261851311, 'train/grad_norm_pre_clip': 1.837592601776123, 'timing/data': 0.00040098559111356735, 'timing/model': 0.6164031811058521, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.12})
+[2026-07-28 07:28:01,112][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.05896049737930298, 'train/grad_norm_pre_clip': 1.6053556203842163, 'timing/data': 0.0003888839855790138, 'timing/model': 0.6189519045874476, 'learning_rate/qwen_vl_interface': 9.567277302238524e-06, 'learning_rate/action_model': 9.567277302238524e-05, 'epoch': 0.12})
+[2026-07-28 07:28:10,440][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.07584767043590546, 'train/grad_norm_pre_clip': 2.3182213306427, 'timing/data': 0.0003248266875743866, 'timing/model': 0.8421419588848948, 'learning_rate/qwen_vl_interface': 9.565680285069104e-06, 'learning_rate/action_model': 9.565680285069103e-05, 'epoch': 0.12})
+[2026-07-28 07:28:19,739][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.06464195251464844, 'train/grad_norm_pre_clip': 1.6619133949279785, 'timing/data': 0.00039406120777130127, 'timing/model': 0.6179973697289824, 'learning_rate/qwen_vl_interface': 9.564080467501601e-06, 'learning_rate/action_model': 9.5640804675016e-05, 'epoch': 0.12})
+[2026-07-28 07:28:28,890][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.04498616233468056, 'train/grad_norm_pre_clip': 1.36664617061615, 'timing/data': 0.00042932014912366867, 'timing/model': 0.617891033180058, 'learning_rate/qwen_vl_interface': 9.562477850574118e-06, 'learning_rate/action_model': 9.562477850574119e-05, 'epoch': 0.12})
+[2026-07-28 07:28:38,706][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.10394002497196198, 'train/grad_norm_pre_clip': 3.769745111465454, 'timing/data': 0.000378570519387722, 'timing/model': 0.6376154366880655, 'learning_rate/qwen_vl_interface': 9.56087243532658e-06, 'learning_rate/action_model': 9.56087243532658e-05, 'epoch': 0.12})
+[2026-07-28 07:28:47,941][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.09915705025196075, 'train/grad_norm_pre_clip': 2.536860704421997, 'timing/data': 0.0003915363922715187, 'timing/model': 0.6214533196762204, 'learning_rate/qwen_vl_interface': 9.55926422280072e-06, 'learning_rate/action_model': 9.55926422280072e-05, 'epoch': 0.12})
+[2026-07-28 07:28:57,415][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.08344244956970215, 'train/grad_norm_pre_clip': 2.8576762676239014, 'timing/data': 0.0003487905487418175, 'timing/model': 0.864302578382194, 'learning_rate/qwen_vl_interface': 9.55765321404009e-06, 'learning_rate/action_model': 9.557653214040091e-05, 'epoch': 0.12})
+[2026-07-28 07:29:06,545][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.05508778616786003, 'train/grad_norm_pre_clip': 1.9083293676376343, 'timing/data': 0.00032094772905111313, 'timing/model': 0.625660770572722, 'learning_rate/qwen_vl_interface': 9.556039410090061e-06, 'learning_rate/action_model': 9.556039410090061e-05, 'epoch': 0.12})
+[2026-07-28 07:29:15,440][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.07616221159696579, 'train/grad_norm_pre_clip': 2.8171918392181396, 'timing/data': 0.0003399541601538658, 'timing/model': 0.6211641980335116, 'learning_rate/qwen_vl_interface': 9.554422811997804e-06, 'learning_rate/action_model': 9.554422811997805e-05, 'epoch': 0.12})
+[2026-07-28 07:29:24,366][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.09615863859653473, 'train/grad_norm_pre_clip': 3.0186758041381836, 'timing/data': 0.00033574458211660385, 'timing/model': 0.6229840219020844, 'learning_rate/qwen_vl_interface': 9.55280342081232e-06, 'learning_rate/action_model': 9.55280342081232e-05, 'epoch': 0.12})
+[2026-07-28 07:29:33,502][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.10320915281772614, 'train/grad_norm_pre_clip': 2.5603857040405273, 'timing/data': 0.00033056456595659256, 'timing/model': 0.6208442766219378, 'learning_rate/qwen_vl_interface': 9.55118123758441e-06, 'learning_rate/action_model': 9.551181237584409e-05, 'epoch': 0.12})
+[2026-07-28 07:29:42,724][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.06304701417684555, 'train/grad_norm_pre_clip': 1.645460844039917, 'timing/data': 0.0003884239122271538, 'timing/model': 0.6367175476625562, 'learning_rate/qwen_vl_interface': 9.549556263366691e-06, 'learning_rate/action_model': 9.549556263366692e-05, 'epoch': 0.12})
+[2026-07-28 07:29:52,474][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.10629390180110931, 'train/grad_norm_pre_clip': 2.8497118949890137, 'timing/data': 0.0003744848072528839, 'timing/model': 0.6282741790637374, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.12})
+[2026-07-28 07:30:02,048][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.07985018193721771, 'train/grad_norm_pre_clip': 3.3854377269744873, 'timing/data': 0.00038719456642866135, 'timing/model': 0.62538799084723, 'learning_rate/qwen_vl_interface': 9.546297946181354e-06, 'learning_rate/action_model': 9.546297946181354e-05, 'epoch': 0.12})
+[2026-07-28 07:30:11,287][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.08221650868654251, 'train/grad_norm_pre_clip': 1.8688483238220215, 'timing/data': 0.00033646728843450546, 'timing/model': 0.6193138873204589, 'learning_rate/qwen_vl_interface': 9.544664605328022e-06, 'learning_rate/action_model': 9.544664605328022e-05, 'epoch': 0.12})
+[2026-07-28 07:30:20,487][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.08560074120759964, 'train/grad_norm_pre_clip': 3.069735050201416, 'timing/data': 0.00033001508563756943, 'timing/model': 0.6193931763991714, 'learning_rate/qwen_vl_interface': 9.543028477713455e-06, 'learning_rate/action_model': 9.543028477713453e-05, 'epoch': 0.12})
+[2026-07-28 07:30:29,686][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.08341071009635925, 'train/grad_norm_pre_clip': 2.7772319316864014, 'timing/data': 0.0003429120406508446, 'timing/model': 0.6188566135242581, 'learning_rate/qwen_vl_interface': 9.541389564399316e-06, 'learning_rate/action_model': 9.541389564399315e-05, 'epoch': 0.12})
+[2026-07-28 07:30:39,072][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.04979171231389046, 'train/grad_norm_pre_clip': 1.126641869544983, 'timing/data': 0.0003947122022509575, 'timing/model': 0.845320101827383, 'learning_rate/qwen_vl_interface': 9.53974786644908e-06, 'learning_rate/action_model': 9.53974786644908e-05, 'epoch': 0.12})
+[2026-07-28 07:30:48,225][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.07319634407758713, 'train/grad_norm_pre_clip': 1.5143111944198608, 'timing/data': 0.000818728469312191, 'timing/model': 0.620254160836339, 'learning_rate/qwen_vl_interface': 9.538103384928029e-06, 'learning_rate/action_model': 9.538103384928027e-05, 'epoch': 0.12})
+[2026-07-28 07:30:57,600][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.07889778912067413, 'train/grad_norm_pre_clip': 1.4701794385910034, 'timing/data': 0.00032512564212083817, 'timing/model': 0.8464834876358509, 'learning_rate/qwen_vl_interface': 9.536456120903244e-06, 'learning_rate/action_model': 9.536456120903244e-05, 'epoch': 0.12})
+[2026-07-28 07:31:06,793][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.06466229259967804, 'train/grad_norm_pre_clip': 1.5286407470703125, 'timing/data': 0.0004521263763308525, 'timing/model': 0.6216146908700466, 'learning_rate/qwen_vl_interface': 9.534806075443619e-06, 'learning_rate/action_model': 9.534806075443619e-05, 'epoch': 0.12})
+[2026-07-28 07:31:15,950][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.04917927458882332, 'train/grad_norm_pre_clip': 1.9552788734436035, 'timing/data': 0.00034934747964143753, 'timing/model': 0.6190657382830977, 'learning_rate/qwen_vl_interface': 9.533153249619853e-06, 'learning_rate/action_model': 9.533153249619853e-05, 'epoch': 0.12})
+[2026-07-28 07:31:25,328][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.0872243344783783, 'train/grad_norm_pre_clip': 1.794357419013977, 'timing/data': 0.0004376377910375595, 'timing/model': 0.6183528658002615, 'learning_rate/qwen_vl_interface': 9.531497644504444e-06, 'learning_rate/action_model': 9.531497644504445e-05, 'epoch': 0.12})
+[2026-07-28 07:31:34,475][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.04789246618747711, 'train/grad_norm_pre_clip': 1.6708881855010986, 'timing/data': 0.00036563165485858917, 'timing/model': 0.6196253690868616, 'learning_rate/qwen_vl_interface': 9.529839261171698e-06, 'learning_rate/action_model': 9.529839261171696e-05, 'epoch': 0.12})
+[2026-07-28 07:31:44,014][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.07938786596059799, 'train/grad_norm_pre_clip': 2.719888687133789, 'timing/data': 0.00041501689702272415, 'timing/model': 0.6233500614762306, 'learning_rate/qwen_vl_interface': 9.52817810069772e-06, 'learning_rate/action_model': 9.52817810069772e-05, 'epoch': 0.12})
+[2026-07-28 07:31:53,422][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.0800580233335495, 'train/grad_norm_pre_clip': 2.8672003746032715, 'timing/data': 0.0003557885065674782, 'timing/model': 0.6212925976142287, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.12})
+[2026-07-28 07:32:02,615][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.09839392453432083, 'train/grad_norm_pre_clip': 4.670693874359131, 'timing/data': 0.0003253258764743805, 'timing/model': 0.618511563166976, 'learning_rate/qwen_vl_interface': 9.524847452639509e-06, 'learning_rate/action_model': 9.524847452639509e-05, 'epoch': 0.12})
+[2026-07-28 07:32:11,766][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.061990994960069656, 'train/grad_norm_pre_clip': 3.827090263366699, 'timing/data': 0.0004074461758136749, 'timing/model': 0.619266758672893, 'learning_rate/qwen_vl_interface': 9.523177967216496e-06, 'learning_rate/action_model': 9.523177967216496e-05, 'epoch': 0.12})
+[2026-07-28 07:32:21,121][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.09158340096473694, 'train/grad_norm_pre_clip': 4.72422456741333, 'timing/data': 0.00043226219713687897, 'timing/model': 0.6181444087997079, 'learning_rate/qwen_vl_interface': 9.521505708974694e-06, 'learning_rate/action_model': 9.521505708974694e-05, 'epoch': 0.12})
+[2026-07-28 07:32:30,463][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.07316438853740692, 'train/grad_norm_pre_clip': 1.8572360277175903, 'timing/data': 0.00043791625648736954, 'timing/model': 0.8472852120175958, 'learning_rate/qwen_vl_interface': 9.519830678999214e-06, 'learning_rate/action_model': 9.519830678999213e-05, 'epoch': 0.12})
+[2026-07-28 07:32:39,608][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.05879770964384079, 'train/grad_norm_pre_clip': 1.4840116500854492, 'timing/data': 0.00031969696283340454, 'timing/model': 0.6183170881122351, 'learning_rate/qwen_vl_interface': 9.518152878376961e-06, 'learning_rate/action_model': 9.51815287837696e-05, 'epoch': 0.12})
+[2026-07-28 07:32:48,995][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.07231026142835617, 'train/grad_norm_pre_clip': 1.5384515523910522, 'timing/data': 0.00043199025094509125, 'timing/model': 0.8660841025412083, 'learning_rate/qwen_vl_interface': 9.516472308196646e-06, 'learning_rate/action_model': 9.516472308196645e-05, 'epoch': 0.12})
+[2026-07-28 07:32:58,151][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.05498369038105011, 'train/grad_norm_pre_clip': 1.0323593616485596, 'timing/data': 0.0003435499966144562, 'timing/model': 0.617912657558918, 'learning_rate/qwen_vl_interface': 9.51478896954877e-06, 'learning_rate/action_model': 9.51478896954877e-05, 'epoch': 0.12})
+[2026-07-28 07:33:07,261][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.07177664339542389, 'train/grad_norm_pre_clip': 1.5166743993759155, 'timing/data': 0.00032480061054229736, 'timing/model': 0.6193938562646508, 'learning_rate/qwen_vl_interface': 9.51310286352563e-06, 'learning_rate/action_model': 9.51310286352563e-05, 'epoch': 0.12})
+[2026-07-28 07:33:16,413][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.07875122129917145, 'train/grad_norm_pre_clip': 1.6741738319396973, 'timing/data': 0.00044353585690259933, 'timing/model': 0.618394703604281, 'learning_rate/qwen_vl_interface': 9.51141399122133e-06, 'learning_rate/action_model': 9.511413991221331e-05, 'epoch': 0.12})
+[2026-07-28 07:33:25,805][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.06656119972467422, 'train/grad_norm_pre_clip': 1.2995203733444214, 'timing/data': 0.00040887482464313507, 'timing/model': 0.6216291561722755, 'learning_rate/qwen_vl_interface': 9.509722353731757e-06, 'learning_rate/action_model': 9.509722353731756e-05, 'epoch': 0.12})
+[2026-07-28 07:33:35,241][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.05247235298156738, 'train/grad_norm_pre_clip': 1.3272130489349365, 'timing/data': 0.0003486406058073044, 'timing/model': 0.8804660197347403, 'learning_rate/qwen_vl_interface': 9.508027952154595e-06, 'learning_rate/action_model': 9.508027952154593e-05, 'epoch': 0.12})
+[2026-07-28 07:33:44,371][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.06597930192947388, 'train/grad_norm_pre_clip': 1.8337327241897583, 'timing/data': 0.0003218557685613632, 'timing/model': 0.6189928567036986, 'learning_rate/qwen_vl_interface': 9.506330787589321e-06, 'learning_rate/action_model': 9.506330787589322e-05, 'epoch': 0.12})
+[2026-07-28 07:33:53,710][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.11398978531360626, 'train/grad_norm_pre_clip': 3.61865234375, 'timing/data': 0.00033491291105747223, 'timing/model': 0.6182500869035721, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.12})
+[2026-07-28 07:34:02,648][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.09007509052753448, 'train/grad_norm_pre_clip': 1.981157898902893, 'timing/data': 0.0003284970298409462, 'timing/model': 0.6207151301205158, 'learning_rate/qwen_vl_interface': 9.50292817390133e-06, 'learning_rate/action_model': 9.502928173901329e-05, 'epoch': 0.12})
+[2026-07-28 07:34:11,791][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.05390015244483948, 'train/grad_norm_pre_clip': 1.2863819599151611, 'timing/data': 0.0003435807302594185, 'timing/model': 0.6170838680118322, 'learning_rate/qwen_vl_interface': 9.501222726986525e-06, 'learning_rate/action_model': 9.501222726986524e-05, 'epoch': 0.12})
+[2026-07-28 07:34:21,105][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.07018294930458069, 'train/grad_norm_pre_clip': 1.7974259853363037, 'timing/data': 0.0003283582627773285, 'timing/model': 0.7858884446322918, 'learning_rate/qwen_vl_interface': 9.49951452149945e-06, 'learning_rate/action_model': 9.499514521499451e-05, 'epoch': 0.12})
+[2026-07-28 07:34:30,297][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.08513075858354568, 'train/grad_norm_pre_clip': 1.666815161705017, 'timing/data': 0.00033314526081085205, 'timing/model': 0.6190427765250206, 'learning_rate/qwen_vl_interface': 9.497803558548539e-06, 'learning_rate/action_model': 9.497803558548539e-05, 'epoch': 0.12})
+[2026-07-28 07:34:39,676][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.10517162084579468, 'train/grad_norm_pre_clip': 1.8908467292785645, 'timing/data': 0.000336524099111557, 'timing/model': 0.6226462889462709, 'learning_rate/qwen_vl_interface': 9.496089839244015e-06, 'learning_rate/action_model': 9.496089839244015e-05, 'epoch': 0.12})
+[2026-07-28 07:34:48,870][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.08815649151802063, 'train/grad_norm_pre_clip': 1.2297395467758179, 'timing/data': 0.0004133954644203186, 'timing/model': 0.6193264424800873, 'learning_rate/qwen_vl_interface': 9.494373364697895e-06, 'learning_rate/action_model': 9.494373364697894e-05, 'epoch': 0.12})
+[2026-07-28 07:34:58,247][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.07202459126710892, 'train/grad_norm_pre_clip': 2.981489419937134, 'timing/data': 0.0007406407967209816, 'timing/model': 0.6188370091840625, 'learning_rate/qwen_vl_interface': 9.492654136023978e-06, 'learning_rate/action_model': 9.492654136023979e-05, 'epoch': 0.12})
+[2026-07-28 07:35:07,412][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.09999845921993256, 'train/grad_norm_pre_clip': 4.964880466461182, 'timing/data': 0.00033162347972393036, 'timing/model': 0.6183696920052171, 'learning_rate/qwen_vl_interface': 9.490932154337855e-06, 'learning_rate/action_model': 9.490932154337855e-05, 'epoch': 0.12})
+[2026-07-28 07:35:16,359][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.0669848620891571, 'train/grad_norm_pre_clip': 1.9526866674423218, 'timing/data': 0.0003815172240138054, 'timing/model': 0.6180920489132404, 'learning_rate/qwen_vl_interface': 9.4892074207569e-06, 'learning_rate/action_model': 9.489207420756899e-05, 'epoch': 0.13})
+[2026-07-28 07:35:25,513][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.08025563508272171, 'train/grad_norm_pre_clip': 2.005192756652832, 'timing/data': 0.0004386827349662781, 'timing/model': 0.8558501815423369, 'learning_rate/qwen_vl_interface': 9.487479936400274e-06, 'learning_rate/action_model': 9.487479936400272e-05, 'epoch': 0.13})
+[2026-07-28 07:35:35,053][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.06372765451669693, 'train/grad_norm_pre_clip': 1.6949812173843384, 'timing/data': 0.0003458615392446518, 'timing/model': 0.6223341161385179, 'learning_rate/qwen_vl_interface': 9.485749702388923e-06, 'learning_rate/action_model': 9.485749702388923e-05, 'epoch': 0.13})
+[2026-07-28 07:35:44,200][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.047569628804922104, 'train/grad_norm_pre_clip': 1.1084696054458618, 'timing/data': 0.00033922865986824036, 'timing/model': 0.6165315145626664, 'learning_rate/qwen_vl_interface': 9.484016719845576e-06, 'learning_rate/action_model': 9.484016719845576e-05, 'epoch': 0.13})
+[2026-07-28 07:35:53,315][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.05194183439016342, 'train/grad_norm_pre_clip': 1.4715027809143066, 'timing/data': 0.00032239407300949097, 'timing/model': 0.6177220102399588, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.13})
+[2026-07-28 07:36:02,428][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.08669289201498032, 'train/grad_norm_pre_clip': 3.200435161590576, 'timing/data': 0.000340266153216362, 'timing/model': 0.6177941709756851, 'learning_rate/qwen_vl_interface': 9.480542513662734e-06, 'learning_rate/action_model': 9.480542513662734e-05, 'epoch': 0.13})
+[2026-07-28 07:36:11,781][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.03892375901341438, 'train/grad_norm_pre_clip': 1.448201060295105, 'timing/data': 0.0003179088234901428, 'timing/model': 0.6208896925672889, 'learning_rate/qwen_vl_interface': 9.478801292277617e-06, 'learning_rate/action_model': 9.478801292277616e-05, 'epoch': 0.13})
+[2026-07-28 07:36:21,129][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.13977017998695374, 'train/grad_norm_pre_clip': 3.1283047199249268, 'timing/data': 0.00032937154173851013, 'timing/model': 0.6179002905264497, 'learning_rate/qwen_vl_interface': 9.477057326869249e-06, 'learning_rate/action_model': 9.477057326869249e-05, 'epoch': 0.13})
+[2026-07-28 07:36:30,243][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.11362858861684799, 'train/grad_norm_pre_clip': 4.583251953125, 'timing/data': 0.00033984333276748657, 'timing/model': 0.6180812586098909, 'learning_rate/qwen_vl_interface': 9.475310618569274e-06, 'learning_rate/action_model': 9.475310618569274e-05, 'epoch': 0.13})
+[2026-07-28 07:36:39,383][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.1264040768146515, 'train/grad_norm_pre_clip': 3.8268704414367676, 'timing/data': 0.00033883005380630493, 'timing/model': 0.617229588329792, 'learning_rate/qwen_vl_interface': 9.473561168511115e-06, 'learning_rate/action_model': 9.473561168511115e-05, 'epoch': 0.13})
+[2026-07-28 07:36:48,484][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.10604158043861389, 'train/grad_norm_pre_clip': 3.352226734161377, 'timing/data': 0.0004261638969182968, 'timing/model': 0.6182605931535363, 'learning_rate/qwen_vl_interface': 9.471808977829967e-06, 'learning_rate/action_model': 9.471808977829967e-05, 'epoch': 0.13})
+[2026-07-28 07:36:58,097][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.10610175877809525, 'train/grad_norm_pre_clip': 2.721096992492676, 'timing/data': 0.000399724580347538, 'timing/model': 0.858221479691565, 'learning_rate/qwen_vl_interface': 9.470054047662808e-06, 'learning_rate/action_model': 9.470054047662808e-05, 'epoch': 0.13})
+[2026-07-28 07:37:07,229][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.11139507591724396, 'train/grad_norm_pre_clip': 1.985338568687439, 'timing/data': 0.0003507640212774277, 'timing/model': 0.6171656679362059, 'learning_rate/qwen_vl_interface': 9.468296379148394e-06, 'learning_rate/action_model': 9.468296379148393e-05, 'epoch': 0.13})
+[2026-07-28 07:37:16,319][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.08716939389705658, 'train/grad_norm_pre_clip': 3.07918643951416, 'timing/data': 0.0003286115825176239, 'timing/model': 0.6180767770856619, 'learning_rate/qwen_vl_interface': 9.466535973427259e-06, 'learning_rate/action_model': 9.466535973427258e-05, 'epoch': 0.13})
+[2026-07-28 07:37:25,686][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.09017301350831985, 'train/grad_norm_pre_clip': 2.9886770248413086, 'timing/data': 0.00033227168023586273, 'timing/model': 0.8643752420321107, 'learning_rate/qwen_vl_interface': 9.464772831641706e-06, 'learning_rate/action_model': 9.464772831641707e-05, 'epoch': 0.13})
+[2026-07-28 07:37:34,818][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.10422352701425552, 'train/grad_norm_pre_clip': 2.1090540885925293, 'timing/data': 0.0003332141786813736, 'timing/model': 0.6179294204339385, 'learning_rate/qwen_vl_interface': 9.463006954935823e-06, 'learning_rate/action_model': 9.463006954935822e-05, 'epoch': 0.13})
+[2026-07-28 07:37:44,151][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.09260332584381104, 'train/grad_norm_pre_clip': 3.773869752883911, 'timing/data': 0.00031068455427885056, 'timing/model': 0.6185280699282885, 'learning_rate/qwen_vl_interface': 9.461238344455468e-06, 'learning_rate/action_model': 9.461238344455468e-05, 'epoch': 0.13})
+[2026-07-28 07:37:53,264][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.09592048823833466, 'train/grad_norm_pre_clip': 2.7433838844299316, 'timing/data': 0.0003351578488945961, 'timing/model': 0.6173478495329618, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.13})
+[2026-07-28 07:38:02,398][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.0898265540599823, 'train/grad_norm_pre_clip': 1.5225001573562622, 'timing/data': 0.0003269882872700691, 'timing/model': 0.6192345814779401, 'learning_rate/qwen_vl_interface': 9.457692926763638e-06, 'learning_rate/action_model': 9.457692926763638e-05, 'epoch': 0.13})
+[2026-07-28 07:38:12,043][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.08773903548717499, 'train/grad_norm_pre_clip': 3.9952211380004883, 'timing/data': 0.0004105325788259506, 'timing/model': 0.6296407030895352, 'learning_rate/qwen_vl_interface': 9.455916121852748e-06, 'learning_rate/action_model': 9.455916121852747e-05, 'epoch': 0.13})
+[2026-07-28 07:38:21,229][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.10051217675209045, 'train/grad_norm_pre_clip': 3.8037967681884766, 'timing/data': 0.00037394557148218155, 'timing/model': 0.6505451379343867, 'learning_rate/qwen_vl_interface': 9.454136587768549e-06, 'learning_rate/action_model': 9.454136587768548e-05, 'epoch': 0.13})
+[2026-07-28 07:38:30,232][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.08782166242599487, 'train/grad_norm_pre_clip': 3.347013473510742, 'timing/data': 0.00039162952452898026, 'timing/model': 0.6431096475571394, 'learning_rate/qwen_vl_interface': 9.452354325665763e-06, 'learning_rate/action_model': 9.452354325665763e-05, 'epoch': 0.13})
+[2026-07-28 07:38:39,217][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.06133650988340378, 'train/grad_norm_pre_clip': 2.1687839031219482, 'timing/data': 0.00034657493233680725, 'timing/model': 0.6267324052751064, 'learning_rate/qwen_vl_interface': 9.450569336700878e-06, 'learning_rate/action_model': 9.450569336700878e-05, 'epoch': 0.13})
+[2026-07-28 07:38:48,627][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.08544886112213135, 'train/grad_norm_pre_clip': 1.4493130445480347, 'timing/data': 0.00043112318962812424, 'timing/model': 0.6211748495697975, 'learning_rate/qwen_vl_interface': 9.448781622032155e-06, 'learning_rate/action_model': 9.448781622032155e-05, 'epoch': 0.13})
+[2026-07-28 07:38:57,785][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.07919354736804962, 'train/grad_norm_pre_clip': 2.912362575531006, 'timing/data': 0.0003935592249035835, 'timing/model': 0.6191369947046041, 'learning_rate/qwen_vl_interface': 9.446991182819624e-06, 'learning_rate/action_model': 9.446991182819623e-05, 'epoch': 0.13})
+[2026-07-28 07:39:06,940][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.05214981734752655, 'train/grad_norm_pre_clip': 1.516000747680664, 'timing/data': 0.0004893532022833824, 'timing/model': 0.6181997805833817, 'learning_rate/qwen_vl_interface': 9.44519802022508e-06, 'learning_rate/action_model': 9.44519802022508e-05, 'epoch': 0.13})
+[2026-07-28 07:39:16,022][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.02350083366036415, 'train/grad_norm_pre_clip': 0.6593997478485107, 'timing/data': 0.00033066701143980026, 'timing/model': 0.7962299222126603, 'learning_rate/qwen_vl_interface': 9.443402135412088e-06, 'learning_rate/action_model': 9.443402135412088e-05, 'epoch': 0.13})
+[2026-07-28 07:39:25,145][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.059581831097602844, 'train/grad_norm_pre_clip': 3.0506365299224854, 'timing/data': 0.0004127407446503639, 'timing/model': 0.6184154618531466, 'learning_rate/qwen_vl_interface': 9.441603529545977e-06, 'learning_rate/action_model': 9.441603529545976e-05, 'epoch': 0.13})
+[2026-07-28 07:39:34,237][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.12549996376037598, 'train/grad_norm_pre_clip': 3.7320821285247803, 'timing/data': 0.00031715724617242813, 'timing/model': 0.6191403195261955, 'learning_rate/qwen_vl_interface': 9.439802203793843e-06, 'learning_rate/action_model': 9.439802203793842e-05, 'epoch': 0.13})
+[2026-07-28 07:39:43,366][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.04853900521993637, 'train/grad_norm_pre_clip': 2.3561723232269287, 'timing/data': 0.00031853094696998596, 'timing/model': 0.6191630717366934, 'learning_rate/qwen_vl_interface': 9.437998159324547e-06, 'learning_rate/action_model': 9.437998159324547e-05, 'epoch': 0.13})
+[2026-07-28 07:39:52,737][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.0691705197095871, 'train/grad_norm_pre_clip': 2.468618154525757, 'timing/data': 0.0003491835668683052, 'timing/model': 0.8504194654524326, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.13})
+[2026-07-28 07:40:02,098][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.06434525549411774, 'train/grad_norm_pre_clip': 2.0714187622070312, 'timing/data': 0.0003605550155043602, 'timing/model': 0.8459793655201793, 'learning_rate/qwen_vl_interface': 9.434381918918731e-06, 'learning_rate/action_model': 9.434381918918731e-05, 'epoch': 0.13})
+[2026-07-28 07:40:11,243][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.09891636669635773, 'train/grad_norm_pre_clip': 2.5858137607574463, 'timing/data': 0.00038523413240909576, 'timing/model': 0.6181805701926351, 'learning_rate/qwen_vl_interface': 9.432569725328753e-06, 'learning_rate/action_model': 9.432569725328753e-05, 'epoch': 0.13})
+[2026-07-28 07:40:20,428][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.07805155962705612, 'train/grad_norm_pre_clip': 2.1683566570281982, 'timing/data': 0.0004090927541255951, 'timing/model': 0.6197449853643775, 'learning_rate/qwen_vl_interface': 9.430754817714686e-06, 'learning_rate/action_model': 9.430754817714687e-05, 'epoch': 0.13})
+[2026-07-28 07:40:29,808][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.14792747795581818, 'train/grad_norm_pre_clip': 4.05524206161499, 'timing/data': 0.0003180829808115959, 'timing/model': 0.6197702130302787, 'learning_rate/qwen_vl_interface': 9.42893719725421e-06, 'learning_rate/action_model': 9.42893719725421e-05, 'epoch': 0.13})
+[2026-07-28 07:40:38,939][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.08847224712371826, 'train/grad_norm_pre_clip': 2.2071382999420166, 'timing/data': 0.00033288076519966125, 'timing/model': 0.6176836332306266, 'learning_rate/qwen_vl_interface': 9.427116865126757e-06, 'learning_rate/action_model': 9.427116865126756e-05, 'epoch': 0.13})
+[2026-07-28 07:40:48,056][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.03560042381286621, 'train/grad_norm_pre_clip': 1.306218147277832, 'timing/data': 0.00039445050060749054, 'timing/model': 0.6185363130643964, 'learning_rate/qwen_vl_interface': 9.42529382251352e-06, 'learning_rate/action_model': 9.42529382251352e-05, 'epoch': 0.13})
+[2026-07-28 07:40:57,415][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.11456592381000519, 'train/grad_norm_pre_clip': 2.23903489112854, 'timing/data': 0.0003246776759624481, 'timing/model': 0.6210345132276416, 'learning_rate/qwen_vl_interface': 9.423468070597453e-06, 'learning_rate/action_model': 9.423468070597452e-05, 'epoch': 0.13})
+[2026-07-28 07:41:06,818][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.048006728291511536, 'train/grad_norm_pre_clip': 2.02211856842041, 'timing/data': 0.0003336053341627121, 'timing/model': 0.6191247748211026, 'learning_rate/qwen_vl_interface': 9.421639610563264e-06, 'learning_rate/action_model': 9.421639610563264e-05, 'epoch': 0.13})
+[2026-07-28 07:41:15,960][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.09729816019535065, 'train/grad_norm_pre_clip': 1.4830564260482788, 'timing/data': 0.00040128640830516815, 'timing/model': 0.6181523445993662, 'learning_rate/qwen_vl_interface': 9.419808443597422e-06, 'learning_rate/action_model': 9.419808443597423e-05, 'epoch': 0.13})
+[2026-07-28 07:41:24,846][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.09987054765224457, 'train/grad_norm_pre_clip': 1.489080548286438, 'timing/data': 0.00038863252848386765, 'timing/model': 0.6185974022373557, 'learning_rate/qwen_vl_interface': 9.417974570888152e-06, 'learning_rate/action_model': 9.417974570888153e-05, 'epoch': 0.13})
+[2026-07-28 07:41:33,989][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.07558763027191162, 'train/grad_norm_pre_clip': 1.9805442094802856, 'timing/data': 0.00041008275002241135, 'timing/model': 0.8550535514950752, 'learning_rate/qwen_vl_interface': 9.416137993625437e-06, 'learning_rate/action_model': 9.416137993625437e-05, 'epoch': 0.13})
+[2026-07-28 07:41:43,142][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.06097424402832985, 'train/grad_norm_pre_clip': 2.721022367477417, 'timing/data': 0.00040525756776332855, 'timing/model': 0.6194860655814409, 'learning_rate/qwen_vl_interface': 9.414298713001007e-06, 'learning_rate/action_model': 9.414298713001006e-05, 'epoch': 0.13})
+[2026-07-28 07:41:52,347][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.04132666438817978, 'train/grad_norm_pre_clip': 0.6576619744300842, 'timing/data': 0.0003659455105662346, 'timing/model': 0.6211039712652564, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.13})
+[2026-07-28 07:42:01,455][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.07042616605758667, 'train/grad_norm_pre_clip': 1.3717567920684814, 'timing/data': 0.0003333119675517082, 'timing/model': 0.6200690120458603, 'learning_rate/qwen_vl_interface': 9.410612046442719e-06, 'learning_rate/action_model': 9.410612046442718e-05, 'epoch': 0.13})
+[2026-07-28 07:42:10,851][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.07910771667957306, 'train/grad_norm_pre_clip': 1.1312750577926636, 'timing/data': 0.000422036275267601, 'timing/model': 0.6196160912513733, 'learning_rate/qwen_vl_interface': 9.408764662901097e-06, 'learning_rate/action_model': 9.408764662901098e-05, 'epoch': 0.13})
+[2026-07-28 07:42:20,208][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.07292360067367554, 'train/grad_norm_pre_clip': 1.6570320129394531, 'timing/data': 0.00032445136457681656, 'timing/model': 0.6169759379699826, 'learning_rate/qwen_vl_interface': 9.406914580782238e-06, 'learning_rate/action_model': 9.406914580782238e-05, 'epoch': 0.13})
+[2026-07-28 07:42:29,357][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.0746309757232666, 'train/grad_norm_pre_clip': 2.9691853523254395, 'timing/data': 0.0003566974774003029, 'timing/model': 0.6179725462570786, 'learning_rate/qwen_vl_interface': 9.405061801286638e-06, 'learning_rate/action_model': 9.405061801286638e-05, 'epoch': 0.13})
+[2026-07-28 07:42:38,500][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.15558767318725586, 'train/grad_norm_pre_clip': 3.5056302547454834, 'timing/data': 0.00038340501487255096, 'timing/model': 0.6195391593500972, 'learning_rate/qwen_vl_interface': 9.403206325616546e-06, 'learning_rate/action_model': 9.403206325616545e-05, 'epoch': 0.13})
+[2026-07-28 07:42:47,869][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.02391722798347473, 'train/grad_norm_pre_clip': 1.1636500358581543, 'timing/data': 0.00040169060230255127, 'timing/model': 0.8557755928486586, 'learning_rate/qwen_vl_interface': 9.40134815497596e-06, 'learning_rate/action_model': 9.40134815497596e-05, 'epoch': 0.13})
+[2026-07-28 07:42:57,079][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.08474993705749512, 'train/grad_norm_pre_clip': 2.6860499382019043, 'timing/data': 0.00033593736588954926, 'timing/model': 0.6185891637578607, 'learning_rate/qwen_vl_interface': 9.399487290570629e-06, 'learning_rate/action_model': 9.399487290570629e-05, 'epoch': 0.13})
+[2026-07-28 07:43:06,229][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.1190410926938057, 'train/grad_norm_pre_clip': 3.1708502769470215, 'timing/data': 0.0003414405509829521, 'timing/model': 0.619742345996201, 'learning_rate/qwen_vl_interface': 9.397623733608042e-06, 'learning_rate/action_model': 9.397623733608042e-05, 'epoch': 0.13})
+[2026-07-28 07:43:15,413][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.06341695785522461, 'train/grad_norm_pre_clip': 1.382280707359314, 'timing/data': 0.0005660774186253548, 'timing/model': 0.6187984813004732, 'learning_rate/qwen_vl_interface': 9.395757485297447e-06, 'learning_rate/action_model': 9.395757485297446e-05, 'epoch': 0.13})
+[2026-07-28 07:43:24,864][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.07538595795631409, 'train/grad_norm_pre_clip': 1.7227007150650024, 'timing/data': 0.00036964286118745804, 'timing/model': 0.6277975980192423, 'learning_rate/qwen_vl_interface': 9.393888546849828e-06, 'learning_rate/action_model': 9.393888546849828e-05, 'epoch': 0.13})
+[2026-07-28 07:43:34,117][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.08651502430438995, 'train/grad_norm_pre_clip': 1.9807555675506592, 'timing/data': 0.0003354167565703392, 'timing/model': 0.6200723182410002, 'learning_rate/qwen_vl_interface': 9.392016919477922e-06, 'learning_rate/action_model': 9.392016919477921e-05, 'epoch': 0.14})
+[2026-07-28 07:43:43,525][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.11118736863136292, 'train/grad_norm_pre_clip': 3.441889762878418, 'timing/data': 0.0003529982641339302, 'timing/model': 0.6246249256655574, 'learning_rate/qwen_vl_interface': 9.390142604396206e-06, 'learning_rate/action_model': 9.390142604396205e-05, 'epoch': 0.14})
+[2026-07-28 07:43:52,688][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.10039761662483215, 'train/grad_norm_pre_clip': 1.8030633926391602, 'timing/data': 0.0003306586295366287, 'timing/model': 0.6175351832062006, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.14})
+[2026-07-28 07:44:02,052][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.09281004220247269, 'train/grad_norm_pre_clip': 2.011645793914795, 'timing/data': 0.00032602716237306595, 'timing/model': 0.8537217564880848, 'learning_rate/qwen_vl_interface': 9.386385915969975e-06, 'learning_rate/action_model': 9.386385915969976e-05, 'epoch': 0.14})
+[2026-07-28 07:44:11,230][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.0785352885723114, 'train/grad_norm_pre_clip': 1.8099535703659058, 'timing/data': 0.0003395611420273781, 'timing/model': 0.6282903552055359, 'learning_rate/qwen_vl_interface': 9.384503545063138e-06, 'learning_rate/action_model': 9.384503545063137e-05, 'epoch': 0.14})
+[2026-07-28 07:44:20,652][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.0791485607624054, 'train/grad_norm_pre_clip': 1.4793154001235962, 'timing/data': 0.0004145270213484764, 'timing/model': 0.6218115165829659, 'learning_rate/qwen_vl_interface': 9.382618491321837e-06, 'learning_rate/action_model': 9.382618491321837e-05, 'epoch': 0.14})
+[2026-07-28 07:44:29,839][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.08828109502792358, 'train/grad_norm_pre_clip': 1.4435890913009644, 'timing/data': 0.00032377615571022034, 'timing/model': 0.6207565208896995, 'learning_rate/qwen_vl_interface': 9.380730755969265e-06, 'learning_rate/action_model': 9.380730755969264e-05, 'epoch': 0.14})
+[2026-07-28 07:44:38,968][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.0637422502040863, 'train/grad_norm_pre_clip': 1.611559510231018, 'timing/data': 0.00038356613367795944, 'timing/model': 0.6183238392695785, 'learning_rate/qwen_vl_interface': 9.378840340230353e-06, 'learning_rate/action_model': 9.378840340230352e-05, 'epoch': 0.14})
+[2026-07-28 07:44:48,351][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.11995501816272736, 'train/grad_norm_pre_clip': 2.856144905090332, 'timing/data': 0.0003343643620610237, 'timing/model': 0.8557203412055969, 'learning_rate/qwen_vl_interface': 9.376947245331769e-06, 'learning_rate/action_model': 9.376947245331767e-05, 'epoch': 0.14})
+[2026-07-28 07:44:57,638][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.055044323205947876, 'train/grad_norm_pre_clip': 0.8938788771629333, 'timing/data': 0.00040033459663391113, 'timing/model': 0.6242441646754742, 'learning_rate/qwen_vl_interface': 9.375051472501922e-06, 'learning_rate/action_model': 9.375051472501923e-05, 'epoch': 0.14})
+[2026-07-28 07:45:06,824][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.08749277889728546, 'train/grad_norm_pre_clip': 1.495093822479248, 'timing/data': 0.00035670679062604904, 'timing/model': 0.6324222618713975, 'learning_rate/qwen_vl_interface': 9.373153022970961e-06, 'learning_rate/action_model': 9.373153022970961e-05, 'epoch': 0.14})
+[2026-07-28 07:45:16,018][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.08078345656394958, 'train/grad_norm_pre_clip': 1.4084774255752563, 'timing/data': 0.0003711450845003128, 'timing/model': 0.6241451511159539, 'learning_rate/qwen_vl_interface': 9.371251897970768e-06, 'learning_rate/action_model': 9.371251897970768e-05, 'epoch': 0.14})
+[2026-07-28 07:45:25,411][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.06709074229001999, 'train/grad_norm_pre_clip': 1.1283406019210815, 'timing/data': 0.0004110652953386307, 'timing/model': 0.6169786220416427, 'learning_rate/qwen_vl_interface': 9.369348098734961e-06, 'learning_rate/action_model': 9.369348098734961e-05, 'epoch': 0.14})
+[2026-07-28 07:45:34,601][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.06394658237695694, 'train/grad_norm_pre_clip': 1.5401718616485596, 'timing/data': 0.00040914304554462433, 'timing/model': 0.6202212041243911, 'learning_rate/qwen_vl_interface': 9.367441626498897e-06, 'learning_rate/action_model': 9.367441626498896e-05, 'epoch': 0.14})
+[2026-07-28 07:45:43,752][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.06142561137676239, 'train/grad_norm_pre_clip': 1.7346415519714355, 'timing/data': 0.00039430800825357437, 'timing/model': 0.6201724288985133, 'learning_rate/qwen_vl_interface': 9.365532482499663e-06, 'learning_rate/action_model': 9.365532482499662e-05, 'epoch': 0.14})
+[2026-07-28 07:45:53,273][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 750
+[2026-07-28 07:46:03,476][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 750
+[2026-07-28 07:57:25,439][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 750
+[2026-07-28 07:57:25,440][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.07336843013763428, 'train/grad_norm_pre_clip': 1.3887391090393066, 'eval/action_loss/seconds': 10.198041471652687, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.07530420273542404, 'eval/latency_3/loss': 0.07530420273542404, 'eval/flappy/loss': 0.07530420273542404, 'eval/flappy/latency_3/loss': 0.07530420273542404, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 267.25000315085055, 'rl_games_eval/mid_train/mean_length': 2171.8, 'rl_games_eval/mid_train/std_reward': 160.98253665124207, 'rl_games_eval/mid_train/std_length': 1298.493419313321, 'rl_games_eval/mid_train/macro_mean_reward': 267.25000315085055, 'rl_games_eval/mid_train/macro_mean_length': 2171.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 267.25000315085055, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2171.8, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 160.98253665124207, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1298.493419313321, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 267.25000315085055, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00032638199627399445, 'timing/model': 0.7876205230131745, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.14})
+[2026-07-28 07:57:34,493][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.11076612770557404, 'train/grad_norm_pre_clip': 3.136133909225464, 'timing/data': 0.00034783873707056046, 'timing/model': 0.623661401681602, 'learning_rate/qwen_vl_interface': 9.361706184168714e-06, 'learning_rate/action_model': 9.361706184168713e-05, 'epoch': 0.14})
+[2026-07-28 07:57:43,664][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.08129740506410599, 'train/grad_norm_pre_clip': 1.7827099561691284, 'timing/data': 0.00048154033720493317, 'timing/model': 0.6167320106178522, 'learning_rate/qwen_vl_interface': 9.359789032319843e-06, 'learning_rate/action_model': 9.359789032319842e-05, 'epoch': 0.14})
+[2026-07-28 07:57:52,835][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.046341776847839355, 'train/grad_norm_pre_clip': 1.132717490196228, 'timing/data': 0.00034068338572978973, 'timing/model': 0.6224533170461655, 'learning_rate/qwen_vl_interface': 9.357869213673484e-06, 'learning_rate/action_model': 9.357869213673485e-05, 'epoch': 0.14})
+[2026-07-28 07:58:02,218][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.13964387774467468, 'train/grad_norm_pre_clip': 3.728337526321411, 'timing/data': 0.0003253920003771782, 'timing/model': 0.8498537503182888, 'learning_rate/qwen_vl_interface': 9.355946729475393e-06, 'learning_rate/action_model': 9.355946729475393e-05, 'epoch': 0.14})
+[2026-07-28 07:58:11,615][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.07147867232561111, 'train/grad_norm_pre_clip': 3.118619680404663, 'timing/data': 0.0003809398040175438, 'timing/model': 0.87024712190032, 'learning_rate/qwen_vl_interface': 9.354021580973049e-06, 'learning_rate/action_model': 9.354021580973048e-05, 'epoch': 0.14})
+[2026-07-28 07:58:20,825][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.0427265465259552, 'train/grad_norm_pre_clip': 1.5766206979751587, 'timing/data': 0.00035254843533039093, 'timing/model': 0.6344254994764924, 'learning_rate/qwen_vl_interface': 9.352093769415655e-06, 'learning_rate/action_model': 9.352093769415655e-05, 'epoch': 0.14})
+[2026-07-28 07:58:29,990][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.06522077322006226, 'train/grad_norm_pre_clip': 2.2276782989501953, 'timing/data': 0.00034127291291952133, 'timing/model': 0.6184469647705555, 'learning_rate/qwen_vl_interface': 9.35016329605415e-06, 'learning_rate/action_model': 9.350163296054151e-05, 'epoch': 0.14})
+[2026-07-28 07:58:39,125][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.03237544000148773, 'train/grad_norm_pre_clip': 0.7757171988487244, 'timing/data': 0.0003340980038046837, 'timing/model': 0.6173929441720247, 'learning_rate/qwen_vl_interface': 9.348230162141198e-06, 'learning_rate/action_model': 9.348230162141198e-05, 'epoch': 0.14})
+[2026-07-28 07:58:48,252][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.07669326663017273, 'train/grad_norm_pre_clip': 2.4011378288269043, 'timing/data': 0.00033167656511068344, 'timing/model': 0.6204907177016139, 'learning_rate/qwen_vl_interface': 9.346294368931188e-06, 'learning_rate/action_model': 9.346294368931188e-05, 'epoch': 0.14})
+[2026-07-28 07:58:57,696][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.07371298968791962, 'train/grad_norm_pre_clip': 1.7537189722061157, 'timing/data': 0.000654817558825016, 'timing/model': 0.6197634702548385, 'learning_rate/qwen_vl_interface': 9.344355917680235e-06, 'learning_rate/action_model': 9.344355917680235e-05, 'epoch': 0.14})
+[2026-07-28 07:59:06,919][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.10446713119745255, 'train/grad_norm_pre_clip': 3.057032823562622, 'timing/data': 0.00044359825551509857, 'timing/model': 0.6207080325111747, 'learning_rate/qwen_vl_interface': 9.342414809646179e-06, 'learning_rate/action_model': 9.342414809646179e-05, 'epoch': 0.14})
+[2026-07-28 07:59:16,420][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.09671363234519958, 'train/grad_norm_pre_clip': 2.301377534866333, 'timing/data': 0.00035168323665857315, 'timing/model': 0.6191522357985377, 'learning_rate/qwen_vl_interface': 9.340471046088581e-06, 'learning_rate/action_model': 9.34047104608858e-05, 'epoch': 0.14})
+[2026-07-28 07:59:25,832][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.08365459740161896, 'train/grad_norm_pre_clip': 1.5503453016281128, 'timing/data': 0.00042148493230342865, 'timing/model': 0.620785791426897, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.14})
+[2026-07-28 07:59:35,013][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.08950651437044144, 'train/grad_norm_pre_clip': 2.4898428916931152, 'timing/data': 0.0003269556909799576, 'timing/model': 0.6176925310865045, 'learning_rate/qwen_vl_interface': 9.336575557449641e-06, 'learning_rate/action_model': 9.336575557449641e-05, 'epoch': 0.14})
+[2026-07-28 07:59:44,166][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.03958010673522949, 'train/grad_norm_pre_clip': 1.094218134880066, 'timing/data': 0.00042268913239240646, 'timing/model': 0.6182065149769187, 'learning_rate/qwen_vl_interface': 9.334623834896039e-06, 'learning_rate/action_model': 9.334623834896038e-05, 'epoch': 0.14})
+[2026-07-28 07:59:53,796][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.08139374852180481, 'train/grad_norm_pre_clip': 1.3300505876541138, 'timing/data': 0.0003564562648534775, 'timing/model': 1.011129324324429, 'learning_rate/qwen_vl_interface': 9.332669461874376e-06, 'learning_rate/action_model': 9.332669461874374e-05, 'epoch': 0.14})
+[2026-07-28 08:00:02,799][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.16192032396793365, 'train/grad_norm_pre_clip': 4.367000102996826, 'timing/data': 0.0004193643108010292, 'timing/model': 0.61887670122087, 'learning_rate/qwen_vl_interface': 9.330712439652824e-06, 'learning_rate/action_model': 9.330712439652823e-05, 'epoch': 0.14})
+[2026-07-28 08:00:11,714][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.08243609219789505, 'train/grad_norm_pre_clip': 1.9525810480117798, 'timing/data': 0.00041702575981616974, 'timing/model': 0.6187398470938206, 'learning_rate/qwen_vl_interface': 9.328752769501272e-06, 'learning_rate/action_model': 9.328752769501272e-05, 'epoch': 0.14})
+[2026-07-28 08:00:20,871][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.10517717897891998, 'train/grad_norm_pre_clip': 2.0812911987304688, 'timing/data': 0.0004371693357825279, 'timing/model': 0.6191135449334979, 'learning_rate/qwen_vl_interface': 9.326790452691332e-06, 'learning_rate/action_model': 9.326790452691332e-05, 'epoch': 0.14})
+[2026-07-28 08:00:30,070][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.0951431542634964, 'train/grad_norm_pre_clip': 1.1305872201919556, 'timing/data': 0.0003523118793964386, 'timing/model': 0.618659078143537, 'learning_rate/qwen_vl_interface': 9.324825490496327e-06, 'learning_rate/action_model': 9.324825490496327e-05, 'epoch': 0.14})
+[2026-07-28 08:00:39,694][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.06204264611005783, 'train/grad_norm_pre_clip': 1.193535327911377, 'timing/data': 0.00033461861312389374, 'timing/model': 0.6212996654212475, 'learning_rate/qwen_vl_interface': 9.322857884191303e-06, 'learning_rate/action_model': 9.322857884191302e-05, 'epoch': 0.14})
+[2026-07-28 08:00:48,888][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.06332947313785553, 'train/grad_norm_pre_clip': 2.371683359146118, 'timing/data': 0.00045165885239839554, 'timing/model': 0.6182630248367786, 'learning_rate/qwen_vl_interface': 9.320887635053013e-06, 'learning_rate/action_model': 9.320887635053014e-05, 'epoch': 0.14})
+[2026-07-28 08:00:58,007][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.12781274318695068, 'train/grad_norm_pre_clip': 3.2551212310791016, 'timing/data': 0.00034541357308626175, 'timing/model': 0.6186812873929739, 'learning_rate/qwen_vl_interface': 9.318914744359936e-06, 'learning_rate/action_model': 9.318914744359937e-05, 'epoch': 0.14})
+[2026-07-28 08:01:07,162][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.13593894243240356, 'train/grad_norm_pre_clip': 4.035465240478516, 'timing/data': 0.0003515845164656639, 'timing/model': 0.6212454931810498, 'learning_rate/qwen_vl_interface': 9.316939213392257e-06, 'learning_rate/action_model': 9.316939213392257e-05, 'epoch': 0.14})
+[2026-07-28 08:01:16,589][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.0808902382850647, 'train/grad_norm_pre_clip': 2.429096221923828, 'timing/data': 0.00040715839713811874, 'timing/model': 0.8658560533076525, 'learning_rate/qwen_vl_interface': 9.314961043431877e-06, 'learning_rate/action_model': 9.314961043431876e-05, 'epoch': 0.14})
+[2026-07-28 08:01:26,023][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.06371662020683289, 'train/grad_norm_pre_clip': 1.5181673765182495, 'timing/data': 0.00038335565477609634, 'timing/model': 0.6176124690100551, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.14})
+[2026-07-28 08:01:35,148][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.07235865294933319, 'train/grad_norm_pre_clip': 1.6630322933197021, 'timing/data': 0.00041364412754774094, 'timing/model': 0.618868961930275, 'learning_rate/qwen_vl_interface': 9.310996791669175e-06, 'learning_rate/action_model': 9.310996791669175e-05, 'epoch': 0.14})
+[2026-07-28 08:01:44,291][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.06782689690589905, 'train/grad_norm_pre_clip': 1.1609668731689453, 'timing/data': 0.00041212793439626694, 'timing/model': 0.6197572955861688, 'learning_rate/qwen_vl_interface': 9.309010712439213e-06, 'learning_rate/action_model': 9.309010712439213e-05, 'epoch': 0.14})
+[2026-07-28 08:01:53,438][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.07322309911251068, 'train/grad_norm_pre_clip': 1.3591480255126953, 'timing/data': 0.00035951193422079086, 'timing/model': 0.6205755537375808, 'learning_rate/qwen_vl_interface': 9.307021999361267e-06, 'learning_rate/action_model': 9.307021999361267e-05, 'epoch': 0.14})
+[2026-07-28 08:02:02,888][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.09179334342479706, 'train/grad_norm_pre_clip': 2.1544623374938965, 'timing/data': 0.0004186555743217468, 'timing/model': 0.6221468839794397, 'learning_rate/qwen_vl_interface': 9.305030653725794e-06, 'learning_rate/action_model': 9.305030653725794e-05, 'epoch': 0.14})
+[2026-07-28 08:02:12,242][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.04635012894868851, 'train/grad_norm_pre_clip': 1.5674793720245361, 'timing/data': 0.0003333520144224167, 'timing/model': 0.8403609786182642, 'learning_rate/qwen_vl_interface': 9.303036676824954e-06, 'learning_rate/action_model': 9.303036676824953e-05, 'epoch': 0.14})
+[2026-07-28 08:02:21,370][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.03790777549147606, 'train/grad_norm_pre_clip': 0.7236562967300415, 'timing/data': 0.00033289846032857895, 'timing/model': 0.6172589231282473, 'learning_rate/qwen_vl_interface': 9.301040069952616e-06, 'learning_rate/action_model': 9.301040069952616e-05, 'epoch': 0.14})
+[2026-07-28 08:02:30,477][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.06318258494138718, 'train/grad_norm_pre_clip': 1.7653560638427734, 'timing/data': 0.00032815616577863693, 'timing/model': 0.6187874022871256, 'learning_rate/qwen_vl_interface': 9.299040834404362e-06, 'learning_rate/action_model': 9.299040834404361e-05, 'epoch': 0.14})
+[2026-07-28 08:02:39,626][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.09083744883537292, 'train/grad_norm_pre_clip': 3.394761800765991, 'timing/data': 0.0005148937925696373, 'timing/model': 0.6214467482641339, 'learning_rate/qwen_vl_interface': 9.297038971477468e-06, 'learning_rate/action_model': 9.297038971477466e-05, 'epoch': 0.14})
+[2026-07-28 08:02:49,063][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.06736037880182266, 'train/grad_norm_pre_clip': 2.6943323612213135, 'timing/data': 0.000418701209127903, 'timing/model': 0.6209881082177162, 'learning_rate/qwen_vl_interface': 9.295034482470922e-06, 'learning_rate/action_model': 9.295034482470923e-05, 'epoch': 0.14})
+[2026-07-28 08:02:58,531][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.05980757623910904, 'train/grad_norm_pre_clip': 1.718532919883728, 'timing/data': 0.0003453930839896202, 'timing/model': 0.6289604948833585, 'learning_rate/qwen_vl_interface': 9.293027368685418e-06, 'learning_rate/action_model': 9.293027368685419e-05, 'epoch': 0.14})
+[2026-07-28 08:03:07,503][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.04085984453558922, 'train/grad_norm_pre_clip': 0.8290616869926453, 'timing/data': 0.0004814649000763893, 'timing/model': 0.6216761423274875, 'learning_rate/qwen_vl_interface': 9.29101763142335e-06, 'learning_rate/action_model': 9.29101763142335e-05, 'epoch': 0.14})
+[2026-07-28 08:03:16,476][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.06637638062238693, 'train/grad_norm_pre_clip': 1.39484703540802, 'timing/data': 0.0003612516447901726, 'timing/model': 0.6204897053539753, 'learning_rate/qwen_vl_interface': 9.28900527198881e-06, 'learning_rate/action_model': 9.289005271988811e-05, 'epoch': 0.14})
+[2026-07-28 08:03:25,956][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.08780801296234131, 'train/grad_norm_pre_clip': 1.294278621673584, 'timing/data': 0.0004274072125554085, 'timing/model': 0.7134032966569066, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.14})
+[2026-07-28 08:03:35,415][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.10088978707790375, 'train/grad_norm_pre_clip': 2.7865543365478516, 'timing/data': 0.00033635087311267853, 'timing/model': 0.6193080302327871, 'learning_rate/qwen_vl_interface': 9.284972691827226e-06, 'learning_rate/action_model': 9.284972691827224e-05, 'epoch': 0.15})
+[2026-07-28 08:03:44,567][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.08381348103284836, 'train/grad_norm_pre_clip': 2.0404553413391113, 'timing/data': 0.00033270474523305893, 'timing/model': 0.6216874066740274, 'learning_rate/qwen_vl_interface': 9.282952473716873e-06, 'learning_rate/action_model': 9.282952473716874e-05, 'epoch': 0.15})
+[2026-07-28 08:03:53,697][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.07065826654434204, 'train/grad_norm_pre_clip': 1.4673548936843872, 'timing/data': 0.00033188797533512115, 'timing/model': 0.6170021770521998, 'learning_rate/qwen_vl_interface': 9.280929638667447e-06, 'learning_rate/action_model': 9.280929638667447e-05, 'epoch': 0.15})
+[2026-07-28 08:04:03,087][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.08490440249443054, 'train/grad_norm_pre_clip': 2.6011056900024414, 'timing/data': 0.00033804215490818024, 'timing/model': 0.8508146572858095, 'learning_rate/qwen_vl_interface': 9.278904187991542e-06, 'learning_rate/action_model': 9.278904187991541e-05, 'epoch': 0.15})
+[2026-07-28 08:04:12,225][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.04733826965093613, 'train/grad_norm_pre_clip': 1.5062134265899658, 'timing/data': 0.00043057650327682495, 'timing/model': 0.6175264650955796, 'learning_rate/qwen_vl_interface': 9.27687612300345e-06, 'learning_rate/action_model': 9.27687612300345e-05, 'epoch': 0.15})
+[2026-07-28 08:04:21,439][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.09445004910230637, 'train/grad_norm_pre_clip': 1.6660457849502563, 'timing/data': 0.00044090766459703445, 'timing/model': 0.6190836913883686, 'learning_rate/qwen_vl_interface': 9.27484544501916e-06, 'learning_rate/action_model': 9.274845445019159e-05, 'epoch': 0.15})
+[2026-07-28 08:04:30,598][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.10740561783313751, 'train/grad_norm_pre_clip': 2.680415391921997, 'timing/data': 0.0003420105203986168, 'timing/model': 0.6178726088255644, 'learning_rate/qwen_vl_interface': 9.27281215535636e-06, 'learning_rate/action_model': 9.272812155356358e-05, 'epoch': 0.15})
+[2026-07-28 08:04:39,951][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.07825399935245514, 'train/grad_norm_pre_clip': 2.4336233139038086, 'timing/data': 0.0003205724060535431, 'timing/model': 0.7962010502815247, 'learning_rate/qwen_vl_interface': 9.270776255334425e-06, 'learning_rate/action_model': 9.270776255334425e-05, 'epoch': 0.15})
+[2026-07-28 08:04:49,123][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.09689567983150482, 'train/grad_norm_pre_clip': 1.2672395706176758, 'timing/data': 0.0003265487030148506, 'timing/model': 0.6184349749237299, 'learning_rate/qwen_vl_interface': 9.268737746274432e-06, 'learning_rate/action_model': 9.268737746274432e-05, 'epoch': 0.15})
+[2026-07-28 08:04:58,491][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.054652050137519836, 'train/grad_norm_pre_clip': 1.198659062385559, 'timing/data': 0.00045247096568346024, 'timing/model': 0.6215537823736668, 'learning_rate/qwen_vl_interface': 9.266696629499146e-06, 'learning_rate/action_model': 9.266696629499145e-05, 'epoch': 0.15})
+[2026-07-28 08:05:07,658][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.08228997886180878, 'train/grad_norm_pre_clip': 0.9299641251564026, 'timing/data': 0.00035081058740615845, 'timing/model': 0.6201327508315444, 'learning_rate/qwen_vl_interface': 9.264652906333026e-06, 'learning_rate/action_model': 9.264652906333026e-05, 'epoch': 0.15})
+[2026-07-28 08:05:17,071][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.046182818710803986, 'train/grad_norm_pre_clip': 0.878710925579071, 'timing/data': 0.00042506586760282516, 'timing/model': 0.8734508352354169, 'learning_rate/qwen_vl_interface': 9.26260657810222e-06, 'learning_rate/action_model': 9.262606578102221e-05, 'epoch': 0.15})
+[2026-07-28 08:05:26,269][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.04182980954647064, 'train/grad_norm_pre_clip': 2.13493013381958, 'timing/data': 0.0004025958478450775, 'timing/model': 0.6189996041357517, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.15})
+[2026-07-28 08:05:35,642][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.04537711292505264, 'train/grad_norm_pre_clip': 1.0849740505218506, 'timing/data': 0.0003370959311723709, 'timing/model': 0.8472723048180342, 'learning_rate/qwen_vl_interface': 9.258506111759608e-06, 'learning_rate/action_model': 9.258506111759609e-05, 'epoch': 0.15})
+[2026-07-28 08:05:44,799][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.0457848496735096, 'train/grad_norm_pre_clip': 1.497774600982666, 'timing/data': 0.0003931531682610512, 'timing/model': 0.6174820903688669, 'learning_rate/qwen_vl_interface': 9.256451976308548e-06, 'learning_rate/action_model': 9.25645197630855e-05, 'epoch': 0.15})
+[2026-07-28 08:05:54,228][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.07552428543567657, 'train/grad_norm_pre_clip': 2.006563901901245, 'timing/data': 0.0003720596432685852, 'timing/model': 0.6200732029974461, 'learning_rate/qwen_vl_interface': 9.2543952411143e-06, 'learning_rate/action_model': 9.254395241114299e-05, 'epoch': 0.15})
+[2026-07-28 08:06:03,364][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.11266537755727768, 'train/grad_norm_pre_clip': 3.624786376953125, 'timing/data': 0.00040087010711431503, 'timing/model': 0.6190589424222708, 'learning_rate/qwen_vl_interface': 9.252335907511452e-06, 'learning_rate/action_model': 9.252335907511453e-05, 'epoch': 0.15})
+[2026-07-28 08:06:12,734][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.04597185552120209, 'train/grad_norm_pre_clip': 1.1958774328231812, 'timing/data': 0.0003451108932495117, 'timing/model': 0.6257853647693992, 'learning_rate/qwen_vl_interface': 9.250273976836289e-06, 'learning_rate/action_model': 9.250273976836289e-05, 'epoch': 0.15})
+[2026-07-28 08:06:22,069][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.04562125355005264, 'train/grad_norm_pre_clip': 1.6604735851287842, 'timing/data': 0.0004382506012916565, 'timing/model': 0.6778494454920292, 'learning_rate/qwen_vl_interface': 9.248209450426772e-06, 'learning_rate/action_model': 9.248209450426772e-05, 'epoch': 0.15})
+[2026-07-28 08:06:31,561][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.04921426251530647, 'train/grad_norm_pre_clip': 2.2898449897766113, 'timing/data': 0.00040596816688776016, 'timing/model': 0.6182621214538813, 'learning_rate/qwen_vl_interface': 9.24614232962255e-06, 'learning_rate/action_model': 9.24614232962255e-05, 'epoch': 0.15})
+[2026-07-28 08:06:41,025][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.12702767550945282, 'train/grad_norm_pre_clip': 5.75615119934082, 'timing/data': 0.0003835288807749748, 'timing/model': 0.851268152706325, 'learning_rate/qwen_vl_interface': 9.244072615764957e-06, 'learning_rate/action_model': 9.244072615764956e-05, 'epoch': 0.15})
+[2026-07-28 08:06:50,189][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.06257140636444092, 'train/grad_norm_pre_clip': 3.61785626411438, 'timing/data': 0.00040863826870918274, 'timing/model': 0.6176332226023078, 'learning_rate/qwen_vl_interface': 9.242000310197005e-06, 'learning_rate/action_model': 9.242000310197004e-05, 'epoch': 0.15})
+[2026-07-28 08:06:59,320][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.03911759331822395, 'train/grad_norm_pre_clip': 0.9680470824241638, 'timing/data': 0.0003831107169389725, 'timing/model': 0.6179485935717821, 'learning_rate/qwen_vl_interface': 9.239925414263392e-06, 'learning_rate/action_model': 9.239925414263392e-05, 'epoch': 0.15})
+[2026-07-28 08:07:08,450][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.12255828082561493, 'train/grad_norm_pre_clip': 2.9796411991119385, 'timing/data': 0.0003895079717040062, 'timing/model': 0.6165449349209666, 'learning_rate/qwen_vl_interface': 9.237847929310497e-06, 'learning_rate/action_model': 9.237847929310496e-05, 'epoch': 0.15})
+[2026-07-28 08:07:17,792][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.07613804936408997, 'train/grad_norm_pre_clip': 3.39928936958313, 'timing/data': 0.0006135217845439911, 'timing/model': 0.6198185281828046, 'learning_rate/qwen_vl_interface': 9.235767856686375e-06, 'learning_rate/action_model': 9.235767856686375e-05, 'epoch': 0.15})
+[2026-07-28 08:07:27,207][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.05543237924575806, 'train/grad_norm_pre_clip': 1.2947406768798828, 'timing/data': 0.0003967154771089554, 'timing/model': 0.6225862093269825, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.15})
+[2026-07-28 08:07:36,318][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.06846130639314651, 'train/grad_norm_pre_clip': 2.2837107181549072, 'timing/data': 0.00042426493018865585, 'timing/model': 0.6174678644165397, 'learning_rate/qwen_vl_interface': 9.23159995382508e-06, 'learning_rate/action_model': 9.231599953825079e-05, 'epoch': 0.15})
+[2026-07-28 08:07:45,443][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.059325382113456726, 'train/grad_norm_pre_clip': 1.1330304145812988, 'timing/data': 0.000392768532037735, 'timing/model': 0.6185298627242446, 'learning_rate/qwen_vl_interface': 9.22951212629241e-06, 'learning_rate/action_model': 9.229512126292411e-05, 'epoch': 0.15})
+[2026-07-28 08:07:54,665][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.10449279844760895, 'train/grad_norm_pre_clip': 1.523922324180603, 'timing/data': 0.00035860389471054077, 'timing/model': 0.6209145598113537, 'learning_rate/qwen_vl_interface': 9.227421716497531e-06, 'learning_rate/action_model': 9.227421716497531e-05, 'epoch': 0.15})
+[2026-07-28 08:08:04,092][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.09017477929592133, 'train/grad_norm_pre_clip': 2.99363374710083, 'timing/data': 0.0004043951630592346, 'timing/model': 0.6204433208331466, 'learning_rate/qwen_vl_interface': 9.225328725796879e-06, 'learning_rate/action_model': 9.22532872579688e-05, 'epoch': 0.15})
+[2026-07-28 08:08:13,532][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.05152967572212219, 'train/grad_norm_pre_clip': 0.9200685620307922, 'timing/data': 0.0004077814519405365, 'timing/model': 0.8574552629143, 'learning_rate/qwen_vl_interface': 9.223233155548579e-06, 'learning_rate/action_model': 9.223233155548578e-05, 'epoch': 0.15})
+[2026-07-28 08:08:22,670][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.04478662088513374, 'train/grad_norm_pre_clip': 1.982418179512024, 'timing/data': 0.0004269639030098915, 'timing/model': 0.6201948178932071, 'learning_rate/qwen_vl_interface': 9.22113500711242e-06, 'learning_rate/action_model': 9.221135007112419e-05, 'epoch': 0.15})
+[2026-07-28 08:08:32,025][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.0699833407998085, 'train/grad_norm_pre_clip': 1.909158706665039, 'timing/data': 0.00033670756965875626, 'timing/model': 0.843558132648468, 'learning_rate/qwen_vl_interface': 9.219034281849867e-06, 'learning_rate/action_model': 9.219034281849867e-05, 'epoch': 0.15})
+[2026-07-28 08:08:41,171][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.08186419308185577, 'train/grad_norm_pre_clip': 3.6755101680755615, 'timing/data': 0.0003547677770256996, 'timing/model': 0.6191982496529818, 'learning_rate/qwen_vl_interface': 9.216930981124062e-06, 'learning_rate/action_model': 9.216930981124062e-05, 'epoch': 0.15})
+[2026-07-28 08:08:50,307][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.08579348772764206, 'train/grad_norm_pre_clip': 2.1493980884552, 'timing/data': 0.0004126857966184616, 'timing/model': 0.6187393944710493, 'learning_rate/qwen_vl_interface': 9.214825106299808e-06, 'learning_rate/action_model': 9.214825106299808e-05, 'epoch': 0.15})
+[2026-07-28 08:08:59,667][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.10848002880811691, 'train/grad_norm_pre_clip': 2.4691503047943115, 'timing/data': 0.0003888336941599846, 'timing/model': 0.6322597125545144, 'learning_rate/qwen_vl_interface': 9.212716658743592e-06, 'learning_rate/action_model': 9.212716658743592e-05, 'epoch': 0.15})
+[2026-07-28 08:09:08,801][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.040525563061237335, 'train/grad_norm_pre_clip': 1.8169612884521484, 'timing/data': 0.00033728964626789093, 'timing/model': 0.61941047757864, 'learning_rate/qwen_vl_interface': 9.210605639823558e-06, 'learning_rate/action_model': 9.210605639823557e-05, 'epoch': 0.15})
+[2026-07-28 08:09:18,282][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.0722985714673996, 'train/grad_norm_pre_clip': 1.494497299194336, 'timing/data': 0.0003979634493589401, 'timing/model': 0.8641464337706566, 'learning_rate/qwen_vl_interface': 9.208492050909523e-06, 'learning_rate/action_model': 9.208492050909523e-05, 'epoch': 0.15})
+[2026-07-28 08:09:27,701][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.04726685956120491, 'train/grad_norm_pre_clip': 1.4200245141983032, 'timing/data': 0.0004703933373093605, 'timing/model': 0.633738411590457, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.15})
+[2026-07-28 08:09:36,876][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.06569384038448334, 'train/grad_norm_pre_clip': 2.096134901046753, 'timing/data': 0.00033136364072561264, 'timing/model': 0.6200004629790783, 'learning_rate/qwen_vl_interface': 9.204257168587061e-06, 'learning_rate/action_model': 9.204257168587061e-05, 'epoch': 0.15})
+[2026-07-28 08:09:46,062][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.10228857398033142, 'train/grad_norm_pre_clip': 2.5309715270996094, 'timing/data': 0.00032859016209840775, 'timing/model': 0.6179956775158644, 'learning_rate/qwen_vl_interface': 9.2021358779266e-06, 'learning_rate/action_model': 9.202135877926601e-05, 'epoch': 0.15})
+[2026-07-28 08:09:55,277][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.09677895903587341, 'train/grad_norm_pre_clip': 3.006647825241089, 'timing/data': 0.00033972691744565964, 'timing/model': 0.6181622073054314, 'learning_rate/qwen_vl_interface': 9.200012022768077e-06, 'learning_rate/action_model': 9.200012022768077e-05, 'epoch': 0.15})
+[2026-07-28 08:10:04,679][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.10562705248594284, 'train/grad_norm_pre_clip': 2.090036630630493, 'timing/data': 0.0003308244049549103, 'timing/model': 0.6177651705220342, 'learning_rate/qwen_vl_interface': 9.19788560448964e-06, 'learning_rate/action_model': 9.197885604489639e-05, 'epoch': 0.15})
+[2026-07-28 08:10:14,057][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.09956761449575424, 'train/grad_norm_pre_clip': 2.769526958465576, 'timing/data': 0.0003937873989343643, 'timing/model': 0.6193417077884078, 'learning_rate/qwen_vl_interface': 9.195756624471094e-06, 'learning_rate/action_model': 9.195756624471094e-05, 'epoch': 0.15})
+[2026-07-28 08:10:23,164][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.07347197830677032, 'train/grad_norm_pre_clip': 1.7293334007263184, 'timing/data': 0.00033577345311641693, 'timing/model': 0.6188225904479623, 'learning_rate/qwen_vl_interface': 9.193625084093912e-06, 'learning_rate/action_model': 9.193625084093911e-05, 'epoch': 0.15})
+[2026-07-28 08:10:32,310][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.06494221836328506, 'train/grad_norm_pre_clip': 0.8578107953071594, 'timing/data': 0.00031925830990076065, 'timing/model': 0.6214936822652817, 'learning_rate/qwen_vl_interface': 9.191490984741227e-06, 'learning_rate/action_model': 9.191490984741227e-05, 'epoch': 0.15})
+[2026-07-28 08:10:41,699][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.04933859407901764, 'train/grad_norm_pre_clip': 1.1019865274429321, 'timing/data': 0.000337919220328331, 'timing/model': 0.6202285997569561, 'learning_rate/qwen_vl_interface': 9.189354327797835e-06, 'learning_rate/action_model': 9.189354327797836e-05, 'epoch': 0.15})
+[2026-07-28 08:10:51,104][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.07212871313095093, 'train/grad_norm_pre_clip': 1.6872749328613281, 'timing/data': 0.00034057069569826126, 'timing/model': 0.85291120223701, 'learning_rate/qwen_vl_interface': 9.18721511465019e-06, 'learning_rate/action_model': 9.187215114650191e-05, 'epoch': 0.15})
+[2026-07-28 08:11:00,275][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.07665195316076279, 'train/grad_norm_pre_clip': 2.1435296535491943, 'timing/data': 0.0006128326058387756, 'timing/model': 0.6274296473711729, 'learning_rate/qwen_vl_interface': 9.185073346686402e-06, 'learning_rate/action_model': 9.185073346686401e-05, 'epoch': 0.15})
+[2026-07-28 08:11:09,413][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.16739630699157715, 'train/grad_norm_pre_clip': 6.286211013793945, 'timing/data': 0.00035929400473833084, 'timing/model': 0.6182170240208507, 'learning_rate/qwen_vl_interface': 9.182929025296242e-06, 'learning_rate/action_model': 9.182929025296242e-05, 'epoch': 0.15})
+[2026-07-28 08:11:18,858][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.15761017799377441, 'train/grad_norm_pre_clip': 2.5913174152374268, 'timing/data': 0.00040267035365104675, 'timing/model': 0.8596248337998986, 'learning_rate/qwen_vl_interface': 9.180782151871138e-06, 'learning_rate/action_model': 9.180782151871137e-05, 'epoch': 0.15})
+[2026-07-28 08:11:28,044][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.12428736686706543, 'train/grad_norm_pre_clip': 1.4566506147384644, 'timing/data': 0.0005559874698519707, 'timing/model': 0.6181086888536811, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.15})
+[2026-07-28 08:11:37,511][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.06579606980085373, 'train/grad_norm_pre_clip': 1.3423298597335815, 'timing/data': 0.00034590624272823334, 'timing/model': 0.8581305453553796, 'learning_rate/qwen_vl_interface': 9.176480754490078e-06, 'learning_rate/action_model': 9.176480754490078e-05, 'epoch': 0.15})
+[2026-07-28 08:11:46,679][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.03043639473617077, 'train/grad_norm_pre_clip': 1.2640385627746582, 'timing/data': 0.0004041232168674469, 'timing/model': 0.6181063950061798, 'learning_rate/qwen_vl_interface': 9.174326233325253e-06, 'learning_rate/action_model': 9.174326233325253e-05, 'epoch': 0.15})
+[2026-07-28 08:11:55,768][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.03805715590715408, 'train/grad_norm_pre_clip': 1.1345075368881226, 'timing/data': 0.00034504756331443787, 'timing/model': 0.6190967913717031, 'learning_rate/qwen_vl_interface': 9.17216916570774e-06, 'learning_rate/action_model': 9.172169165707741e-05, 'epoch': 0.16})
+[2026-07-28 08:12:04,902][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.09860049188137054, 'train/grad_norm_pre_clip': 1.5736123323440552, 'timing/data': 0.0003408268094062805, 'timing/model': 0.6185046955943108, 'learning_rate/qwen_vl_interface': 9.170009553037239e-06, 'learning_rate/action_model': 9.170009553037238e-05, 'epoch': 0.16})
+[2026-07-28 08:12:14,265][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.10105479508638382, 'train/grad_norm_pre_clip': 2.1389801502227783, 'timing/data': 0.00041857827454805374, 'timing/model': 0.6187261575832963, 'learning_rate/qwen_vl_interface': 9.167847396715098e-06, 'learning_rate/action_model': 9.167847396715098e-05, 'epoch': 0.16})
+[2026-07-28 08:12:23,394][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.042522016912698746, 'train/grad_norm_pre_clip': 0.9390051960945129, 'timing/data': 0.00033339671790599823, 'timing/model': 0.6171190589666367, 'learning_rate/qwen_vl_interface': 9.165682698144317e-06, 'learning_rate/action_model': 9.165682698144317e-05, 'epoch': 0.16})
+[2026-07-28 08:12:32,504][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.06236938014626503, 'train/grad_norm_pre_clip': 1.1861368417739868, 'timing/data': 0.0003431113436818123, 'timing/model': 0.6188915725797415, 'learning_rate/qwen_vl_interface': 9.163515458729545e-06, 'learning_rate/action_model': 9.163515458729545e-05, 'epoch': 0.16})
+[2026-07-28 08:12:42,036][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.1001528799533844, 'train/grad_norm_pre_clip': 1.9489549398422241, 'timing/data': 0.00033510103821754456, 'timing/model': 0.6187163628637791, 'learning_rate/qwen_vl_interface': 9.16134567987708e-06, 'learning_rate/action_model': 9.16134567987708e-05, 'epoch': 0.16})
+[2026-07-28 08:12:51,118][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.05152402073144913, 'train/grad_norm_pre_clip': 1.1408543586730957, 'timing/data': 0.00034147128462791443, 'timing/model': 0.6195501144975424, 'learning_rate/qwen_vl_interface': 9.159173362994868e-06, 'learning_rate/action_model': 9.159173362994869e-05, 'epoch': 0.16})
+[2026-07-28 08:13:00,257][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.051003821194171906, 'train/grad_norm_pre_clip': 1.2006062269210815, 'timing/data': 0.0008996762335300446, 'timing/model': 0.6198132839053869, 'learning_rate/qwen_vl_interface': 9.156998509492503e-06, 'learning_rate/action_model': 9.156998509492503e-05, 'epoch': 0.16})
+[2026-07-28 08:13:09,664][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.0669175535440445, 'train/grad_norm_pre_clip': 0.9192629456520081, 'timing/data': 0.00034385453909635544, 'timing/model': 0.7889091083779931, 'learning_rate/qwen_vl_interface': 9.154821120781221e-06, 'learning_rate/action_model': 9.154821120781222e-05, 'epoch': 0.16})
+[2026-07-28 08:13:18,967][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.059086404740810394, 'train/grad_norm_pre_clip': 1.2798094749450684, 'timing/data': 0.00038646068423986435, 'timing/model': 0.6177609879523516, 'learning_rate/qwen_vl_interface': 9.152641198273911e-06, 'learning_rate/action_model': 9.15264119827391e-05, 'epoch': 0.16})
+[2026-07-28 08:13:28,253][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.0941745936870575, 'train/grad_norm_pre_clip': 2.843312978744507, 'timing/data': 0.00032541435211896896, 'timing/model': 0.6176253538578749, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.16})
+[2026-07-28 08:13:37,832][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.07386486232280731, 'train/grad_norm_pre_clip': 1.768151044845581, 'timing/data': 0.0003294786438345909, 'timing/model': 0.664617083966732, 'learning_rate/qwen_vl_interface': 9.148273757530948e-06, 'learning_rate/action_model': 9.148273757530948e-05, 'epoch': 0.16})
+[2026-07-28 08:13:47,200][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.08316578716039658, 'train/grad_norm_pre_clip': 1.703037142753601, 'timing/data': 0.0003611324355006218, 'timing/model': 0.6175655657425523, 'learning_rate/qwen_vl_interface': 9.146086242129283e-06, 'learning_rate/action_model': 9.146086242129282e-05, 'epoch': 0.16})
+[2026-07-28 08:13:56,164][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.04518115147948265, 'train/grad_norm_pre_clip': 1.547460913658142, 'timing/data': 0.00039356574416160583, 'timing/model': 0.6179533684626222, 'learning_rate/qwen_vl_interface': 9.143896198599553e-06, 'learning_rate/action_model': 9.143896198599553e-05, 'epoch': 0.16})
+[2026-07-28 08:14:05,105][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.06409307569265366, 'train/grad_norm_pre_clip': 1.9872797727584839, 'timing/data': 0.00033361464738845825, 'timing/model': 0.6164167374372482, 'learning_rate/qwen_vl_interface': 9.141703628362858e-06, 'learning_rate/action_model': 9.141703628362858e-05, 'epoch': 0.16})
+[2026-07-28 08:14:14,022][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.046516042202711105, 'train/grad_norm_pre_clip': 1.4105002880096436, 'timing/data': 0.0003301771357655525, 'timing/model': 0.6178084788843989, 'learning_rate/qwen_vl_interface': 9.139508532841928e-06, 'learning_rate/action_model': 9.139508532841927e-05, 'epoch': 0.16})
+[2026-07-28 08:14:23,635][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.07434538751840591, 'train/grad_norm_pre_clip': 2.52439546585083, 'timing/data': 0.0003362540155649185, 'timing/model': 0.8534533148631454, 'learning_rate/qwen_vl_interface': 9.137310913461139e-06, 'learning_rate/action_model': 9.137310913461139e-05, 'epoch': 0.16})
+[2026-07-28 08:14:32,804][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.1249217540025711, 'train/grad_norm_pre_clip': 6.714706897735596, 'timing/data': 0.0003600018098950386, 'timing/model': 0.6183893540874124, 'learning_rate/qwen_vl_interface': 9.135110771646502e-06, 'learning_rate/action_model': 9.135110771646502e-05, 'epoch': 0.16})
+[2026-07-28 08:14:41,972][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.12824399769306183, 'train/grad_norm_pre_clip': 5.611670017242432, 'timing/data': 0.0004494376480579376, 'timing/model': 0.6184448283165693, 'learning_rate/qwen_vl_interface': 9.132908108825667e-06, 'learning_rate/action_model': 9.132908108825667e-05, 'epoch': 0.16})
+[2026-07-28 08:14:51,162][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.07502228021621704, 'train/grad_norm_pre_clip': 2.462770700454712, 'timing/data': 0.0003425413742661476, 'timing/model': 0.6188714373856783, 'learning_rate/qwen_vl_interface': 9.130702926427915e-06, 'learning_rate/action_model': 9.130702926427914e-05, 'epoch': 0.16})
+[2026-07-28 08:15:00,579][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.09796585887670517, 'train/grad_norm_pre_clip': 3.2741458415985107, 'timing/data': 0.00034720171242952347, 'timing/model': 0.6181085593998432, 'learning_rate/qwen_vl_interface': 9.128495225884167e-06, 'learning_rate/action_model': 9.128495225884166e-05, 'epoch': 0.16})
+[2026-07-28 08:15:09,757][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.09952399134635925, 'train/grad_norm_pre_clip': 2.0886101722717285, 'timing/data': 0.00032173097133636475, 'timing/model': 0.6177833881229162, 'learning_rate/qwen_vl_interface': 9.126285008626975e-06, 'learning_rate/action_model': 9.126285008626974e-05, 'epoch': 0.16})
+[2026-07-28 08:15:19,122][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.12070675194263458, 'train/grad_norm_pre_clip': 1.9679304361343384, 'timing/data': 0.0003271494060754776, 'timing/model': 0.6230282979086041, 'learning_rate/qwen_vl_interface': 9.124072276090524e-06, 'learning_rate/action_model': 9.124072276090524e-05, 'epoch': 0.16})
+[2026-07-28 08:15:28,484][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.10521020740270615, 'train/grad_norm_pre_clip': 3.028994083404541, 'timing/data': 0.00033194106072187424, 'timing/model': 0.6179553484544158, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.16})
+[2026-07-28 08:15:37,686][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.07912208884954453, 'train/grad_norm_pre_clip': 4.24958610534668, 'timing/data': 0.0004261918365955353, 'timing/model': 0.6175179649144411, 'learning_rate/qwen_vl_interface': 9.119639270924754e-06, 'learning_rate/action_model': 9.119639270924754e-05, 'epoch': 0.16})
+[2026-07-28 08:15:46,878][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.12391076982021332, 'train/grad_norm_pre_clip': 4.3033270835876465, 'timing/data': 0.00034377723932266235, 'timing/model': 0.6190902460366488, 'learning_rate/qwen_vl_interface': 9.11741900117196e-06, 'learning_rate/action_model': 9.11741900117196e-05, 'epoch': 0.16})
+[2026-07-28 08:15:56,316][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.09841081500053406, 'train/grad_norm_pre_clip': 3.0916812419891357, 'timing/data': 0.00033789873123168945, 'timing/model': 0.852990829385817, 'learning_rate/qwen_vl_interface': 9.115196221892969e-06, 'learning_rate/action_model': 9.115196221892969e-05, 'epoch': 0.16})
+[2026-07-28 08:16:05,719][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.04217292368412018, 'train/grad_norm_pre_clip': 1.1765497922897339, 'timing/data': 0.00034052785485982895, 'timing/model': 0.8542400868609548, 'learning_rate/qwen_vl_interface': 9.112970934530111e-06, 'learning_rate/action_model': 9.112970934530111e-05, 'epoch': 0.16})
+[2026-07-28 08:16:14,853][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.060319073498249054, 'train/grad_norm_pre_clip': 0.9264382719993591, 'timing/data': 0.000323479063808918, 'timing/model': 0.6197536699473858, 'learning_rate/qwen_vl_interface': 9.110743140527353e-06, 'learning_rate/action_model': 9.110743140527352e-05, 'epoch': 0.16})
+[2026-07-28 08:16:24,081][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.07034842669963837, 'train/grad_norm_pre_clip': 1.5749592781066895, 'timing/data': 0.000353851355612278, 'timing/model': 0.6344048334285617, 'learning_rate/qwen_vl_interface': 9.108512841330288e-06, 'learning_rate/action_model': 9.108512841330287e-05, 'epoch': 0.16})
+[2026-07-28 08:16:33,461][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.0692262277007103, 'train/grad_norm_pre_clip': 2.2061657905578613, 'timing/data': 0.0006001042202115059, 'timing/model': 0.8459165012463927, 'learning_rate/qwen_vl_interface': 9.106280038386129e-06, 'learning_rate/action_model': 9.10628003838613e-05, 'epoch': 0.16})
+[2026-07-28 08:16:42,672][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.06835483014583588, 'train/grad_norm_pre_clip': 2.455202579498291, 'timing/data': 0.0003351457417011261, 'timing/model': 0.6181569527834654, 'learning_rate/qwen_vl_interface': 9.104044733143723e-06, 'learning_rate/action_model': 9.104044733143723e-05, 'epoch': 0.16})
+[2026-07-28 08:16:52,081][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.0783468559384346, 'train/grad_norm_pre_clip': 1.5974555015563965, 'timing/data': 0.000337737612426281, 'timing/model': 0.883916993625462, 'learning_rate/qwen_vl_interface': 9.10180692705353e-06, 'learning_rate/action_model': 9.10180692705353e-05, 'epoch': 0.16})
+[2026-07-28 08:17:01,213][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.07135879993438721, 'train/grad_norm_pre_clip': 1.7863117456436157, 'timing/data': 0.00037301424890756607, 'timing/model': 0.617525584064424, 'learning_rate/qwen_vl_interface': 9.099566621567643e-06, 'learning_rate/action_model': 9.099566621567642e-05, 'epoch': 0.16})
+[2026-07-28 08:17:10,557][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.07980407774448395, 'train/grad_norm_pre_clip': 1.1899811029434204, 'timing/data': 0.00033306702971458435, 'timing/model': 0.6186751313507557, 'learning_rate/qwen_vl_interface': 9.097323818139767e-06, 'learning_rate/action_model': 9.097323818139767e-05, 'epoch': 0.16})
+[2026-07-28 08:17:19,697][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.05179643630981445, 'train/grad_norm_pre_clip': 1.7917232513427734, 'timing/data': 0.00034310854971408844, 'timing/model': 0.6192975146695971, 'learning_rate/qwen_vl_interface': 9.095078518225237e-06, 'learning_rate/action_model': 9.095078518225236e-05, 'epoch': 0.16})
+[2026-07-28 08:17:28,878][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.08680973201990128, 'train/grad_norm_pre_clip': 2.783562421798706, 'timing/data': 0.0003338530659675598, 'timing/model': 0.6183848567306995, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.16})
+[2026-07-28 08:17:38,320][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.03750212490558624, 'train/grad_norm_pre_clip': 0.9133380055427551, 'timing/data': 0.00033837463706731796, 'timing/model': 0.8567939894273877, 'learning_rate/qwen_vl_interface': 9.09058043476563e-06, 'learning_rate/action_model': 9.09058043476563e-05, 'epoch': 0.16})
+[2026-07-28 08:17:47,645][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.05823718011379242, 'train/grad_norm_pre_clip': 1.6225303411483765, 'timing/data': 0.0003449469804763794, 'timing/model': 0.6202204385772347, 'learning_rate/qwen_vl_interface': 9.088327654139311e-06, 'learning_rate/action_model': 9.08832765413931e-05, 'epoch': 0.16})
+[2026-07-28 08:17:57,045][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.045579783618450165, 'train/grad_norm_pre_clip': 0.8568181395530701, 'timing/data': 0.0003414452075958252, 'timing/model': 0.6191616291180253, 'learning_rate/qwen_vl_interface': 9.08607238286385e-06, 'learning_rate/action_model': 9.08607238286385e-05, 'epoch': 0.16})
+[2026-07-28 08:18:06,201][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.06278111785650253, 'train/grad_norm_pre_clip': 2.394190788269043, 'timing/data': 0.0003906199708580971, 'timing/model': 0.6193265840411186, 'learning_rate/qwen_vl_interface': 9.083814622402667e-06, 'learning_rate/action_model': 9.083814622402666e-05, 'epoch': 0.16})
+[2026-07-28 08:18:15,322][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.053419992327690125, 'train/grad_norm_pre_clip': 2.0187160968780518, 'timing/data': 0.00037031620740890503, 'timing/model': 0.6196689633652568, 'learning_rate/qwen_vl_interface': 9.0815543742208e-06, 'learning_rate/action_model': 9.081554374220799e-05, 'epoch': 0.16})
+[2026-07-28 08:18:24,715][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.10666176676750183, 'train/grad_norm_pre_clip': 3.1604552268981934, 'timing/data': 0.00043829530477523804, 'timing/model': 0.6221060194075108, 'learning_rate/qwen_vl_interface': 9.079291639784897e-06, 'learning_rate/action_model': 9.079291639784896e-05, 'epoch': 0.16})
+[2026-07-28 08:18:34,088][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.043505892157554626, 'train/grad_norm_pre_clip': 2.2875688076019287, 'timing/data': 0.0003459108993411064, 'timing/model': 0.6768972622230649, 'learning_rate/qwen_vl_interface': 9.077026420563223e-06, 'learning_rate/action_model': 9.077026420563223e-05, 'epoch': 0.16})
+[2026-07-28 08:18:43,450][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.10295277833938599, 'train/grad_norm_pre_clip': 3.520665168762207, 'timing/data': 0.0003449488431215286, 'timing/model': 0.6197392651811242, 'learning_rate/qwen_vl_interface': 9.074758718025652e-06, 'learning_rate/action_model': 9.074758718025652e-05, 'epoch': 0.16})
+[2026-07-28 08:18:52,967][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.07118597626686096, 'train/grad_norm_pre_clip': 1.8948171138763428, 'timing/data': 0.00034599844366312027, 'timing/model': 0.6157824900001287, 'learning_rate/qwen_vl_interface': 9.072488533643678e-06, 'learning_rate/action_model': 9.072488533643677e-05, 'epoch': 0.16})
+[2026-07-28 08:19:02,116][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.07031925022602081, 'train/grad_norm_pre_clip': 2.19242000579834, 'timing/data': 0.0004337998107075691, 'timing/model': 0.61875962279737, 'learning_rate/qwen_vl_interface': 9.070215868890391e-06, 'learning_rate/action_model': 9.070215868890391e-05, 'epoch': 0.16})
+[2026-07-28 08:19:11,569][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.07333016395568848, 'train/grad_norm_pre_clip': 3.608948230743408, 'timing/data': 0.00034851860255002975, 'timing/model': 0.8535737441852689, 'learning_rate/qwen_vl_interface': 9.067940725240503e-06, 'learning_rate/action_model': 9.067940725240503e-05, 'epoch': 0.16})
+[2026-07-28 08:19:20,748][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.05774220824241638, 'train/grad_norm_pre_clip': 2.332867383956909, 'timing/data': 0.00035401806235313416, 'timing/model': 0.618681219406426, 'learning_rate/qwen_vl_interface': 9.06566310417033e-06, 'learning_rate/action_model': 9.06566310417033e-05, 'epoch': 0.16})
+[2026-07-28 08:19:29,913][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.06986495852470398, 'train/grad_norm_pre_clip': 3.246354818344116, 'timing/data': 0.00043286755681037903, 'timing/model': 0.6199151454493403, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.16})
+[2026-07-28 08:19:39,339][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.085623599588871, 'train/grad_norm_pre_clip': 1.4253301620483398, 'timing/data': 0.00044088251888751984, 'timing/model': 0.6165285054594278, 'learning_rate/qwen_vl_interface': 9.061100435682429e-06, 'learning_rate/action_model': 9.061100435682429e-05, 'epoch': 0.16})
+[2026-07-28 08:19:48,511][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.07239197194576263, 'train/grad_norm_pre_clip': 1.4560692310333252, 'timing/data': 0.0003561936318874359, 'timing/model': 0.6188317593187094, 'learning_rate/qwen_vl_interface': 9.058815391225363e-06, 'learning_rate/action_model': 9.058815391225362e-05, 'epoch': 0.16})
+[2026-07-28 08:19:57,859][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.08346312493085861, 'train/grad_norm_pre_clip': 2.8122832775115967, 'timing/data': 0.0003858795389533043, 'timing/model': 0.6189200524240732, 'learning_rate/qwen_vl_interface': 9.056527875269344e-06, 'learning_rate/action_model': 9.056527875269344e-05, 'epoch': 0.16})
+[2026-07-28 08:20:06,991][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.10609734058380127, 'train/grad_norm_pre_clip': 5.132511138916016, 'timing/data': 0.00034677889198064804, 'timing/model': 0.6186898676678538, 'learning_rate/qwen_vl_interface': 9.054237889298714e-06, 'learning_rate/action_model': 9.054237889298714e-05, 'epoch': 0.16})
+[2026-07-28 08:20:16,430][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.08569861948490143, 'train/grad_norm_pre_clip': 4.2866315841674805, 'timing/data': 0.00041147973388433456, 'timing/model': 0.6206247247755527, 'learning_rate/qwen_vl_interface': 9.05194543479942e-06, 'learning_rate/action_model': 9.05194543479942e-05, 'epoch': 0.16})
+[2026-07-28 08:20:25,883][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.04808609187602997, 'train/grad_norm_pre_clip': 1.2432559728622437, 'timing/data': 0.000348920002579689, 'timing/model': 0.6224457249045372, 'learning_rate/qwen_vl_interface': 9.049650513259007e-06, 'learning_rate/action_model': 9.049650513259008e-05, 'epoch': 0.17})
+[2026-07-28 08:20:35,012][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.09191833436489105, 'train/grad_norm_pre_clip': 3.675806760787964, 'timing/data': 0.0004172651097178459, 'timing/model': 0.6183672398328781, 'learning_rate/qwen_vl_interface': 9.047353126166632e-06, 'learning_rate/action_model': 9.047353126166632e-05, 'epoch': 0.17})
+[2026-07-28 08:20:44,401][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.056164491921663284, 'train/grad_norm_pre_clip': 1.402959942817688, 'timing/data': 0.00038428325206041336, 'timing/model': 0.8551389016211033, 'learning_rate/qwen_vl_interface': 9.045053275013038e-06, 'learning_rate/action_model': 9.045053275013038e-05, 'epoch': 0.17})
+[2026-07-28 08:20:53,542][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.09951381385326385, 'train/grad_norm_pre_clip': 1.3699640035629272, 'timing/data': 0.0004120776429772377, 'timing/model': 0.6200211429968476, 'learning_rate/qwen_vl_interface': 9.042750961290578e-06, 'learning_rate/action_model': 9.042750961290578e-05, 'epoch': 0.17})
+[2026-07-28 08:21:02,891][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.05092129856348038, 'train/grad_norm_pre_clip': 1.446613073348999, 'timing/data': 0.0003598267212510109, 'timing/model': 0.6182450829073787, 'learning_rate/qwen_vl_interface': 9.040446186493195e-06, 'learning_rate/action_model': 9.040446186493194e-05, 'epoch': 0.17})
+[2026-07-28 08:21:12,118][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.0756663978099823, 'train/grad_norm_pre_clip': 0.937816321849823, 'timing/data': 0.0004142262041568756, 'timing/model': 0.6263618804514408, 'learning_rate/qwen_vl_interface': 9.038138952116435e-06, 'learning_rate/action_model': 9.038138952116435e-05, 'epoch': 0.17})
+[2026-07-28 08:21:21,541][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.059374209493398666, 'train/grad_norm_pre_clip': 0.9172983765602112, 'timing/data': 0.0003387555480003357, 'timing/model': 0.8879874097183347, 'learning_rate/qwen_vl_interface': 9.035829259657437e-06, 'learning_rate/action_model': 9.035829259657436e-05, 'epoch': 0.17})
+[2026-07-28 08:21:30,741][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.04271933063864708, 'train/grad_norm_pre_clip': 2.2787601947784424, 'timing/data': 0.00035904254764318466, 'timing/model': 0.6201253049075603, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.17})
+[2026-07-28 08:21:40,161][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.0522361621260643, 'train/grad_norm_pre_clip': 1.2655643224716187, 'timing/data': 0.00034744013100862503, 'timing/model': 0.617252254858613, 'learning_rate/qwen_vl_interface': 9.031202506489254e-06, 'learning_rate/action_model': 9.031202506489253e-05, 'epoch': 0.17})
+[2026-07-28 08:21:49,286][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.06678572297096252, 'train/grad_norm_pre_clip': 1.9619382619857788, 'timing/data': 0.00039750058203935623, 'timing/model': 0.6179654495790601, 'learning_rate/qwen_vl_interface': 9.02888544878232e-06, 'learning_rate/action_model': 9.028885448782319e-05, 'epoch': 0.17})
+[2026-07-28 08:21:58,666][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.10878220200538635, 'train/grad_norm_pre_clip': 2.098029613494873, 'timing/data': 0.00039555225521326065, 'timing/model': 0.6201362777501345, 'learning_rate/qwen_vl_interface': 9.026565938997644e-06, 'learning_rate/action_model': 9.026565938997645e-05, 'epoch': 0.17})
+[2026-07-28 08:22:07,886][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.09778909385204315, 'train/grad_norm_pre_clip': 2.6255013942718506, 'timing/data': 0.00036290939897298813, 'timing/model': 0.6166414609178901, 'learning_rate/qwen_vl_interface': 9.024243978640334e-06, 'learning_rate/action_model': 9.024243978640333e-05, 'epoch': 0.17})
+[2026-07-28 08:22:17,248][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.09353631734848022, 'train/grad_norm_pre_clip': 2.902907133102417, 'timing/data': 0.0004107290878891945, 'timing/model': 0.8430027086287737, 'learning_rate/qwen_vl_interface': 9.02191956921708e-06, 'learning_rate/action_model': 9.02191956921708e-05, 'epoch': 0.17})
+[2026-07-28 08:22:26,483][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.0769055187702179, 'train/grad_norm_pre_clip': 3.4339966773986816, 'timing/data': 0.0003945976495742798, 'timing/model': 0.6185753969475627, 'learning_rate/qwen_vl_interface': 9.019592712236172e-06, 'learning_rate/action_model': 9.019592712236171e-05, 'epoch': 0.17})
+[2026-07-28 08:22:35,883][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.09862563014030457, 'train/grad_norm_pre_clip': 1.6151292324066162, 'timing/data': 0.00038169976323843, 'timing/model': 0.6188901290297508, 'learning_rate/qwen_vl_interface': 9.017263409207478e-06, 'learning_rate/action_model': 9.017263409207478e-05, 'epoch': 0.17})
+[2026-07-28 08:22:45,041][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.058925677090883255, 'train/grad_norm_pre_clip': 1.5111403465270996, 'timing/data': 0.0003810301423072815, 'timing/model': 0.6198379043489695, 'learning_rate/qwen_vl_interface': 9.014931661642455e-06, 'learning_rate/action_model': 9.014931661642455e-05, 'epoch': 0.17})
+[2026-07-28 08:22:54,206][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.07246430218219757, 'train/grad_norm_pre_clip': 2.191530466079712, 'timing/data': 0.0004030531272292137, 'timing/model': 0.6163034848868847, 'learning_rate/qwen_vl_interface': 9.012597471054154e-06, 'learning_rate/action_model': 9.012597471054154e-05, 'epoch': 0.17})
+[2026-07-28 08:23:03,620][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.06812206655740738, 'train/grad_norm_pre_clip': 2.0064682960510254, 'timing/data': 0.00044859573245048523, 'timing/model': 0.8497229535132647, 'learning_rate/qwen_vl_interface': 9.010260838957206e-06, 'learning_rate/action_model': 9.010260838957205e-05, 'epoch': 0.17})
+[2026-07-28 08:23:13,102][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.08543406426906586, 'train/grad_norm_pre_clip': 1.4924172163009644, 'timing/data': 0.00040018465369939804, 'timing/model': 0.6258340366184711, 'learning_rate/qwen_vl_interface': 9.007921766867823e-06, 'learning_rate/action_model': 9.007921766867822e-05, 'epoch': 0.17})
+[2026-07-28 08:23:22,319][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.07339498400688171, 'train/grad_norm_pre_clip': 1.0440424680709839, 'timing/data': 0.00044116005301475525, 'timing/model': 0.6180181568488479, 'learning_rate/qwen_vl_interface': 9.005580256303804e-06, 'learning_rate/action_model': 9.005580256303803e-05, 'epoch': 0.17})
+[2026-07-28 08:23:31,557][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.05100023001432419, 'train/grad_norm_pre_clip': 0.82853102684021, 'timing/data': 0.0005317796021699905, 'timing/model': 0.6196975922212005, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.17})
+[2026-07-28 08:23:41,113][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.06609396636486053, 'train/grad_norm_pre_clip': 2.0576179027557373, 'timing/data': 0.00037433765828609467, 'timing/model': 0.9191555129364133, 'learning_rate/qwen_vl_interface': 9.000889925830964e-06, 'learning_rate/action_model': 9.000889925830965e-05, 'epoch': 0.17})
+[2026-07-28 08:23:50,334][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.05778339505195618, 'train/grad_norm_pre_clip': 1.0656498670578003, 'timing/data': 0.00046822335571050644, 'timing/model': 0.6203359737992287, 'learning_rate/qwen_vl_interface': 8.998541108965648e-06, 'learning_rate/action_model': 8.998541108965648e-05, 'epoch': 0.17})
+[2026-07-28 08:23:59,859][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.0382576547563076, 'train/grad_norm_pre_clip': 0.9248843193054199, 'timing/data': 0.0007001683115959167, 'timing/model': 0.6441075783222914, 'learning_rate/qwen_vl_interface': 8.996189859712704e-06, 'learning_rate/action_model': 8.996189859712703e-05, 'epoch': 0.17})
+[2026-07-28 08:24:09,134][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.0721154659986496, 'train/grad_norm_pre_clip': 2.9062306880950928, 'timing/data': 0.0004512760788202286, 'timing/model': 0.6210203152149916, 'learning_rate/qwen_vl_interface': 8.993836179597833e-06, 'learning_rate/action_model': 8.993836179597833e-05, 'epoch': 0.17})
+[2026-07-28 08:24:18,351][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.06610782444477081, 'train/grad_norm_pre_clip': 1.8666772842407227, 'timing/data': 0.00041313469409942627, 'timing/model': 0.6188278440386057, 'learning_rate/qwen_vl_interface': 8.99148007014831e-06, 'learning_rate/action_model': 8.99148007014831e-05, 'epoch': 0.17})
+[2026-07-28 08:24:27,878][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.028362292796373367, 'train/grad_norm_pre_clip': 0.7274014949798584, 'timing/data': 0.00035880226641893387, 'timing/model': 0.9109183726832271, 'learning_rate/qwen_vl_interface': 8.989121532892989e-06, 'learning_rate/action_model': 8.989121532892988e-05, 'epoch': 0.17})
+[2026-07-28 08:24:37,061][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.02858481928706169, 'train/grad_norm_pre_clip': 0.8153139352798462, 'timing/data': 0.0004073241725564003, 'timing/model': 0.6179891498759389, 'learning_rate/qwen_vl_interface': 8.986760569362302e-06, 'learning_rate/action_model': 8.986760569362301e-05, 'epoch': 0.17})
+[2026-07-28 08:24:46,430][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.08781459927558899, 'train/grad_norm_pre_clip': 3.270062208175659, 'timing/data': 0.0004836898297071457, 'timing/model': 0.8468721648678184, 'learning_rate/qwen_vl_interface': 8.984397181088251e-06, 'learning_rate/action_model': 8.984397181088251e-05, 'epoch': 0.17})
+[2026-07-28 08:24:55,616][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.035614013671875, 'train/grad_norm_pre_clip': 0.8968585729598999, 'timing/data': 0.00042024534195661545, 'timing/model': 0.6178413713350892, 'learning_rate/qwen_vl_interface': 8.982031369604412e-06, 'learning_rate/action_model': 8.982031369604411e-05, 'epoch': 0.17})
+[2026-07-28 08:25:04,819][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.01871630549430847, 'train/grad_norm_pre_clip': 0.7765460014343262, 'timing/data': 0.0004521152004599571, 'timing/model': 0.6180386748164892, 'learning_rate/qwen_vl_interface': 8.979663136445936e-06, 'learning_rate/action_model': 8.979663136445935e-05, 'epoch': 0.17})
+[2026-07-28 08:25:13,997][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.033913396298885345, 'train/grad_norm_pre_clip': 1.6395978927612305, 'timing/data': 0.00044083036482334137, 'timing/model': 0.6188959535211325, 'learning_rate/qwen_vl_interface': 8.977292483149542e-06, 'learning_rate/action_model': 8.977292483149542e-05, 'epoch': 0.17})
+[2026-07-28 08:25:23,437][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.10077330470085144, 'train/grad_norm_pre_clip': 3.2349934577941895, 'timing/data': 0.0004231436178088188, 'timing/model': 0.8573584174737334, 'learning_rate/qwen_vl_interface': 8.974919411253523e-06, 'learning_rate/action_model': 8.974919411253523e-05, 'epoch': 0.17})
+[2026-07-28 08:25:32,569][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.07844428718090057, 'train/grad_norm_pre_clip': 6.2065839767456055, 'timing/data': 0.0004903823137283325, 'timing/model': 0.6182678295299411, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.17})
+[2026-07-28 08:25:41,754][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.02493552304804325, 'train/grad_norm_pre_clip': 1.4810826778411865, 'timing/data': 0.00034961942583322525, 'timing/model': 0.625509375706315, 'learning_rate/qwen_vl_interface': 8.970166017823621e-06, 'learning_rate/action_model': 8.97016601782362e-05, 'epoch': 0.17})
+[2026-07-28 08:25:50,891][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.06765371561050415, 'train/grad_norm_pre_clip': 2.4227406978607178, 'timing/data': 0.0003794711083173752, 'timing/model': 0.6179891852661967, 'learning_rate/qwen_vl_interface': 8.96778569937416e-06, 'learning_rate/action_model': 8.96778569937416e-05, 'epoch': 0.17})
+[2026-07-28 08:26:00,267][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.054061710834503174, 'train/grad_norm_pre_clip': 3.033277988433838, 'timing/data': 0.00038565322756767273, 'timing/model': 0.6239939145743847, 'learning_rate/qwen_vl_interface': 8.965402968493922e-06, 'learning_rate/action_model': 8.965402968493922e-05, 'epoch': 0.17})
+[2026-07-28 08:26:09,692][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.14167392253875732, 'train/grad_norm_pre_clip': 3.7390053272247314, 'timing/data': 0.0003829197958111763, 'timing/model': 0.6216397443786263, 'learning_rate/qwen_vl_interface': 8.96301782672904e-06, 'learning_rate/action_model': 8.963017826729038e-05, 'epoch': 0.17})
+[2026-07-28 08:26:18,816][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.05946784466505051, 'train/grad_norm_pre_clip': 2.719264030456543, 'timing/data': 0.0004322454333305359, 'timing/model': 0.6187843028455973, 'learning_rate/qwen_vl_interface': 8.960630275627197e-06, 'learning_rate/action_model': 8.960630275627197e-05, 'epoch': 0.17})
+[2026-07-28 08:26:28,008][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.057418037205934525, 'train/grad_norm_pre_clip': 2.0864417552948, 'timing/data': 0.0003645308315753937, 'timing/model': 0.6186929335817695, 'learning_rate/qwen_vl_interface': 8.958240316737655e-06, 'learning_rate/action_model': 8.958240316737656e-05, 'epoch': 0.17})
+[2026-07-28 08:26:37,351][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.06289571523666382, 'train/grad_norm_pre_clip': 1.6987228393554688, 'timing/data': 0.00038321875035762787, 'timing/model': 0.8470253497362137, 'learning_rate/qwen_vl_interface': 8.955847951611233e-06, 'learning_rate/action_model': 8.955847951611232e-05, 'epoch': 0.17})
+[2026-07-28 08:26:46,712][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.15926426649093628, 'train/grad_norm_pre_clip': 4.7155866622924805, 'timing/data': 0.00043616537004709244, 'timing/model': 0.620315682142973, 'learning_rate/qwen_vl_interface': 8.953453181800307e-06, 'learning_rate/action_model': 8.953453181800306e-05, 'epoch': 0.17})
+[2026-07-28 08:26:55,892][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.05315159261226654, 'train/grad_norm_pre_clip': 1.532173752784729, 'timing/data': 0.00041111744940280914, 'timing/model': 0.617800279520452, 'learning_rate/qwen_vl_interface': 8.951056008858819e-06, 'learning_rate/action_model': 8.95105600885882e-05, 'epoch': 0.17})
+[2026-07-28 08:27:04,822][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.030724911019206047, 'train/grad_norm_pre_clip': 1.2306993007659912, 'timing/data': 0.0003832736983895302, 'timing/model': 0.6173886554315686, 'learning_rate/qwen_vl_interface': 8.94865643434227e-06, 'learning_rate/action_model': 8.948656434342269e-05, 'epoch': 0.17})
+[2026-07-28 08:27:13,976][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.09151547402143478, 'train/grad_norm_pre_clip': 1.9149280786514282, 'timing/data': 0.00038569699972867966, 'timing/model': 0.6193709205836058, 'learning_rate/qwen_vl_interface': 8.946254459807717e-06, 'learning_rate/action_model': 8.946254459807717e-05, 'epoch': 0.17})
+[2026-07-28 08:27:23,357][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.06946147978305817, 'train/grad_norm_pre_clip': 2.2924160957336426, 'timing/data': 0.0004098769277334213, 'timing/model': 0.6183481086045504, 'learning_rate/qwen_vl_interface': 8.943850086813773e-06, 'learning_rate/action_model': 8.943850086813773e-05, 'epoch': 0.17})
+[2026-07-28 08:27:32,755][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.1177583634853363, 'train/grad_norm_pre_clip': 1.594245195388794, 'timing/data': 0.00040343310683965683, 'timing/model': 0.6178765231743455, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.17})
+[2026-07-28 08:27:41,930][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.09001599252223969, 'train/grad_norm_pre_clip': 3.530751943588257, 'timing/data': 0.000340399332344532, 'timing/model': 0.6172347404062748, 'learning_rate/qwen_vl_interface': 8.93903415168996e-06, 'learning_rate/action_model': 8.93903415168996e-05, 'epoch': 0.17})
+[2026-07-28 08:27:51,336][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.06474681943655014, 'train/grad_norm_pre_clip': 1.6244268417358398, 'timing/data': 0.000492841936647892, 'timing/model': 0.8353984728455544, 'learning_rate/qwen_vl_interface': 8.9366225926851e-06, 'learning_rate/action_model': 8.936622592685098e-05, 'epoch': 0.17})
+[2026-07-28 08:28:00,510][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.12826386094093323, 'train/grad_norm_pre_clip': 2.66410231590271, 'timing/data': 0.00043825339525938034, 'timing/model': 0.6265089819207788, 'learning_rate/qwen_vl_interface': 8.934208641470863e-06, 'learning_rate/action_model': 8.934208641470863e-05, 'epoch': 0.17})
+[2026-07-28 08:28:09,914][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.0725712776184082, 'train/grad_norm_pre_clip': 1.3759970664978027, 'timing/data': 0.000435415655374527, 'timing/model': 0.853063328191638, 'learning_rate/qwen_vl_interface': 8.931792299613637e-06, 'learning_rate/action_model': 8.931792299613636e-05, 'epoch': 0.17})
+[2026-07-28 08:28:19,134][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.06190897524356842, 'train/grad_norm_pre_clip': 2.7394917011260986, 'timing/data': 0.00032080430537462234, 'timing/model': 0.6175411278381944, 'learning_rate/qwen_vl_interface': 8.92937356868136e-06, 'learning_rate/action_model': 8.92937356868136e-05, 'epoch': 0.17})
+[2026-07-28 08:28:28,692][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.07433850318193436, 'train/grad_norm_pre_clip': 3.4737484455108643, 'timing/data': 0.0003264276310801506, 'timing/model': 0.8621701672673225, 'learning_rate/qwen_vl_interface': 8.926952450243522e-06, 'learning_rate/action_model': 8.926952450243522e-05, 'epoch': 0.17})
+[2026-07-28 08:28:37,858][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.10542602837085724, 'train/grad_norm_pre_clip': 5.119333744049072, 'timing/data': 0.0003313440829515457, 'timing/model': 0.6216048421338201, 'learning_rate/qwen_vl_interface': 8.924528945871157e-06, 'learning_rate/action_model': 8.924528945871156e-05, 'epoch': 0.17})
+[2026-07-28 08:28:47,199][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.06629602611064911, 'train/grad_norm_pre_clip': 1.303416132926941, 'timing/data': 0.0003582434728741646, 'timing/model': 0.6170137990266085, 'learning_rate/qwen_vl_interface': 8.922103057136857e-06, 'learning_rate/action_model': 8.922103057136856e-05, 'epoch': 0.18})
+[2026-07-28 08:28:56,343][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.06470903754234314, 'train/grad_norm_pre_clip': 1.7311524152755737, 'timing/data': 0.0003886837512254715, 'timing/model': 0.6203796453773975, 'learning_rate/qwen_vl_interface': 8.919674785614747e-06, 'learning_rate/action_model': 8.919674785614747e-05, 'epoch': 0.18})
+[2026-07-28 08:29:05,814][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.10282394289970398, 'train/grad_norm_pre_clip': 1.3476301431655884, 'timing/data': 0.00035344716161489487, 'timing/model': 0.655455163680017, 'learning_rate/qwen_vl_interface': 8.91724413288051e-06, 'learning_rate/action_model': 8.91724413288051e-05, 'epoch': 0.18})
+[2026-07-28 08:29:15,241][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.09627197682857513, 'train/grad_norm_pre_clip': 1.912079095840454, 'timing/data': 0.0003313319757580757, 'timing/model': 0.8856096705421805, 'learning_rate/qwen_vl_interface': 8.91481110051137e-06, 'learning_rate/action_model': 8.91481110051137e-05, 'epoch': 0.18})
+[2026-07-28 08:29:24,600][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.038047369569540024, 'train/grad_norm_pre_clip': 0.686943769454956, 'timing/data': 0.0004906514659523964, 'timing/model': 0.6215304732322693, 'learning_rate/qwen_vl_interface': 8.912375690086097e-06, 'learning_rate/action_model': 8.912375690086096e-05, 'epoch': 0.18})
+[2026-07-28 08:29:33,823][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.06350725144147873, 'train/grad_norm_pre_clip': 1.504723072052002, 'timing/data': 0.00046181585639715195, 'timing/model': 0.622197144664824, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.18})
+[2026-07-28 08:29:43,380][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.056539781391620636, 'train/grad_norm_pre_clip': 1.1444307565689087, 'timing/data': 0.0005487799644470215, 'timing/model': 0.6188197005540133, 'learning_rate/qwen_vl_interface': 8.907497741389932e-06, 'learning_rate/action_model': 8.907497741389932e-05, 'epoch': 0.18})
+[2026-07-28 08:29:52,611][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.05819854140281677, 'train/grad_norm_pre_clip': 1.4449703693389893, 'timing/data': 0.0004525771364569664, 'timing/model': 0.6196572231128812, 'learning_rate/qwen_vl_interface': 8.905055206284289e-06, 'learning_rate/action_model': 8.905055206284288e-05, 'epoch': 0.18})
+[2026-07-28 08:30:02,096][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.048545241355895996, 'train/grad_norm_pre_clip': 0.9196591377258301, 'timing/data': 0.0003338176757097244, 'timing/model': 0.6231799237430096, 'learning_rate/qwen_vl_interface': 8.902610299453005e-06, 'learning_rate/action_model': 8.902610299453004e-05, 'epoch': 0.18})
+[2026-07-28 08:30:11,601][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.052104659378528595, 'train/grad_norm_pre_clip': 0.9851862788200378, 'timing/data': 0.00036064162850379944, 'timing/model': 0.8696176344528794, 'learning_rate/qwen_vl_interface': 8.900163022482554e-06, 'learning_rate/action_model': 8.900163022482553e-05, 'epoch': 0.18})
+[2026-07-28 08:30:20,764][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.05109057575464249, 'train/grad_norm_pre_clip': 1.1684640645980835, 'timing/data': 0.0003664530813694, 'timing/model': 0.6207566484808922, 'learning_rate/qwen_vl_interface': 8.897713376960946e-06, 'learning_rate/action_model': 8.897713376960946e-05, 'epoch': 0.18})
+[2026-07-28 08:30:30,024][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.06129223853349686, 'train/grad_norm_pre_clip': 1.4526036977767944, 'timing/data': 0.00033913739025592804, 'timing/model': 0.6201042514294386, 'learning_rate/qwen_vl_interface': 8.895261364477731e-06, 'learning_rate/action_model': 8.895261364477731e-05, 'epoch': 0.18})
+[2026-07-28 08:30:39,120][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.10128097981214523, 'train/grad_norm_pre_clip': 2.544275999069214, 'timing/data': 0.00034199655055999756, 'timing/model': 0.6180046806111932, 'learning_rate/qwen_vl_interface': 8.892806986623993e-06, 'learning_rate/action_model': 8.892806986623992e-05, 'epoch': 0.18})
+[2026-07-28 08:30:48,512][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.1159045547246933, 'train/grad_norm_pre_clip': 2.4803500175476074, 'timing/data': 0.0004325937479734421, 'timing/model': 0.8605806455016136, 'learning_rate/qwen_vl_interface': 8.890350244992349e-06, 'learning_rate/action_model': 8.890350244992348e-05, 'epoch': 0.18})
+[2026-07-28 08:30:57,907][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.05218811705708504, 'train/grad_norm_pre_clip': 1.47368323802948, 'timing/data': 0.0003323787823319435, 'timing/model': 0.6174798030406237, 'learning_rate/qwen_vl_interface': 8.887891141176954e-06, 'learning_rate/action_model': 8.887891141176954e-05, 'epoch': 0.18})
+[2026-07-28 08:31:07,039][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.06944070756435394, 'train/grad_norm_pre_clip': 1.731926679611206, 'timing/data': 0.0003944532945752144, 'timing/model': 0.6204831330105662, 'learning_rate/qwen_vl_interface': 8.885429676773495e-06, 'learning_rate/action_model': 8.885429676773495e-05, 'epoch': 0.18})
+[2026-07-28 08:31:16,259][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.07113887369632721, 'train/grad_norm_pre_clip': 1.830851674079895, 'timing/data': 0.0003557540476322174, 'timing/model': 0.6179915126413107, 'learning_rate/qwen_vl_interface': 8.882965853379188e-06, 'learning_rate/action_model': 8.882965853379187e-05, 'epoch': 0.18})
+[2026-07-28 08:31:25,416][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.0640166625380516, 'train/grad_norm_pre_clip': 1.9731571674346924, 'timing/data': 0.00035940762609243393, 'timing/model': 0.6186280520632863, 'learning_rate/qwen_vl_interface': 8.88049967259278e-06, 'learning_rate/action_model': 8.880499672592778e-05, 'epoch': 0.18})
+[2026-07-28 08:31:34,829][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.03248400241136551, 'train/grad_norm_pre_clip': 0.8526889681816101, 'timing/data': 0.0004217345267534256, 'timing/model': 0.6195448143407702, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.18})
+[2026-07-28 08:31:43,753][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.0761934444308281, 'train/grad_norm_pre_clip': 2.3338677883148193, 'timing/data': 0.0003304453566670418, 'timing/model': 0.6169941462576389, 'learning_rate/qwen_vl_interface': 8.875560245246307e-06, 'learning_rate/action_model': 8.875560245246307e-05, 'epoch': 0.18})
+[2026-07-28 08:31:52,874][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.07825332880020142, 'train/grad_norm_pre_clip': 2.0608487129211426, 'timing/data': 0.0003124196082353592, 'timing/model': 0.621056872420013, 'learning_rate/qwen_vl_interface': 8.873087001891381e-06, 'learning_rate/action_model': 8.87308700189138e-05, 'epoch': 0.18})
+[2026-07-28 08:32:02,492][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.04949893802404404, 'train/grad_norm_pre_clip': 1.8983008861541748, 'timing/data': 0.0004420764744281769, 'timing/model': 0.803698968142271, 'learning_rate/qwen_vl_interface': 8.870611407554635e-06, 'learning_rate/action_model': 8.870611407554636e-05, 'epoch': 0.18})
+[2026-07-28 08:32:11,642][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.07828404754400253, 'train/grad_norm_pre_clip': 1.4149366617202759, 'timing/data': 0.00039757881313562393, 'timing/model': 0.6170827727764845, 'learning_rate/qwen_vl_interface': 8.868133463842456e-06, 'learning_rate/action_model': 8.868133463842457e-05, 'epoch': 0.18})
+[2026-07-28 08:32:20,528][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.0654379203915596, 'train/grad_norm_pre_clip': 1.2305803298950195, 'timing/data': 0.000588754191994667, 'timing/model': 0.6169931953772902, 'learning_rate/qwen_vl_interface': 8.865653172362755e-06, 'learning_rate/action_model': 8.865653172362754e-05, 'epoch': 0.18})
+[2026-07-28 08:32:29,670][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.03949282318353653, 'train/grad_norm_pre_clip': 0.8466789722442627, 'timing/data': 0.000375550240278244, 'timing/model': 0.6181259993463755, 'learning_rate/qwen_vl_interface': 8.863170534724962e-06, 'learning_rate/action_model': 8.863170534724961e-05, 'epoch': 0.18})
+[2026-07-28 08:32:39,299][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.0806623101234436, 'train/grad_norm_pre_clip': 1.6264152526855469, 'timing/data': 0.0003872644156217575, 'timing/model': 0.8732661074027419, 'learning_rate/qwen_vl_interface': 8.860685552540038e-06, 'learning_rate/action_model': 8.860685552540037e-05, 'epoch': 0.18})
+[2026-07-28 08:32:48,465][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.0391809344291687, 'train/grad_norm_pre_clip': 0.8833826184272766, 'timing/data': 0.0003945492208003998, 'timing/model': 0.615481554530561, 'learning_rate/qwen_vl_interface': 8.858198227420461e-06, 'learning_rate/action_model': 8.85819822742046e-05, 'epoch': 0.18})
+[2026-07-28 08:32:57,584][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.06823669373989105, 'train/grad_norm_pre_clip': 1.2188597917556763, 'timing/data': 0.0003801221027970314, 'timing/model': 0.6176286693662405, 'learning_rate/qwen_vl_interface': 8.855708560980225e-06, 'learning_rate/action_model': 8.855708560980225e-05, 'epoch': 0.18})
+[2026-07-28 08:33:06,706][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.04704282432794571, 'train/grad_norm_pre_clip': 1.1656237840652466, 'timing/data': 0.00038543064147233963, 'timing/model': 0.6171021312475204, 'learning_rate/qwen_vl_interface': 8.853216554834852e-06, 'learning_rate/action_model': 8.85321655483485e-05, 'epoch': 0.18})
+[2026-07-28 08:33:16,087][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.0677318423986435, 'train/grad_norm_pre_clip': 2.1950526237487793, 'timing/data': 0.00041084829717874527, 'timing/model': 0.6204831311479211, 'learning_rate/qwen_vl_interface': 8.850722210601376e-06, 'learning_rate/action_model': 8.850722210601377e-05, 'epoch': 0.18})
+[2026-07-28 08:33:25,261][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.0846935585141182, 'train/grad_norm_pre_clip': 1.1764907836914062, 'timing/data': 0.0004043048247694969, 'timing/model': 0.6188012761995196, 'learning_rate/qwen_vl_interface': 8.848225529898351e-06, 'learning_rate/action_model': 8.84822552989835e-05, 'epoch': 0.18})
+[2026-07-28 08:33:34,179][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.08042966574430466, 'train/grad_norm_pre_clip': 1.3142926692962646, 'timing/data': 0.0003935769200325012, 'timing/model': 0.6274390947073698, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.18})
+[2026-07-28 08:33:43,843][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.03776386380195618, 'train/grad_norm_pre_clip': 1.168128252029419, 'timing/data': 0.0003815405070781708, 'timing/model': 0.6252488568425179, 'learning_rate/qwen_vl_interface': 8.84322516556544e-06, 'learning_rate/action_model': 8.84322516556544e-05, 'epoch': 0.18})
+[2026-07-28 08:33:53,403][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.07602277398109436, 'train/grad_norm_pre_clip': 2.782048225402832, 'timing/data': 0.0003959517925977707, 'timing/model': 0.8483637347817421, 'learning_rate/qwen_vl_interface': 8.840721485180237e-06, 'learning_rate/action_model': 8.840721485180236e-05, 'epoch': 0.18})
+[2026-07-28 08:34:02,551][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.08780030906200409, 'train/grad_norm_pre_clip': 2.7210259437561035, 'timing/data': 0.000440903939306736, 'timing/model': 0.6178419049829245, 'learning_rate/qwen_vl_interface': 8.838215474814848e-06, 'learning_rate/action_model': 8.838215474814847e-05, 'epoch': 0.18})
+[2026-07-28 08:34:11,732][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.058516085147857666, 'train/grad_norm_pre_clip': 1.4849371910095215, 'timing/data': 0.0004074098542332649, 'timing/model': 0.6199309322983027, 'learning_rate/qwen_vl_interface': 8.835707136095392e-06, 'learning_rate/action_model': 8.835707136095391e-05, 'epoch': 0.18})
+[2026-07-28 08:34:21,158][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.07031212002038956, 'train/grad_norm_pre_clip': 3.356494903564453, 'timing/data': 0.00038174819201231003, 'timing/model': 0.8642534473910928, 'learning_rate/qwen_vl_interface': 8.833196470649505e-06, 'learning_rate/action_model': 8.833196470649505e-05, 'epoch': 0.18})
+[2026-07-28 08:34:30,279][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.12011417001485825, 'train/grad_norm_pre_clip': 3.8386383056640625, 'timing/data': 0.00039931759238243103, 'timing/model': 0.6175917908549309, 'learning_rate/qwen_vl_interface': 8.83068348010633e-06, 'learning_rate/action_model': 8.83068348010633e-05, 'epoch': 0.18})
+[2026-07-28 08:34:39,424][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.08413893729448318, 'train/grad_norm_pre_clip': 2.772340774536133, 'timing/data': 0.0003673601895570755, 'timing/model': 0.6211198605597019, 'learning_rate/qwen_vl_interface': 8.82816816609652e-06, 'learning_rate/action_model': 8.828168166096521e-05, 'epoch': 0.18})
+[2026-07-28 08:34:48,800][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.07392910122871399, 'train/grad_norm_pre_clip': 1.0870271921157837, 'timing/data': 0.00042664818465709686, 'timing/model': 0.6171122621744871, 'learning_rate/qwen_vl_interface': 8.825650530252238e-06, 'learning_rate/action_model': 8.825650530252238e-05, 'epoch': 0.18})
+[2026-07-28 08:34:57,920][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.06279812008142471, 'train/grad_norm_pre_clip': 1.093421459197998, 'timing/data': 0.0003783358260989189, 'timing/model': 0.6178794978186488, 'learning_rate/qwen_vl_interface': 8.823130574207144e-06, 'learning_rate/action_model': 8.823130574207143e-05, 'epoch': 0.18})
+[2026-07-28 08:35:07,347][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.08665680140256882, 'train/grad_norm_pre_clip': 1.4182188510894775, 'timing/data': 0.0003333473578095436, 'timing/model': 0.6176511785015464, 'learning_rate/qwen_vl_interface': 8.82060829959641e-06, 'learning_rate/action_model': 8.820608299596411e-05, 'epoch': 0.18})
+[2026-07-28 08:35:16,718][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.05933789163827896, 'train/grad_norm_pre_clip': 0.855056643486023, 'timing/data': 0.00037803687155246735, 'timing/model': 0.617269728332758, 'learning_rate/qwen_vl_interface': 8.81808370805672e-06, 'learning_rate/action_model': 8.818083708056721e-05, 'epoch': 0.18})
+[2026-07-28 08:35:25,840][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.07912816107273102, 'train/grad_norm_pre_clip': 2.0681772232055664, 'timing/data': 0.0003989366814494133, 'timing/model': 0.6202420238405466, 'learning_rate/qwen_vl_interface': 8.815556801226251e-06, 'learning_rate/action_model': 8.815556801226251e-05, 'epoch': 0.18})
+[2026-07-28 08:35:34,977][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.04913712665438652, 'train/grad_norm_pre_clip': 0.8850576877593994, 'timing/data': 0.0003381771966814995, 'timing/model': 0.6175261158496141, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.18})
+[2026-07-28 08:35:44,523][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.05452348291873932, 'train/grad_norm_pre_clip': 1.9512399435043335, 'timing/data': 0.0003290977329015732, 'timing/model': 0.7860261695459485, 'learning_rate/qwen_vl_interface': 8.810496048253203e-06, 'learning_rate/action_model': 8.810496048253202e-05, 'epoch': 0.18})
+[2026-07-28 08:35:53,668][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.037658851593732834, 'train/grad_norm_pre_clip': 1.5141873359680176, 'timing/data': 0.00034085754305124283, 'timing/model': 0.6174665912985802, 'learning_rate/qwen_vl_interface': 8.80796220539449e-06, 'learning_rate/action_model': 8.80796220539449e-05, 'epoch': 0.18})
+[2026-07-28 08:36:02,814][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1000
+[2026-07-28 08:36:13,092][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1000
+[2026-07-28 08:47:07,010][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000
+[2026-07-28 08:47:07,011][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.0738668441772461, 'train/grad_norm_pre_clip': 1.9099187850952148, 'eval/action_loss/seconds': 10.2729747062549, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.04753677546977997, 'eval/latency_3/loss': 0.04753677546977997, 'eval/flappy/loss': 0.04753677546977997, 'eval/flappy/latency_3/loss': 0.04753677546977997, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 189.86000225022434, 'rl_games_eval/mid_train/mean_length': 1550.75, 'rl_games_eval/mid_train/std_reward': 127.10599033053973, 'rl_games_eval/mid_train/std_length': 1024.5753693604, 'rl_games_eval/mid_train/macro_mean_reward': 189.86000225022434, 'rl_games_eval/mid_train/macro_mean_length': 1550.75, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 189.86000225022434, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 1550.75, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 127.10599033053973, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1024.5753693604, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 189.86000225022434, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0004169447347521782, 'timing/model': 0.616511701606214, 'learning_rate/qwen_vl_interface': 8.805426053812731e-06, 'learning_rate/action_model': 8.805426053812731e-05, 'epoch': 0.18})
+[2026-07-28 08:47:07,014][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1000_state
+[2026-07-28 08:47:07,014][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 08:47:45,056][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1000_state/pytorch_model
+[2026-07-28 08:47:45,057][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1000_state/sampler.bin
+[2026-07-28 08:47:45,057][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1000_state/sampler_1.bin
+[2026-07-28 08:47:45,060][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1000_state/random_states_0.pkl
+[2026-07-28 08:47:53,283][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000
+[2026-07-28 08:47:53,283][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 08:47:53,305][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 08:48:02,709][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.06695088744163513, 'train/grad_norm_pre_clip': 1.3193409442901611, 'timing/data': 0.00038144271820783615, 'timing/model': 0.6193570094183087, 'learning_rate/qwen_vl_interface': 8.802887595153607e-06, 'learning_rate/action_model': 8.802887595153606e-05, 'epoch': 0.18})
+[2026-07-28 08:48:12,144][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.063310906291008, 'train/grad_norm_pre_clip': 1.9887895584106445, 'timing/data': 0.0003698151558637619, 'timing/model': 0.6207071719691157, 'learning_rate/qwen_vl_interface': 8.800346831064294e-06, 'learning_rate/action_model': 8.800346831064294e-05, 'epoch': 0.18})
+[2026-07-28 08:48:21,601][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.13886550068855286, 'train/grad_norm_pre_clip': 2.242368698120117, 'timing/data': 0.0005111116915941238, 'timing/model': 0.6561215249821544, 'learning_rate/qwen_vl_interface': 8.797803763193471e-06, 'learning_rate/action_model': 8.797803763193471e-05, 'epoch': 0.18})
+[2026-07-28 08:48:30,822][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.07114821672439575, 'train/grad_norm_pre_clip': 3.1685850620269775, 'timing/data': 0.000387607142329216, 'timing/model': 0.6171477744355798, 'learning_rate/qwen_vl_interface': 8.795258393191302e-06, 'learning_rate/action_model': 8.795258393191301e-05, 'epoch': 0.18})
+[2026-07-28 08:48:39,997][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.056057728826999664, 'train/grad_norm_pre_clip': 1.403313398361206, 'timing/data': 0.0003316979855298996, 'timing/model': 0.6194265717640519, 'learning_rate/qwen_vl_interface': 8.792710722709451e-06, 'learning_rate/action_model': 8.792710722709451e-05, 'epoch': 0.18})
+[2026-07-28 08:48:49,397][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.06922943145036697, 'train/grad_norm_pre_clip': 1.6236695051193237, 'timing/data': 0.00032887421548366547, 'timing/model': 0.8460387261584401, 'learning_rate/qwen_vl_interface': 8.790160753401076e-06, 'learning_rate/action_model': 8.790160753401077e-05, 'epoch': 0.18})
+[2026-07-28 08:48:58,542][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.0640307143330574, 'train/grad_norm_pre_clip': 2.4443788528442383, 'timing/data': 0.00034996960312128067, 'timing/model': 0.6176991686224937, 'learning_rate/qwen_vl_interface': 8.787608486920824e-06, 'learning_rate/action_model': 8.787608486920824e-05, 'epoch': 0.18})
+[2026-07-28 08:49:07,735][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.11055026203393936, 'train/grad_norm_pre_clip': 1.786839246749878, 'timing/data': 0.0003366684541106224, 'timing/model': 0.6203215075656772, 'learning_rate/qwen_vl_interface': 8.78505392492483e-06, 'learning_rate/action_model': 8.78505392492483e-05, 'epoch': 0.19})
+[2026-07-28 08:49:16,713][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.05299210548400879, 'train/grad_norm_pre_clip': 1.5639855861663818, 'timing/data': 0.00031582731753587723, 'timing/model': 0.6193686341866851, 'learning_rate/qwen_vl_interface': 8.782497069070725e-06, 'learning_rate/action_model': 8.782497069070725e-05, 'epoch': 0.19})
+[2026-07-28 08:49:26,075][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.06061690300703049, 'train/grad_norm_pre_clip': 1.2422467470169067, 'timing/data': 0.00039478112012147903, 'timing/model': 0.7903172625228763, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.19})
+[2026-07-28 08:49:35,224][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.07252544164657593, 'train/grad_norm_pre_clip': 1.154735803604126, 'timing/data': 0.00036202650517225266, 'timing/model': 0.619094755500555, 'learning_rate/qwen_vl_interface': 8.777376482426129e-06, 'learning_rate/action_model': 8.777376482426128e-05, 'epoch': 0.19})
+[2026-07-28 08:49:44,584][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.058101803064346313, 'train/grad_norm_pre_clip': 0.9846367239952087, 'timing/data': 0.0003708014264702797, 'timing/model': 0.6191308805719018, 'learning_rate/qwen_vl_interface': 8.774812754958332e-06, 'learning_rate/action_model': 8.774812754958331e-05, 'epoch': 0.19})
+[2026-07-28 08:49:53,744][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.06691012531518936, 'train/grad_norm_pre_clip': 2.577698230743408, 'timing/data': 0.0003559580072760582, 'timing/model': 0.6187293482944369, 'learning_rate/qwen_vl_interface': 8.772246740277805e-06, 'learning_rate/action_model': 8.772246740277805e-05, 'epoch': 0.19})
+[2026-07-28 08:50:03,160][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.06294552981853485, 'train/grad_norm_pre_clip': 2.027498722076416, 'timing/data': 0.00033340975642204285, 'timing/model': 0.8578415960073471, 'learning_rate/qwen_vl_interface': 8.769678440049608e-06, 'learning_rate/action_model': 8.769678440049608e-05, 'epoch': 0.19})
+[2026-07-28 08:50:12,531][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.055099211633205414, 'train/grad_norm_pre_clip': 1.1169474124908447, 'timing/data': 0.0003253398463129997, 'timing/model': 0.6171747194603086, 'learning_rate/qwen_vl_interface': 8.767107855940287e-06, 'learning_rate/action_model': 8.767107855940286e-05, 'epoch': 0.19})
+[2026-07-28 08:50:21,651][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.0801534652709961, 'train/grad_norm_pre_clip': 1.979405403137207, 'timing/data': 0.00032581109553575516, 'timing/model': 0.6182932406663895, 'learning_rate/qwen_vl_interface': 8.764534989617861e-06, 'learning_rate/action_model': 8.76453498961786e-05, 'epoch': 0.19})
+[2026-07-28 08:50:30,787][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.05108339712023735, 'train/grad_norm_pre_clip': 0.9212256669998169, 'timing/data': 0.00032292306423187256, 'timing/model': 0.6183628449216485, 'learning_rate/qwen_vl_interface': 8.761959842751837e-06, 'learning_rate/action_model': 8.761959842751837e-05, 'epoch': 0.19})
+[2026-07-28 08:50:40,352][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.05116121470928192, 'train/grad_norm_pre_clip': 1.2660528421401978, 'timing/data': 0.0003427555784583092, 'timing/model': 0.7901368532329798, 'learning_rate/qwen_vl_interface': 8.7593824170132e-06, 'learning_rate/action_model': 8.7593824170132e-05, 'epoch': 0.19})
+[2026-07-28 08:50:49,500][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.04880044609308243, 'train/grad_norm_pre_clip': 0.8090922832489014, 'timing/data': 0.00034035835415124893, 'timing/model': 0.6212671883404255, 'learning_rate/qwen_vl_interface': 8.756802714074412e-06, 'learning_rate/action_model': 8.756802714074412e-05, 'epoch': 0.19})
+[2026-07-28 08:50:58,631][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.047893229871988297, 'train/grad_norm_pre_clip': 2.3543951511383057, 'timing/data': 0.0003604032099246979, 'timing/model': 0.6193424994125962, 'learning_rate/qwen_vl_interface': 8.75422073560942e-06, 'learning_rate/action_model': 8.754220735609419e-05, 'epoch': 0.19})
+[2026-07-28 08:51:07,800][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.04420766979455948, 'train/grad_norm_pre_clip': 1.850987195968628, 'timing/data': 0.0004608873277902603, 'timing/model': 0.6223279973492026, 'learning_rate/qwen_vl_interface': 8.751636483293635e-06, 'learning_rate/action_model': 8.751636483293634e-05, 'epoch': 0.19})
+[2026-07-28 08:51:17,226][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.10592713952064514, 'train/grad_norm_pre_clip': 2.7336018085479736, 'timing/data': 0.0004018237814307213, 'timing/model': 0.6212797937914729, 'learning_rate/qwen_vl_interface': 8.749049958803952e-06, 'learning_rate/action_model': 8.749049958803951e-05, 'epoch': 0.19})
+[2026-07-28 08:51:26,480][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.05069170147180557, 'train/grad_norm_pre_clip': 1.6132882833480835, 'timing/data': 0.00033685844391584396, 'timing/model': 0.6227842774242163, 'learning_rate/qwen_vl_interface': 8.74646116381874e-06, 'learning_rate/action_model': 8.74646116381874e-05, 'epoch': 0.19})
+[2026-07-28 08:51:35,709][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.03166668862104416, 'train/grad_norm_pre_clip': 1.1730568408966064, 'timing/data': 0.0004095938056707382, 'timing/model': 0.6217992017045617, 'learning_rate/qwen_vl_interface': 8.74387010001784e-06, 'learning_rate/action_model': 8.743870100017838e-05, 'epoch': 0.19})
+[2026-07-28 08:51:45,427][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.1217830628156662, 'train/grad_norm_pre_clip': 3.49460506439209, 'timing/data': 0.00033120904117822647, 'timing/model': 0.8503407370299101, 'learning_rate/qwen_vl_interface': 8.741276769082562e-06, 'learning_rate/action_model': 8.741276769082562e-05, 'epoch': 0.19})
+[2026-07-28 08:51:54,471][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.07710753381252289, 'train/grad_norm_pre_clip': 2.0171170234680176, 'timing/data': 0.00042762886732816696, 'timing/model': 0.6202092804014683, 'learning_rate/qwen_vl_interface': 8.738681172695694e-06, 'learning_rate/action_model': 8.738681172695694e-05, 'epoch': 0.19})
+[2026-07-28 08:52:03,714][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.05415736138820648, 'train/grad_norm_pre_clip': 1.087997317314148, 'timing/data': 0.0003390256315469742, 'timing/model': 0.6382403066381812, 'learning_rate/qwen_vl_interface': 8.73608331254149e-06, 'learning_rate/action_model': 8.736083312541489e-05, 'epoch': 0.19})
+[2026-07-28 08:52:12,730][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.05403173342347145, 'train/grad_norm_pre_clip': 1.155808448791504, 'timing/data': 0.0003307834267616272, 'timing/model': 0.6170520223677158, 'learning_rate/qwen_vl_interface': 8.733483190305671e-06, 'learning_rate/action_model': 8.733483190305671e-05, 'epoch': 0.19})
+[2026-07-28 08:52:22,192][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.08251303434371948, 'train/grad_norm_pre_clip': 1.328922152519226, 'timing/data': 0.00032820459455251694, 'timing/model': 0.8544664587825537, 'learning_rate/qwen_vl_interface': 8.730880807675431e-06, 'learning_rate/action_model': 8.73088080767543e-05, 'epoch': 0.19})
+[2026-07-28 08:52:31,627][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.04681050032377243, 'train/grad_norm_pre_clip': 1.4149264097213745, 'timing/data': 0.000561228021979332, 'timing/model': 0.6199044045060873, 'learning_rate/qwen_vl_interface': 8.728276166339426e-06, 'learning_rate/action_model': 8.728276166339426e-05, 'epoch': 0.19})
+[2026-07-28 08:52:40,853][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.08062257617712021, 'train/grad_norm_pre_clip': 1.5281933546066284, 'timing/data': 0.00034148991107940674, 'timing/model': 0.6308857183903456, 'learning_rate/qwen_vl_interface': 8.725669267987779e-06, 'learning_rate/action_model': 8.725669267987778e-05, 'epoch': 0.19})
+[2026-07-28 08:52:50,113][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.0737706869840622, 'train/grad_norm_pre_clip': 1.3642175197601318, 'timing/data': 0.00034505780786275864, 'timing/model': 0.6196532370522618, 'learning_rate/qwen_vl_interface': 8.723060114312079e-06, 'learning_rate/action_model': 8.723060114312079e-05, 'epoch': 0.19})
+[2026-07-28 08:52:59,418][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.04347764700651169, 'train/grad_norm_pre_clip': 1.4799344539642334, 'timing/data': 0.00033462420105934143, 'timing/model': 0.6196580538526177, 'learning_rate/qwen_vl_interface': 8.720448707005377e-06, 'learning_rate/action_model': 8.720448707005376e-05, 'epoch': 0.19})
+[2026-07-28 08:53:08,830][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.05320556461811066, 'train/grad_norm_pre_clip': 1.1577452421188354, 'timing/data': 0.00033132266253232956, 'timing/model': 0.8884487245231867, 'learning_rate/qwen_vl_interface': 8.717835047762188e-06, 'learning_rate/action_model': 8.717835047762187e-05, 'epoch': 0.19})
+[2026-07-28 08:53:18,270][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.0471215657889843, 'train/grad_norm_pre_clip': 0.8245106339454651, 'timing/data': 0.0003347732126712799, 'timing/model': 0.8656058702617884, 'learning_rate/qwen_vl_interface': 8.715219138278486e-06, 'learning_rate/action_model': 8.715219138278486e-05, 'epoch': 0.19})
+[2026-07-28 08:53:27,403][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.05986173450946808, 'train/grad_norm_pre_clip': 2.077608346939087, 'timing/data': 0.0003292961046099663, 'timing/model': 0.6205975627526641, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.19})
+[2026-07-28 08:53:36,327][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.05018347129225731, 'train/grad_norm_pre_clip': 1.4607930183410645, 'timing/data': 0.0003367466852068901, 'timing/model': 0.6173119200393558, 'learning_rate/qwen_vl_interface': 8.709980575380741e-06, 'learning_rate/action_model': 8.709980575380741e-05, 'epoch': 0.19})
+[2026-07-28 08:53:45,472][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.03655600920319557, 'train/grad_norm_pre_clip': 2.2973275184631348, 'timing/data': 0.0003345096483826637, 'timing/model': 0.6222693705931306, 'learning_rate/qwen_vl_interface': 8.707357925365949e-06, 'learning_rate/action_model': 8.707357925365948e-05, 'epoch': 0.19})
+[2026-07-28 08:53:54,638][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.043932270258665085, 'train/grad_norm_pre_clip': 1.337040662765503, 'timing/data': 0.00032609887421131134, 'timing/model': 0.6185767604038119, 'learning_rate/qwen_vl_interface': 8.704733031909134e-06, 'learning_rate/action_model': 8.704733031909133e-05, 'epoch': 0.19})
+[2026-07-28 08:54:04,053][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.12462303042411804, 'train/grad_norm_pre_clip': 1.8990981578826904, 'timing/data': 0.0003402121365070343, 'timing/model': 0.6206944035366178, 'learning_rate/qwen_vl_interface': 8.70210589671356e-06, 'learning_rate/action_model': 8.702105896713561e-05, 'epoch': 0.19})
+[2026-07-28 08:54:13,469][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.09460420906543732, 'train/grad_norm_pre_clip': 2.1676933765411377, 'timing/data': 0.00033554807305336, 'timing/model': 0.6216871459037066, 'learning_rate/qwen_vl_interface': 8.699476521483952e-06, 'learning_rate/action_model': 8.699476521483952e-05, 'epoch': 0.19})
+[2026-07-28 08:54:22,595][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.0832684338092804, 'train/grad_norm_pre_clip': 2.038362503051758, 'timing/data': 0.0003451388329267502, 'timing/model': 0.6189422467723489, 'learning_rate/qwen_vl_interface': 8.696844907926478e-06, 'learning_rate/action_model': 8.696844907926477e-05, 'epoch': 0.19})
+[2026-07-28 08:54:32,033][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.08636490255594254, 'train/grad_norm_pre_clip': 1.5373643636703491, 'timing/data': 0.00045049097388982773, 'timing/model': 0.6229500528424978, 'learning_rate/qwen_vl_interface': 8.694211057748766e-06, 'learning_rate/action_model': 8.694211057748765e-05, 'epoch': 0.19})
+[2026-07-28 08:54:41,210][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.04742567241191864, 'train/grad_norm_pre_clip': 1.2837523221969604, 'timing/data': 0.0004332335665822029, 'timing/model': 0.6200765036046505, 'learning_rate/qwen_vl_interface': 8.69157497265989e-06, 'learning_rate/action_model': 8.69157497265989e-05, 'epoch': 0.19})
+[2026-07-28 08:54:50,428][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.09469886124134064, 'train/grad_norm_pre_clip': 1.6350164413452148, 'timing/data': 0.0003460599109530449, 'timing/model': 0.6196257947012782, 'learning_rate/qwen_vl_interface': 8.68893665437038e-06, 'learning_rate/action_model': 8.688936654370378e-05, 'epoch': 0.19})
+[2026-07-28 08:54:59,568][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.0484355166554451, 'train/grad_norm_pre_clip': 1.0448930263519287, 'timing/data': 0.0003379685804247856, 'timing/model': 0.6191988624632359, 'learning_rate/qwen_vl_interface': 8.68629610459221e-06, 'learning_rate/action_model': 8.686296104592209e-05, 'epoch': 0.19})
+[2026-07-28 08:55:09,270][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.04722117260098457, 'train/grad_norm_pre_clip': 0.7877568602561951, 'timing/data': 0.00033109914511442184, 'timing/model': 0.646186551079154, 'learning_rate/qwen_vl_interface': 8.683653325038805e-06, 'learning_rate/action_model': 8.683653325038804e-05, 'epoch': 0.19})
+[2026-07-28 08:55:18,480][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.09085601568222046, 'train/grad_norm_pre_clip': 1.440667986869812, 'timing/data': 0.0003620907664299011, 'timing/model': 0.6204899977892637, 'learning_rate/qwen_vl_interface': 8.681008317425037e-06, 'learning_rate/action_model': 8.681008317425037e-05, 'epoch': 0.19})
+[2026-07-28 08:55:27,614][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.06540939211845398, 'train/grad_norm_pre_clip': 0.8809062838554382, 'timing/data': 0.000334874726831913, 'timing/model': 0.6168358474969864, 'learning_rate/qwen_vl_interface': 8.678361083467221e-06, 'learning_rate/action_model': 8.67836108346722e-05, 'epoch': 0.19})
+[2026-07-28 08:55:36,733][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.0625588446855545, 'train/grad_norm_pre_clip': 1.0024120807647705, 'timing/data': 0.00033967848867177963, 'timing/model': 0.6174876391887665, 'learning_rate/qwen_vl_interface': 8.67571162488312e-06, 'learning_rate/action_model': 8.67571162488312e-05, 'epoch': 0.19})
+[2026-07-28 08:55:46,045][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.04933414235711098, 'train/grad_norm_pre_clip': 2.147062301635742, 'timing/data': 0.000332588329911232, 'timing/model': 0.619013492949307, 'learning_rate/qwen_vl_interface': 8.673059943391937e-06, 'learning_rate/action_model': 8.673059943391936e-05, 'epoch': 0.19})
+[2026-07-28 08:55:55,416][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.06552913784980774, 'train/grad_norm_pre_clip': 1.01808762550354, 'timing/data': 0.00033205561339855194, 'timing/model': 0.6220829337835312, 'learning_rate/qwen_vl_interface': 8.670406040714321e-06, 'learning_rate/action_model': 8.670406040714321e-05, 'epoch': 0.19})
+[2026-07-28 08:56:04,582][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.04306817799806595, 'train/grad_norm_pre_clip': 1.6794885396957397, 'timing/data': 0.0004050368443131447, 'timing/model': 0.6188547052443027, 'learning_rate/qwen_vl_interface': 8.667749918572361e-06, 'learning_rate/action_model': 8.667749918572361e-05, 'epoch': 0.19})
+[2026-07-28 08:56:13,710][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.03705696389079094, 'train/grad_norm_pre_clip': 0.7746362686157227, 'timing/data': 0.000376393087208271, 'timing/model': 0.6197112174704671, 'learning_rate/qwen_vl_interface': 8.665091578689584e-06, 'learning_rate/action_model': 8.665091578689584e-05, 'epoch': 0.19})
+[2026-07-28 08:56:23,080][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.10286226868629456, 'train/grad_norm_pre_clip': 2.3283631801605225, 'timing/data': 0.0003717299550771713, 'timing/model': 0.847800669260323, 'learning_rate/qwen_vl_interface': 8.662431022790958e-06, 'learning_rate/action_model': 8.662431022790958e-05, 'epoch': 0.19})
+[2026-07-28 08:56:32,249][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.07262949645519257, 'train/grad_norm_pre_clip': 1.9781060218811035, 'timing/data': 0.00036398228257894516, 'timing/model': 0.6174998236820102, 'learning_rate/qwen_vl_interface': 8.65976825260289e-06, 'learning_rate/action_model': 8.659768252602888e-05, 'epoch': 0.19})
+[2026-07-28 08:56:41,429][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.09237553179264069, 'train/grad_norm_pre_clip': 2.618131637573242, 'timing/data': 0.0005691023543477058, 'timing/model': 0.6352493762969971, 'learning_rate/qwen_vl_interface': 8.657103269853222e-06, 'learning_rate/action_model': 8.657103269853222e-05, 'epoch': 0.19})
+[2026-07-28 08:56:50,381][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.044331230223178864, 'train/grad_norm_pre_clip': 1.4928086996078491, 'timing/data': 0.0003382135182619095, 'timing/model': 0.6398883732035756, 'learning_rate/qwen_vl_interface': 8.654436076271232e-06, 'learning_rate/action_model': 8.654436076271232e-05, 'epoch': 0.19})
+[2026-07-28 08:56:59,781][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.05920115113258362, 'train/grad_norm_pre_clip': 2.2692792415618896, 'timing/data': 0.0003611734136939049, 'timing/model': 0.8186770146712661, 'learning_rate/qwen_vl_interface': 8.651766673587635e-06, 'learning_rate/action_model': 8.651766673587635e-05, 'epoch': 0.19})
+[2026-07-28 08:57:09,010][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.08646886795759201, 'train/grad_norm_pre_clip': 1.6948678493499756, 'timing/data': 0.0004135146737098694, 'timing/model': 0.6408571423962712, 'learning_rate/qwen_vl_interface': 8.649095063534576e-06, 'learning_rate/action_model': 8.649095063534576e-05, 'epoch': 0.19})
+[2026-07-28 08:57:18,433][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.07683321833610535, 'train/grad_norm_pre_clip': 1.376612901687622, 'timing/data': 0.0003403685986995697, 'timing/model': 0.6211465438827872, 'learning_rate/qwen_vl_interface': 8.646421247845633e-06, 'learning_rate/action_model': 8.646421247845633e-05, 'epoch': 0.19})
+[2026-07-28 08:57:27,568][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.030901141464710236, 'train/grad_norm_pre_clip': 2.1196541786193848, 'timing/data': 0.00041536707431077957, 'timing/model': 0.6254029460251331, 'learning_rate/qwen_vl_interface': 8.643745228255816e-06, 'learning_rate/action_model': 8.643745228255817e-05, 'epoch': 0.2})
+[2026-07-28 08:57:36,986][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.03993615508079529, 'train/grad_norm_pre_clip': 1.6143673658370972, 'timing/data': 0.00040870625525712967, 'timing/model': 0.859323282726109, 'learning_rate/qwen_vl_interface': 8.641067006501568e-06, 'learning_rate/action_model': 8.641067006501567e-05, 'epoch': 0.2})
+[2026-07-28 08:57:46,111][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.0704600065946579, 'train/grad_norm_pre_clip': 2.0279018878936768, 'timing/data': 0.0004104059189558029, 'timing/model': 0.6192266531288624, 'learning_rate/qwen_vl_interface': 8.638386584320753e-06, 'learning_rate/action_model': 8.638386584320752e-05, 'epoch': 0.2})
+[2026-07-28 08:57:55,355][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.05857144296169281, 'train/grad_norm_pre_clip': 2.5384633541107178, 'timing/data': 0.0003874748945236206, 'timing/model': 0.6190782943740487, 'learning_rate/qwen_vl_interface': 8.635703963452673e-06, 'learning_rate/action_model': 8.635703963452673e-05, 'epoch': 0.2})
+[2026-07-28 08:58:04,730][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.07925648242235184, 'train/grad_norm_pre_clip': 2.341150999069214, 'timing/data': 0.00039421021938323975, 'timing/model': 0.6252405773848295, 'learning_rate/qwen_vl_interface': 8.633019145638048e-06, 'learning_rate/action_model': 8.633019145638048e-05, 'epoch': 0.2})
+[2026-07-28 08:58:14,105][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.08010730147361755, 'train/grad_norm_pre_clip': 2.5368175506591797, 'timing/data': 0.0004093088209629059, 'timing/model': 0.6358008347451687, 'learning_rate/qwen_vl_interface': 8.630332132619029e-06, 'learning_rate/action_model': 8.630332132619028e-05, 'epoch': 0.2})
+[2026-07-28 08:58:23,010][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.07045750319957733, 'train/grad_norm_pre_clip': 2.236208200454712, 'timing/data': 0.0003415662795305252, 'timing/model': 0.6199495177716017, 'learning_rate/qwen_vl_interface': 8.627642926139187e-06, 'learning_rate/action_model': 8.627642926139187e-05, 'epoch': 0.2})
+[2026-07-28 08:58:32,136][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.0767725259065628, 'train/grad_norm_pre_clip': 1.354814887046814, 'timing/data': 0.0003226920962333679, 'timing/model': 0.619423920288682, 'learning_rate/qwen_vl_interface': 8.624951527943522e-06, 'learning_rate/action_model': 8.624951527943522e-05, 'epoch': 0.2})
+[2026-07-28 08:58:41,274][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.05636746436357498, 'train/grad_norm_pre_clip': 1.8557921648025513, 'timing/data': 0.00040386710315942764, 'timing/model': 0.6177466996014118, 'learning_rate/qwen_vl_interface': 8.622257939778453e-06, 'learning_rate/action_model': 8.622257939778453e-05, 'epoch': 0.2})
+[2026-07-28 08:58:50,869][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.0516878217458725, 'train/grad_norm_pre_clip': 1.6168156862258911, 'timing/data': 0.0004426259547472, 'timing/model': 0.844372701831162, 'learning_rate/qwen_vl_interface': 8.619562163391818e-06, 'learning_rate/action_model': 8.619562163391817e-05, 'epoch': 0.2})
+[2026-07-28 08:59:00,012][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.05997215956449509, 'train/grad_norm_pre_clip': 2.4620120525360107, 'timing/data': 0.00031558237969875336, 'timing/model': 0.6190143255516887, 'learning_rate/qwen_vl_interface': 8.616864200532877e-06, 'learning_rate/action_model': 8.616864200532877e-05, 'epoch': 0.2})
+[2026-07-28 08:59:09,199][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.04924032464623451, 'train/grad_norm_pre_clip': 1.6035188436508179, 'timing/data': 0.0003247484564781189, 'timing/model': 0.6166034648194909, 'learning_rate/qwen_vl_interface': 8.61416405295231e-06, 'learning_rate/action_model': 8.61416405295231e-05, 'epoch': 0.2})
+[2026-07-28 08:59:18,564][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.07592014223337173, 'train/grad_norm_pre_clip': 1.343122124671936, 'timing/data': 0.00044850632548332214, 'timing/model': 0.7903059124946594, 'learning_rate/qwen_vl_interface': 8.611461722402214e-06, 'learning_rate/action_model': 8.611461722402215e-05, 'epoch': 0.2})
+[2026-07-28 08:59:27,712][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.08863649517297745, 'train/grad_norm_pre_clip': 2.0551857948303223, 'timing/data': 0.0003968263044953346, 'timing/model': 0.618746317923069, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.2})
+[2026-07-28 08:59:37,079][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.1020449846982956, 'train/grad_norm_pre_clip': 2.257829189300537, 'timing/data': 0.00038784928619861603, 'timing/model': 0.6208496754989028, 'learning_rate/qwen_vl_interface': 8.606050519408901e-06, 'learning_rate/action_model': 8.606050519408901e-05, 'epoch': 0.2})
+[2026-07-28 08:59:46,230][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.05831544101238251, 'train/grad_norm_pre_clip': 0.7923833131790161, 'timing/data': 0.00034244265407323837, 'timing/model': 0.6189293321222067, 'learning_rate/qwen_vl_interface': 8.603341650476953e-06, 'learning_rate/action_model': 8.603341650476952e-05, 'epoch': 0.2})
+[2026-07-28 08:59:55,666][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.07308467477560043, 'train/grad_norm_pre_clip': 2.6673367023468018, 'timing/data': 0.0003208909183740616, 'timing/model': 0.8551669511944056, 'learning_rate/qwen_vl_interface': 8.600630605598016e-06, 'learning_rate/action_model': 8.600630605598015e-05, 'epoch': 0.2})
+[2026-07-28 09:00:04,778][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.16321918368339539, 'train/grad_norm_pre_clip': 5.699605464935303, 'timing/data': 0.0003285277634859085, 'timing/model': 0.617712470702827, 'learning_rate/qwen_vl_interface': 8.597917386531255e-06, 'learning_rate/action_model': 8.597917386531254e-05, 'epoch': 0.2})
+[2026-07-28 09:00:14,140][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.09318405389785767, 'train/grad_norm_pre_clip': 4.383075714111328, 'timing/data': 0.00033903028815984726, 'timing/model': 0.8493520189076662, 'learning_rate/qwen_vl_interface': 8.595201995037252e-06, 'learning_rate/action_model': 8.595201995037251e-05, 'epoch': 0.2})
+[2026-07-28 09:00:23,287][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.10352388024330139, 'train/grad_norm_pre_clip': 3.4847795963287354, 'timing/data': 0.0003885934129357338, 'timing/model': 0.6181400399655104, 'learning_rate/qwen_vl_interface': 8.59248443287799e-06, 'learning_rate/action_model': 8.592484432877989e-05, 'epoch': 0.2})
+[2026-07-28 09:00:32,443][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.07558847218751907, 'train/grad_norm_pre_clip': 1.8100038766860962, 'timing/data': 0.00041475892066955566, 'timing/model': 0.6191558465361595, 'learning_rate/qwen_vl_interface': 8.589764701816868e-06, 'learning_rate/action_model': 8.589764701816868e-05, 'epoch': 0.2})
+[2026-07-28 09:00:41,830][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.06505933403968811, 'train/grad_norm_pre_clip': 1.8237839937210083, 'timing/data': 0.00038777850568294525, 'timing/model': 0.8483578804880381, 'learning_rate/qwen_vl_interface': 8.587042803618693e-06, 'learning_rate/action_model': 8.587042803618693e-05, 'epoch': 0.2})
+[2026-07-28 09:00:50,975][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.0401066318154335, 'train/grad_norm_pre_clip': 1.0024124383926392, 'timing/data': 0.0006258059293031693, 'timing/model': 0.6242865482345223, 'learning_rate/qwen_vl_interface': 8.584318740049672e-06, 'learning_rate/action_model': 8.584318740049672e-05, 'epoch': 0.2})
+[2026-07-28 09:01:00,111][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.059648144990205765, 'train/grad_norm_pre_clip': 1.834928274154663, 'timing/data': 0.00034501031041145325, 'timing/model': 0.6367137208580971, 'learning_rate/qwen_vl_interface': 8.581592512877419e-06, 'learning_rate/action_model': 8.581592512877418e-05, 'epoch': 0.2})
+[2026-07-28 09:01:09,286][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.05280674993991852, 'train/grad_norm_pre_clip': 1.5613529682159424, 'timing/data': 0.00033378414809703827, 'timing/model': 0.6182170202955604, 'learning_rate/qwen_vl_interface': 8.578864123870955e-06, 'learning_rate/action_model': 8.578864123870955e-05, 'epoch': 0.2})
+[2026-07-28 09:01:18,474][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.04036153852939606, 'train/grad_norm_pre_clip': 1.9794445037841797, 'timing/data': 0.0004331497475504875, 'timing/model': 0.618914170190692, 'learning_rate/qwen_vl_interface': 8.576133574800703e-06, 'learning_rate/action_model': 8.576133574800703e-05, 'epoch': 0.2})
+[2026-07-28 09:01:27,965][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.11008892953395844, 'train/grad_norm_pre_clip': 4.071423530578613, 'timing/data': 0.00032520201057195663, 'timing/model': 0.8531206743791699, 'learning_rate/qwen_vl_interface': 8.573400867438486e-06, 'learning_rate/action_model': 8.573400867438486e-05, 'epoch': 0.2})
+[2026-07-28 09:01:37,186][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.09557253867387772, 'train/grad_norm_pre_clip': 1.6339259147644043, 'timing/data': 0.00043705571442842484, 'timing/model': 0.6205673143267632, 'learning_rate/qwen_vl_interface': 8.570666003557528e-06, 'learning_rate/action_model': 8.570666003557528e-05, 'epoch': 0.2})
+[2026-07-28 09:01:46,536][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.05781950429081917, 'train/grad_norm_pre_clip': 1.793654203414917, 'timing/data': 0.0004058806225657463, 'timing/model': 0.6196794975548983, 'learning_rate/qwen_vl_interface': 8.567928984932452e-06, 'learning_rate/action_model': 8.567928984932452e-05, 'epoch': 0.2})
+[2026-07-28 09:01:55,678][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.04870377480983734, 'train/grad_norm_pre_clip': 0.7524921894073486, 'timing/data': 0.0004142569378018379, 'timing/model': 0.6291671320796013, 'learning_rate/qwen_vl_interface': 8.565189813339278e-06, 'learning_rate/action_model': 8.565189813339278e-05, 'epoch': 0.2})
+[2026-07-28 09:02:04,786][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.08632992208003998, 'train/grad_norm_pre_clip': 1.7843583822250366, 'timing/data': 0.0004067756235599518, 'timing/model': 0.6178561458364129, 'learning_rate/qwen_vl_interface': 8.562448490555429e-06, 'learning_rate/action_model': 8.562448490555429e-05, 'epoch': 0.2})
+[2026-07-28 09:02:13,910][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.05625130236148834, 'train/grad_norm_pre_clip': 1.146359920501709, 'timing/data': 0.0003998046740889549, 'timing/model': 0.6174309635534883, 'learning_rate/qwen_vl_interface': 8.559705018359716e-06, 'learning_rate/action_model': 8.559705018359716e-05, 'epoch': 0.2})
+[2026-07-28 09:02:23,320][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.06374948471784592, 'train/grad_norm_pre_clip': 1.788796305656433, 'timing/data': 0.0004862463101744652, 'timing/model': 0.619178588502109, 'learning_rate/qwen_vl_interface': 8.55695939853235e-06, 'learning_rate/action_model': 8.55695939853235e-05, 'epoch': 0.2})
+[2026-07-28 09:02:32,802][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.052178509533405304, 'train/grad_norm_pre_clip': 2.1359922885894775, 'timing/data': 0.00045902933925390244, 'timing/model': 0.6240857122465968, 'learning_rate/qwen_vl_interface': 8.554211632854932e-06, 'learning_rate/action_model': 8.554211632854932e-05, 'epoch': 0.2})
+[2026-07-28 09:02:41,941][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.029162926599383354, 'train/grad_norm_pre_clip': 0.9770863056182861, 'timing/data': 0.00040529482066631317, 'timing/model': 0.6179977674037218, 'learning_rate/qwen_vl_interface': 8.551461723110461e-06, 'learning_rate/action_model': 8.551461723110461e-05, 'epoch': 0.2})
+[2026-07-28 09:02:51,070][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.07566206157207489, 'train/grad_norm_pre_clip': 1.7476933002471924, 'timing/data': 0.00033385399729013443, 'timing/model': 0.618615242652595, 'learning_rate/qwen_vl_interface': 8.548709671083321e-06, 'learning_rate/action_model': 8.54870967108332e-05, 'epoch': 0.2})
+[2026-07-28 09:03:00,537][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.11299408972263336, 'train/grad_norm_pre_clip': 2.363093137741089, 'timing/data': 0.00033909454941749573, 'timing/model': 0.9938735123723745, 'learning_rate/qwen_vl_interface': 8.545955478559288e-06, 'learning_rate/action_model': 8.545955478559288e-05, 'epoch': 0.2})
+[2026-07-28 09:03:09,451][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.07783087342977524, 'train/grad_norm_pre_clip': 3.467195987701416, 'timing/data': 0.00039875973016023636, 'timing/model': 0.618035283870995, 'learning_rate/qwen_vl_interface': 8.54319914732553e-06, 'learning_rate/action_model': 8.543199147325529e-05, 'epoch': 0.2})
+[2026-07-28 09:03:18,547][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.061901845037937164, 'train/grad_norm_pre_clip': 1.6232316493988037, 'timing/data': 0.00033262092620134354, 'timing/model': 0.6180984675884247, 'learning_rate/qwen_vl_interface': 8.540440679170599e-06, 'learning_rate/action_model': 8.540440679170598e-05, 'epoch': 0.2})
+[2026-07-28 09:03:27,702][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.06368312239646912, 'train/grad_norm_pre_clip': 1.5286133289337158, 'timing/data': 0.0004329858347773552, 'timing/model': 0.6191788129508495, 'learning_rate/qwen_vl_interface': 8.537680075884433e-06, 'learning_rate/action_model': 8.537680075884433e-05, 'epoch': 0.2})
+[2026-07-28 09:03:36,844][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.08110540360212326, 'train/grad_norm_pre_clip': 1.6544034481048584, 'timing/data': 0.00032880622893571854, 'timing/model': 0.8551484644412994, 'learning_rate/qwen_vl_interface': 8.53491733925836e-06, 'learning_rate/action_model': 8.53491733925836e-05, 'epoch': 0.2})
+[2026-07-28 09:03:45,801][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.06500960886478424, 'train/grad_norm_pre_clip': 3.020939826965332, 'timing/data': 0.0004123235121369362, 'timing/model': 0.625530949793756, 'learning_rate/qwen_vl_interface': 8.532152471085092e-06, 'learning_rate/action_model': 8.53215247108509e-05, 'epoch': 0.2})
+[2026-07-28 09:03:54,943][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.05402456223964691, 'train/grad_norm_pre_clip': 0.7581020593643188, 'timing/data': 0.0003328900784254074, 'timing/model': 0.6262406716123223, 'learning_rate/qwen_vl_interface': 8.529385473158717e-06, 'learning_rate/action_model': 8.529385473158717e-05, 'epoch': 0.2})
+[2026-07-28 09:04:03,860][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.03887024521827698, 'train/grad_norm_pre_clip': 0.8975074291229248, 'timing/data': 0.0003234902396798134, 'timing/model': 0.6193285556510091, 'learning_rate/qwen_vl_interface': 8.526616347274712e-06, 'learning_rate/action_model': 8.526616347274712e-05, 'epoch': 0.2})
+[2026-07-28 09:04:13,424][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.061302974820137024, 'train/grad_norm_pre_clip': 1.0885072946548462, 'timing/data': 0.0004287939518690109, 'timing/model': 0.6206728275865316, 'learning_rate/qwen_vl_interface': 8.523845095229935e-06, 'learning_rate/action_model': 8.523845095229935e-05, 'epoch': 0.2})
+[2026-07-28 09:04:22,592][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.06745936721563339, 'train/grad_norm_pre_clip': 1.1760374307632446, 'timing/data': 0.0003866078332066536, 'timing/model': 0.6803636848926544, 'learning_rate/qwen_vl_interface': 8.521071718822616e-06, 'learning_rate/action_model': 8.521071718822616e-05, 'epoch': 0.2})
+[2026-07-28 09:04:31,479][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.03025282919406891, 'train/grad_norm_pre_clip': 1.1215064525604248, 'timing/data': 0.00040220096707344055, 'timing/model': 0.6189753031358123, 'learning_rate/qwen_vl_interface': 8.518296219852375e-06, 'learning_rate/action_model': 8.518296219852375e-05, 'epoch': 0.2})
+[2026-07-28 09:04:40,810][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.041686151176691055, 'train/grad_norm_pre_clip': 1.0600030422210693, 'timing/data': 0.00040758587419986725, 'timing/model': 0.6325689405202866, 'learning_rate/qwen_vl_interface': 8.515518600120201e-06, 'learning_rate/action_model': 8.5155186001202e-05, 'epoch': 0.2})
+[2026-07-28 09:04:50,315][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.06867657601833344, 'train/grad_norm_pre_clip': 1.9250439405441284, 'timing/data': 0.0003273915499448776, 'timing/model': 0.6333943577483296, 'learning_rate/qwen_vl_interface': 8.512738861428458e-06, 'learning_rate/action_model': 8.512738861428458e-05, 'epoch': 0.2})
+[2026-07-28 09:04:59,498][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.04810541495680809, 'train/grad_norm_pre_clip': 1.9281046390533447, 'timing/data': 0.000580398365855217, 'timing/model': 0.6198067711666226, 'learning_rate/qwen_vl_interface': 8.509957005580892e-06, 'learning_rate/action_model': 8.509957005580892e-05, 'epoch': 0.2})
+[2026-07-28 09:05:08,630][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.09347241371870041, 'train/grad_norm_pre_clip': 3.199143409729004, 'timing/data': 0.0004026377573609352, 'timing/model': 0.6188903441652656, 'learning_rate/qwen_vl_interface': 8.507173034382617e-06, 'learning_rate/action_model': 8.507173034382617e-05, 'epoch': 0.2})
+[2026-07-28 09:05:17,875][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.02648652158677578, 'train/grad_norm_pre_clip': 1.1588131189346313, 'timing/data': 0.00032356660813093185, 'timing/model': 0.8940629903227091, 'learning_rate/qwen_vl_interface': 8.504386949640124e-06, 'learning_rate/action_model': 8.504386949640124e-05, 'epoch': 0.2})
+[2026-07-28 09:05:27,021][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.05467024818062782, 'train/grad_norm_pre_clip': 2.1924922466278076, 'timing/data': 0.00032094772905111313, 'timing/model': 0.6192011022940278, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.2})
+[2026-07-28 09:05:36,175][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.09079857915639877, 'train/grad_norm_pre_clip': 2.19441556930542, 'timing/data': 0.0003382861614227295, 'timing/model': 0.6179556297138333, 'learning_rate/qwen_vl_interface': 8.498808446755291e-06, 'learning_rate/action_model': 8.498808446755291e-05, 'epoch': 0.2})
+[2026-07-28 09:05:45,355][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.035711802542209625, 'train/grad_norm_pre_clip': 1.695577621459961, 'timing/data': 0.00031197071075439453, 'timing/model': 0.6275838082656264, 'learning_rate/qwen_vl_interface': 8.49601603223278e-06, 'learning_rate/action_model': 8.49601603223278e-05, 'epoch': 0.2})
+[2026-07-28 09:05:54,611][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.06228022277355194, 'train/grad_norm_pre_clip': 1.2384004592895508, 'timing/data': 0.000340927392244339, 'timing/model': 0.624536782503128, 'learning_rate/qwen_vl_interface': 8.493221511405708e-06, 'learning_rate/action_model': 8.493221511405707e-05, 'epoch': 0.21})
+[2026-07-28 09:06:03,999][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.031790174543857574, 'train/grad_norm_pre_clip': 1.5899802446365356, 'timing/data': 0.0003386642783880234, 'timing/model': 0.6174577316269279, 'learning_rate/qwen_vl_interface': 8.490424886087407e-06, 'learning_rate/action_model': 8.490424886087407e-05, 'epoch': 0.21})
+[2026-07-28 09:06:13,390][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.0314924381673336, 'train/grad_norm_pre_clip': 1.099987506866455, 'timing/data': 0.00032836198806762695, 'timing/model': 0.617470171302557, 'learning_rate/qwen_vl_interface': 8.487626158092578e-06, 'learning_rate/action_model': 8.487626158092579e-05, 'epoch': 0.21})
+[2026-07-28 09:06:22,500][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.08222892135381699, 'train/grad_norm_pre_clip': 2.1636548042297363, 'timing/data': 0.00032115355134010315, 'timing/model': 0.6173878367990255, 'learning_rate/qwen_vl_interface': 8.484825329237287e-06, 'learning_rate/action_model': 8.484825329237287e-05, 'epoch': 0.21})
+[2026-07-28 09:06:31,620][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.06247521936893463, 'train/grad_norm_pre_clip': 2.5522220134735107, 'timing/data': 0.0003162343055009842, 'timing/model': 0.6207528915256262, 'learning_rate/qwen_vl_interface': 8.48202240133896e-06, 'learning_rate/action_model': 8.482022401338959e-05, 'epoch': 0.21})
+[2026-07-28 09:06:40,992][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.07415636628866196, 'train/grad_norm_pre_clip': 3.591090202331543, 'timing/data': 0.00039368215948343277, 'timing/model': 0.6185685452073812, 'learning_rate/qwen_vl_interface': 8.479217376216385e-06, 'learning_rate/action_model': 8.479217376216385e-05, 'epoch': 0.21})
+[2026-07-28 09:06:50,148][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.05245557427406311, 'train/grad_norm_pre_clip': 1.4410743713378906, 'timing/data': 0.0003712764009833336, 'timing/model': 0.618101472966373, 'learning_rate/qwen_vl_interface': 8.476410255689714e-06, 'learning_rate/action_model': 8.476410255689713e-05, 'epoch': 0.21})
+[2026-07-28 09:06:59,496][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.09431703388690948, 'train/grad_norm_pre_clip': 1.8967666625976562, 'timing/data': 0.0003708871081471443, 'timing/model': 0.8386494480073452, 'learning_rate/qwen_vl_interface': 8.473601041580457e-06, 'learning_rate/action_model': 8.473601041580457e-05, 'epoch': 0.21})
+[2026-07-28 09:07:08,866][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.04955216497182846, 'train/grad_norm_pre_clip': 1.4227067232131958, 'timing/data': 0.0004222327843308449, 'timing/model': 0.619209355674684, 'learning_rate/qwen_vl_interface': 8.470789735711483e-06, 'learning_rate/action_model': 8.470789735711483e-05, 'epoch': 0.21})
+[2026-07-28 09:07:18,039][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.07661522924900055, 'train/grad_norm_pre_clip': 1.205734133720398, 'timing/data': 0.00039429590106010437, 'timing/model': 0.6178880762308836, 'learning_rate/qwen_vl_interface': 8.467976339907015e-06, 'learning_rate/action_model': 8.467976339907016e-05, 'epoch': 0.21})
+[2026-07-28 09:07:26,976][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.03823462873697281, 'train/grad_norm_pre_clip': 1.0958845615386963, 'timing/data': 0.00040328316390514374, 'timing/model': 0.617893154732883, 'learning_rate/qwen_vl_interface': 8.465160855992639e-06, 'learning_rate/action_model': 8.465160855992638e-05, 'epoch': 0.21})
+[2026-07-28 09:07:36,132][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.06174500286579132, 'train/grad_norm_pre_clip': 1.310096025466919, 'timing/data': 0.00032621994614601135, 'timing/model': 0.6170751564204693, 'learning_rate/qwen_vl_interface': 8.462343285795289e-06, 'learning_rate/action_model': 8.462343285795288e-05, 'epoch': 0.21})
+[2026-07-28 09:07:45,293][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.05153968185186386, 'train/grad_norm_pre_clip': 2.3313369750976562, 'timing/data': 0.00041589513421058655, 'timing/model': 0.8619071152061224, 'learning_rate/qwen_vl_interface': 8.459523631143259e-06, 'learning_rate/action_model': 8.459523631143258e-05, 'epoch': 0.21})
+[2026-07-28 09:07:54,649][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.03940548002719879, 'train/grad_norm_pre_clip': 0.7813822627067566, 'timing/data': 0.00030906684696674347, 'timing/model': 0.6171892620623112, 'learning_rate/qwen_vl_interface': 8.456701893866188e-06, 'learning_rate/action_model': 8.456701893866187e-05, 'epoch': 0.21})
+[2026-07-28 09:08:03,789][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.055500615388154984, 'train/grad_norm_pre_clip': 1.4157344102859497, 'timing/data': 0.0003288146108388901, 'timing/model': 0.6186315435916185, 'learning_rate/qwen_vl_interface': 8.453878075795072e-06, 'learning_rate/action_model': 8.453878075795072e-05, 'epoch': 0.21})
+[2026-07-28 09:08:12,954][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.05028769001364708, 'train/grad_norm_pre_clip': 1.5945353507995605, 'timing/data': 0.0004069209098815918, 'timing/model': 0.6175385629758239, 'learning_rate/qwen_vl_interface': 8.451052178762258e-06, 'learning_rate/action_model': 8.451052178762258e-05, 'epoch': 0.21})
+[2026-07-28 09:08:22,104][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.05639934539794922, 'train/grad_norm_pre_clip': 1.2516025304794312, 'timing/data': 0.00038593634963035583, 'timing/model': 0.617867992259562, 'learning_rate/qwen_vl_interface': 8.448224204601439e-06, 'learning_rate/action_model': 8.448224204601438e-05, 'epoch': 0.21})
+[2026-07-28 09:08:31,520][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.10484689474105835, 'train/grad_norm_pre_clip': 2.135422706604004, 'timing/data': 0.00041024573147296906, 'timing/model': 0.8459667162969708, 'learning_rate/qwen_vl_interface': 8.445394155147655e-06, 'learning_rate/action_model': 8.445394155147653e-05, 'epoch': 0.21})
+[2026-07-28 09:08:40,693][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.04259927570819855, 'train/grad_norm_pre_clip': 0.9156941771507263, 'timing/data': 0.0005245860666036606, 'timing/model': 0.6233399817720056, 'learning_rate/qwen_vl_interface': 8.442562032237295e-06, 'learning_rate/action_model': 8.442562032237296e-05, 'epoch': 0.21})
+[2026-07-28 09:08:49,899][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.07370008528232574, 'train/grad_norm_pre_clip': 1.0793628692626953, 'timing/data': 0.0003468785434961319, 'timing/model': 0.6183870295062661, 'learning_rate/qwen_vl_interface': 8.439727837708094e-06, 'learning_rate/action_model': 8.439727837708095e-05, 'epoch': 0.21})
+[2026-07-28 09:08:59,329][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.07455240190029144, 'train/grad_norm_pre_clip': 1.6321098804473877, 'timing/data': 0.0003610597923398018, 'timing/model': 0.8411483690142632, 'learning_rate/qwen_vl_interface': 8.436891573399132e-06, 'learning_rate/action_model': 8.436891573399131e-05, 'epoch': 0.21})
+[2026-07-28 09:09:08,553][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.05145283788442612, 'train/grad_norm_pre_clip': 1.090505838394165, 'timing/data': 0.00033922865986824036, 'timing/model': 0.6174935642629862, 'learning_rate/qwen_vl_interface': 8.434053241150826e-06, 'learning_rate/action_model': 8.434053241150825e-05, 'epoch': 0.21})
+[2026-07-28 09:09:17,705][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.04753969609737396, 'train/grad_norm_pre_clip': 1.069930076599121, 'timing/data': 0.00039793550968170166, 'timing/model': 0.6189625356346369, 'learning_rate/qwen_vl_interface': 8.43121284280494e-06, 'learning_rate/action_model': 8.431212842804939e-05, 'epoch': 0.21})
+[2026-07-28 09:09:26,841][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.0853065699338913, 'train/grad_norm_pre_clip': 3.134852170944214, 'timing/data': 0.00033560581505298615, 'timing/model': 0.619160913862288, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.21})
+[2026-07-28 09:09:36,201][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.076692134141922, 'train/grad_norm_pre_clip': 2.3099772930145264, 'timing/data': 0.0003190813586115837, 'timing/model': 0.6170574929565191, 'learning_rate/qwen_vl_interface': 8.425525855194184e-06, 'learning_rate/action_model': 8.425525855194183e-05, 'epoch': 0.21})
+[2026-07-28 09:09:45,594][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.07440520077943802, 'train/grad_norm_pre_clip': 1.2027630805969238, 'timing/data': 0.0003308393061161041, 'timing/model': 0.6209000488743186, 'learning_rate/qwen_vl_interface': 8.422679269619538e-06, 'learning_rate/action_model': 8.422679269619538e-05, 'epoch': 0.21})
+[2026-07-28 09:09:54,779][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.07844790816307068, 'train/grad_norm_pre_clip': 2.2887020111083984, 'timing/data': 0.0004058685153722763, 'timing/model': 0.620021915063262, 'learning_rate/qwen_vl_interface': 8.419830625327758e-06, 'learning_rate/action_model': 8.419830625327758e-05, 'epoch': 0.21})
+[2026-07-28 09:10:03,906][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.06648412346839905, 'train/grad_norm_pre_clip': 1.6093915700912476, 'timing/data': 0.0004085218533873558, 'timing/model': 0.6200020322576165, 'learning_rate/qwen_vl_interface': 8.416979924167303e-06, 'learning_rate/action_model': 8.416979924167303e-05, 'epoch': 0.21})
+[2026-07-28 09:10:13,090][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.059999991208314896, 'train/grad_norm_pre_clip': 0.8091080188751221, 'timing/data': 0.0004229666665196419, 'timing/model': 0.61719651799649, 'learning_rate/qwen_vl_interface': 8.414127167987957e-06, 'learning_rate/action_model': 8.414127167987957e-05, 'epoch': 0.21})
+[2026-07-28 09:10:22,360][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.10481621325016022, 'train/grad_norm_pre_clip': 3.567883014678955, 'timing/data': 0.00033551547676324844, 'timing/model': 0.6191594200208783, 'learning_rate/qwen_vl_interface': 8.411272358640843e-06, 'learning_rate/action_model': 8.411272358640844e-05, 'epoch': 0.21})
+[2026-07-28 09:10:32,021][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.06801407039165497, 'train/grad_norm_pre_clip': 2.301715850830078, 'timing/data': 0.00034075044095516205, 'timing/model': 0.8426556577906013, 'learning_rate/qwen_vl_interface': 8.408415497978418e-06, 'learning_rate/action_model': 8.408415497978418e-05, 'epoch': 0.21})
+[2026-07-28 09:10:40,943][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.0785055160522461, 'train/grad_norm_pre_clip': 1.948225975036621, 'timing/data': 0.0003345906734466553, 'timing/model': 0.619338508695364, 'learning_rate/qwen_vl_interface': 8.405556587854468e-06, 'learning_rate/action_model': 8.405556587854467e-05, 'epoch': 0.21})
+[2026-07-28 09:10:49,848][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.04180710390210152, 'train/grad_norm_pre_clip': 1.5913199186325073, 'timing/data': 0.0003802590072154999, 'timing/model': 0.6173493256792426, 'learning_rate/qwen_vl_interface': 8.402695630124106e-06, 'learning_rate/action_model': 8.402695630124106e-05, 'epoch': 0.21})
+[2026-07-28 09:10:58,726][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.07089132070541382, 'train/grad_norm_pre_clip': 2.653423309326172, 'timing/data': 0.0003892742097377777, 'timing/model': 0.6173058906570077, 'learning_rate/qwen_vl_interface': 8.399832626643778e-06, 'learning_rate/action_model': 8.399832626643778e-05, 'epoch': 0.21})
+[2026-07-28 09:11:07,682][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.08408726751804352, 'train/grad_norm_pre_clip': 1.1050752401351929, 'timing/data': 0.0003439290449023247, 'timing/model': 0.6189041305333376, 'learning_rate/qwen_vl_interface': 8.396967579271256e-06, 'learning_rate/action_model': 8.396967579271257e-05, 'epoch': 0.21})
+[2026-07-28 09:11:17,036][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.04791107028722763, 'train/grad_norm_pre_clip': 0.9668158292770386, 'timing/data': 0.00034272484481334686, 'timing/model': 0.6190769206732512, 'learning_rate/qwen_vl_interface': 8.394100489865638e-06, 'learning_rate/action_model': 8.394100489865638e-05, 'epoch': 0.21})
+[2026-07-28 09:11:26,190][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.043893493711948395, 'train/grad_norm_pre_clip': 1.8527576923370361, 'timing/data': 0.0003298018127679825, 'timing/model': 0.6198087120428681, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.21})
+[2026-07-28 09:11:35,301][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.08155299723148346, 'train/grad_norm_pre_clip': 3.424574136734009, 'timing/data': 0.00035288557410240173, 'timing/model': 0.6181826870888472, 'learning_rate/qwen_vl_interface': 8.388360192398131e-06, 'learning_rate/action_model': 8.38836019239813e-05, 'epoch': 0.21})
+[2026-07-28 09:11:44,695][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.0696677565574646, 'train/grad_norm_pre_clip': 1.9690279960632324, 'timing/data': 0.0003411732614040375, 'timing/model': 0.6194188054651022, 'learning_rate/qwen_vl_interface': 8.385486988061058e-06, 'learning_rate/action_model': 8.385486988061058e-05, 'epoch': 0.21})
+[2026-07-28 09:11:54,075][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.04012889787554741, 'train/grad_norm_pre_clip': 1.0075393915176392, 'timing/data': 0.00034397467970848083, 'timing/model': 0.6186488037928939, 'learning_rate/qwen_vl_interface': 8.38261174914052e-06, 'learning_rate/action_model': 8.38261174914052e-05, 'epoch': 0.21})
+[2026-07-28 09:12:03,234][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.0645742118358612, 'train/grad_norm_pre_clip': 1.2609317302703857, 'timing/data': 0.0003372812643647194, 'timing/model': 0.6179475989192724, 'learning_rate/qwen_vl_interface': 8.379734477502229e-06, 'learning_rate/action_model': 8.379734477502228e-05, 'epoch': 0.21})
+[2026-07-28 09:12:12,412][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.03162521868944168, 'train/grad_norm_pre_clip': 1.0759886503219604, 'timing/data': 0.0003429986536502838, 'timing/model': 0.6193366087973118, 'learning_rate/qwen_vl_interface': 8.376855175013214e-06, 'learning_rate/action_model': 8.376855175013213e-05, 'epoch': 0.21})
+[2026-07-28 09:12:22,058][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.07252460718154907, 'train/grad_norm_pre_clip': 2.083740234375, 'timing/data': 0.0003876481205224991, 'timing/model': 0.8604146353900433, 'learning_rate/qwen_vl_interface': 8.373973843541822e-06, 'learning_rate/action_model': 8.373973843541822e-05, 'epoch': 0.21})
+[2026-07-28 09:12:31,185][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.06792968511581421, 'train/grad_norm_pre_clip': 2.3794713020324707, 'timing/data': 0.00039800163358449936, 'timing/model': 0.6182262748479843, 'learning_rate/qwen_vl_interface': 8.371090484957722e-06, 'learning_rate/action_model': 8.371090484957722e-05, 'epoch': 0.21})
+[2026-07-28 09:12:40,333][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.04754986613988876, 'train/grad_norm_pre_clip': 1.8275879621505737, 'timing/data': 0.00033481698483228683, 'timing/model': 0.6202960219234228, 'learning_rate/qwen_vl_interface': 8.368205101131891e-06, 'learning_rate/action_model': 8.368205101131891e-05, 'epoch': 0.21})
+[2026-07-28 09:12:49,467][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.0725603997707367, 'train/grad_norm_pre_clip': 1.9018068313598633, 'timing/data': 0.00039397645741701126, 'timing/model': 0.6174400057643652, 'learning_rate/qwen_vl_interface': 8.365317693936624e-06, 'learning_rate/action_model': 8.365317693936623e-05, 'epoch': 0.21})
+[2026-07-28 09:12:59,045][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.016385497525334358, 'train/grad_norm_pre_clip': 0.8378438949584961, 'timing/data': 0.0003255316987633705, 'timing/model': 0.8600426148623228, 'learning_rate/qwen_vl_interface': 8.362428265245526e-06, 'learning_rate/action_model': 8.362428265245526e-05, 'epoch': 0.21})
+[2026-07-28 09:13:08,170][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.053239330649375916, 'train/grad_norm_pre_clip': 1.1753367185592651, 'timing/data': 0.0003311876207590103, 'timing/model': 0.6192503217607737, 'learning_rate/qwen_vl_interface': 8.359536816933518e-06, 'learning_rate/action_model': 8.359536816933518e-05, 'epoch': 0.21})
+[2026-07-28 09:13:17,468][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.0914393812417984, 'train/grad_norm_pre_clip': 1.9897701740264893, 'timing/data': 0.0004907771944999695, 'timing/model': 0.6179468035697937, 'learning_rate/qwen_vl_interface': 8.356643350876832e-06, 'learning_rate/action_model': 8.356643350876832e-05, 'epoch': 0.21})
+[2026-07-28 09:13:26,693][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.1324259638786316, 'train/grad_norm_pre_clip': 3.150197982788086, 'timing/data': 0.00032967329025268555, 'timing/model': 0.6182606657966971, 'learning_rate/qwen_vl_interface': 8.353747868953002e-06, 'learning_rate/action_model': 8.353747868953003e-05, 'epoch': 0.21})
+[2026-07-28 09:13:36,198][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.04229845851659775, 'train/grad_norm_pre_clip': 1.752628207206726, 'timing/data': 0.0003818720579147339, 'timing/model': 0.9191070236265659, 'learning_rate/qwen_vl_interface': 8.350850373040876e-06, 'learning_rate/action_model': 8.350850373040876e-05, 'epoch': 0.21})
+[2026-07-28 09:13:45,352][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.06952279806137085, 'train/grad_norm_pre_clip': 2.162248134613037, 'timing/data': 0.0003614630550146103, 'timing/model': 0.6177675742655993, 'learning_rate/qwen_vl_interface': 8.34795086502061e-06, 'learning_rate/action_model': 8.347950865020609e-05, 'epoch': 0.21})
+[2026-07-28 09:13:54,567][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.040670134127140045, 'train/grad_norm_pre_clip': 1.244747519493103, 'timing/data': 0.0003492133691906929, 'timing/model': 0.6212255135178566, 'learning_rate/qwen_vl_interface': 8.34504934677366e-06, 'learning_rate/action_model': 8.34504934677366e-05, 'epoch': 0.21})
+[2026-07-28 09:14:03,866][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.07173745334148407, 'train/grad_norm_pre_clip': 2.1663591861724854, 'timing/data': 0.0003327634185552597, 'timing/model': 0.6236912030726671, 'learning_rate/qwen_vl_interface': 8.342145820182795e-06, 'learning_rate/action_model': 8.342145820182793e-05, 'epoch': 0.21})
+[2026-07-28 09:14:13,406][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.07197420299053192, 'train/grad_norm_pre_clip': 1.1316375732421875, 'timing/data': 0.00035573914647102356, 'timing/model': 0.6241342946887016, 'learning_rate/qwen_vl_interface': 8.339240287132073e-06, 'learning_rate/action_model': 8.339240287132073e-05, 'epoch': 0.22})
+[2026-07-28 09:14:22,714][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.044838398694992065, 'train/grad_norm_pre_clip': 1.1249909400939941, 'timing/data': 0.0003497414290904999, 'timing/model': 0.6240617381408811, 'learning_rate/qwen_vl_interface': 8.336332749506872e-06, 'learning_rate/action_model': 8.336332749506872e-05, 'epoch': 0.22})
+[2026-07-28 09:14:32,132][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.09043758362531662, 'train/grad_norm_pre_clip': 3.208266019821167, 'timing/data': 0.000462925061583519, 'timing/model': 0.6237343456596136, 'learning_rate/qwen_vl_interface': 8.333423209193857e-06, 'learning_rate/action_model': 8.333423209193856e-05, 'epoch': 0.22})
+[2026-07-28 09:14:41,679][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.05954243987798691, 'train/grad_norm_pre_clip': 1.1008342504501343, 'timing/data': 0.00035462062805891037, 'timing/model': 0.8106947373598814, 'learning_rate/qwen_vl_interface': 8.330511668080997e-06, 'learning_rate/action_model': 8.330511668080997e-05, 'epoch': 0.22})
+[2026-07-28 09:14:50,924][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.050015076994895935, 'train/grad_norm_pre_clip': 0.717184841632843, 'timing/data': 0.0004001390188932419, 'timing/model': 0.6236723363399506, 'learning_rate/qwen_vl_interface': 8.32759812805756e-06, 'learning_rate/action_model': 8.32759812805756e-05, 'epoch': 0.22})
+[2026-07-28 09:15:00,381][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.060623571276664734, 'train/grad_norm_pre_clip': 1.6015186309814453, 'timing/data': 0.0003816969692707062, 'timing/model': 0.6294450890272856, 'learning_rate/qwen_vl_interface': 8.324682591014111e-06, 'learning_rate/action_model': 8.32468259101411e-05, 'epoch': 0.22})
+[2026-07-28 09:15:09,649][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.06799697130918503, 'train/grad_norm_pre_clip': 1.1720956563949585, 'timing/data': 0.0003409469500184059, 'timing/model': 0.6208676556125283, 'learning_rate/qwen_vl_interface': 8.32176505884251e-06, 'learning_rate/action_model': 8.32176505884251e-05, 'epoch': 0.22})
+[2026-07-28 09:15:19,190][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.0710013285279274, 'train/grad_norm_pre_clip': 1.6149929761886597, 'timing/data': 0.000357980839908123, 'timing/model': 0.8744581276550889, 'learning_rate/qwen_vl_interface': 8.318845533435913e-06, 'learning_rate/action_model': 8.318845533435913e-05, 'epoch': 0.22})
+[2026-07-28 09:15:28,723][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.04325399175286293, 'train/grad_norm_pre_clip': 1.1357629299163818, 'timing/data': 0.000348871573805809, 'timing/model': 0.7868070313706994, 'learning_rate/qwen_vl_interface': 8.315924016688767e-06, 'learning_rate/action_model': 8.315924016688766e-05, 'epoch': 0.22})
+[2026-07-28 09:15:38,055][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.0468270406126976, 'train/grad_norm_pre_clip': 0.6826428771018982, 'timing/data': 0.00039114709943532944, 'timing/model': 0.6182448212057352, 'learning_rate/qwen_vl_interface': 8.313000510496811e-06, 'learning_rate/action_model': 8.313000510496811e-05, 'epoch': 0.22})
+[2026-07-28 09:15:47,435][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.0704268366098404, 'train/grad_norm_pre_clip': 1.4216071367263794, 'timing/data': 0.0004294170066714287, 'timing/model': 0.668702038936317, 'learning_rate/qwen_vl_interface': 8.31007501675708e-06, 'learning_rate/action_model': 8.310075016757079e-05, 'epoch': 0.22})
+[2026-07-28 09:15:56,758][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.06918464601039886, 'train/grad_norm_pre_clip': 2.3906304836273193, 'timing/data': 0.0004289606586098671, 'timing/model': 0.6190283000469208, 'learning_rate/qwen_vl_interface': 8.307147537367894e-06, 'learning_rate/action_model': 8.307147537367893e-05, 'epoch': 0.22})
+[2026-07-28 09:16:06,241][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.023175418376922607, 'train/grad_norm_pre_clip': 0.621971845626831, 'timing/data': 0.0004305504262447357, 'timing/model': 0.8680109540000558, 'learning_rate/qwen_vl_interface': 8.30421807422886e-06, 'learning_rate/action_model': 8.30421807422886e-05, 'epoch': 0.22})
+[2026-07-28 09:16:15,626][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.03162446618080139, 'train/grad_norm_pre_clip': 0.8186124563217163, 'timing/data': 0.0004523247480392456, 'timing/model': 0.6178881647065282, 'learning_rate/qwen_vl_interface': 8.301286629240876e-06, 'learning_rate/action_model': 8.301286629240876e-05, 'epoch': 0.22})
+[2026-07-28 09:16:25,072][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.052149832248687744, 'train/grad_norm_pre_clip': 1.8606758117675781, 'timing/data': 0.0003957552835345268, 'timing/model': 0.6578615270555019, 'learning_rate/qwen_vl_interface': 8.298353204306127e-06, 'learning_rate/action_model': 8.298353204306127e-05, 'epoch': 0.22})
+[2026-07-28 09:16:34,502][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.09300583600997925, 'train/grad_norm_pre_clip': 1.9823843240737915, 'timing/data': 0.00034296419471502304, 'timing/model': 0.8725240705534816, 'learning_rate/qwen_vl_interface': 8.295417801328081e-06, 'learning_rate/action_model': 8.29541780132808e-05, 'epoch': 0.22})
+[2026-07-28 09:16:43,697][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.049067422747612, 'train/grad_norm_pre_clip': 1.772150993347168, 'timing/data': 0.0003763604909181595, 'timing/model': 0.6181418560445309, 'learning_rate/qwen_vl_interface': 8.292480422211486e-06, 'learning_rate/action_model': 8.292480422211485e-05, 'epoch': 0.22})
+[2026-07-28 09:16:52,940][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.057212136685848236, 'train/grad_norm_pre_clip': 1.9814730882644653, 'timing/data': 0.000361097976565361, 'timing/model': 0.6273729186505079, 'learning_rate/qwen_vl_interface': 8.289541068862377e-06, 'learning_rate/action_model': 8.289541068862377e-05, 'epoch': 0.22})
+[2026-07-28 09:17:02,131][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.06324497610330582, 'train/grad_norm_pre_clip': 1.7424592971801758, 'timing/data': 0.0003445865586400032, 'timing/model': 0.622238096781075, 'learning_rate/qwen_vl_interface': 8.286599743188068e-06, 'learning_rate/action_model': 8.286599743188069e-05, 'epoch': 0.22})
+[2026-07-28 09:17:11,656][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.05218828469514847, 'train/grad_norm_pre_clip': 2.093409776687622, 'timing/data': 0.0004349369555711746, 'timing/model': 0.6291574584320188, 'learning_rate/qwen_vl_interface': 8.283656447097154e-06, 'learning_rate/action_model': 8.283656447097153e-05, 'epoch': 0.22})
+[2026-07-28 09:17:21,077][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.013533875346183777, 'train/grad_norm_pre_clip': 0.8692969679832458, 'timing/data': 0.0003544529899954796, 'timing/model': 0.6223214827477932, 'learning_rate/qwen_vl_interface': 8.280711182499507e-06, 'learning_rate/action_model': 8.280711182499506e-05, 'epoch': 0.22})
+[2026-07-28 09:17:30,276][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.07041051238775253, 'train/grad_norm_pre_clip': 3.277578592300415, 'timing/data': 0.00043461285531520844, 'timing/model': 0.6193980975076556, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.22})
+[2026-07-28 09:17:39,395][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.05033768713474274, 'train/grad_norm_pre_clip': 1.696146011352539, 'timing/data': 0.0003344491124153137, 'timing/model': 0.618409221060574, 'learning_rate/qwen_vl_interface': 8.274814755429892e-06, 'learning_rate/action_model': 8.274814755429891e-05, 'epoch': 0.22})
+[2026-07-28 09:17:48,655][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.04367046430706978, 'train/grad_norm_pre_clip': 0.8607578873634338, 'timing/data': 0.0003725001588463783, 'timing/model': 0.6216276446357369, 'learning_rate/qwen_vl_interface': 8.271863596784053e-06, 'learning_rate/action_model': 8.271863596784051e-05, 'epoch': 0.22})
+[2026-07-28 09:17:57,668][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.02639632672071457, 'train/grad_norm_pre_clip': 0.911345899105072, 'timing/data': 0.00048081763088703156, 'timing/model': 0.6263337107375264, 'learning_rate/qwen_vl_interface': 8.268910477283732e-06, 'learning_rate/action_model': 8.26891047728373e-05, 'epoch': 0.22})
+[2026-07-28 09:18:07,207][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.014148244634270668, 'train/grad_norm_pre_clip': 0.8707817196846008, 'timing/data': 0.00035282131284475327, 'timing/model': 0.8643178744241595, 'learning_rate/qwen_vl_interface': 8.265955398845177e-06, 'learning_rate/action_model': 8.265955398845177e-05, 'epoch': 0.22})
+[2026-07-28 09:18:16,440][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.0670211911201477, 'train/grad_norm_pre_clip': 1.5976051092147827, 'timing/data': 0.00040614884346723557, 'timing/model': 0.6221423568204045, 'learning_rate/qwen_vl_interface': 8.26299836338591e-06, 'learning_rate/action_model': 8.26299836338591e-05, 'epoch': 0.22})
+[2026-07-28 09:18:25,601][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.022571926936507225, 'train/grad_norm_pre_clip': 0.841655433177948, 'timing/data': 0.00043472833931446075, 'timing/model': 0.6188142187893391, 'learning_rate/qwen_vl_interface': 8.260039372824715e-06, 'learning_rate/action_model': 8.260039372824715e-05, 'epoch': 0.22})
+[2026-07-28 09:18:34,782][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.07109330594539642, 'train/grad_norm_pre_clip': 2.0393025875091553, 'timing/data': 0.00034676212817430496, 'timing/model': 0.6179773900657892, 'learning_rate/qwen_vl_interface': 8.257078429081652e-06, 'learning_rate/action_model': 8.257078429081651e-05, 'epoch': 0.22})
+[2026-07-28 09:18:43,905][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.07379760593175888, 'train/grad_norm_pre_clip': 2.897024393081665, 'timing/data': 0.00033361464738845825, 'timing/model': 0.6183324484154582, 'learning_rate/qwen_vl_interface': 8.254115534078041e-06, 'learning_rate/action_model': 8.254115534078042e-05, 'epoch': 0.22})
+[2026-07-28 09:18:53,474][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.137538343667984, 'train/grad_norm_pre_clip': 3.515392303466797, 'timing/data': 0.00032818503677845, 'timing/model': 0.6225658087059855, 'learning_rate/qwen_vl_interface': 8.25115068973648e-06, 'learning_rate/action_model': 8.25115068973648e-05, 'epoch': 0.22})
+[2026-07-28 09:19:02,611][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.037181586027145386, 'train/grad_norm_pre_clip': 1.8387995958328247, 'timing/data': 0.0003801397979259491, 'timing/model': 0.6204544333741069, 'learning_rate/qwen_vl_interface': 8.248183897980815e-06, 'learning_rate/action_model': 8.248183897980815e-05, 'epoch': 0.22})
+[2026-07-28 09:19:11,836][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.1229294016957283, 'train/grad_norm_pre_clip': 3.11169171333313, 'timing/data': 0.00045190099626779556, 'timing/model': 0.6191461086273193, 'learning_rate/qwen_vl_interface': 8.24521516073617e-06, 'learning_rate/action_model': 8.24521516073617e-05, 'epoch': 0.22})
+[2026-07-28 09:19:21,355][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.05550607293844223, 'train/grad_norm_pre_clip': 2.121792793273926, 'timing/data': 0.00037711113691329956, 'timing/model': 0.6212635412812233, 'learning_rate/qwen_vl_interface': 8.24224447992893e-06, 'learning_rate/action_model': 8.242244479928928e-05, 'epoch': 0.22})
+[2026-07-28 09:19:30,726][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.08185543864965439, 'train/grad_norm_pre_clip': 3.60569429397583, 'timing/data': 0.00039060134440660477, 'timing/model': 0.6181955207139254, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.22})
+[2026-07-28 09:19:39,851][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.022949974983930588, 'train/grad_norm_pre_clip': 0.9141924381256104, 'timing/data': 0.00035617128014564514, 'timing/model': 0.618742149323225, 'learning_rate/qwen_vl_interface': 8.236297295338474e-06, 'learning_rate/action_model': 8.236297295338474e-05, 'epoch': 0.22})
+[2026-07-28 09:19:49,016][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.03888062387704849, 'train/grad_norm_pre_clip': 1.0312864780426025, 'timing/data': 0.0003415727987885475, 'timing/model': 0.6215973012149334, 'learning_rate/qwen_vl_interface': 8.233320795414327e-06, 'learning_rate/action_model': 8.233320795414327e-05, 'epoch': 0.22})
+[2026-07-28 09:19:58,638][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.05732032284140587, 'train/grad_norm_pre_clip': 1.412521243095398, 'timing/data': 0.0003453316166996956, 'timing/model': 0.8394930241629481, 'learning_rate/qwen_vl_interface': 8.230342359645705e-06, 'learning_rate/action_model': 8.230342359645705e-05, 'epoch': 0.22})
+[2026-07-28 09:20:07,865][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.07150372117757797, 'train/grad_norm_pre_clip': 3.1130120754241943, 'timing/data': 0.0003569042310118675, 'timing/model': 0.629974446259439, 'learning_rate/qwen_vl_interface': 8.227361989965285e-06, 'learning_rate/action_model': 8.227361989965284e-05, 'epoch': 0.22})
+[2026-07-28 09:20:17,015][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.09853126853704453, 'train/grad_norm_pre_clip': 1.5990816354751587, 'timing/data': 0.00040062423795461655, 'timing/model': 0.6197903556749225, 'learning_rate/qwen_vl_interface': 8.22437968830699e-06, 'learning_rate/action_model': 8.22437968830699e-05, 'epoch': 0.22})
+[2026-07-28 09:20:26,148][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.07088667899370193, 'train/grad_norm_pre_clip': 1.7175089120864868, 'timing/data': 0.0003390973433852196, 'timing/model': 0.6176214087754488, 'learning_rate/qwen_vl_interface': 8.221395456606012e-06, 'learning_rate/action_model': 8.221395456606011e-05, 'epoch': 0.22})
+[2026-07-28 09:20:35,482][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.05478745698928833, 'train/grad_norm_pre_clip': 1.4767539501190186, 'timing/data': 0.00033982284367084503, 'timing/model': 0.6176360780373216, 'learning_rate/qwen_vl_interface': 8.21840929679878e-06, 'learning_rate/action_model': 8.218409296798778e-05, 'epoch': 0.22})
+[2026-07-28 09:20:44,839][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.08430330455303192, 'train/grad_norm_pre_clip': 1.361129641532898, 'timing/data': 0.0003390982747077942, 'timing/model': 0.8500237856060266, 'learning_rate/qwen_vl_interface': 8.215421210822982e-06, 'learning_rate/action_model': 8.215421210822982e-05, 'epoch': 0.22})
+[2026-07-28 09:20:53,934][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.0795137956738472, 'train/grad_norm_pre_clip': 4.17917537689209, 'timing/data': 0.0003935443237423897, 'timing/model': 0.6184575399383903, 'learning_rate/qwen_vl_interface': 8.212431200617557e-06, 'learning_rate/action_model': 8.212431200617556e-05, 'epoch': 0.22})
+[2026-07-28 09:21:03,083][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.04210387542843819, 'train/grad_norm_pre_clip': 1.6769063472747803, 'timing/data': 0.0005184579640626907, 'timing/model': 0.6242210296913981, 'learning_rate/qwen_vl_interface': 8.209439268122689e-06, 'learning_rate/action_model': 8.209439268122688e-05, 'epoch': 0.22})
+[2026-07-28 09:21:12,555][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.05130990222096443, 'train/grad_norm_pre_clip': 1.6853461265563965, 'timing/data': 0.00045631173998117447, 'timing/model': 0.8663156311959028, 'learning_rate/qwen_vl_interface': 8.206445415279808e-06, 'learning_rate/action_model': 8.206445415279807e-05, 'epoch': 0.22})
+[2026-07-28 09:21:21,739][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.05936910957098007, 'train/grad_norm_pre_clip': 1.9315317869186401, 'timing/data': 0.0003302181139588356, 'timing/model': 0.6185346189886332, 'learning_rate/qwen_vl_interface': 8.203449644031595e-06, 'learning_rate/action_model': 8.203449644031596e-05, 'epoch': 0.22})
+[2026-07-28 09:21:30,861][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.1086200401186943, 'train/grad_norm_pre_clip': 3.978194236755371, 'timing/data': 0.0003398144617676735, 'timing/model': 0.618485533632338, 'learning_rate/qwen_vl_interface': 8.200451956321974e-06, 'learning_rate/action_model': 8.200451956321974e-05, 'epoch': 0.22})
+[2026-07-28 09:21:40,001][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.03752773255109787, 'train/grad_norm_pre_clip': 0.8441939353942871, 'timing/data': 0.000598520040512085, 'timing/model': 0.6193456836044788, 'learning_rate/qwen_vl_interface': 8.197452354096112e-06, 'learning_rate/action_model': 8.197452354096112e-05, 'epoch': 0.22})
+[2026-07-28 09:21:49,381][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.09802314639091492, 'train/grad_norm_pre_clip': 2.3255374431610107, 'timing/data': 0.0003653820604085922, 'timing/model': 0.6215394083410501, 'learning_rate/qwen_vl_interface': 8.194450839300418e-06, 'learning_rate/action_model': 8.194450839300417e-05, 'epoch': 0.22})
+[2026-07-28 09:21:58,565][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.057293180376291275, 'train/grad_norm_pre_clip': 2.247380018234253, 'timing/data': 0.00038261711597442627, 'timing/model': 0.6215544464066625, 'learning_rate/qwen_vl_interface': 8.191447413882539e-06, 'learning_rate/action_model': 8.191447413882539e-05, 'epoch': 0.22})
+[2026-07-28 09:22:07,969][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.0728989839553833, 'train/grad_norm_pre_clip': 1.294630527496338, 'timing/data': 0.0004617879167199135, 'timing/model': 0.6181560596451163, 'learning_rate/qwen_vl_interface': 8.188442079791369e-06, 'learning_rate/action_model': 8.188442079791369e-05, 'epoch': 0.22})
+[2026-07-28 09:22:17,103][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.04893450811505318, 'train/grad_norm_pre_clip': 1.8422822952270508, 'timing/data': 0.00043008849024772644, 'timing/model': 0.6193816680461168, 'learning_rate/qwen_vl_interface': 8.185434838977036e-06, 'learning_rate/action_model': 8.185434838977037e-05, 'epoch': 0.22})
+[2026-07-28 09:22:26,277][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.04063737392425537, 'train/grad_norm_pre_clip': 1.329875111579895, 'timing/data': 0.0003302060067653656, 'timing/model': 0.6182884052395821, 'learning_rate/qwen_vl_interface': 8.182425693390905e-06, 'learning_rate/action_model': 8.182425693390905e-05, 'epoch': 0.22})
+[2026-07-28 09:22:35,646][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.051313988864421844, 'train/grad_norm_pre_clip': 2.2191855907440186, 'timing/data': 0.00032863207161426544, 'timing/model': 0.6186586720868945, 'learning_rate/qwen_vl_interface': 8.179414644985582e-06, 'learning_rate/action_model': 8.179414644985582e-05, 'epoch': 0.23})
+[2026-07-28 09:22:44,975][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.04286254942417145, 'train/grad_norm_pre_clip': 0.9040931463241577, 'timing/data': 0.0003396160900592804, 'timing/model': 0.8410252574831247, 'learning_rate/qwen_vl_interface': 8.176401695714897e-06, 'learning_rate/action_model': 8.176401695714895e-05, 'epoch': 0.23})
+[2026-07-28 09:22:54,093][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.06732699275016785, 'train/grad_norm_pre_clip': 2.0138537883758545, 'timing/data': 0.0003498280420899391, 'timing/model': 0.6210865778848529, 'learning_rate/qwen_vl_interface': 8.173386847533922e-06, 'learning_rate/action_model': 8.173386847533921e-05, 'epoch': 0.23})
+[2026-07-28 09:23:03,389][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.0456472709774971, 'train/grad_norm_pre_clip': 1.3823069334030151, 'timing/data': 0.00035347137600183487, 'timing/model': 0.7941702920943499, 'learning_rate/qwen_vl_interface': 8.170370102398962e-06, 'learning_rate/action_model': 8.17037010239896e-05, 'epoch': 0.23})
+[2026-07-28 09:23:12,524][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.053527720272541046, 'train/grad_norm_pre_clip': 1.2493711709976196, 'timing/data': 0.00034261494874954224, 'timing/model': 0.6200386499986053, 'learning_rate/qwen_vl_interface': 8.167351462267548e-06, 'learning_rate/action_model': 8.167351462267548e-05, 'epoch': 0.23})
+[2026-07-28 09:23:21,654][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.06512598693370819, 'train/grad_norm_pre_clip': 1.7345917224884033, 'timing/data': 0.00036146771162748337, 'timing/model': 0.6187713285908103, 'learning_rate/qwen_vl_interface': 8.164330929098444e-06, 'learning_rate/action_model': 8.164330929098443e-05, 'epoch': 0.23})
+[2026-07-28 09:23:31,112][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.03542851284146309, 'train/grad_norm_pre_clip': 0.6806804537773132, 'timing/data': 0.0003402046859264374, 'timing/model': 0.6197604024782777, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.23})
+[2026-07-28 09:23:40,494][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.0382390022277832, 'train/grad_norm_pre_clip': 1.1135603189468384, 'timing/data': 0.00034577399492263794, 'timing/model': 0.6158687565475702, 'learning_rate/qwen_vl_interface': 8.158284191488355e-06, 'learning_rate/action_model': 8.158284191488355e-05, 'epoch': 0.23})
+[2026-07-28 09:23:49,715][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.05319840461015701, 'train/grad_norm_pre_clip': 1.2916375398635864, 'timing/data': 0.00033198483288288116, 'timing/model': 0.6168646765872836, 'learning_rate/qwen_vl_interface': 8.155257990971032e-06, 'learning_rate/action_model': 8.155257990971031e-05, 'epoch': 0.23})
+[2026-07-28 09:23:58,883][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.05917275696992874, 'train/grad_norm_pre_clip': 3.6390762329101562, 'timing/data': 0.00034610461443662643, 'timing/model': 0.6183441979810596, 'learning_rate/qwen_vl_interface': 8.152229905263341e-06, 'learning_rate/action_model': 8.152229905263342e-05, 'epoch': 0.23})
+[2026-07-28 09:24:08,484][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.03187878057360649, 'train/grad_norm_pre_clip': 0.7936422228813171, 'timing/data': 0.0003397557884454727, 'timing/model': 0.8356749303638935, 'learning_rate/qwen_vl_interface': 8.149199936330175e-06, 'learning_rate/action_model': 8.149199936330174e-05, 'epoch': 0.23})
+[2026-07-28 09:24:17,595][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.028110869228839874, 'train/grad_norm_pre_clip': 1.2150770425796509, 'timing/data': 0.00034934934228658676, 'timing/model': 0.6186800813302398, 'learning_rate/qwen_vl_interface': 8.146168086137646e-06, 'learning_rate/action_model': 8.146168086137646e-05, 'epoch': 0.23})
+[2026-07-28 09:24:26,749][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.047654684633016586, 'train/grad_norm_pre_clip': 1.4026622772216797, 'timing/data': 0.00033021531999111176, 'timing/model': 0.6194831943139434, 'learning_rate/qwen_vl_interface': 8.143134356653087e-06, 'learning_rate/action_model': 8.143134356653087e-05, 'epoch': 0.23})
+[2026-07-28 09:24:36,101][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.04784165322780609, 'train/grad_norm_pre_clip': 1.6071139574050903, 'timing/data': 0.0003326674923300743, 'timing/model': 0.8364036846905947, 'learning_rate/qwen_vl_interface': 8.140098749845056e-06, 'learning_rate/action_model': 8.140098749845055e-05, 'epoch': 0.23})
+[2026-07-28 09:24:45,247][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.04617246240377426, 'train/grad_norm_pre_clip': 1.581336259841919, 'timing/data': 0.0003772713243961334, 'timing/model': 0.6193275339901447, 'learning_rate/qwen_vl_interface': 8.137061267683323e-06, 'learning_rate/action_model': 8.137061267683323e-05, 'epoch': 0.23})
+[2026-07-28 09:24:54,612][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.0871213749051094, 'train/grad_norm_pre_clip': 2.67246413230896, 'timing/data': 0.0004242928698658943, 'timing/model': 0.6215962432324886, 'learning_rate/qwen_vl_interface': 8.134021912138879e-06, 'learning_rate/action_model': 8.134021912138879e-05, 'epoch': 0.23})
+[2026-07-28 09:25:03,793][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.04144754260778427, 'train/grad_norm_pre_clip': 1.352819561958313, 'timing/data': 0.0003443891182541847, 'timing/model': 0.6178704062476754, 'learning_rate/qwen_vl_interface': 8.130980685183927e-06, 'learning_rate/action_model': 8.130980685183926e-05, 'epoch': 0.23})
+[2026-07-28 09:25:12,943][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.0343102291226387, 'train/grad_norm_pre_clip': 1.431548833847046, 'timing/data': 0.000355633907020092, 'timing/model': 0.6185994883999228, 'learning_rate/qwen_vl_interface': 8.127937588791887e-06, 'learning_rate/action_model': 8.127937588791886e-05, 'epoch': 0.23})
+[2026-07-28 09:25:22,303][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.059999238699674606, 'train/grad_norm_pre_clip': 1.2189109325408936, 'timing/data': 0.0003320425748825073, 'timing/model': 0.8507593749091029, 'learning_rate/qwen_vl_interface': 8.12489262493739e-06, 'learning_rate/action_model': 8.12489262493739e-05, 'epoch': 0.23})
+[2026-07-28 09:25:31,436][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.040323950350284576, 'train/grad_norm_pre_clip': 1.1278949975967407, 'timing/data': 0.00033517926931381226, 'timing/model': 0.6177848940715194, 'learning_rate/qwen_vl_interface': 8.121845795596282e-06, 'learning_rate/action_model': 8.121845795596282e-05, 'epoch': 0.23})
+[2026-07-28 09:25:40,803][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.05027313530445099, 'train/grad_norm_pre_clip': 1.1922284364700317, 'timing/data': 0.00039867497980594635, 'timing/model': 0.8516175858676434, 'learning_rate/qwen_vl_interface': 8.118797102745614e-06, 'learning_rate/action_model': 8.118797102745615e-05, 'epoch': 0.23})
+[2026-07-28 09:25:49,971][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.019446901977062225, 'train/grad_norm_pre_clip': 1.4048305749893188, 'timing/data': 0.0004937499761581421, 'timing/model': 0.616442690603435, 'learning_rate/qwen_vl_interface': 8.115746548363655e-06, 'learning_rate/action_model': 8.115746548363655e-05, 'epoch': 0.23})
+[2026-07-28 09:25:59,115][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.06413508951663971, 'train/grad_norm_pre_clip': 1.5105198621749878, 'timing/data': 0.00035109370946884155, 'timing/model': 0.6175865726545453, 'learning_rate/qwen_vl_interface': 8.112694134429873e-06, 'learning_rate/action_model': 8.112694134429873e-05, 'epoch': 0.23})
+[2026-07-28 09:26:08,221][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.047714993357658386, 'train/grad_norm_pre_clip': 2.345852851867676, 'timing/data': 0.0003453278914093971, 'timing/model': 0.6171253183856606, 'learning_rate/qwen_vl_interface': 8.109639862924947e-06, 'learning_rate/action_model': 8.109639862924945e-05, 'epoch': 0.23})
+[2026-07-28 09:26:17,390][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.030482744798064232, 'train/grad_norm_pre_clip': 0.8906347751617432, 'timing/data': 0.0003910493105649948, 'timing/model': 0.6156489597633481, 'learning_rate/qwen_vl_interface': 8.10658373583076e-06, 'learning_rate/action_model': 8.106583735830761e-05, 'epoch': 0.23})
+[2026-07-28 09:26:26,751][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1250
+[2026-07-28 09:26:37,160][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1250
+[2026-07-28 09:42:34,180][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1250
+[2026-07-28 09:42:34,182][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.04323573783040047, 'train/grad_norm_pre_clip': 1.638857364654541, 'eval/action_loss/seconds': 10.405981979332864, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.05426392704248428, 'eval/latency_3/loss': 0.05426392704248428, 'eval/flappy/loss': 0.05426392704248428, 'eval/flappy/latency_3/loss': 0.05426392704248428, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 370.0700043536723, 'rl_games_eval/mid_train/mean_length': 3000.3, 'rl_games_eval/mid_train/std_reward': 120.02981894853505, 'rl_games_eval/mid_train/std_length': 966.6442520389805, 'rl_games_eval/mid_train/macro_mean_reward': 370.0700043536723, 'rl_games_eval/mid_train/macro_mean_length': 3000.3, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 370.0700043536723, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 3000.3, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 120.02981894853505, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 966.6442520389805, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 370.0700043536723, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003082975745201111, 'timing/model': 0.6181721147149801, 'learning_rate/qwen_vl_interface': 8.1035257551304e-06, 'learning_rate/action_model': 8.1035257551304e-05, 'epoch': 0.23})
+[2026-07-28 09:42:43,571][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.03416423127055168, 'train/grad_norm_pre_clip': 1.3885451555252075, 'timing/data': 0.0004215193912386894, 'timing/model': 0.6212939284741879, 'learning_rate/qwen_vl_interface': 8.100465922808158e-06, 'learning_rate/action_model': 8.100465922808159e-05, 'epoch': 0.23})
+[2026-07-28 09:42:52,928][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.056349609047174454, 'train/grad_norm_pre_clip': 1.7401922941207886, 'timing/data': 0.00031286105513572693, 'timing/model': 0.6188444644212723, 'learning_rate/qwen_vl_interface': 8.097404240849527e-06, 'learning_rate/action_model': 8.097404240849527e-05, 'epoch': 0.23})
+[2026-07-28 09:43:02,273][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.06936061382293701, 'train/grad_norm_pre_clip': 1.5527597665786743, 'timing/data': 0.0003424137830734253, 'timing/model': 0.8571344697847962, 'learning_rate/qwen_vl_interface': 8.094340711241198e-06, 'learning_rate/action_model': 8.094340711241198e-05, 'epoch': 0.23})
+[2026-07-28 09:43:11,371][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.06619646400213242, 'train/grad_norm_pre_clip': 1.9953445196151733, 'timing/data': 0.0003295512869954109, 'timing/model': 0.6179059715941548, 'learning_rate/qwen_vl_interface': 8.09127533597106e-06, 'learning_rate/action_model': 8.091275335971061e-05, 'epoch': 0.23})
+[2026-07-28 09:43:20,702][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.06281378865242004, 'train/grad_norm_pre_clip': 2.1458637714385986, 'timing/data': 0.00032309722155332565, 'timing/model': 0.618990944698453, 'learning_rate/qwen_vl_interface': 8.088208117028205e-06, 'learning_rate/action_model': 8.088208117028206e-05, 'epoch': 0.23})
+[2026-07-28 09:43:29,819][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.06317050009965897, 'train/grad_norm_pre_clip': 1.6977945566177368, 'timing/data': 0.00032801553606987, 'timing/model': 0.6185941034927964, 'learning_rate/qwen_vl_interface': 8.085139056402918e-06, 'learning_rate/action_model': 8.085139056402917e-05, 'epoch': 0.23})
+[2026-07-28 09:43:38,973][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.044316694140434265, 'train/grad_norm_pre_clip': 1.7087924480438232, 'timing/data': 0.0003236830234527588, 'timing/model': 0.6177949905395508, 'learning_rate/qwen_vl_interface': 8.082068156086675e-06, 'learning_rate/action_model': 8.082068156086675e-05, 'epoch': 0.23})
+[2026-07-28 09:43:48,333][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.062207721173763275, 'train/grad_norm_pre_clip': 1.526681900024414, 'timing/data': 0.0003129029646515846, 'timing/model': 0.6225951239466667, 'learning_rate/qwen_vl_interface': 8.078995418072156e-06, 'learning_rate/action_model': 8.078995418072157e-05, 'epoch': 0.23})
+[2026-07-28 09:43:57,727][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.06196129694581032, 'train/grad_norm_pre_clip': 1.7758102416992188, 'timing/data': 0.0003210185095667839, 'timing/model': 0.6305430103093386, 'learning_rate/qwen_vl_interface': 8.075920844353223e-06, 'learning_rate/action_model': 8.075920844353223e-05, 'epoch': 0.23})
+[2026-07-28 09:44:06,632][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.035337187349796295, 'train/grad_norm_pre_clip': 1.1959471702575684, 'timing/data': 0.00031928718090057373, 'timing/model': 0.6184983346611261, 'learning_rate/qwen_vl_interface': 8.072844436924934e-06, 'learning_rate/action_model': 8.072844436924934e-05, 'epoch': 0.23})
+[2026-07-28 09:44:15,764][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.03579207509756088, 'train/grad_norm_pre_clip': 1.5411179065704346, 'timing/data': 0.00032943207770586014, 'timing/model': 0.618926196359098, 'learning_rate/qwen_vl_interface': 8.069766197783537e-06, 'learning_rate/action_model': 8.069766197783537e-05, 'epoch': 0.23})
+[2026-07-28 09:44:24,657][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.036017123609781265, 'train/grad_norm_pre_clip': 1.319211721420288, 'timing/data': 0.000312192365527153, 'timing/model': 0.6175666423514485, 'learning_rate/qwen_vl_interface': 8.066686128926468e-06, 'learning_rate/action_model': 8.066686128926468e-05, 'epoch': 0.23})
+[2026-07-28 09:44:34,246][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.0498015359044075, 'train/grad_norm_pre_clip': 1.1958364248275757, 'timing/data': 0.00032322388142347336, 'timing/model': 0.6192991714924574, 'learning_rate/qwen_vl_interface': 8.063604232352352e-06, 'learning_rate/action_model': 8.06360423235235e-05, 'epoch': 0.23})
+[2026-07-28 09:44:43,168][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.051497094333171844, 'train/grad_norm_pre_clip': 1.276816725730896, 'timing/data': 0.0003265542909502983, 'timing/model': 0.6204432435333729, 'learning_rate/qwen_vl_interface': 8.060520510060996e-06, 'learning_rate/action_model': 8.060520510060995e-05, 'epoch': 0.23})
+[2026-07-28 09:44:52,315][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.05190371349453926, 'train/grad_norm_pre_clip': 1.259170413017273, 'timing/data': 0.00033114105463027954, 'timing/model': 0.617912539280951, 'learning_rate/qwen_vl_interface': 8.057434964053391e-06, 'learning_rate/action_model': 8.057434964053392e-05, 'epoch': 0.23})
+[2026-07-28 09:45:01,211][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.051524437963962555, 'train/grad_norm_pre_clip': 1.1852633953094482, 'timing/data': 0.00032703671604394913, 'timing/model': 0.6189559670165181, 'learning_rate/qwen_vl_interface': 8.05434759633172e-06, 'learning_rate/action_model': 8.054347596331721e-05, 'epoch': 0.23})
+[2026-07-28 09:45:10,617][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.03577343001961708, 'train/grad_norm_pre_clip': 1.0733176469802856, 'timing/data': 0.0004436727613210678, 'timing/model': 0.8404793031513691, 'learning_rate/qwen_vl_interface': 8.05125840889934e-06, 'learning_rate/action_model': 8.05125840889934e-05, 'epoch': 0.23})
+[2026-07-28 09:45:19,770][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.04729587957262993, 'train/grad_norm_pre_clip': 1.6576189994812012, 'timing/data': 0.0004180213436484337, 'timing/model': 0.6180056221783161, 'learning_rate/qwen_vl_interface': 8.048167403760792e-06, 'learning_rate/action_model': 8.048167403760792e-05, 'epoch': 0.23})
+[2026-07-28 09:45:28,893][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.05663830786943436, 'train/grad_norm_pre_clip': 1.7656989097595215, 'timing/data': 0.0003847023472189903, 'timing/model': 0.6181821627542377, 'learning_rate/qwen_vl_interface': 8.045074582921797e-06, 'learning_rate/action_model': 8.045074582921797e-05, 'epoch': 0.23})
+[2026-07-28 10:36:34,787][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21'
+[2026-07-28 10:36:34,787][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 10:36:34,788][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 10:36:35,010][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 10:36:35,010][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 10:36:35,012][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 10:36:35,012][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 10:36:35,745][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 10:36:35,760][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 10:36:38,303][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 10:36:39,240][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 10:36:39,241][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 10:36:39,241][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 10:36:39,241][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 10:36:39,241][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 10:36:39,241][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 10:36:39,241][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 10:36:39,242][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 10:36:39,242][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 10:36:39,242][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 10:36:39,243][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 10:36:39,243][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 10:36:39,243][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 10:36:39,244][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 10:36:39,244][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 10:36:39,244][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 10:36:39,245][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 10:36:39,245][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 10:36:39,245][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 10:36:39,246][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 10:36:39,246][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 10:36:39,246][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 10:36:39,247][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 10:36:39,247][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 10:36:39,247][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 10:36:39,247][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 10:36:39,247][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 10:36:39,248][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 10:36:39,248][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 10:36:39,248][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 10:36:39,248][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 10:36:39,522][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 10:36:39,537][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 10:36:41,121][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 10:36:42,048][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 10:36:42,049][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 10:36:42,049][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 10:36:42,049][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 10:36:42,050][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 10:36:42,050][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 10:36:42,050][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 10:36:42,051][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 10:36:42,051][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 10:36:42,051][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 10:36:42,052][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 10:36:42,052][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 10:36:42,052][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 10:36:42,053][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 10:36:42,053][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 10:36:42,053][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 10:36:42,054][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 10:36:42,054][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 10:36:42,054][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 10:36:42,055][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 10:36:42,055][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 10:36:42,055][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 10:36:42,055][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 10:36:42,056][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 10:36:42,056][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 10:36:42,056][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 10:36:42,056][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 10:36:42,056][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 10:36:42,057][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 10:36:42,057][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 10:36:42,057][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 10:36:43,424][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 10:36:43,424][starVLA.dataloader][INFO] - Created train VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 10:36:43,427][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 10:36:43,445][starVLA.dataloader][INFO] - Created eval VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 10:36:43,454][starVLA.training.train_starvla][INFO] - LR Group qwen_vl_interface: lr=1e-05, num_params=713
+[2026-07-28 10:36:43,454][starVLA.training.train_starvla][INFO] - LR Group action_model: lr=0.0001, num_params=16
+[2026-07-28 10:36:44,800][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1000_state, steps: 1000
+[2026-07-28 10:36:44,811][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/config.full.yaml
+[2026-07-28 10:36:44,824][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/config.yaml
+[2026-07-28 10:36:44,825][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1000_state, steps: 1000
+[2026-07-28 10:36:44,830][accelerate.accelerator][INFO] - Since you passed both train and evaluation dataloader, `is_train_batch_min` (here True will decide the `train_batch_size` (8).
+[2026-07-28 10:37:19,857][accelerate.accelerator][INFO] - Loading states from /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1000_state
+[2026-07-28 10:37:19,857][accelerate.accelerator][INFO] - Loading DeepSpeed Model and Optimizer
+[2026-07-28 10:37:54,876][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer loaded from input dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1000_state/pytorch_model
+[2026-07-28 10:37:54,877][accelerate.checkpointing][INFO] - All model weights loaded successfully
+[2026-07-28 10:37:54,877][accelerate.checkpointing][INFO] - All optimizer states loaded successfully
+[2026-07-28 10:37:54,877][accelerate.checkpointing][INFO] - All scheduler states loaded successfully
+[2026-07-28 10:37:54,877][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully
+[2026-07-28 10:37:54,882][accelerate.checkpointing][INFO] - All random states loaded successfully
+[2026-07-28 10:37:54,883][accelerate.accelerator][INFO] - Loading in 0 custom states
+[2026-07-28 10:37:59,279][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-07-28 10:37:59,279][starVLA.training.train_starvla][INFO] -   Total optimization steps = 4000
+[2026-07-28 10:37:59,280][starVLA.training.train_starvla][INFO] -   Per device batch size = 8
+[2026-07-28 10:37:59,280][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-07-28 10:37:59,280][starVLA.training.train_starvla][INFO] -   Total batch size = 256
+[2026-07-28 10:39:19,323][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.03186050057411194, 'train/grad_norm_pre_clip': 1.0122541189193726, 'timing/data': 0.0003172159194946289, 'timing/model': 0.783578023314476, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.18})
+[2026-07-28 10:39:26,579][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.060927316546440125, 'train/grad_norm_pre_clip': 1.1362478733062744, 'timing/data': 0.0005213068798184395, 'timing/model': 0.5285662095993757, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.18})
+[2026-07-28 10:39:33,783][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.06972655653953552, 'train/grad_norm_pre_clip': 1.3235447406768799, 'timing/data': 0.00034187454730272293, 'timing/model': 0.5258514629676938, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.18})
+[2026-07-28 10:39:41,138][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.042149823158979416, 'train/grad_norm_pre_clip': 1.4479641914367676, 'timing/data': 0.0006289510056376457, 'timing/model': 0.522676401771605, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.18})
+[2026-07-28 10:39:48,476][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.0395512580871582, 'train/grad_norm_pre_clip': 1.5000091791152954, 'timing/data': 0.00032132212072610855, 'timing/model': 0.525311297737062, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.18})
+[2026-07-28 10:39:55,746][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.09027591347694397, 'train/grad_norm_pre_clip': 2.4204232692718506, 'timing/data': 0.00041415728628635406, 'timing/model': 0.5273220371454954, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.18})
+[2026-07-28 10:40:03,016][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.07561652362346649, 'train/grad_norm_pre_clip': 1.47344172000885, 'timing/data': 0.0004559876397252083, 'timing/model': 0.5261920681223273, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.18})
+[2026-07-28 10:40:10,361][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.07413910329341888, 'train/grad_norm_pre_clip': 2.6696724891662598, 'timing/data': 0.00034552253782749176, 'timing/model': 0.5258554480969906, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.19})
+[2026-07-28 10:40:17,712][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.08641114085912704, 'train/grad_norm_pre_clip': 1.4549094438552856, 'timing/data': 0.00040537770837545395, 'timing/model': 0.5304695349186659, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.19})
+[2026-07-28 10:40:25,056][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.06891269981861115, 'train/grad_norm_pre_clip': 0.9797093272209167, 'timing/data': 0.000331636518239975, 'timing/model': 0.5266556888818741, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.19})
+[2026-07-28 10:40:32,378][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.09466347098350525, 'train/grad_norm_pre_clip': 2.5872507095336914, 'timing/data': 0.0003611678257584572, 'timing/model': 0.5241468297317624, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.19})
+[2026-07-28 10:40:39,702][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.10376641154289246, 'train/grad_norm_pre_clip': 3.6634819507598877, 'timing/data': 0.00035735033452510834, 'timing/model': 0.5239542834460735, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.19})
+[2026-07-28 10:40:46,880][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.0350828543305397, 'train/grad_norm_pre_clip': 1.327942132949829, 'timing/data': 0.00042569637298583984, 'timing/model': 0.5310365976765752, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.19})
+[2026-07-28 10:40:54,066][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.10628829896450043, 'train/grad_norm_pre_clip': 1.4751056432724, 'timing/data': 0.00043792184442281723, 'timing/model': 0.5259700855240226, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.19})
+[2026-07-28 10:41:01,271][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.07405754923820496, 'train/grad_norm_pre_clip': 1.8629428148269653, 'timing/data': 0.000349314883351326, 'timing/model': 0.5295841600745916, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.19})
+[2026-07-28 10:41:08,468][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.10612856596708298, 'train/grad_norm_pre_clip': 3.543052911758423, 'timing/data': 0.00048195570707321167, 'timing/model': 0.5259958142414689, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.19})
+[2026-07-28 10:41:15,670][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.09680478274822235, 'train/grad_norm_pre_clip': 1.2865058183670044, 'timing/data': 0.00039824750274419785, 'timing/model': 0.5288531240075827, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.19})
+[2026-07-28 10:41:22,869][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.04853289574384689, 'train/grad_norm_pre_clip': 1.5684865713119507, 'timing/data': 0.00033688079565763474, 'timing/model': 0.5254174247384071, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.19})
+[2026-07-28 10:41:30,047][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.05275958403944969, 'train/grad_norm_pre_clip': 0.904090404510498, 'timing/data': 0.0003232154995203018, 'timing/model': 0.5263227177783847, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.19})
+[2026-07-28 10:41:37,235][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.06051177531480789, 'train/grad_norm_pre_clip': 1.1262649297714233, 'timing/data': 0.0005521178245544434, 'timing/model': 0.5240476420149207, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.19})
+[2026-07-28 10:41:44,480][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.08768171072006226, 'train/grad_norm_pre_clip': 1.418836236000061, 'timing/data': 0.0005523925647139549, 'timing/model': 0.5319978818297386, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.19})
+[2026-07-28 10:41:51,696][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.04336594045162201, 'train/grad_norm_pre_clip': 1.4563593864440918, 'timing/data': 0.00033243559300899506, 'timing/model': 0.5312069309875369, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.19})
+[2026-07-28 10:41:58,894][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.05823640525341034, 'train/grad_norm_pre_clip': 1.5647929906845093, 'timing/data': 0.0004214458167552948, 'timing/model': 0.5266946395859122, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.19})
+[2026-07-28 10:42:06,087][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.061835166066884995, 'train/grad_norm_pre_clip': 1.3686143159866333, 'timing/data': 0.0004036491736769676, 'timing/model': 0.5291694104671478, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.19})
+[2026-07-28 10:42:13,261][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.06587688624858856, 'train/grad_norm_pre_clip': 1.5354634523391724, 'timing/data': 0.0003305422142148018, 'timing/model': 0.5267624156549573, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.19})
+[2026-07-28 10:42:20,484][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.06061974912881851, 'train/grad_norm_pre_clip': 1.1272393465042114, 'timing/data': 0.0003376202657818794, 'timing/model': 0.5279626548290253, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.19})
+[2026-07-28 10:42:27,889][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.05386266112327576, 'train/grad_norm_pre_clip': 1.3730074167251587, 'timing/data': 0.0004609934985637665, 'timing/model': 0.5371568687260151, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.19})
+[2026-07-28 10:42:35,526][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.06785964220762253, 'train/grad_norm_pre_clip': 1.224644660949707, 'timing/data': 0.00039565470069646835, 'timing/model': 0.5390200158581138, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.19})
+[2026-07-28 10:42:43,126][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.06535397469997406, 'train/grad_norm_pre_clip': 0.8269012570381165, 'timing/data': 0.0003349967300891876, 'timing/model': 0.5336346821859479, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.19})
+[2026-07-28 10:42:50,297][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.05628617852926254, 'train/grad_norm_pre_clip': 1.6811318397521973, 'timing/data': 0.00037363823503255844, 'timing/model': 0.5250974977388978, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.19})
+[2026-07-28 10:42:57,478][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.0385240875184536, 'train/grad_norm_pre_clip': 1.054690957069397, 'timing/data': 0.0007100822404026985, 'timing/model': 0.526435287669301, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.19})
+[2026-07-28 10:43:04,669][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.05094984173774719, 'train/grad_norm_pre_clip': 0.9271349310874939, 'timing/data': 0.00033359136432409286, 'timing/model': 0.5284931939095259, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.19})
+[2026-07-28 10:43:11,912][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.05686379224061966, 'train/grad_norm_pre_clip': 1.760553240776062, 'timing/data': 0.0004736604169011116, 'timing/model': 0.5304086646065116, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.19})
+[2026-07-28 10:43:19,205][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.053116872906684875, 'train/grad_norm_pre_clip': 1.0937120914459229, 'timing/data': 0.0003649480640888214, 'timing/model': 0.5334728378802538, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.19})
+[2026-07-28 10:43:26,454][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.03615252301096916, 'train/grad_norm_pre_clip': 0.939107358455658, 'timing/data': 0.0003321627154946327, 'timing/model': 0.5280262315645814, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.19})
+[2026-07-28 10:43:33,726][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.030625512823462486, 'train/grad_norm_pre_clip': 0.7090281844139099, 'timing/data': 0.0003976309671998024, 'timing/model': 0.5262910379096866, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.19})
+[2026-07-28 10:43:40,930][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.03637862205505371, 'train/grad_norm_pre_clip': 1.8469345569610596, 'timing/data': 0.00034655071794986725, 'timing/model': 0.5266000563278794, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.19})
+[2026-07-28 10:43:48,128][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.05260089039802551, 'train/grad_norm_pre_clip': 2.6900765895843506, 'timing/data': 0.0003645382821559906, 'timing/model': 0.5272107757627964, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.19})
+[2026-07-28 10:43:55,325][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.09625744819641113, 'train/grad_norm_pre_clip': 3.921698570251465, 'timing/data': 0.00037300679832696915, 'timing/model': 0.5288175800815225, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.19})
+[2026-07-28 10:44:02,520][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.04683200269937515, 'train/grad_norm_pre_clip': 2.6534907817840576, 'timing/data': 0.00033605005592107773, 'timing/model': 0.527022740803659, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.19})
+[2026-07-28 10:44:09,711][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.03896729648113251, 'train/grad_norm_pre_clip': 1.9852664470672607, 'timing/data': 0.00036343932151794434, 'timing/model': 0.528392038308084, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.19})
+[2026-07-28 10:44:16,891][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.07105941325426102, 'train/grad_norm_pre_clip': 2.4127485752105713, 'timing/data': 0.00035529863089323044, 'timing/model': 0.5257064225152135, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.19})
+[2026-07-28 10:44:24,083][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.03954112529754639, 'train/grad_norm_pre_clip': 1.3519505262374878, 'timing/data': 0.00038138497620821, 'timing/model': 0.5271614687517285, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.19})
+[2026-07-28 10:44:31,274][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.027689283713698387, 'train/grad_norm_pre_clip': 1.435041069984436, 'timing/data': 0.00036206934601068497, 'timing/model': 0.5274197198450565, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.19})
+[2026-07-28 10:44:38,465][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.03472331911325455, 'train/grad_norm_pre_clip': 1.4336729049682617, 'timing/data': 0.0003329552710056305, 'timing/model': 0.5265366584062576, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.19})
+[2026-07-28 10:44:45,655][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.032105229794979095, 'train/grad_norm_pre_clip': 1.563651204109192, 'timing/data': 0.0003283517435193062, 'timing/model': 0.5312913302332163, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.19})
+[2026-07-28 10:44:52,853][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.0823911651968956, 'train/grad_norm_pre_clip': 1.9088058471679688, 'timing/data': 0.0003399820998311043, 'timing/model': 0.5293633518740535, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.19})
+[2026-07-28 10:45:00,037][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.10012972354888916, 'train/grad_norm_pre_clip': 3.0775275230407715, 'timing/data': 0.00040044449269771576, 'timing/model': 0.5281054126098752, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.19})
+[2026-07-28 10:45:07,382][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.06926625967025757, 'train/grad_norm_pre_clip': 2.883289098739624, 'timing/data': 0.0004965281113982201, 'timing/model': 0.5523023400455713, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.19})
+[2026-07-28 10:45:14,924][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.046809546649456024, 'train/grad_norm_pre_clip': 1.3878165483474731, 'timing/data': 0.0005596345290541649, 'timing/model': 0.5496949460357428, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.19})
+[2026-07-28 10:45:22,120][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.08776281028985977, 'train/grad_norm_pre_clip': 2.110520601272583, 'timing/data': 0.0003474568948149681, 'timing/model': 0.5277239819988608, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.19})
+[2026-07-28 10:45:29,323][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.05168367549777031, 'train/grad_norm_pre_clip': 2.272846221923828, 'timing/data': 0.0003578411415219307, 'timing/model': 0.5275245197117329, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.19})
+[2026-07-28 10:45:36,509][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.04904630035161972, 'train/grad_norm_pre_clip': 2.9924211502075195, 'timing/data': 0.0003409655764698982, 'timing/model': 0.5277426848188043, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.19})
+[2026-07-28 10:45:43,687][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.07887782901525497, 'train/grad_norm_pre_clip': 2.3975324630737305, 'timing/data': 0.00035519711673259735, 'timing/model': 0.5255273310467601, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.19})
+[2026-07-28 10:45:50,863][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.04338429123163223, 'train/grad_norm_pre_clip': 0.9367244243621826, 'timing/data': 0.00034149736166000366, 'timing/model': 0.5264614289626479, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.19})
+[2026-07-28 10:45:58,082][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.03787928447127342, 'train/grad_norm_pre_clip': 0.9932609796524048, 'timing/data': 0.0003636954352259636, 'timing/model': 0.5287080900743604, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.19})
+[2026-07-28 10:46:05,289][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.043583378195762634, 'train/grad_norm_pre_clip': 1.4717601537704468, 'timing/data': 0.00041134003549814224, 'timing/model': 0.529277853667736, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.19})
+[2026-07-28 10:46:12,484][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.07020372897386551, 'train/grad_norm_pre_clip': 1.60235595703125, 'timing/data': 0.00038947444409132004, 'timing/model': 0.526859836652875, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.19})
+[2026-07-28 10:46:19,666][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.05879737809300423, 'train/grad_norm_pre_clip': 1.9986038208007812, 'timing/data': 0.00031934957951307297, 'timing/model': 0.5281331297010183, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.19})
+[2026-07-28 10:46:26,848][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.05934056639671326, 'train/grad_norm_pre_clip': 1.5578703880310059, 'timing/data': 0.0003720950335264206, 'timing/model': 0.5274113211780787, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.19})
+[2026-07-28 10:46:34,194][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.06071825325489044, 'train/grad_norm_pre_clip': 2.1893107891082764, 'timing/data': 0.0003458913415670395, 'timing/model': 0.5540398675948381, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.19})
+[2026-07-28 10:46:41,600][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.03897465020418167, 'train/grad_norm_pre_clip': 0.9533362984657288, 'timing/data': 0.0003815526142716408, 'timing/model': 0.5250467089936137, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.2})
+[2026-07-28 10:46:48,795][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.02904842607676983, 'train/grad_norm_pre_clip': 1.4855105876922607, 'timing/data': 0.00035473238676786423, 'timing/model': 0.5285700056701899, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.2})
+[2026-07-28 10:46:55,989][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.04344084486365318, 'train/grad_norm_pre_clip': 1.4399906396865845, 'timing/data': 0.0004470162093639374, 'timing/model': 0.5281440326943994, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.2})
+[2026-07-28 10:47:03,178][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.044910870492458344, 'train/grad_norm_pre_clip': 1.0320549011230469, 'timing/data': 0.0004324633628129959, 'timing/model': 0.526763535104692, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.2})
+[2026-07-28 10:47:10,359][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.07838358730077744, 'train/grad_norm_pre_clip': 1.6710022687911987, 'timing/data': 0.00033538229763507843, 'timing/model': 0.5269329203292727, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.2})
+[2026-07-28 10:47:18,110][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.04065656661987305, 'train/grad_norm_pre_clip': 1.4587208032608032, 'timing/data': 0.0003258902579545975, 'timing/model': 0.5269822655245662, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.2})
+[2026-07-28 10:47:25,309][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.04800812527537346, 'train/grad_norm_pre_clip': 1.5590194463729858, 'timing/data': 0.0003356225788593292, 'timing/model': 0.5302760088816285, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.2})
+[2026-07-28 10:47:32,490][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.0477188304066658, 'train/grad_norm_pre_clip': 1.0582985877990723, 'timing/data': 0.000346173532307148, 'timing/model': 0.5289176860824227, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.2})
+[2026-07-28 10:47:39,675][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.04780140519142151, 'train/grad_norm_pre_clip': 1.7475141286849976, 'timing/data': 0.00039463210850954056, 'timing/model': 0.5272758649662137, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.2})
+[2026-07-28 10:47:46,857][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.07082225382328033, 'train/grad_norm_pre_clip': 5.7296295166015625, 'timing/data': 0.0003909114748239517, 'timing/model': 0.5302779851481318, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.2})
+[2026-07-28 10:47:54,039][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.09786217659711838, 'train/grad_norm_pre_clip': 1.910325288772583, 'timing/data': 0.0004032626748085022, 'timing/model': 0.5285864397883415, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.2})
+[2026-07-28 10:48:01,215][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.05943118408322334, 'train/grad_norm_pre_clip': 1.6183704137802124, 'timing/data': 0.00033943355083465576, 'timing/model': 0.5259652491658926, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.2})
+[2026-07-28 10:48:08,396][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.056377753615379333, 'train/grad_norm_pre_clip': 0.864525318145752, 'timing/data': 0.00036733318120241165, 'timing/model': 0.5281498599797487, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.2})
+[2026-07-28 10:48:15,585][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.035977430641651154, 'train/grad_norm_pre_clip': 1.3384283781051636, 'timing/data': 0.0003731055185198784, 'timing/model': 0.5263643898069859, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.2})
+[2026-07-28 10:48:22,760][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.06586810201406479, 'train/grad_norm_pre_clip': 3.983400583267212, 'timing/data': 0.0003233710303902626, 'timing/model': 0.5252820178866386, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.2})
+[2026-07-28 10:48:29,934][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.05545765161514282, 'train/grad_norm_pre_clip': 1.69304621219635, 'timing/data': 0.00031808577477931976, 'timing/model': 0.5276880320161581, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.2})
+[2026-07-28 10:48:37,102][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.04642855376005173, 'train/grad_norm_pre_clip': 1.1720324754714966, 'timing/data': 0.00035140570253133774, 'timing/model': 0.5248355381190777, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.2})
+[2026-07-28 10:48:44,288][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.037499479949474335, 'train/grad_norm_pre_clip': 0.8875709176063538, 'timing/data': 0.000359436497092247, 'timing/model': 0.528417737223208, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.2})
+[2026-07-28 10:48:51,465][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.10083846747875214, 'train/grad_norm_pre_clip': 2.672125816345215, 'timing/data': 0.0003806706517934799, 'timing/model': 0.5275549683719873, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.2})
+[2026-07-28 10:48:58,647][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.08198749274015427, 'train/grad_norm_pre_clip': 2.1746890544891357, 'timing/data': 0.00040400493890047073, 'timing/model': 0.529558090493083, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.2})
+[2026-07-28 10:49:05,820][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.05108311027288437, 'train/grad_norm_pre_clip': 1.6517680883407593, 'timing/data': 0.0003325389698147774, 'timing/model': 0.5262372894212604, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.2})
+[2026-07-28 10:49:12,997][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.03558601438999176, 'train/grad_norm_pre_clip': 1.2930152416229248, 'timing/data': 0.00032569747418165207, 'timing/model': 0.528832720592618, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.2})
+[2026-07-28 10:49:20,169][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.04210131615400314, 'train/grad_norm_pre_clip': 1.3584517240524292, 'timing/data': 0.00033460743725299835, 'timing/model': 0.5260814344510436, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.2})
+[2026-07-28 10:49:27,339][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.043029360473155975, 'train/grad_norm_pre_clip': 0.7897263765335083, 'timing/data': 0.0006558140739798546, 'timing/model': 0.5274992072954774, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.2})
+[2026-07-28 10:49:34,523][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.04697169363498688, 'train/grad_norm_pre_clip': 1.5003979206085205, 'timing/data': 0.0003268551081418991, 'timing/model': 0.5262242155149579, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.2})
+[2026-07-28 10:49:41,697][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.05187630653381348, 'train/grad_norm_pre_clip': 1.2399846315383911, 'timing/data': 0.0003716368228197098, 'timing/model': 0.5272255837917328, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.2})
+[2026-07-28 10:49:48,873][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.05560354143381119, 'train/grad_norm_pre_clip': 1.1488723754882812, 'timing/data': 0.0003203675150871277, 'timing/model': 0.5268352050334215, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.2})
+[2026-07-28 10:49:56,043][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.025519952178001404, 'train/grad_norm_pre_clip': 0.8195155262947083, 'timing/data': 0.0003554457798600197, 'timing/model': 0.5282635111361742, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.2})
+[2026-07-28 10:50:03,223][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.06728550046682358, 'train/grad_norm_pre_clip': 2.0752058029174805, 'timing/data': 0.0003609396517276764, 'timing/model': 0.5263150716200471, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.2})
+[2026-07-28 10:50:10,407][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.11405102908611298, 'train/grad_norm_pre_clip': 2.2498483657836914, 'timing/data': 0.0004101702943444252, 'timing/model': 0.5272046867758036, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.2})
+[2026-07-28 10:50:17,578][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.0912259966135025, 'train/grad_norm_pre_clip': 3.114164352416992, 'timing/data': 0.0003408687189221382, 'timing/model': 0.5255013387650251, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.2})
+[2026-07-28 10:50:24,757][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.09824614226818085, 'train/grad_norm_pre_clip': 2.0027496814727783, 'timing/data': 0.0003726622089743614, 'timing/model': 0.5273473681882024, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.2})
+[2026-07-28 10:50:31,923][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.07699239253997803, 'train/grad_norm_pre_clip': 1.1208940744400024, 'timing/data': 0.00035371165722608566, 'timing/model': 0.5268566943705082, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.2})
+[2026-07-28 10:50:39,095][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.07887870818376541, 'train/grad_norm_pre_clip': 2.026392936706543, 'timing/data': 0.00035170838236808777, 'timing/model': 0.5252622421830893, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.2})
+[2026-07-28 10:50:46,275][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.0611770898103714, 'train/grad_norm_pre_clip': 1.2031954526901245, 'timing/data': 0.0003455597907304764, 'timing/model': 0.5278198877349496, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.2})
+[2026-07-28 10:50:53,454][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.07773838192224503, 'train/grad_norm_pre_clip': 1.2615171670913696, 'timing/data': 0.00032751914113759995, 'timing/model': 0.5291556231677532, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.2})
+[2026-07-28 10:51:00,626][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.05113277584314346, 'train/grad_norm_pre_clip': 2.0339951515197754, 'timing/data': 0.00042189285159111023, 'timing/model': 0.5291525069624186, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.2})
+[2026-07-28 10:51:07,807][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.0759134441614151, 'train/grad_norm_pre_clip': 0.9329752326011658, 'timing/data': 0.0004157964140176773, 'timing/model': 0.527009311132133, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.2})
+[2026-07-28 10:51:14,998][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.06256906688213348, 'train/grad_norm_pre_clip': 1.4293537139892578, 'timing/data': 0.0003458559513092041, 'timing/model': 0.5267759645357728, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.2})
+[2026-07-28 10:51:22,178][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.038591716438531876, 'train/grad_norm_pre_clip': 0.7557775378227234, 'timing/data': 0.0003376742824912071, 'timing/model': 0.5278795249760151, 'learning_rate/qwen_vl_interface': 9.999998458888285e-06, 'learning_rate/action_model': 9.999998458888285e-05, 'epoch': 0.2})
+[2026-07-28 10:51:29,355][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.06664198637008667, 'train/grad_norm_pre_clip': 1.304031252861023, 'timing/data': 0.00032573845237493515, 'timing/model': 0.5259382342919707, 'learning_rate/qwen_vl_interface': 9.999993835554141e-06, 'learning_rate/action_model': 9.999993835554141e-05, 'epoch': 0.2})
+[2026-07-28 10:51:36,547][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.07902297377586365, 'train/grad_norm_pre_clip': 1.903378963470459, 'timing/data': 0.00047951284795999527, 'timing/model': 0.5295287258923054, 'learning_rate/qwen_vl_interface': 9.999986130000566e-06, 'learning_rate/action_model': 9.999986130000566e-05, 'epoch': 0.2})
+[2026-07-28 10:51:43,721][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.02841619960963726, 'train/grad_norm_pre_clip': 0.8503535985946655, 'timing/data': 0.00036215223371982574, 'timing/model': 0.5286519536748528, 'learning_rate/qwen_vl_interface': 9.99997534223256e-06, 'learning_rate/action_model': 9.99997534223256e-05, 'epoch': 0.2})
+[2026-07-28 10:51:51,468][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.06058400869369507, 'train/grad_norm_pre_clip': 1.777661681175232, 'timing/data': 0.00034306198358535767, 'timing/model': 0.8081573750823736, 'learning_rate/qwen_vl_interface': 9.999961472257127e-06, 'learning_rate/action_model': 9.999961472257127e-05, 'epoch': 0.2})
+[2026-07-28 10:51:58,637][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.050716057419776917, 'train/grad_norm_pre_clip': 1.5565682649612427, 'timing/data': 0.000364670529961586, 'timing/model': 0.5287933545187116, 'learning_rate/qwen_vl_interface': 9.999944520083262e-06, 'learning_rate/action_model': 9.999944520083261e-05, 'epoch': 0.2})
+[2026-07-28 10:52:05,821][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.061338961124420166, 'train/grad_norm_pre_clip': 2.876594305038452, 'timing/data': 0.00036853086203336716, 'timing/model': 0.5260000359266996, 'learning_rate/qwen_vl_interface': 9.999924485721968e-06, 'learning_rate/action_model': 9.999924485721968e-05, 'epoch': 0.2})
+[2026-07-28 10:52:12,997][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.08267034590244293, 'train/grad_norm_pre_clip': 2.743828058242798, 'timing/data': 0.00038441363722085953, 'timing/model': 0.5308338003233075, 'learning_rate/qwen_vl_interface': 9.999901369186244e-06, 'learning_rate/action_model': 9.999901369186244e-05, 'epoch': 0.2})
+[2026-07-28 10:52:20,169][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.042772769927978516, 'train/grad_norm_pre_clip': 2.1909799575805664, 'timing/data': 0.00037743523716926575, 'timing/model': 0.5275738853961229, 'learning_rate/qwen_vl_interface': 9.999875170491092e-06, 'learning_rate/action_model': 9.999875170491091e-05, 'epoch': 0.2})
+[2026-07-28 10:52:27,349][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.060146696865558624, 'train/grad_norm_pre_clip': 1.870550274848938, 'timing/data': 0.00033280160278081894, 'timing/model': 0.527047854848206, 'learning_rate/qwen_vl_interface': 9.999845889653509e-06, 'learning_rate/action_model': 9.999845889653508e-05, 'epoch': 0.2})
+[2026-07-28 10:52:34,539][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.10014665126800537, 'train/grad_norm_pre_clip': 4.338907718658447, 'timing/data': 0.0004862360656261444, 'timing/model': 0.5282137338072062, 'learning_rate/qwen_vl_interface': 9.999813526692496e-06, 'learning_rate/action_model': 9.999813526692495e-05, 'epoch': 0.2})
+[2026-07-28 10:52:41,718][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.012327921576797962, 'train/grad_norm_pre_clip': 0.512012779712677, 'timing/data': 0.0007985914126038551, 'timing/model': 0.5262497905641794, 'learning_rate/qwen_vl_interface': 9.999778081629055e-06, 'learning_rate/action_model': 9.999778081629054e-05, 'epoch': 0.2})
+[2026-07-28 10:52:48,892][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.04013475030660629, 'train/grad_norm_pre_clip': 1.1149344444274902, 'timing/data': 0.00037221238017082214, 'timing/model': 0.5253144055604935, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.2})
+[2026-07-28 10:52:56,098][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.016044851392507553, 'train/grad_norm_pre_clip': 0.9622796773910522, 'timing/data': 0.0003805123269557953, 'timing/model': 0.5259137004613876, 'learning_rate/qwen_vl_interface': 9.999697945288881e-06, 'learning_rate/action_model': 9.999697945288881e-05, 'epoch': 0.2})
+[2026-07-28 10:53:03,289][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.05257114768028259, 'train/grad_norm_pre_clip': 1.3288689851760864, 'timing/data': 0.0003408370539546013, 'timing/model': 0.5260132104158401, 'learning_rate/qwen_vl_interface': 9.999653254064153e-06, 'learning_rate/action_model': 9.999653254064152e-05, 'epoch': 0.2})
+[2026-07-28 10:53:10,485][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.0569722019135952, 'train/grad_norm_pre_clip': 1.888231873512268, 'timing/data': 0.0003535225987434387, 'timing/model': 0.527469128370285, 'learning_rate/qwen_vl_interface': 9.99960548084099e-06, 'learning_rate/action_model': 9.99960548084099e-05, 'epoch': 0.2})
+[2026-07-28 10:53:17,680][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.07933194190263748, 'train/grad_norm_pre_clip': 3.182786226272583, 'timing/data': 0.0003703925758600235, 'timing/model': 0.5251245507970452, 'learning_rate/qwen_vl_interface': 9.999554625650398e-06, 'learning_rate/action_model': 9.999554625650399e-05, 'epoch': 0.21})
+[2026-07-28 10:53:24,861][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.07595109939575195, 'train/grad_norm_pre_clip': 2.112917184829712, 'timing/data': 0.0003294972702860832, 'timing/model': 0.5268248030915856, 'learning_rate/qwen_vl_interface': 9.999500688525376e-06, 'learning_rate/action_model': 9.999500688525375e-05, 'epoch': 0.21})
+[2026-07-28 10:53:32,033][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.11873164027929306, 'train/grad_norm_pre_clip': 2.559349298477173, 'timing/data': 0.0004192488268017769, 'timing/model': 0.529260735027492, 'learning_rate/qwen_vl_interface': 9.99944366950092e-06, 'learning_rate/action_model': 9.999443669500919e-05, 'epoch': 0.21})
+[2026-07-28 10:53:39,205][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.06421713531017303, 'train/grad_norm_pre_clip': 1.852493166923523, 'timing/data': 0.00042793434113264084, 'timing/model': 0.5254583321511745, 'learning_rate/qwen_vl_interface': 9.99938356861403e-06, 'learning_rate/action_model': 9.99938356861403e-05, 'epoch': 0.21})
+[2026-07-28 10:53:46,389][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.06649955362081528, 'train/grad_norm_pre_clip': 1.5487037897109985, 'timing/data': 0.0003370186313986778, 'timing/model': 0.5304002920165658, 'learning_rate/qwen_vl_interface': 9.999320385903709e-06, 'learning_rate/action_model': 9.999320385903708e-05, 'epoch': 0.21})
+[2026-07-28 10:53:53,578][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.03408759459853172, 'train/grad_norm_pre_clip': 1.1707412004470825, 'timing/data': 0.00036135222762823105, 'timing/model': 0.528031182475388, 'learning_rate/qwen_vl_interface': 9.999254121410948e-06, 'learning_rate/action_model': 9.999254121410949e-05, 'epoch': 0.21})
+[2026-07-28 10:54:00,764][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.05820842087268829, 'train/grad_norm_pre_clip': 1.5891326665878296, 'timing/data': 0.0003753434866666794, 'timing/model': 0.5278178630396724, 'learning_rate/qwen_vl_interface': 9.999184775178751e-06, 'learning_rate/action_model': 9.999184775178751e-05, 'epoch': 0.21})
+[2026-07-28 10:54:07,950][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.09344640374183655, 'train/grad_norm_pre_clip': 4.042906284332275, 'timing/data': 0.0003310777246952057, 'timing/model': 0.5302844755351543, 'learning_rate/qwen_vl_interface': 9.999112347252117e-06, 'learning_rate/action_model': 9.999112347252116e-05, 'epoch': 0.21})
+[2026-07-28 10:54:15,123][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.0695011094212532, 'train/grad_norm_pre_clip': 3.8312008380889893, 'timing/data': 0.0003390330821275711, 'timing/model': 0.5264126481488347, 'learning_rate/qwen_vl_interface': 9.99903683767804e-06, 'learning_rate/action_model': 9.99903683767804e-05, 'epoch': 0.21})
+[2026-07-28 10:54:22,293][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.036164943128824234, 'train/grad_norm_pre_clip': 1.6312212944030762, 'timing/data': 0.00036236178129911423, 'timing/model': 0.5269518243148923, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.21})
+[2026-07-28 10:54:29,468][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.03843850642442703, 'train/grad_norm_pre_clip': 0.817905843257904, 'timing/data': 0.0003335867077112198, 'timing/model': 0.526402928866446, 'learning_rate/qwen_vl_interface': 9.998876573785547e-06, 'learning_rate/action_model': 9.998876573785547e-05, 'epoch': 0.21})
+[2026-07-28 10:54:36,647][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.10652541369199753, 'train/grad_norm_pre_clip': 3.5074126720428467, 'timing/data': 0.000354706309735775, 'timing/model': 0.5278675314038992, 'learning_rate/qwen_vl_interface': 9.998791819571126e-06, 'learning_rate/action_model': 9.998791819571127e-05, 'epoch': 0.21})
+[2026-07-28 10:54:43,821][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.09447043389081955, 'train/grad_norm_pre_clip': 1.992786169052124, 'timing/data': 0.00033049751073122025, 'timing/model': 0.5261875251308084, 'learning_rate/qwen_vl_interface': 9.998703983917252e-06, 'learning_rate/action_model': 9.998703983917252e-05, 'epoch': 0.21})
+[2026-07-28 10:54:50,985][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.06910435110330582, 'train/grad_norm_pre_clip': 1.7791742086410522, 'timing/data': 0.0004008635878562927, 'timing/model': 0.5249883309006691, 'learning_rate/qwen_vl_interface': 9.998613066880919e-06, 'learning_rate/action_model': 9.998613066880918e-05, 'epoch': 0.21})
+[2026-07-28 10:54:58,163][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.04152858257293701, 'train/grad_norm_pre_clip': 1.3426434993743896, 'timing/data': 0.0003909291699528694, 'timing/model': 0.5261861467733979, 'learning_rate/qwen_vl_interface': 9.99851906852112e-06, 'learning_rate/action_model': 9.998519068521119e-05, 'epoch': 0.21})
+[2026-07-28 10:55:05,338][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.10183703154325485, 'train/grad_norm_pre_clip': 1.9386634826660156, 'timing/data': 0.0005042999982833862, 'timing/model': 0.5267175221815705, 'learning_rate/qwen_vl_interface': 9.998421988898851e-06, 'learning_rate/action_model': 9.998421988898851e-05, 'epoch': 0.21})
+[2026-07-28 10:55:12,522][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.04163810983300209, 'train/grad_norm_pre_clip': 1.1194767951965332, 'timing/data': 0.0005124760791659355, 'timing/model': 0.5267559206113219, 'learning_rate/qwen_vl_interface': 9.998321828077107e-06, 'learning_rate/action_model': 9.998321828077106e-05, 'epoch': 0.21})
+[2026-07-28 10:55:19,722][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.04659191146492958, 'train/grad_norm_pre_clip': 1.3295338153839111, 'timing/data': 0.00037351157516241074, 'timing/model': 0.5285637192428112, 'learning_rate/qwen_vl_interface': 9.99821858612088e-06, 'learning_rate/action_model': 9.99821858612088e-05, 'epoch': 0.21})
+[2026-07-28 10:55:26,945][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.04987723007798195, 'train/grad_norm_pre_clip': 0.8603854179382324, 'timing/data': 0.00037275347858667374, 'timing/model': 0.5265997974202037, 'learning_rate/qwen_vl_interface': 9.998112263097164e-06, 'learning_rate/action_model': 9.998112263097165e-05, 'epoch': 0.21})
+[2026-07-28 10:55:34,174][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.0437149740755558, 'train/grad_norm_pre_clip': 2.295966863632202, 'timing/data': 0.00039303209632635117, 'timing/model': 0.5282551003620028, 'learning_rate/qwen_vl_interface': 9.99800285907495e-06, 'learning_rate/action_model': 9.99800285907495e-05, 'epoch': 0.21})
+[2026-07-28 10:55:41,383][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.06153643876314163, 'train/grad_norm_pre_clip': 2.7620840072631836, 'timing/data': 0.00038584135472774506, 'timing/model': 0.5268456386402249, 'learning_rate/qwen_vl_interface': 9.997890374125228e-06, 'learning_rate/action_model': 9.997890374125227e-05, 'epoch': 0.21})
+[2026-07-28 10:55:48,587][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.12410172820091248, 'train/grad_norm_pre_clip': 2.9108471870422363, 'timing/data': 0.00032903533428907394, 'timing/model': 0.528018762357533, 'learning_rate/qwen_vl_interface': 9.997774808320992e-06, 'learning_rate/action_model': 9.997774808320992e-05, 'epoch': 0.21})
+[2026-07-28 10:55:55,799][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.04583931714296341, 'train/grad_norm_pre_clip': 2.7959837913513184, 'timing/data': 0.0008507156744599342, 'timing/model': 0.526796574704349, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.21})
+[2026-07-28 10:56:03,034][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.052939724177122116, 'train/grad_norm_pre_clip': 1.2772763967514038, 'timing/data': 0.00035124924033880234, 'timing/model': 0.5334390765056014, 'learning_rate/qwen_vl_interface': 9.997534434450922e-06, 'learning_rate/action_model': 9.997534434450921e-05, 'epoch': 0.21})
+[2026-07-28 10:56:10,332][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.06280916929244995, 'train/grad_norm_pre_clip': 2.2658369541168213, 'timing/data': 0.00040265824645757675, 'timing/model': 0.5408824644982815, 'learning_rate/qwen_vl_interface': 9.997409626541065e-06, 'learning_rate/action_model': 9.997409626541066e-05, 'epoch': 0.21})
+[2026-07-28 10:56:17,678][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.06567156314849854, 'train/grad_norm_pre_clip': 2.745492696762085, 'timing/data': 0.0003496883437037468, 'timing/model': 0.5258394060656428, 'learning_rate/qwen_vl_interface': 9.997281738088647e-06, 'learning_rate/action_model': 9.997281738088646e-05, 'epoch': 0.21})
+[2026-07-28 10:56:24,942][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.03376566618680954, 'train/grad_norm_pre_clip': 1.5447248220443726, 'timing/data': 0.0004976037889719009, 'timing/model': 0.526710988022387, 'learning_rate/qwen_vl_interface': 9.997150769176648e-06, 'learning_rate/action_model': 9.997150769176647e-05, 'epoch': 0.21})
+[2026-07-28 10:56:32,453][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.04703926295042038, 'train/grad_norm_pre_clip': 0.9188175797462463, 'timing/data': 0.0003964584320783615, 'timing/model': 0.5265805134549737, 'learning_rate/qwen_vl_interface': 9.997016719890052e-06, 'learning_rate/action_model': 9.997016719890052e-05, 'epoch': 0.21})
+[2026-07-28 10:56:39,657][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.05022268742322922, 'train/grad_norm_pre_clip': 1.9878535270690918, 'timing/data': 0.0003447262570261955, 'timing/model': 0.5273903403431177, 'learning_rate/qwen_vl_interface': 9.996879590315847e-06, 'learning_rate/action_model': 9.996879590315846e-05, 'epoch': 0.21})
+[2026-07-28 10:56:46,847][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.07671995460987091, 'train/grad_norm_pre_clip': 1.7017935514450073, 'timing/data': 0.0003093332052230835, 'timing/model': 0.5263765687122941, 'learning_rate/qwen_vl_interface': 9.99673938054301e-06, 'learning_rate/action_model': 9.996739380543009e-05, 'epoch': 0.21})
+[2026-07-28 10:56:54,027][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.08651237189769745, 'train/grad_norm_pre_clip': 1.1943371295928955, 'timing/data': 0.00039180275052785873, 'timing/model': 0.525935965590179, 'learning_rate/qwen_vl_interface': 9.996596090662525e-06, 'learning_rate/action_model': 9.996596090662525e-05, 'epoch': 0.21})
+[2026-07-28 10:57:01,217][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.03905528783798218, 'train/grad_norm_pre_clip': 1.8315788507461548, 'timing/data': 0.00041907932609319687, 'timing/model': 0.5281499130651355, 'learning_rate/qwen_vl_interface': 9.996449720767369e-06, 'learning_rate/action_model': 9.996449720767368e-05, 'epoch': 0.21})
+[2026-07-28 10:57:08,425][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.06683064997196198, 'train/grad_norm_pre_clip': 1.2270772457122803, 'timing/data': 0.00035088881850242615, 'timing/model': 0.5250917505472898, 'learning_rate/qwen_vl_interface': 9.99630027095252e-06, 'learning_rate/action_model': 9.99630027095252e-05, 'epoch': 0.21})
+[2026-07-28 10:57:15,602][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.05221574753522873, 'train/grad_norm_pre_clip': 0.9135621786117554, 'timing/data': 0.00033092964440584183, 'timing/model': 0.5251837261021137, 'learning_rate/qwen_vl_interface': 9.996147741314956e-06, 'learning_rate/action_model': 9.996147741314957e-05, 'epoch': 0.21})
+[2026-07-28 10:57:22,812][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.06311450898647308, 'train/grad_norm_pre_clip': 2.225593090057373, 'timing/data': 0.0003184545785188675, 'timing/model': 0.5265471003949642, 'learning_rate/qwen_vl_interface': 9.99599213195365e-06, 'learning_rate/action_model': 9.99599213195365e-05, 'epoch': 0.21})
+[2026-07-28 10:57:29,999][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.06315743923187256, 'train/grad_norm_pre_clip': 2.101675510406494, 'timing/data': 0.0003950074315071106, 'timing/model': 0.5261110048741102, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.21})
+[2026-07-28 10:57:37,196][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.06441117823123932, 'train/grad_norm_pre_clip': 1.2702080011367798, 'timing/data': 0.0003823228180408478, 'timing/model': 0.5297221960499883, 'learning_rate/qwen_vl_interface': 9.995671674465709e-06, 'learning_rate/action_model': 9.995671674465709e-05, 'epoch': 0.21})
+[2026-07-28 10:57:44,386][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.049152180552482605, 'train/grad_norm_pre_clip': 1.6210628747940063, 'timing/data': 0.0003887908533215523, 'timing/model': 0.5291207628324628, 'learning_rate/qwen_vl_interface': 9.99550682654701e-06, 'learning_rate/action_model': 9.995506826547011e-05, 'epoch': 0.21})
+[2026-07-28 10:57:51,569][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.10012486577033997, 'train/grad_norm_pre_clip': 2.3762025833129883, 'timing/data': 0.0003890162333846092, 'timing/model': 0.5300924507901073, 'learning_rate/qwen_vl_interface': 9.995338899320457e-06, 'learning_rate/action_model': 9.995338899320456e-05, 'epoch': 0.21})
+[2026-07-28 10:57:58,754][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.07988140732049942, 'train/grad_norm_pre_clip': 1.306910753250122, 'timing/data': 0.00035067368298768997, 'timing/model': 0.5262832520529628, 'learning_rate/qwen_vl_interface': 9.995167892895011e-06, 'learning_rate/action_model': 9.99516789289501e-05, 'epoch': 0.21})
+[2026-07-28 10:58:05,968][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.06361114978790283, 'train/grad_norm_pre_clip': 2.2617921829223633, 'timing/data': 0.00036541279405355453, 'timing/model': 0.5270196618512273, 'learning_rate/qwen_vl_interface': 9.994993807381635e-06, 'learning_rate/action_model': 9.994993807381635e-05, 'epoch': 0.21})
+[2026-07-28 10:58:13,169][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.06925880908966064, 'train/grad_norm_pre_clip': 1.2598766088485718, 'timing/data': 0.0005626808851957321, 'timing/model': 0.5266760075464845, 'learning_rate/qwen_vl_interface': 9.994816642893294e-06, 'learning_rate/action_model': 9.994816642893293e-05, 'epoch': 0.21})
+[2026-07-28 10:58:20,385][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.032848406583070755, 'train/grad_norm_pre_clip': 1.7869226932525635, 'timing/data': 0.00032367371022701263, 'timing/model': 0.5267468811944127, 'learning_rate/qwen_vl_interface': 9.994636399544948e-06, 'learning_rate/action_model': 9.994636399544947e-05, 'epoch': 0.21})
+[2026-07-28 10:58:27,587][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.08592747896909714, 'train/grad_norm_pre_clip': 1.721614956855774, 'timing/data': 0.0003969995304942131, 'timing/model': 0.5284194415435195, 'learning_rate/qwen_vl_interface': 9.994453077453552e-06, 'learning_rate/action_model': 9.994453077453552e-05, 'epoch': 0.21})
+[2026-07-28 10:58:34,805][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.052730560302734375, 'train/grad_norm_pre_clip': 1.5285974740982056, 'timing/data': 0.00037061329931020737, 'timing/model': 0.5351045457646251, 'learning_rate/qwen_vl_interface': 9.994266676738065e-06, 'learning_rate/action_model': 9.994266676738065e-05, 'epoch': 0.21})
+[2026-07-28 10:58:42,004][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.06702833622694016, 'train/grad_norm_pre_clip': 1.6616075038909912, 'timing/data': 0.0004232088103890419, 'timing/model': 0.5267402166500688, 'learning_rate/qwen_vl_interface': 9.99407719751944e-06, 'learning_rate/action_model': 9.994077197519439e-05, 'epoch': 0.21})
+[2026-07-28 10:58:49,197][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.046401314437389374, 'train/grad_norm_pre_clip': 2.057299852371216, 'timing/data': 0.00034318119287490845, 'timing/model': 0.5293075256049633, 'learning_rate/qwen_vl_interface': 9.993884639920626e-06, 'learning_rate/action_model': 9.993884639920626e-05, 'epoch': 0.21})
+[2026-07-28 10:58:56,389][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.06820157170295715, 'train/grad_norm_pre_clip': 2.564399003982544, 'timing/data': 0.0003165053203701973, 'timing/model': 0.5253069987520576, 'learning_rate/qwen_vl_interface': 9.993689004066573e-06, 'learning_rate/action_model': 9.993689004066573e-05, 'epoch': 0.21})
+[2026-07-28 10:59:03,580][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.034377168864011765, 'train/grad_norm_pre_clip': 1.3892781734466553, 'timing/data': 0.0003775758668780327, 'timing/model': 0.5265653785318136, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.21})
+[2026-07-28 10:59:10,775][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.03436034917831421, 'train/grad_norm_pre_clip': 0.9088868498802185, 'timing/data': 0.0007698843255639076, 'timing/model': 0.5279112681746483, 'learning_rate/qwen_vl_interface': 9.993288498102531e-06, 'learning_rate/action_model': 9.993288498102532e-05, 'epoch': 0.21})
+[2026-07-28 10:59:17,985][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.07153604924678802, 'train/grad_norm_pre_clip': 1.5182857513427734, 'timing/data': 0.00032353494316339493, 'timing/model': 0.5283692739903927, 'learning_rate/qwen_vl_interface': 9.993083628252424e-06, 'learning_rate/action_model': 9.993083628252424e-05, 'epoch': 0.21})
+[2026-07-28 10:59:25,185][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.08248388022184372, 'train/grad_norm_pre_clip': 3.2391083240509033, 'timing/data': 0.00040082912892103195, 'timing/model': 0.5278628980740905, 'learning_rate/qwen_vl_interface': 9.992875680666849e-06, 'learning_rate/action_model': 9.992875680666849e-05, 'epoch': 0.21})
+[2026-07-28 10:59:32,392][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.07701623439788818, 'train/grad_norm_pre_clip': 3.10793137550354, 'timing/data': 0.0003877626731991768, 'timing/model': 0.526468294672668, 'learning_rate/qwen_vl_interface': 9.992664655480736e-06, 'learning_rate/action_model': 9.992664655480735e-05, 'epoch': 0.21})
+[2026-07-28 10:59:39,581][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.10735058784484863, 'train/grad_norm_pre_clip': 3.3972203731536865, 'timing/data': 0.0003702538087964058, 'timing/model': 0.5270460350438952, 'learning_rate/qwen_vl_interface': 9.992450552831018e-06, 'learning_rate/action_model': 9.992450552831017e-05, 'epoch': 0.21})
+[2026-07-28 10:59:46,790][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.05733806639909744, 'train/grad_norm_pre_clip': 2.9128000736236572, 'timing/data': 0.0003345990553498268, 'timing/model': 0.5270932717248797, 'learning_rate/qwen_vl_interface': 9.992233372856623e-06, 'learning_rate/action_model': 9.992233372856624e-05, 'epoch': 0.22})
+[2026-07-28 10:59:54,045][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.09530027955770493, 'train/grad_norm_pre_clip': 1.987004041671753, 'timing/data': 0.0003635454922914505, 'timing/model': 0.5349571006372571, 'learning_rate/qwen_vl_interface': 9.99201311569848e-06, 'learning_rate/action_model': 9.99201311569848e-05, 'epoch': 0.22})
+[2026-07-28 11:00:01,274][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.061862677335739136, 'train/grad_norm_pre_clip': 1.2693411111831665, 'timing/data': 0.00040008407086133957, 'timing/model': 0.5276325615122914, 'learning_rate/qwen_vl_interface': 9.99178978149951e-06, 'learning_rate/action_model': 9.99178978149951e-05, 'epoch': 0.22})
+[2026-07-28 11:00:08,518][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.03032534010708332, 'train/grad_norm_pre_clip': 1.6044262647628784, 'timing/data': 0.00033758673816919327, 'timing/model': 0.5272753108292818, 'learning_rate/qwen_vl_interface': 9.99156337040463e-06, 'learning_rate/action_model': 9.991563370404629e-05, 'epoch': 0.22})
+[2026-07-28 11:00:15,700][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.04110664874315262, 'train/grad_norm_pre_clip': 1.1708837747573853, 'timing/data': 0.0004068790003657341, 'timing/model': 0.5270859021693468, 'learning_rate/qwen_vl_interface': 9.991333882560759e-06, 'learning_rate/action_model': 9.991333882560759e-05, 'epoch': 0.22})
+[2026-07-28 11:00:22,867][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.12427042424678802, 'train/grad_norm_pre_clip': 2.4493634700775146, 'timing/data': 0.0003835698589682579, 'timing/model': 0.5265380972996354, 'learning_rate/qwen_vl_interface': 9.991101318116806e-06, 'learning_rate/action_model': 9.991101318116806e-05, 'epoch': 0.22})
+[2026-07-28 11:00:30,030][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.11557761579751968, 'train/grad_norm_pre_clip': 1.6687737703323364, 'timing/data': 0.00041341502219438553, 'timing/model': 0.5266543598845601, 'learning_rate/qwen_vl_interface': 9.990865677223684e-06, 'learning_rate/action_model': 9.990865677223684e-05, 'epoch': 0.22})
+[2026-07-28 11:00:37,206][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.08321599662303925, 'train/grad_norm_pre_clip': 1.9353984594345093, 'timing/data': 0.0003467826172709465, 'timing/model': 0.5274901939556003, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.22})
+[2026-07-28 11:00:44,388][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.042328909039497375, 'train/grad_norm_pre_clip': 1.378039002418518, 'timing/data': 0.0006431238725781441, 'timing/model': 0.5299234418198466, 'learning_rate/qwen_vl_interface': 9.990385166703534e-06, 'learning_rate/action_model': 9.990385166703533e-05, 'epoch': 0.22})
+[2026-07-28 11:00:51,583][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.049765944480895996, 'train/grad_norm_pre_clip': 1.6284152269363403, 'timing/data': 0.0003434019163250923, 'timing/model': 0.5364944590255618, 'learning_rate/qwen_vl_interface': 9.990140297388306e-06, 'learning_rate/action_model': 9.990140297388307e-05, 'epoch': 0.22})
+[2026-07-28 11:00:58,784][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.0698174238204956, 'train/grad_norm_pre_clip': 1.932742714881897, 'timing/data': 0.0003215894103050232, 'timing/model': 0.5397738562896848, 'learning_rate/qwen_vl_interface': 9.989892352247505e-06, 'learning_rate/action_model': 9.989892352247504e-05, 'epoch': 0.22})
+[2026-07-28 11:01:06,280][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.05659704655408859, 'train/grad_norm_pre_clip': 1.0470247268676758, 'timing/data': 0.0003873175010085106, 'timing/model': 0.5360680520534515, 'learning_rate/qwen_vl_interface': 9.989641331442014e-06, 'learning_rate/action_model': 9.989641331442013e-05, 'epoch': 0.22})
+[2026-07-28 11:01:13,469][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.1018618792295456, 'train/grad_norm_pre_clip': 2.1822104454040527, 'timing/data': 0.0003452785313129425, 'timing/model': 0.5394491096958518, 'learning_rate/qwen_vl_interface': 9.989387235134722e-06, 'learning_rate/action_model': 9.989387235134722e-05, 'epoch': 0.22})
+[2026-07-28 11:01:20,958][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.06642498075962067, 'train/grad_norm_pre_clip': 3.3055813312530518, 'timing/data': 0.0003295373171567917, 'timing/model': 0.5325239794328809, 'learning_rate/qwen_vl_interface': 9.98913006349051e-06, 'learning_rate/action_model': 9.989130063490508e-05, 'epoch': 0.22})
+[2026-07-28 11:01:28,152][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.08037722110748291, 'train/grad_norm_pre_clip': 1.8363896608352661, 'timing/data': 0.000365525484085083, 'timing/model': 0.5339854471385479, 'learning_rate/qwen_vl_interface': 9.988869816676247e-06, 'learning_rate/action_model': 9.988869816676246e-05, 'epoch': 0.22})
+[2026-07-28 11:01:35,398][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.03401154652237892, 'train/grad_norm_pre_clip': 1.5309892892837524, 'timing/data': 0.00033977627754211426, 'timing/model': 0.5378253916278481, 'learning_rate/qwen_vl_interface': 9.988606494860811e-06, 'learning_rate/action_model': 9.988606494860812e-05, 'epoch': 0.22})
+[2026-07-28 11:01:42,589][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.09237498790025711, 'train/grad_norm_pre_clip': 2.078354597091675, 'timing/data': 0.00034422799944877625, 'timing/model': 0.5323526263237, 'learning_rate/qwen_vl_interface': 9.98834009821507e-06, 'learning_rate/action_model': 9.988340098215069e-05, 'epoch': 0.22})
+[2026-07-28 11:01:49,820][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.061518967151641846, 'train/grad_norm_pre_clip': 1.848406195640564, 'timing/data': 0.0005622301250696182, 'timing/model': 0.5355294784530997, 'learning_rate/qwen_vl_interface': 9.98807062691188e-06, 'learning_rate/action_model': 9.988070626911879e-05, 'epoch': 0.22})
+[2026-07-28 11:01:57,016][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.08169561624526978, 'train/grad_norm_pre_clip': 1.9687526226043701, 'timing/data': 0.0005158446729183197, 'timing/model': 0.5339710703119636, 'learning_rate/qwen_vl_interface': 9.987798081126099e-06, 'learning_rate/action_model': 9.987798081126099e-05, 'epoch': 0.22})
+[2026-07-28 11:02:04,229][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.04457172006368637, 'train/grad_norm_pre_clip': 1.4568005800247192, 'timing/data': 0.00042997486889362335, 'timing/model': 0.5350277628749609, 'learning_rate/qwen_vl_interface': 9.987522461034584e-06, 'learning_rate/action_model': 9.987522461034584e-05, 'epoch': 0.22})
+[2026-07-28 11:02:11,467][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.043174926191568375, 'train/grad_norm_pre_clip': 0.7293850183486938, 'timing/data': 0.0004198867827653885, 'timing/model': 0.5334658976644278, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.22})
+[2026-07-28 11:02:18,669][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.058286577463150024, 'train/grad_norm_pre_clip': 1.3526499271392822, 'timing/data': 0.0004556169733405113, 'timing/model': 0.5337613439187407, 'learning_rate/qwen_vl_interface': 9.98696199865172e-06, 'learning_rate/action_model': 9.98696199865172e-05, 'epoch': 0.22})
+[2026-07-28 11:02:25,871][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.10069508850574493, 'train/grad_norm_pre_clip': 1.9181841611862183, 'timing/data': 0.0009027859196066856, 'timing/model': 0.5281521482393146, 'learning_rate/qwen_vl_interface': 9.986677156724054e-06, 'learning_rate/action_model': 9.986677156724054e-05, 'epoch': 0.22})
+[2026-07-28 11:02:33,059][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.05129551887512207, 'train/grad_norm_pre_clip': 2.21990704536438, 'timing/data': 0.0004027746617794037, 'timing/model': 0.5305113261565566, 'learning_rate/qwen_vl_interface': 9.986389241218003e-06, 'learning_rate/action_model': 9.986389241218003e-05, 'epoch': 0.22})
+[2026-07-28 11:02:40,243][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.13820886611938477, 'train/grad_norm_pre_clip': 2.36745548248291, 'timing/data': 0.0004083095118403435, 'timing/model': 0.527103959582746, 'learning_rate/qwen_vl_interface': 9.986098252320397e-06, 'learning_rate/action_model': 9.986098252320397e-05, 'epoch': 0.22})
+[2026-07-28 11:02:47,429][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.08286331593990326, 'train/grad_norm_pre_clip': 2.413701295852661, 'timing/data': 0.0003324141725897789, 'timing/model': 0.5344065828248858, 'learning_rate/qwen_vl_interface': 9.985804190220054e-06, 'learning_rate/action_model': 9.985804190220053e-05, 'epoch': 0.22})
+[2026-07-28 11:02:54,614][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.04940196871757507, 'train/grad_norm_pre_clip': 1.4083102941513062, 'timing/data': 0.00041503459215164185, 'timing/model': 0.5354291284456849, 'learning_rate/qwen_vl_interface': 9.985507055107788e-06, 'learning_rate/action_model': 9.985507055107788e-05, 'epoch': 0.22})
+[2026-07-28 11:03:01,805][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.08117108047008514, 'train/grad_norm_pre_clip': 2.1281213760375977, 'timing/data': 0.00039206352084875107, 'timing/model': 0.5363200400024652, 'learning_rate/qwen_vl_interface': 9.985206847176406e-06, 'learning_rate/action_model': 9.985206847176406e-05, 'epoch': 0.22})
+[2026-07-28 11:03:09,003][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.0774047002196312, 'train/grad_norm_pre_clip': 1.8448609113693237, 'timing/data': 0.00037412717938423157, 'timing/model': 0.53598219063133, 'learning_rate/qwen_vl_interface': 9.984903566620711e-06, 'learning_rate/action_model': 9.984903566620712e-05, 'epoch': 0.22})
+[2026-07-28 11:03:16,214][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.06490988284349442, 'train/grad_norm_pre_clip': 1.5215343236923218, 'timing/data': 0.0004851091653108597, 'timing/model': 0.5415857573971152, 'learning_rate/qwen_vl_interface': 9.984597213637497e-06, 'learning_rate/action_model': 9.984597213637497e-05, 'epoch': 0.22})
+[2026-07-28 11:03:23,408][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.07110035419464111, 'train/grad_norm_pre_clip': 1.8361917734146118, 'timing/data': 0.0003415476530790329, 'timing/model': 0.5375276124104857, 'learning_rate/qwen_vl_interface': 9.984287788425555e-06, 'learning_rate/action_model': 9.984287788425555e-05, 'epoch': 0.22})
+[2026-07-28 11:03:30,605][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.04600736126303673, 'train/grad_norm_pre_clip': 0.8090894222259521, 'timing/data': 0.0003770347684621811, 'timing/model': 0.5358162438496947, 'learning_rate/qwen_vl_interface': 9.983975291185666e-06, 'learning_rate/action_model': 9.983975291185667e-05, 'epoch': 0.22})
+[2026-07-28 11:03:37,789][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.051804348826408386, 'train/grad_norm_pre_clip': 1.0758298635482788, 'timing/data': 0.00048191845417022705, 'timing/model': 0.5371435601264238, 'learning_rate/qwen_vl_interface': 9.983659722120607e-06, 'learning_rate/action_model': 9.983659722120608e-05, 'epoch': 0.22})
+[2026-07-28 11:03:44,984][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.052421532571315765, 'train/grad_norm_pre_clip': 1.346105694770813, 'timing/data': 0.0003413548693060875, 'timing/model': 0.5346064837649465, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.22})
+[2026-07-28 11:03:52,164][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.06658235937356949, 'train/grad_norm_pre_clip': 1.3599518537521362, 'timing/data': 0.0003491705283522606, 'timing/model': 0.5274769552052021, 'learning_rate/qwen_vl_interface': 9.983019369336048e-06, 'learning_rate/action_model': 9.983019369336048e-05, 'epoch': 0.22})
+[2026-07-28 11:03:59,344][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.08082849532365799, 'train/grad_norm_pre_clip': 1.6655749082565308, 'timing/data': 0.00037131737917661667, 'timing/model': 0.5283400360494852, 'learning_rate/qwen_vl_interface': 9.982694586032066e-06, 'learning_rate/action_model': 9.982694586032065e-05, 'epoch': 0.22})
+[2026-07-28 11:04:06,525][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.07776849716901779, 'train/grad_norm_pre_clip': 1.8251726627349854, 'timing/data': 0.00034011900424957275, 'timing/model': 0.5266931280493736, 'learning_rate/qwen_vl_interface': 9.982366731733948e-06, 'learning_rate/action_model': 9.982366731733947e-05, 'epoch': 0.22})
+[2026-07-28 11:04:13,704][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.06554214656352997, 'train/grad_norm_pre_clip': 1.4196995496749878, 'timing/data': 0.00039989594370126724, 'timing/model': 0.5279038120061159, 'learning_rate/qwen_vl_interface': 9.982035806654437e-06, 'learning_rate/action_model': 9.982035806654437e-05, 'epoch': 0.22})
+[2026-07-28 11:04:20,884][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.07022399455308914, 'train/grad_norm_pre_clip': 3.5288195610046387, 'timing/data': 0.0003511635586619377, 'timing/model': 0.5293328631669283, 'learning_rate/qwen_vl_interface': 9.981701811008267e-06, 'learning_rate/action_model': 9.981701811008267e-05, 'epoch': 0.22})
+[2026-07-28 11:04:28,067][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.05229945480823517, 'train/grad_norm_pre_clip': 2.641327142715454, 'timing/data': 0.0004297783598303795, 'timing/model': 0.5271604005247355, 'learning_rate/qwen_vl_interface': 9.981364745012163e-06, 'learning_rate/action_model': 9.981364745012163e-05, 'epoch': 0.22})
+[2026-07-28 11:04:35,248][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.06570887565612793, 'train/grad_norm_pre_clip': 1.7330559492111206, 'timing/data': 0.00040421634912490845, 'timing/model': 0.5311449496075511, 'learning_rate/qwen_vl_interface': 9.981024608884842e-06, 'learning_rate/action_model': 9.981024608884843e-05, 'epoch': 0.22})
+[2026-07-28 11:04:42,435][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.05204566568136215, 'train/grad_norm_pre_clip': 2.112562417984009, 'timing/data': 0.00033519789576530457, 'timing/model': 0.5271473238244653, 'learning_rate/qwen_vl_interface': 9.980681402847017e-06, 'learning_rate/action_model': 9.980681402847016e-05, 'epoch': 0.22})
+[2026-07-28 11:04:49,619][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.05145310238003731, 'train/grad_norm_pre_clip': 1.309108853340149, 'timing/data': 0.00043451692909002304, 'timing/model': 0.5287868892773986, 'learning_rate/qwen_vl_interface': 9.980335127121391e-06, 'learning_rate/action_model': 9.98033512712139e-05, 'epoch': 0.22})
+[2026-07-28 11:04:56,797][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.039297088980674744, 'train/grad_norm_pre_clip': 1.6976075172424316, 'timing/data': 0.00042619742453098297, 'timing/model': 0.5272535299882293, 'learning_rate/qwen_vl_interface': 9.979985781932658e-06, 'learning_rate/action_model': 9.979985781932657e-05, 'epoch': 0.22})
+[2026-07-28 11:05:03,978][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.058747775852680206, 'train/grad_norm_pre_clip': 2.4918856620788574, 'timing/data': 0.00041225459426641464, 'timing/model': 0.5263481866568327, 'learning_rate/qwen_vl_interface': 9.979633367507501e-06, 'learning_rate/action_model': 9.979633367507502e-05, 'epoch': 0.22})
+[2026-07-28 11:05:11,164][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.0784686729311943, 'train/grad_norm_pre_clip': 2.18973970413208, 'timing/data': 0.0003743991255760193, 'timing/model': 0.531085486523807, 'learning_rate/qwen_vl_interface': 9.979277884074604e-06, 'learning_rate/action_model': 9.979277884074604e-05, 'epoch': 0.22})
+[2026-07-28 11:05:18,346][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.054568301886320114, 'train/grad_norm_pre_clip': 1.2902804613113403, 'timing/data': 0.00034230295568704605, 'timing/model': 0.5271287020295858, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.22})
+[2026-07-28 11:05:25,529][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.06297940760850906, 'train/grad_norm_pre_clip': 1.8972930908203125, 'timing/data': 0.00036804378032684326, 'timing/model': 0.5295079713687301, 'learning_rate/qwen_vl_interface': 9.978557711110245e-06, 'learning_rate/action_model': 9.978557711110244e-05, 'epoch': 0.22})
+[2026-07-28 11:05:32,709][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.04798087850213051, 'train/grad_norm_pre_clip': 0.7319579124450684, 'timing/data': 0.0003509931266307831, 'timing/model': 0.5263382457196712, 'learning_rate/qwen_vl_interface': 9.978193022046097e-06, 'learning_rate/action_model': 9.978193022046097e-05, 'epoch': 0.22})
+[2026-07-28 11:05:39,891][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.0830916315317154, 'train/grad_norm_pre_clip': 1.5191500186920166, 'timing/data': 0.0008718734607100487, 'timing/model': 0.5306940339505672, 'learning_rate/qwen_vl_interface': 9.977825264908832e-06, 'learning_rate/action_model': 9.977825264908832e-05, 'epoch': 0.22})
+[2026-07-28 11:05:47,370][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.09406762570142746, 'train/grad_norm_pre_clip': 2.605323076248169, 'timing/data': 0.0003921324387192726, 'timing/model': 0.5275487238541245, 'learning_rate/qwen_vl_interface': 9.97745443993708e-06, 'learning_rate/action_model': 9.97745443993708e-05, 'epoch': 0.22})
+[2026-07-28 11:05:54,554][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.04271099716424942, 'train/grad_norm_pre_clip': 1.3867651224136353, 'timing/data': 0.00040635839104652405, 'timing/model': 0.5266394075006247, 'learning_rate/qwen_vl_interface': 9.977080547371468e-06, 'learning_rate/action_model': 9.977080547371468e-05, 'epoch': 0.22})
+[2026-07-28 11:06:02,000][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.055892348289489746, 'train/grad_norm_pre_clip': 1.6817482709884644, 'timing/data': 0.00037935469299554825, 'timing/model': 0.5289316764101386, 'learning_rate/qwen_vl_interface': 9.97670358745461e-06, 'learning_rate/action_model': 9.97670358745461e-05, 'epoch': 0.22})
+[2026-07-28 11:06:09,318][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.08125066757202148, 'train/grad_norm_pre_clip': 1.1328973770141602, 'timing/data': 0.0003925105556845665, 'timing/model': 0.5387749010697007, 'learning_rate/qwen_vl_interface': 9.976323560431112e-06, 'learning_rate/action_model': 9.976323560431111e-05, 'epoch': 0.22})
+[2026-07-28 11:06:16,532][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.05668546259403229, 'train/grad_norm_pre_clip': 1.6677905321121216, 'timing/data': 0.0003521079197525978, 'timing/model': 0.5380005994811654, 'learning_rate/qwen_vl_interface': 9.975940466547567e-06, 'learning_rate/action_model': 9.975940466547567e-05, 'epoch': 0.23})
+[2026-07-28 11:06:23,741][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.043118324130773544, 'train/grad_norm_pre_clip': 1.0355119705200195, 'timing/data': 0.00032800063490867615, 'timing/model': 0.5379741806536913, 'learning_rate/qwen_vl_interface': 9.975554306052563e-06, 'learning_rate/action_model': 9.975554306052562e-05, 'epoch': 0.23})
+[2026-07-28 11:06:30,935][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.05299614369869232, 'train/grad_norm_pre_clip': 0.8874066472053528, 'timing/data': 0.00033169984817504883, 'timing/model': 0.5348827224224806, 'learning_rate/qwen_vl_interface': 9.975165079196674e-06, 'learning_rate/action_model': 9.975165079196673e-05, 'epoch': 0.23})
+[2026-07-28 11:06:38,122][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.048948924988508224, 'train/grad_norm_pre_clip': 1.0622304677963257, 'timing/data': 0.00036769360303878784, 'timing/model': 0.5376201290637255, 'learning_rate/qwen_vl_interface': 9.974772786232466e-06, 'learning_rate/action_model': 9.974772786232466e-05, 'epoch': 0.23})
+[2026-07-28 11:06:45,309][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.08176052570343018, 'train/grad_norm_pre_clip': 2.6391959190368652, 'timing/data': 0.00041173677891492844, 'timing/model': 0.5358867412433028, 'learning_rate/qwen_vl_interface': 9.974377427414493e-06, 'learning_rate/action_model': 9.974377427414493e-05, 'epoch': 0.23})
+[2026-07-28 11:06:52,552][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.032293856143951416, 'train/grad_norm_pre_clip': 1.2854973077774048, 'timing/data': 0.00035164598375558853, 'timing/model': 0.5380394868552685, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.23})
+[2026-07-28 11:06:59,730][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.05618030205368996, 'train/grad_norm_pre_clip': 2.128574848175049, 'timing/data': 0.0004294253885746002, 'timing/model': 0.5282384846359491, 'learning_rate/qwen_vl_interface': 9.973577513245417e-06, 'learning_rate/action_model': 9.973577513245417e-05, 'epoch': 0.23})
+[2026-07-28 11:07:06,897][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.06495565921068192, 'train/grad_norm_pre_clip': 1.8793503046035767, 'timing/data': 0.00036095548421144485, 'timing/model': 0.5279727764427662, 'learning_rate/qwen_vl_interface': 9.973172958413372e-06, 'learning_rate/action_model': 9.97317295841337e-05, 'epoch': 0.23})
+[2026-07-28 11:07:14,071][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.057163141667842865, 'train/grad_norm_pre_clip': 1.7744046449661255, 'timing/data': 0.0004912605509161949, 'timing/model': 0.5257021989673376, 'learning_rate/qwen_vl_interface': 9.97276533876567e-06, 'learning_rate/action_model': 9.97276533876567e-05, 'epoch': 0.23})
+[2026-07-28 11:07:21,224][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.049670808017253876, 'train/grad_norm_pre_clip': 1.2321598529815674, 'timing/data': 0.0004316819831728935, 'timing/model': 0.5259681278839707, 'learning_rate/qwen_vl_interface': 9.972354654566816e-06, 'learning_rate/action_model': 9.972354654566816e-05, 'epoch': 0.23})
+[2026-07-28 11:07:28,396][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.025687407702207565, 'train/grad_norm_pre_clip': 1.3559575080871582, 'timing/data': 0.00036441534757614136, 'timing/model': 0.5272853141650558, 'learning_rate/qwen_vl_interface': 9.971940906083295e-06, 'learning_rate/action_model': 9.971940906083295e-05, 'epoch': 0.23})
+[2026-07-28 11:07:35,560][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.08195210993289948, 'train/grad_norm_pre_clip': 1.4377802610397339, 'timing/data': 0.0003904532641172409, 'timing/model': 0.5253206547349691, 'learning_rate/qwen_vl_interface': 9.971524093583586e-06, 'learning_rate/action_model': 9.971524093583586e-05, 'epoch': 0.23})
+[2026-07-28 11:07:42,728][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.07460261136293411, 'train/grad_norm_pre_clip': 1.5731509923934937, 'timing/data': 0.0004039611667394638, 'timing/model': 0.5264022704213858, 'learning_rate/qwen_vl_interface': 9.971104217338154e-06, 'learning_rate/action_model': 9.971104217338153e-05, 'epoch': 0.23})
+[2026-07-28 11:07:49,900][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.057564303278923035, 'train/grad_norm_pre_clip': 1.6137501001358032, 'timing/data': 0.0004052072763442993, 'timing/model': 0.5282435836270452, 'learning_rate/qwen_vl_interface': 9.970681277619452e-06, 'learning_rate/action_model': 9.970681277619452e-05, 'epoch': 0.23})
+[2026-07-28 11:07:57,067][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.028018418699502945, 'train/grad_norm_pre_clip': 0.7252383828163147, 'timing/data': 0.0003709271550178528, 'timing/model': 0.5268977861851454, 'learning_rate/qwen_vl_interface': 9.97025527470192e-06, 'learning_rate/action_model': 9.970255274701919e-05, 'epoch': 0.23})
+[2026-07-28 11:08:04,237][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.05901758372783661, 'train/grad_norm_pre_clip': 1.2742369174957275, 'timing/data': 0.0003298865631222725, 'timing/model': 0.526689451187849, 'learning_rate/qwen_vl_interface': 9.969826208861987e-06, 'learning_rate/action_model': 9.969826208861987e-05, 'epoch': 0.23})
+[2026-07-28 11:08:11,394][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.0872756764292717, 'train/grad_norm_pre_clip': 2.220933198928833, 'timing/data': 0.0003443043678998947, 'timing/model': 0.5250707762315869, 'learning_rate/qwen_vl_interface': 9.96939408037807e-06, 'learning_rate/action_model': 9.96939408037807e-05, 'epoch': 0.23})
+[2026-07-28 11:08:18,588][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.0409914068877697, 'train/grad_norm_pre_clip': 1.4539988040924072, 'timing/data': 0.0004142429679632187, 'timing/model': 0.5353137301281095, 'learning_rate/qwen_vl_interface': 9.968958889530573e-06, 'learning_rate/action_model': 9.968958889530573e-05, 'epoch': 0.23})
+[2026-07-28 11:08:25,796][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.07050517946481705, 'train/grad_norm_pre_clip': 1.1559380292892456, 'timing/data': 0.00043430179357528687, 'timing/model': 0.5392069462686777, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.23})
+[2026-07-28 11:08:33,062][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.0413513109087944, 'train/grad_norm_pre_clip': 1.1262128353118896, 'timing/data': 0.0004349770024418831, 'timing/model': 0.5545448949560523, 'learning_rate/qwen_vl_interface': 9.968079321876382e-06, 'learning_rate/action_model': 9.968079321876382e-05, 'epoch': 0.23})
+[2026-07-28 11:08:40,386][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.0350324921309948, 'train/grad_norm_pre_clip': 0.8682870268821716, 'timing/data': 0.00040695536881685257, 'timing/model': 0.5568862585350871, 'learning_rate/qwen_vl_interface': 9.967634945640432e-06, 'learning_rate/action_model': 9.967634945640431e-05, 'epoch': 0.23})
+[2026-07-28 11:08:47,681][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.0856625884771347, 'train/grad_norm_pre_clip': 2.4729418754577637, 'timing/data': 0.0003520650789141655, 'timing/model': 0.5341812558472157, 'learning_rate/qwen_vl_interface': 9.967187508182385e-06, 'learning_rate/action_model': 9.967187508182385e-05, 'epoch': 0.23})
+[2026-07-28 11:08:54,895][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.085641048848629, 'train/grad_norm_pre_clip': 1.078310489654541, 'timing/data': 0.0006830403581261635, 'timing/model': 0.537104700691998, 'learning_rate/qwen_vl_interface': 9.966737009792577e-06, 'learning_rate/action_model': 9.966737009792577e-05, 'epoch': 0.23})
+[2026-07-28 11:09:02,119][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.06721135228872299, 'train/grad_norm_pre_clip': 2.4775588512420654, 'timing/data': 0.00034022238105535507, 'timing/model': 0.541403821669519, 'learning_rate/qwen_vl_interface': 9.966283450763333e-06, 'learning_rate/action_model': 9.966283450763333e-05, 'epoch': 0.23})
+[2026-07-28 11:09:09,333][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.10362420976161957, 'train/grad_norm_pre_clip': 2.693633794784546, 'timing/data': 0.00039658695459365845, 'timing/model': 0.5383803993463516, 'learning_rate/qwen_vl_interface': 9.965826831388962e-06, 'learning_rate/action_model': 9.965826831388961e-05, 'epoch': 0.23})
+[2026-07-28 11:09:16,524][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1250
+[2026-07-28 11:10:36,756][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1250
+[2026-07-28 11:25:21,160][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1250
+[2026-07-28 11:25:21,161][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.05513937771320343, 'train/grad_norm_pre_clip': 1.489422082901001, 'eval/action_loss/seconds': 80.22831240482628, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.0750194638967514, 'eval/latency_3/loss': 0.0750194638967514, 'eval/flappy/loss': 0.0750194638967514, 'eval/flappy/latency_3/loss': 0.0750194638967514, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 273.75500322692096, 'rl_games_eval/mid_train/mean_length': 2223.75, 'rl_games_eval/mid_train/std_reward': 186.2417630023561, 'rl_games_eval/mid_train/std_length': 1500.723254800831, 'rl_games_eval/mid_train/macro_mean_reward': 273.75500322692096, 'rl_games_eval/mid_train/macro_mean_length': 2223.75, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 273.75500322692096, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2223.75, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 186.2417630023561, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1500.723254800831, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 273.75500322692096, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00033932551741600037, 'timing/model': 0.5286895111203194, 'learning_rate/qwen_vl_interface': 9.965367151965757e-06, 'learning_rate/action_model': 9.965367151965756e-05, 'epoch': 0.23})
+[2026-07-28 11:25:30,443][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.06296326220035553, 'train/grad_norm_pre_clip': 1.256990671157837, 'timing/data': 0.0004030875861644745, 'timing/model': 0.6165398107841611, 'learning_rate/qwen_vl_interface': 9.964904412792002e-06, 'learning_rate/action_model': 9.964904412792002e-05, 'epoch': 0.23})
+[2026-07-28 11:25:39,820][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.06253871321678162, 'train/grad_norm_pre_clip': 2.500171184539795, 'timing/data': 0.0003613242879509926, 'timing/model': 0.6203714543953538, 'learning_rate/qwen_vl_interface': 9.964438614167964e-06, 'learning_rate/action_model': 9.964438614167964e-05, 'epoch': 0.23})
+[2026-07-28 11:25:48,933][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.05480755493044853, 'train/grad_norm_pre_clip': 4.171367645263672, 'timing/data': 0.0003888215869665146, 'timing/model': 0.6182152964174747, 'learning_rate/qwen_vl_interface': 9.963969756395892e-06, 'learning_rate/action_model': 9.96396975639589e-05, 'epoch': 0.23})
+[2026-07-28 11:25:58,272][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.09117579460144043, 'train/grad_norm_pre_clip': 1.571664571762085, 'timing/data': 0.0004558144137263298, 'timing/model': 0.6172005580738187, 'learning_rate/qwen_vl_interface': 9.963497839780025e-06, 'learning_rate/action_model': 9.963497839780024e-05, 'epoch': 0.23})
+[2026-07-28 11:26:07,512][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.05883670225739479, 'train/grad_norm_pre_clip': 1.7624311447143555, 'timing/data': 0.00036306027323007584, 'timing/model': 0.6180482162162662, 'learning_rate/qwen_vl_interface': 9.963022864626582e-06, 'learning_rate/action_model': 9.963022864626581e-05, 'epoch': 0.23})
+[2026-07-28 11:26:16,900][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.07703626900911331, 'train/grad_norm_pre_clip': 1.461211919784546, 'timing/data': 0.0003861803561449051, 'timing/model': 0.6182919768616557, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.23})
+[2026-07-28 11:26:26,238][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.0781952366232872, 'train/grad_norm_pre_clip': 0.8271040320396423, 'timing/data': 0.0003450848162174225, 'timing/model': 0.8448878219351172, 'learning_rate/qwen_vl_interface': 9.962063739941783e-06, 'learning_rate/action_model': 9.962063739941783e-05, 'epoch': 0.23})
+[2026-07-28 11:26:35,344][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.05736095830798149, 'train/grad_norm_pre_clip': 1.437662124633789, 'timing/data': 0.0003481311723589897, 'timing/model': 0.6176793482154608, 'learning_rate/qwen_vl_interface': 9.961579591032791e-06, 'learning_rate/action_model': 9.96157959103279e-05, 'epoch': 0.23})
+[2026-07-28 11:26:44,621][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.04820980876684189, 'train/grad_norm_pre_clip': 1.117745280265808, 'timing/data': 0.00036881305277347565, 'timing/model': 0.7947367690503597, 'learning_rate/qwen_vl_interface': 9.961092384830956e-06, 'learning_rate/action_model': 9.961092384830956e-05, 'epoch': 0.23})
+[2026-07-28 11:26:53,730][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.0875580906867981, 'train/grad_norm_pre_clip': 2.6366734504699707, 'timing/data': 0.0003373539075255394, 'timing/model': 0.6175108077004552, 'learning_rate/qwen_vl_interface': 9.96060212165242e-06, 'learning_rate/action_model': 9.96060212165242e-05, 'epoch': 0.23})
+[2026-07-28 11:27:02,835][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.03385484218597412, 'train/grad_norm_pre_clip': 1.05166494846344, 'timing/data': 0.00033324770629405975, 'timing/model': 0.6177539089694619, 'learning_rate/qwen_vl_interface': 9.96010880181531e-06, 'learning_rate/action_model': 9.960108801815309e-05, 'epoch': 0.23})
+[2026-07-28 11:27:12,198][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.07108891010284424, 'train/grad_norm_pre_clip': 1.3179415464401245, 'timing/data': 0.00035651959478855133, 'timing/model': 0.6203799340873957, 'learning_rate/qwen_vl_interface': 9.959612425639736e-06, 'learning_rate/action_model': 9.959612425639735e-05, 'epoch': 0.23})
+[2026-07-28 11:27:21,541][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.04448781535029411, 'train/grad_norm_pre_clip': 1.173985481262207, 'timing/data': 0.00038150791078805923, 'timing/model': 0.8588833222165704, 'learning_rate/qwen_vl_interface': 9.95911299344779e-06, 'learning_rate/action_model': 9.959112993447789e-05, 'epoch': 0.23})
+[2026-07-28 11:27:30,634][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.047660138458013535, 'train/grad_norm_pre_clip': 1.9692445993423462, 'timing/data': 0.0003547715023159981, 'timing/model': 0.6281815795227885, 'learning_rate/qwen_vl_interface': 9.958610505563546e-06, 'learning_rate/action_model': 9.958610505563546e-05, 'epoch': 0.23})
+[2026-07-28 11:27:39,726][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.08009158819913864, 'train/grad_norm_pre_clip': 1.7521790266036987, 'timing/data': 0.00039234012365341187, 'timing/model': 0.6189306667074561, 'learning_rate/qwen_vl_interface': 9.95810496231307e-06, 'learning_rate/action_model': 9.95810496231307e-05, 'epoch': 0.23})
+[2026-07-28 11:27:48,833][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.021869992837309837, 'train/grad_norm_pre_clip': 0.9983699321746826, 'timing/data': 0.0003936123102903366, 'timing/model': 0.6183564327657223, 'learning_rate/qwen_vl_interface': 9.957596364024396e-06, 'learning_rate/action_model': 9.957596364024396e-05, 'epoch': 0.23})
+[2026-07-28 11:27:58,450][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.04903997480869293, 'train/grad_norm_pre_clip': 2.1705901622772217, 'timing/data': 0.00039584841579198837, 'timing/model': 0.8409471614286304, 'learning_rate/qwen_vl_interface': 9.957084711027553e-06, 'learning_rate/action_model': 9.957084711027552e-05, 'epoch': 0.23})
+[2026-07-28 11:28:07,543][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.044406458735466, 'train/grad_norm_pre_clip': 2.351665735244751, 'timing/data': 0.00033521372824907303, 'timing/model': 0.6181887835264206, 'learning_rate/qwen_vl_interface': 9.956570003654544e-06, 'learning_rate/action_model': 9.956570003654544e-05, 'epoch': 0.23})
+[2026-07-28 11:28:16,652][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.08071795850992203, 'train/grad_norm_pre_clip': 2.1409616470336914, 'timing/data': 0.0004160739481449127, 'timing/model': 0.6169687183573842, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.23})
+[2026-07-28 11:28:25,890][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.10859648883342743, 'train/grad_norm_pre_clip': 2.8983054161071777, 'timing/data': 0.000340120866894722, 'timing/model': 0.6180201871320605, 'learning_rate/qwen_vl_interface': 9.955531427117964e-06, 'learning_rate/action_model': 9.955531427117964e-05, 'epoch': 0.23})
+[2026-07-28 11:28:35,253][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.01825122721493244, 'train/grad_norm_pre_clip': 1.236547589302063, 'timing/data': 0.0003335345536470413, 'timing/model': 0.6182614881545305, 'learning_rate/qwen_vl_interface': 9.955007558628316e-06, 'learning_rate/action_model': 9.955007558628316e-05, 'epoch': 0.23})
+[2026-07-28 11:28:44,604][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.026603441685438156, 'train/grad_norm_pre_clip': 0.7981135845184326, 'timing/data': 0.00035521481186151505, 'timing/model': 0.6167129566892982, 'learning_rate/qwen_vl_interface': 9.954480637110346e-06, 'learning_rate/action_model': 9.954480637110344e-05, 'epoch': 0.23})
+[2026-07-28 11:28:53,742][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.11296094208955765, 'train/grad_norm_pre_clip': 3.05291748046875, 'timing/data': 0.00033026188611984253, 'timing/model': 0.6168844932690263, 'learning_rate/qwen_vl_interface': 9.953950662905963e-06, 'learning_rate/action_model': 9.953950662905964e-05, 'epoch': 0.23})
+[2026-07-28 11:29:02,840][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.05551132559776306, 'train/grad_norm_pre_clip': 1.9306343793869019, 'timing/data': 0.0003358554095029831, 'timing/model': 0.6181617258116603, 'learning_rate/qwen_vl_interface': 9.953417636359067e-06, 'learning_rate/action_model': 9.953417636359067e-05, 'epoch': 0.23})
+[2026-07-28 11:29:11,949][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.03713624179363251, 'train/grad_norm_pre_clip': 1.697575569152832, 'timing/data': 0.00032438524067401886, 'timing/model': 0.6197156002745032, 'learning_rate/qwen_vl_interface': 9.95288155781553e-06, 'learning_rate/action_model': 9.95288155781553e-05, 'epoch': 0.23})
+[2026-07-28 11:29:21,500][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.04789317026734352, 'train/grad_norm_pre_clip': 2.246211290359497, 'timing/data': 0.0004931995645165443, 'timing/model': 0.6234004739671946, 'learning_rate/qwen_vl_interface': 9.952342427623209e-06, 'learning_rate/action_model': 9.952342427623208e-05, 'epoch': 0.23})
+[2026-07-28 11:29:30,786][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.05446906387805939, 'train/grad_norm_pre_clip': 1.5276581048965454, 'timing/data': 0.00033789966255426407, 'timing/model': 0.6182360649108887, 'learning_rate/qwen_vl_interface': 9.95180024613194e-06, 'learning_rate/action_model': 9.951800246131939e-05, 'epoch': 0.23})
+[2026-07-28 11:29:39,921][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.06556189805269241, 'train/grad_norm_pre_clip': 1.7053890228271484, 'timing/data': 0.0003372281789779663, 'timing/model': 0.6178831774741411, 'learning_rate/qwen_vl_interface': 9.951255013693537e-06, 'learning_rate/action_model': 9.951255013693536e-05, 'epoch': 0.23})
+[2026-07-28 11:29:49,287][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.0392790362238884, 'train/grad_norm_pre_clip': 1.0388951301574707, 'timing/data': 0.00032181479036808014, 'timing/model': 0.6171314911916852, 'learning_rate/qwen_vl_interface': 9.950706730661797e-06, 'learning_rate/action_model': 9.950706730661797e-05, 'epoch': 0.23})
+[2026-07-28 11:29:58,406][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.038517143577337265, 'train/grad_norm_pre_clip': 1.8051646947860718, 'timing/data': 0.0003897259011864662, 'timing/model': 0.6179631007835269, 'learning_rate/qwen_vl_interface': 9.950155397392494e-06, 'learning_rate/action_model': 9.950155397392493e-05, 'epoch': 0.24})
+[2026-07-28 11:30:07,801][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.037090905010700226, 'train/grad_norm_pre_clip': 0.9366478323936462, 'timing/data': 0.0003802431747317314, 'timing/model': 0.6168806944042444, 'learning_rate/qwen_vl_interface': 9.949601014243382e-06, 'learning_rate/action_model': 9.949601014243382e-05, 'epoch': 0.24})
+[2026-07-28 11:30:16,936][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.04820026457309723, 'train/grad_norm_pre_clip': 1.1577202081680298, 'timing/data': 0.00041803717613220215, 'timing/model': 0.6216042889282107, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.24})
+[2026-07-28 11:30:26,272][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.05682441592216492, 'train/grad_norm_pre_clip': 1.0062769651412964, 'timing/data': 0.0004606274887919426, 'timing/model': 0.6198749113827944, 'learning_rate/qwen_vl_interface': 9.948483099746645e-06, 'learning_rate/action_model': 9.948483099746645e-05, 'epoch': 0.24})
+[2026-07-28 11:30:35,609][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.10263523459434509, 'train/grad_norm_pre_clip': 1.9926561117172241, 'timing/data': 0.0003563258796930313, 'timing/model': 0.8468371545895934, 'learning_rate/qwen_vl_interface': 9.947919569124423e-06, 'learning_rate/action_model': 9.947919569124422e-05, 'epoch': 0.24})
+[2026-07-28 11:30:44,730][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.0515877828001976, 'train/grad_norm_pre_clip': 2.3525567054748535, 'timing/data': 0.0003714505583047867, 'timing/model': 0.6231045043095946, 'learning_rate/qwen_vl_interface': 9.947352990073196e-06, 'learning_rate/action_model': 9.947352990073195e-05, 'epoch': 0.24})
+[2026-07-28 11:30:53,884][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.06450171023607254, 'train/grad_norm_pre_clip': 2.796926736831665, 'timing/data': 0.00044563785195350647, 'timing/model': 0.623949977569282, 'learning_rate/qwen_vl_interface': 9.946783362960612e-06, 'learning_rate/action_model': 9.946783362960612e-05, 'epoch': 0.24})
+[2026-07-28 11:31:03,287][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.06402069330215454, 'train/grad_norm_pre_clip': 2.785982608795166, 'timing/data': 0.0004294170066714287, 'timing/model': 0.8547434890642762, 'learning_rate/qwen_vl_interface': 9.946210688156297e-06, 'learning_rate/action_model': 9.946210688156297e-05, 'epoch': 0.24})
+[2026-07-28 11:31:12,477][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.1335696578025818, 'train/grad_norm_pre_clip': 3.7384164333343506, 'timing/data': 0.0003337329253554344, 'timing/model': 0.6200910704210401, 'learning_rate/qwen_vl_interface': 9.945634966031852e-06, 'learning_rate/action_model': 9.945634966031851e-05, 'epoch': 0.24})
+[2026-07-28 11:31:21,921][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.04196414351463318, 'train/grad_norm_pre_clip': 0.6536312699317932, 'timing/data': 0.0003354093059897423, 'timing/model': 0.862119073048234, 'learning_rate/qwen_vl_interface': 9.945056196960858e-06, 'learning_rate/action_model': 9.945056196960858e-05, 'epoch': 0.24})
+[2026-07-28 11:31:31,034][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.04607301205396652, 'train/grad_norm_pre_clip': 0.7716085314750671, 'timing/data': 0.00031946878880262375, 'timing/model': 0.6332851806655526, 'learning_rate/qwen_vl_interface': 9.944474381318871e-06, 'learning_rate/action_model': 9.94447438131887e-05, 'epoch': 0.24})
+[2026-07-28 11:31:40,144][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.06713202595710754, 'train/grad_norm_pre_clip': 2.236368417739868, 'timing/data': 0.00034259818494319916, 'timing/model': 0.6174662513658404, 'learning_rate/qwen_vl_interface': 9.943889519483424e-06, 'learning_rate/action_model': 9.943889519483424e-05, 'epoch': 0.24})
+[2026-07-28 11:31:49,277][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.057572849094867706, 'train/grad_norm_pre_clip': 1.4286236763000488, 'timing/data': 0.00032968074083328247, 'timing/model': 0.6180677693337202, 'learning_rate/qwen_vl_interface': 9.94330161183403e-06, 'learning_rate/action_model': 9.94330161183403e-05, 'epoch': 0.24})
+[2026-07-28 11:31:58,658][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.0741531103849411, 'train/grad_norm_pre_clip': 1.930959939956665, 'timing/data': 0.0003288993611931801, 'timing/model': 0.6195060154423118, 'learning_rate/qwen_vl_interface': 9.942710658752175e-06, 'learning_rate/action_model': 9.942710658752174e-05, 'epoch': 0.24})
+[2026-07-28 11:32:07,754][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.044668249785900116, 'train/grad_norm_pre_clip': 1.5056047439575195, 'timing/data': 0.0003826506435871124, 'timing/model': 0.6177595229819417, 'learning_rate/qwen_vl_interface': 9.942116660621319e-06, 'learning_rate/action_model': 9.94211666062132e-05, 'epoch': 0.24})
+[2026-07-28 11:32:16,839][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.04779449477791786, 'train/grad_norm_pre_clip': 2.078571319580078, 'timing/data': 0.0003849901258945465, 'timing/model': 0.619834166020155, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.24})
+[2026-07-28 11:32:26,184][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.04318337142467499, 'train/grad_norm_pre_clip': 0.7794729471206665, 'timing/data': 0.00037879496812820435, 'timing/model': 0.6174557060003281, 'learning_rate/qwen_vl_interface': 9.940919530756347e-06, 'learning_rate/action_model': 9.940919530756347e-05, 'epoch': 0.24})
+[2026-07-28 11:32:35,532][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.06795218586921692, 'train/grad_norm_pre_clip': 1.8522825241088867, 'timing/data': 0.0004001231864094734, 'timing/model': 0.8413664503023028, 'learning_rate/qwen_vl_interface': 9.940316399799032e-06, 'learning_rate/action_model': 9.940316399799033e-05, 'epoch': 0.24})
+[2026-07-28 11:32:44,637][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.07002907991409302, 'train/grad_norm_pre_clip': 1.9728586673736572, 'timing/data': 0.0003284430131316185, 'timing/model': 0.6174729578197002, 'learning_rate/qwen_vl_interface': 9.939710225346329e-06, 'learning_rate/action_model': 9.939710225346329e-05, 'epoch': 0.24})
+[2026-07-28 11:32:53,754][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.08844033628702164, 'train/grad_norm_pre_clip': 3.040457010269165, 'timing/data': 0.00032011233270168304, 'timing/model': 0.6184927672147751, 'learning_rate/qwen_vl_interface': 9.939101007791575e-06, 'learning_rate/action_model': 9.939101007791575e-05, 'epoch': 0.24})
+[2026-07-28 11:33:02,835][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.04386705905199051, 'train/grad_norm_pre_clip': 2.3441405296325684, 'timing/data': 0.00038821715861558914, 'timing/model': 0.6167217390611768, 'learning_rate/qwen_vl_interface': 9.938488747530087e-06, 'learning_rate/action_model': 9.938488747530086e-05, 'epoch': 0.24})
+[2026-07-28 11:33:12,180][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.05087236315011978, 'train/grad_norm_pre_clip': 1.2464615106582642, 'timing/data': 0.0003748452290892601, 'timing/model': 0.6225400287657976, 'learning_rate/qwen_vl_interface': 9.937873444959151e-06, 'learning_rate/action_model': 9.937873444959151e-05, 'epoch': 0.24})
+[2026-07-28 11:33:21,261][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.050974197685718536, 'train/grad_norm_pre_clip': 1.3892452716827393, 'timing/data': 0.0003529312089085579, 'timing/model': 0.6173961823806167, 'learning_rate/qwen_vl_interface': 9.937255100478036e-06, 'learning_rate/action_model': 9.937255100478036e-05, 'epoch': 0.24})
+[2026-07-28 11:33:30,388][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.09756582975387573, 'train/grad_norm_pre_clip': 2.18039608001709, 'timing/data': 0.0003492673859000206, 'timing/model': 0.616192470304668, 'learning_rate/qwen_vl_interface': 9.93663371448797e-06, 'learning_rate/action_model': 9.93663371448797e-05, 'epoch': 0.24})
+[2026-07-28 11:33:39,737][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.06676693260669708, 'train/grad_norm_pre_clip': 1.258072018623352, 'timing/data': 0.0003538532182574272, 'timing/model': 0.8573970217257738, 'learning_rate/qwen_vl_interface': 9.936009287392173e-06, 'learning_rate/action_model': 9.936009287392171e-05, 'epoch': 0.24})
+[2026-07-28 11:33:49,097][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.05950754135847092, 'train/grad_norm_pre_clip': 1.7063891887664795, 'timing/data': 0.0003769025206565857, 'timing/model': 0.6199822099879384, 'learning_rate/qwen_vl_interface': 9.935381819595822e-06, 'learning_rate/action_model': 9.935381819595822e-05, 'epoch': 0.24})
+[2026-07-28 11:33:58,210][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.06383468210697174, 'train/grad_norm_pre_clip': 1.12388277053833, 'timing/data': 0.00033437181264162064, 'timing/model': 0.6169218746945262, 'learning_rate/qwen_vl_interface': 9.934751311506078e-06, 'learning_rate/action_model': 9.934751311506077e-05, 'epoch': 0.24})
+[2026-07-28 11:34:07,316][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.06849253177642822, 'train/grad_norm_pre_clip': 1.656836748123169, 'timing/data': 0.0003911154344677925, 'timing/model': 0.6173072243109345, 'learning_rate/qwen_vl_interface': 9.93411776353207e-06, 'learning_rate/action_model': 9.934117763532069e-05, 'epoch': 0.24})
+[2026-07-28 11:34:16,480][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.044449254870414734, 'train/grad_norm_pre_clip': 0.8045485019683838, 'timing/data': 0.00033314619213342667, 'timing/model': 0.61871119402349, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.24})
+[2026-07-28 11:34:25,838][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.05555979534983635, 'train/grad_norm_pre_clip': 1.0641729831695557, 'timing/data': 0.00037024635821580887, 'timing/model': 0.6200967915356159, 'learning_rate/qwen_vl_interface': 9.932841549577642e-06, 'learning_rate/action_model': 9.932841549577641e-05, 'epoch': 0.24})
+[2026-07-28 11:34:35,271][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.10152273625135422, 'train/grad_norm_pre_clip': 1.1391266584396362, 'timing/data': 0.00033517275005578995, 'timing/model': 0.6663646325469017, 'learning_rate/qwen_vl_interface': 9.932198884425343e-06, 'learning_rate/action_model': 9.932198884425342e-05, 'epoch': 0.24})
+[2026-07-28 11:34:44,317][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.06040159612894058, 'train/grad_norm_pre_clip': 1.2956798076629639, 'timing/data': 0.0003642914816737175, 'timing/model': 0.6648684414103627, 'learning_rate/qwen_vl_interface': 9.931553181045022e-06, 'learning_rate/action_model': 9.931553181045022e-05, 'epoch': 0.24})
+[2026-07-28 11:34:53,205][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.055170685052871704, 'train/grad_norm_pre_clip': 1.9050391912460327, 'timing/data': 0.0003493586555123329, 'timing/model': 0.6183746661990881, 'learning_rate/qwen_vl_interface': 9.930904439855667e-06, 'learning_rate/action_model': 9.930904439855667e-05, 'epoch': 0.24})
+[2026-07-28 11:35:02,071][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.0571165531873703, 'train/grad_norm_pre_clip': 0.9741333723068237, 'timing/data': 0.00033552106469869614, 'timing/model': 0.6172187123447657, 'learning_rate/qwen_vl_interface': 9.930252661278245e-06, 'learning_rate/action_model': 9.930252661278244e-05, 'epoch': 0.24})
+[2026-07-28 11:35:11,215][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.04277010262012482, 'train/grad_norm_pre_clip': 0.8937231302261353, 'timing/data': 0.0003431588411331177, 'timing/model': 0.8769362149760127, 'learning_rate/qwen_vl_interface': 9.92959784573568e-06, 'learning_rate/action_model': 9.92959784573568e-05, 'epoch': 0.24})
+[2026-07-28 11:35:20,551][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.05617549270391464, 'train/grad_norm_pre_clip': 2.264233112335205, 'timing/data': 0.0004127407446503639, 'timing/model': 0.6177005730569363, 'learning_rate/qwen_vl_interface': 9.92893999365288e-06, 'learning_rate/action_model': 9.92893999365288e-05, 'epoch': 0.24})
+[2026-07-28 11:35:29,629][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.039817266166210175, 'train/grad_norm_pre_clip': 1.569014310836792, 'timing/data': 0.00036005955189466476, 'timing/model': 0.6196328066289425, 'learning_rate/qwen_vl_interface': 9.928279105456717e-06, 'learning_rate/action_model': 9.928279105456718e-05, 'epoch': 0.24})
+[2026-07-28 11:35:38,741][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.048645272850990295, 'train/grad_norm_pre_clip': 1.9045013189315796, 'timing/data': 0.0004162294790148735, 'timing/model': 0.6188316568732262, 'learning_rate/qwen_vl_interface': 9.927615181576034e-06, 'learning_rate/action_model': 9.927615181576034e-05, 'epoch': 0.24})
+[2026-07-28 11:35:47,856][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.04172208905220032, 'train/grad_norm_pre_clip': 1.7762916088104248, 'timing/data': 0.00035351887345314026, 'timing/model': 0.6190320085734129, 'learning_rate/qwen_vl_interface': 9.926948222441644e-06, 'learning_rate/action_model': 9.926948222441643e-05, 'epoch': 0.24})
+[2026-07-28 11:35:57,210][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.023916520178318024, 'train/grad_norm_pre_clip': 0.9118744730949402, 'timing/data': 0.0003837253898382187, 'timing/model': 0.617626384831965, 'learning_rate/qwen_vl_interface': 9.926278228486328e-06, 'learning_rate/action_model': 9.926278228486327e-05, 'epoch': 0.24})
+[2026-07-28 11:36:06,579][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.08366483449935913, 'train/grad_norm_pre_clip': 3.473223924636841, 'timing/data': 0.000390373170375824, 'timing/model': 0.6175544410943985, 'learning_rate/qwen_vl_interface': 9.92560520014484e-06, 'learning_rate/action_model': 9.92560520014484e-05, 'epoch': 0.24})
+[2026-07-28 11:36:15,679][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.049192916601896286, 'train/grad_norm_pre_clip': 2.1431632041931152, 'timing/data': 0.0003942251205444336, 'timing/model': 0.6171522755175829, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.24})
+[2026-07-28 11:36:24,998][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.07761574536561966, 'train/grad_norm_pre_clip': 2.409992218017578, 'timing/data': 0.0003197593614459038, 'timing/model': 0.6188978748396039, 'learning_rate/qwen_vl_interface': 9.924250042052197e-06, 'learning_rate/action_model': 9.924250042052197e-05, 'epoch': 0.24})
+[2026-07-28 11:36:34,091][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.10175690799951553, 'train/grad_norm_pre_clip': 2.707009792327881, 'timing/data': 0.0003536064177751541, 'timing/model': 0.6175501495599747, 'learning_rate/qwen_vl_interface': 9.92356791318039e-06, 'learning_rate/action_model': 9.92356791318039e-05, 'epoch': 0.24})
+[2026-07-28 11:36:43,190][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.05887993797659874, 'train/grad_norm_pre_clip': 1.758433222770691, 'timing/data': 0.00032876525074243546, 'timing/model': 0.6176551133394241, 'learning_rate/qwen_vl_interface': 9.922882751681104e-06, 'learning_rate/action_model': 9.922882751681104e-05, 'epoch': 0.24})
+[2026-07-28 11:36:52,554][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.03612393140792847, 'train/grad_norm_pre_clip': 1.1183202266693115, 'timing/data': 0.00033594761043787, 'timing/model': 0.6207856116816401, 'learning_rate/qwen_vl_interface': 9.922194557998932e-06, 'learning_rate/action_model': 9.922194557998933e-05, 'epoch': 0.24})
+[2026-07-28 11:37:01,891][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.03429531306028366, 'train/grad_norm_pre_clip': 1.165737271308899, 'timing/data': 0.00037922803312540054, 'timing/model': 0.6169762462377548, 'learning_rate/qwen_vl_interface': 9.921503332580439e-06, 'learning_rate/action_model': 9.921503332580439e-05, 'epoch': 0.24})
+[2026-07-28 11:37:10,980][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.06655295938253403, 'train/grad_norm_pre_clip': 1.4337244033813477, 'timing/data': 0.00034097954630851746, 'timing/model': 0.6183964982628822, 'learning_rate/qwen_vl_interface': 9.92080907587415e-06, 'learning_rate/action_model': 9.92080907587415e-05, 'epoch': 0.24})
+[2026-07-28 11:37:20,093][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.059248700737953186, 'train/grad_norm_pre_clip': 2.1408474445343018, 'timing/data': 0.000798138789832592, 'timing/model': 0.6190693462267518, 'learning_rate/qwen_vl_interface': 9.92011178833056e-06, 'learning_rate/action_model': 9.92011178833056e-05, 'epoch': 0.24})
+[2026-07-28 11:37:29,483][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.07597610354423523, 'train/grad_norm_pre_clip': 1.4717422723770142, 'timing/data': 0.0005098609253764153, 'timing/model': 0.6180646298453212, 'learning_rate/qwen_vl_interface': 9.919411470402136e-06, 'learning_rate/action_model': 9.919411470402136e-05, 'epoch': 0.24})
+[2026-07-28 11:37:38,823][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.05659310519695282, 'train/grad_norm_pre_clip': 0.9352469444274902, 'timing/data': 0.0006900522857904434, 'timing/model': 0.6185146775096655, 'learning_rate/qwen_vl_interface': 9.918708122543305e-06, 'learning_rate/action_model': 9.918708122543305e-05, 'epoch': 0.24})
+[2026-07-28 11:37:47,940][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.033079832792282104, 'train/grad_norm_pre_clip': 1.356475591659546, 'timing/data': 0.0003740442916750908, 'timing/model': 0.6173951234668493, 'learning_rate/qwen_vl_interface': 9.918001745210458e-06, 'learning_rate/action_model': 9.918001745210456e-05, 'epoch': 0.24})
+[2026-07-28 11:37:57,206][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.09777325391769409, 'train/grad_norm_pre_clip': 1.935595154762268, 'timing/data': 0.000413430854678154, 'timing/model': 0.7915951060131192, 'learning_rate/qwen_vl_interface': 9.917292338861958e-06, 'learning_rate/action_model': 9.917292338861957e-05, 'epoch': 0.24})
+[2026-07-28 11:38:06,366][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.057598792016506195, 'train/grad_norm_pre_clip': 3.123173952102661, 'timing/data': 0.0003582816570997238, 'timing/model': 0.6172895915806293, 'learning_rate/qwen_vl_interface': 9.916579903958134e-06, 'learning_rate/action_model': 9.916579903958133e-05, 'epoch': 0.24})
+[2026-07-28 11:38:15,770][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.05200382322072983, 'train/grad_norm_pre_clip': 2.2766568660736084, 'timing/data': 0.00033348798751831055, 'timing/model': 0.837344161234796, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.25})
+[2026-07-28 11:38:24,900][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.06948613375425339, 'train/grad_norm_pre_clip': 2.1116549968719482, 'timing/data': 0.00034389737993478775, 'timing/model': 0.6185201918706298, 'learning_rate/qwen_vl_interface': 9.915145950335631e-06, 'learning_rate/action_model': 9.915145950335632e-05, 'epoch': 0.25})
+[2026-07-28 11:38:34,266][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.047318510711193085, 'train/grad_norm_pre_clip': 1.1686357259750366, 'timing/data': 0.0003784187138080597, 'timing/model': 0.8581830766052008, 'learning_rate/qwen_vl_interface': 9.914424432547434e-06, 'learning_rate/action_model': 9.914424432547433e-05, 'epoch': 0.25})
+[2026-07-28 11:38:43,352][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.08440232276916504, 'train/grad_norm_pre_clip': 2.464752674102783, 'timing/data': 0.00040233973413705826, 'timing/model': 0.6186980018392205, 'learning_rate/qwen_vl_interface': 9.913699888064862e-06, 'learning_rate/action_model': 9.913699888064861e-05, 'epoch': 0.25})
+[2026-07-28 11:38:52,501][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.08474305272102356, 'train/grad_norm_pre_clip': 1.1144449710845947, 'timing/data': 0.00035081896930933, 'timing/model': 0.618060490116477, 'learning_rate/qwen_vl_interface': 9.912972317358066e-06, 'learning_rate/action_model': 9.912972317358066e-05, 'epoch': 0.25})
+[2026-07-28 11:39:01,647][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.06428245455026627, 'train/grad_norm_pre_clip': 1.4840168952941895, 'timing/data': 0.00038911309093236923, 'timing/model': 0.6188009334728122, 'learning_rate/qwen_vl_interface': 9.912241720899157e-06, 'learning_rate/action_model': 9.912241720899157e-05, 'epoch': 0.25})
+[2026-07-28 11:39:10,811][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.034564003348350525, 'train/grad_norm_pre_clip': 0.8924664258956909, 'timing/data': 0.0003650626167654991, 'timing/model': 0.6162250228226185, 'learning_rate/qwen_vl_interface': 9.911508099162215e-06, 'learning_rate/action_model': 9.911508099162214e-05, 'epoch': 0.25})
+[2026-07-28 11:39:19,971][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.04876110702753067, 'train/grad_norm_pre_clip': 1.1623592376708984, 'timing/data': 0.00040985550731420517, 'timing/model': 0.6182076679542661, 'learning_rate/qwen_vl_interface': 9.910771452623276e-06, 'learning_rate/action_model': 9.910771452623275e-05, 'epoch': 0.25})
+[2026-07-28 11:39:28,934][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.08955243229866028, 'train/grad_norm_pre_clip': 1.4424656629562378, 'timing/data': 0.0004474278539419174, 'timing/model': 0.6186289247125387, 'learning_rate/qwen_vl_interface': 9.910031781760342e-06, 'learning_rate/action_model': 9.910031781760342e-05, 'epoch': 0.25})
+[2026-07-28 11:39:37,847][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.04863593727350235, 'train/grad_norm_pre_clip': 2.0194544792175293, 'timing/data': 0.0003912942484021187, 'timing/model': 0.6166485194116831, 'learning_rate/qwen_vl_interface': 9.909289087053378e-06, 'learning_rate/action_model': 9.909289087053378e-05, 'epoch': 0.25})
+[2026-07-28 11:39:46,967][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.05929639935493469, 'train/grad_norm_pre_clip': 1.201253890991211, 'timing/data': 0.000388171523809433, 'timing/model': 0.8384614270180464, 'learning_rate/qwen_vl_interface': 9.90854336898431e-06, 'learning_rate/action_model': 9.90854336898431e-05, 'epoch': 0.25})
+[2026-07-28 11:39:56,322][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.05229427292943001, 'train/grad_norm_pre_clip': 1.7614468336105347, 'timing/data': 0.00040283892303705215, 'timing/model': 0.8509830506518483, 'learning_rate/qwen_vl_interface': 9.907794628037028e-06, 'learning_rate/action_model': 9.907794628037028e-05, 'epoch': 0.25})
+[2026-07-28 11:40:05,442][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.0822717472910881, 'train/grad_norm_pre_clip': 3.0799663066864014, 'timing/data': 0.000332527793943882, 'timing/model': 0.6178236957639456, 'learning_rate/qwen_vl_interface': 9.90704286469738e-06, 'learning_rate/action_model': 9.90704286469738e-05, 'epoch': 0.25})
+[2026-07-28 11:40:14,564][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.07072747498750687, 'train/grad_norm_pre_clip': 2.098100185394287, 'timing/data': 0.00041958969086408615, 'timing/model': 0.6180654214695096, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.25})
+[2026-07-28 11:40:23,896][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.06887605786323547, 'train/grad_norm_pre_clip': 2.4287984371185303, 'timing/data': 0.00038099661469459534, 'timing/model': 0.6185696311295033, 'learning_rate/qwen_vl_interface': 9.905530272794195e-06, 'learning_rate/action_model': 9.905530272794195e-05, 'epoch': 0.25})
+[2026-07-28 11:40:33,022][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.049058686941862106, 'train/grad_norm_pre_clip': 0.7853021025657654, 'timing/data': 0.00038208533078432083, 'timing/model': 0.6187185477465391, 'learning_rate/qwen_vl_interface': 9.904769445212163e-06, 'learning_rate/action_model': 9.904769445212162e-05, 'epoch': 0.25})
+[2026-07-28 11:40:42,156][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.04427884519100189, 'train/grad_norm_pre_clip': 0.8728733658790588, 'timing/data': 0.0004639718681573868, 'timing/model': 0.6200886340811849, 'learning_rate/qwen_vl_interface': 9.904005597200772e-06, 'learning_rate/action_model': 9.904005597200772e-05, 'epoch': 0.25})
+[2026-07-28 11:40:51,603][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.02532377280294895, 'train/grad_norm_pre_clip': 0.9060235619544983, 'timing/data': 0.0004054764285683632, 'timing/model': 0.8774597309529781, 'learning_rate/qwen_vl_interface': 9.903238729255678e-06, 'learning_rate/action_model': 9.903238729255677e-05, 'epoch': 0.25})
+[2026-07-28 11:41:01,009][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.04062020778656006, 'train/grad_norm_pre_clip': 1.8204368352890015, 'timing/data': 0.00033392850309610367, 'timing/model': 0.6181371342390776, 'learning_rate/qwen_vl_interface': 9.902468841874492e-06, 'learning_rate/action_model': 9.902468841874492e-05, 'epoch': 0.25})
+[2026-07-28 11:41:10,134][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.11714843660593033, 'train/grad_norm_pre_clip': 3.1664135456085205, 'timing/data': 0.00042842328548431396, 'timing/model': 0.6193285575136542, 'learning_rate/qwen_vl_interface': 9.901695935556784e-06, 'learning_rate/action_model': 9.901695935556783e-05, 'epoch': 0.25})
+[2026-07-28 11:41:19,259][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.040874283760786057, 'train/grad_norm_pre_clip': 2.0082411766052246, 'timing/data': 0.00035812146961688995, 'timing/model': 0.6164234494790435, 'learning_rate/qwen_vl_interface': 9.900920010804088e-06, 'learning_rate/action_model': 9.900920010804088e-05, 'epoch': 0.25})
+[2026-07-28 11:41:28,360][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.09301459789276123, 'train/grad_norm_pre_clip': 2.3650341033935547, 'timing/data': 0.0007732231169939041, 'timing/model': 0.6195477386936545, 'learning_rate/qwen_vl_interface': 9.900141068119891e-06, 'learning_rate/action_model': 9.900141068119891e-05, 'epoch': 0.25})
+[2026-07-28 11:41:37,774][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.035528913140296936, 'train/grad_norm_pre_clip': 1.5097945928573608, 'timing/data': 0.00038807839155197144, 'timing/model': 0.6182758258655667, 'learning_rate/qwen_vl_interface': 9.89935910800964e-06, 'learning_rate/action_model': 9.89935910800964e-05, 'epoch': 0.25})
+[2026-07-28 11:41:47,134][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.050245728343725204, 'train/grad_norm_pre_clip': 0.92434161901474, 'timing/data': 0.0004263026639819145, 'timing/model': 0.6202647071331739, 'learning_rate/qwen_vl_interface': 9.898574130980743e-06, 'learning_rate/action_model': 9.898574130980742e-05, 'epoch': 0.25})
+[2026-07-28 11:41:56,266][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.04344705492258072, 'train/grad_norm_pre_clip': 0.9544516801834106, 'timing/data': 0.0004132501780986786, 'timing/model': 0.6200318224728107, 'learning_rate/qwen_vl_interface': 9.89778613754256e-06, 'learning_rate/action_model': 9.89778613754256e-05, 'epoch': 0.25})
+[2026-07-28 11:42:05,415][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.04445000737905502, 'train/grad_norm_pre_clip': 2.7541115283966064, 'timing/data': 0.0003563864156603813, 'timing/model': 0.6200395133346319, 'learning_rate/qwen_vl_interface': 9.896995128206412e-06, 'learning_rate/action_model': 9.896995128206411e-05, 'epoch': 0.25})
+[2026-07-28 11:42:14,563][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.09471248090267181, 'train/grad_norm_pre_clip': 3.9765493869781494, 'timing/data': 0.0003490978851914406, 'timing/model': 0.620716892182827, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.25})
+[2026-07-28 11:42:23,969][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.07602895051240921, 'train/grad_norm_pre_clip': 3.1270554065704346, 'timing/data': 0.00042649079114198685, 'timing/model': 0.849158613011241, 'learning_rate/qwen_vl_interface': 9.89540406389529e-06, 'learning_rate/action_model': 9.89540406389529e-05, 'epoch': 0.25})
+[2026-07-28 11:42:33,387][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.07512696087360382, 'train/grad_norm_pre_clip': 1.8936280012130737, 'timing/data': 0.00033405981957912445, 'timing/model': 0.6189562305808067, 'learning_rate/qwen_vl_interface': 9.894604009952742e-06, 'learning_rate/action_model': 9.89460400995274e-05, 'epoch': 0.25})
+[2026-07-28 11:42:42,537][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.027895841747522354, 'train/grad_norm_pre_clip': 1.3633383512496948, 'timing/data': 0.0003774929791688919, 'timing/model': 0.6227866793051362, 'learning_rate/qwen_vl_interface': 9.893800942177073e-06, 'learning_rate/action_model': 9.893800942177073e-05, 'epoch': 0.25})
+[2026-07-28 11:42:51,742][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.04843420907855034, 'train/grad_norm_pre_clip': 1.8232948780059814, 'timing/data': 0.0003513237461447716, 'timing/model': 0.6216949848458171, 'learning_rate/qwen_vl_interface': 9.892994861089393e-06, 'learning_rate/action_model': 9.892994861089393e-05, 'epoch': 0.25})
+[2026-07-28 11:43:00,938][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.06594837456941605, 'train/grad_norm_pre_clip': 1.6537853479385376, 'timing/data': 0.0006332844495773315, 'timing/model': 0.6226417953148484, 'learning_rate/qwen_vl_interface': 9.892185767212756e-06, 'learning_rate/action_model': 9.892185767212756e-05, 'epoch': 0.25})
+[2026-07-28 11:43:10,378][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.05592185631394386, 'train/grad_norm_pre_clip': 2.281081199645996, 'timing/data': 0.0003288211300969124, 'timing/model': 0.6183981671929359, 'learning_rate/qwen_vl_interface': 9.891373661072172e-06, 'learning_rate/action_model': 9.891373661072173e-05, 'epoch': 0.25})
+[2026-07-28 11:43:19,285][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.05802230164408684, 'train/grad_norm_pre_clip': 2.676506996154785, 'timing/data': 0.0003877701237797737, 'timing/model': 0.6200738111510873, 'learning_rate/qwen_vl_interface': 9.89055854319461e-06, 'learning_rate/action_model': 9.890558543194611e-05, 'epoch': 0.25})
+[2026-07-28 11:43:28,619][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.030739903450012207, 'train/grad_norm_pre_clip': 0.9546852111816406, 'timing/data': 0.000439484603703022, 'timing/model': 0.6225928198546171, 'learning_rate/qwen_vl_interface': 9.889740414108994e-06, 'learning_rate/action_model': 9.889740414108995e-05, 'epoch': 0.25})
+[2026-07-28 11:43:38,270][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.07048308849334717, 'train/grad_norm_pre_clip': 1.7769957780838013, 'timing/data': 0.0003598863258957863, 'timing/model': 0.8192601297050714, 'learning_rate/qwen_vl_interface': 9.888919274346196e-06, 'learning_rate/action_model': 9.888919274346195e-05, 'epoch': 0.25})
+[2026-07-28 11:43:47,414][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.05786050856113434, 'train/grad_norm_pre_clip': 1.2070256471633911, 'timing/data': 0.0003353925421833992, 'timing/model': 0.6180363995954394, 'learning_rate/qwen_vl_interface': 9.888095124439043e-06, 'learning_rate/action_model': 9.888095124439043e-05, 'epoch': 0.25})
+[2026-07-28 11:43:56,522][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.03560212254524231, 'train/grad_norm_pre_clip': 1.5900553464889526, 'timing/data': 0.00034566503018140793, 'timing/model': 0.6205484829843044, 'learning_rate/qwen_vl_interface': 9.88726796492232e-06, 'learning_rate/action_model': 9.88726796492232e-05, 'epoch': 0.25})
+[2026-07-28 11:44:05,654][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.05289719998836517, 'train/grad_norm_pre_clip': 1.2738100290298462, 'timing/data': 0.0004820851609110832, 'timing/model': 0.619945952668786, 'learning_rate/qwen_vl_interface': 9.886437796332759e-06, 'learning_rate/action_model': 9.88643779633276e-05, 'epoch': 0.25})
+[2026-07-28 11:44:15,255][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.05600074678659439, 'train/grad_norm_pre_clip': 2.137619733810425, 'timing/data': 0.0003444422036409378, 'timing/model': 0.8558528143912554, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.25})
+[2026-07-28 11:44:24,354][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.05265900492668152, 'train/grad_norm_pre_clip': 2.0191004276275635, 'timing/data': 0.00032955408096313477, 'timing/model': 0.6182560725137591, 'learning_rate/qwen_vl_interface': 9.884768434091832e-06, 'learning_rate/action_model': 9.884768434091832e-05, 'epoch': 0.25})
+[2026-07-28 11:44:33,446][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.025011705234646797, 'train/grad_norm_pre_clip': 0.7592759728431702, 'timing/data': 0.0005272431299090385, 'timing/model': 0.6179720051586628, 'learning_rate/qwen_vl_interface': 9.883929241523698e-06, 'learning_rate/action_model': 9.883929241523697e-05, 'epoch': 0.25})
+[2026-07-28 11:44:42,567][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.09189838171005249, 'train/grad_norm_pre_clip': 1.4857405424118042, 'timing/data': 0.00035788025707006454, 'timing/model': 0.6183816194534302, 'learning_rate/qwen_vl_interface': 9.883087042049185e-06, 'learning_rate/action_model': 9.883087042049186e-05, 'epoch': 0.25})
+[2026-07-28 11:44:51,935][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.057397592812776566, 'train/grad_norm_pre_clip': 1.160849928855896, 'timing/data': 0.0004497598856687546, 'timing/model': 0.6175730908289552, 'learning_rate/qwen_vl_interface': 9.882241836214794e-06, 'learning_rate/action_model': 9.882241836214794e-05, 'epoch': 0.25})
+[2026-07-28 11:45:01,315][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.1011054515838623, 'train/grad_norm_pre_clip': 2.208038806915283, 'timing/data': 0.00046727433800697327, 'timing/model': 0.8749175388365984, 'learning_rate/qwen_vl_interface': 9.881393624568965e-06, 'learning_rate/action_model': 9.881393624568965e-05, 'epoch': 0.25})
+[2026-07-28 11:45:10,413][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.0711120218038559, 'train/grad_norm_pre_clip': 1.3529666662216187, 'timing/data': 0.00041597988456487656, 'timing/model': 0.6187936728820205, 'learning_rate/qwen_vl_interface': 9.880542407662098e-06, 'learning_rate/action_model': 9.880542407662097e-05, 'epoch': 0.25})
+[2026-07-28 11:45:19,532][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.0400552861392498, 'train/grad_norm_pre_clip': 1.216799259185791, 'timing/data': 0.0004185996949672699, 'timing/model': 0.6182423140853643, 'learning_rate/qwen_vl_interface': 9.879688186046535e-06, 'learning_rate/action_model': 9.879688186046534e-05, 'epoch': 0.25})
+[2026-07-28 11:45:28,880][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.07028526067733765, 'train/grad_norm_pre_clip': 2.271599054336548, 'timing/data': 0.0003545256331562996, 'timing/model': 0.841131804510951, 'learning_rate/qwen_vl_interface': 9.87883096027657e-06, 'learning_rate/action_model': 9.87883096027657e-05, 'epoch': 0.25})
+[2026-07-28 11:45:37,996][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.07830743491649628, 'train/grad_norm_pre_clip': 1.2246836423873901, 'timing/data': 0.0005583809688687325, 'timing/model': 0.6152490396052599, 'learning_rate/qwen_vl_interface': 9.87797073090845e-06, 'learning_rate/action_model': 9.87797073090845e-05, 'epoch': 0.25})
+[2026-07-28 11:45:47,119][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.08001437783241272, 'train/grad_norm_pre_clip': 1.6183210611343384, 'timing/data': 0.00034659262746572495, 'timing/model': 0.6170663703233004, 'learning_rate/qwen_vl_interface': 9.877107498500368e-06, 'learning_rate/action_model': 9.877107498500367e-05, 'epoch': 0.25})
+[2026-07-28 11:45:56,224][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.07003416121006012, 'train/grad_norm_pre_clip': 1.1835323572158813, 'timing/data': 0.0003338092938065529, 'timing/model': 0.6192805068567395, 'learning_rate/qwen_vl_interface': 9.876241263612467e-06, 'learning_rate/action_model': 9.876241263612465e-05, 'epoch': 0.25})
+[2026-07-28 11:46:05,565][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.08056002855300903, 'train/grad_norm_pre_clip': 1.0532011985778809, 'timing/data': 0.00034489482641220093, 'timing/model': 0.6186902858316898, 'learning_rate/qwen_vl_interface': 9.875372026806835e-06, 'learning_rate/action_model': 9.875372026806835e-05, 'epoch': 0.25})
+[2026-07-28 11:46:14,686][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.05248185247182846, 'train/grad_norm_pre_clip': 0.8595852255821228, 'timing/data': 0.0003359830006957054, 'timing/model': 0.619584989733994, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.25})
+[2026-07-28 11:46:24,009][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.043098334223032, 'train/grad_norm_pre_clip': 1.3109928369522095, 'timing/data': 0.00041356030851602554, 'timing/model': 0.6200445033609867, 'learning_rate/qwen_vl_interface': 9.873624549700485e-06, 'learning_rate/action_model': 9.873624549700484e-05, 'epoch': 0.25})
+[2026-07-28 11:46:33,145][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.09906788170337677, 'train/grad_norm_pre_clip': 2.9382355213165283, 'timing/data': 0.0003728941082954407, 'timing/model': 0.61819077283144, 'learning_rate/qwen_vl_interface': 9.872746310533683e-06, 'learning_rate/action_model': 9.872746310533684e-05, 'epoch': 0.25})
+[2026-07-28 11:46:42,264][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.048985764384269714, 'train/grad_norm_pre_clip': 1.8294140100479126, 'timing/data': 0.00034852419048547745, 'timing/model': 0.6190136531367898, 'learning_rate/qwen_vl_interface': 9.871865071716991e-06, 'learning_rate/action_model': 9.87186507171699e-05, 'epoch': 0.26})
+[2026-07-28 11:46:51,646][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.06919227540493011, 'train/grad_norm_pre_clip': 2.6182456016540527, 'timing/data': 0.0003490131348371506, 'timing/model': 0.6184663074091077, 'learning_rate/qwen_vl_interface': 9.870980833822231e-06, 'learning_rate/action_model': 9.87098083382223e-05, 'epoch': 0.26})
+[2026-07-28 11:47:00,588][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.07148212194442749, 'train/grad_norm_pre_clip': 1.4630759954452515, 'timing/data': 0.0004660123959183693, 'timing/model': 0.6194665879011154, 'learning_rate/qwen_vl_interface': 9.87009359742318e-06, 'learning_rate/action_model': 9.87009359742318e-05, 'epoch': 0.26})
+[2026-07-28 11:47:09,739][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.069499172270298, 'train/grad_norm_pre_clip': 0.892708420753479, 'timing/data': 0.0003306465223431587, 'timing/model': 0.619527842849493, 'learning_rate/qwen_vl_interface': 9.86920336309555e-06, 'learning_rate/action_model': 9.86920336309555e-05, 'epoch': 0.26})
+[2026-07-28 11:47:18,836][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.05545446649193764, 'train/grad_norm_pre_clip': 2.0619328022003174, 'timing/data': 0.00038734450936317444, 'timing/model': 0.7830540500581264, 'learning_rate/qwen_vl_interface': 9.86831013141701e-06, 'learning_rate/action_model': 9.86831013141701e-05, 'epoch': 0.26})
+[2026-07-28 11:47:28,220][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.04914748668670654, 'train/grad_norm_pre_clip': 3.172529697418213, 'timing/data': 0.0003402857109904289, 'timing/model': 0.6197890061885118, 'learning_rate/qwen_vl_interface': 9.867413902967166e-06, 'learning_rate/action_model': 9.867413902967166e-05, 'epoch': 0.26})
+[2026-07-28 11:47:37,362][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.06689918041229248, 'train/grad_norm_pre_clip': 1.7361549139022827, 'timing/data': 0.0003187870606780052, 'timing/model': 0.616945612244308, 'learning_rate/qwen_vl_interface': 9.86651467832757e-06, 'learning_rate/action_model': 9.86651467832757e-05, 'epoch': 0.26})
+[2026-07-28 11:47:46,478][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.061900876462459564, 'train/grad_norm_pre_clip': 3.200503349304199, 'timing/data': 0.00033457204699516296, 'timing/model': 0.6186084691435099, 'learning_rate/qwen_vl_interface': 9.865612458081718e-06, 'learning_rate/action_model': 9.865612458081719e-05, 'epoch': 0.26})
+[2026-07-28 11:47:55,809][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.07063087075948715, 'train/grad_norm_pre_clip': 3.5838189125061035, 'timing/data': 0.00033406540751457214, 'timing/model': 0.61648980807513, 'learning_rate/qwen_vl_interface': 9.864707242815057e-06, 'learning_rate/action_model': 9.864707242815056e-05, 'epoch': 0.26})
+[2026-07-28 11:48:04,936][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.048764802515506744, 'train/grad_norm_pre_clip': 1.2283244132995605, 'timing/data': 0.0003804033622145653, 'timing/model': 0.6188693661242723, 'learning_rate/qwen_vl_interface': 9.863799033114966e-06, 'learning_rate/action_model': 9.863799033114966e-05, 'epoch': 0.26})
+[2026-07-28 11:48:14,403][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.02197599783539772, 'train/grad_norm_pre_clip': 1.2105368375778198, 'timing/data': 0.000330866314470768, 'timing/model': 0.6184175480157137, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.26})
+[2026-07-28 11:48:23,866][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.1288740038871765, 'train/grad_norm_pre_clip': 2.6476335525512695, 'timing/data': 0.00036672409623861313, 'timing/model': 0.830180655233562, 'learning_rate/qwen_vl_interface': 9.861973632773748e-06, 'learning_rate/action_model': 9.861973632773748e-05, 'epoch': 0.26})
+[2026-07-28 11:48:32,975][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.0785890743136406, 'train/grad_norm_pre_clip': 2.9890098571777344, 'timing/data': 0.00033514760434627533, 'timing/model': 0.6182604460045695, 'learning_rate/qwen_vl_interface': 9.861056443317105e-06, 'learning_rate/action_model': 9.861056443317105e-05, 'epoch': 0.26})
+[2026-07-28 11:48:42,134][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.07687649130821228, 'train/grad_norm_pre_clip': 2.6561951637268066, 'timing/data': 0.0003470899537205696, 'timing/model': 0.6215086737647653, 'learning_rate/qwen_vl_interface': 9.860136261796e-06, 'learning_rate/action_model': 9.860136261795998e-05, 'epoch': 0.26})
+[2026-07-28 11:48:51,542][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.09987350553274155, 'train/grad_norm_pre_clip': 3.8864169120788574, 'timing/data': 0.0003499528393149376, 'timing/model': 0.6265047611668706, 'learning_rate/qwen_vl_interface': 9.859213088807521e-06, 'learning_rate/action_model': 9.85921308880752e-05, 'epoch': 0.26})
+[2026-07-28 11:49:00,715][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.07060298323631287, 'train/grad_norm_pre_clip': 2.7879958152770996, 'timing/data': 0.000337955541908741, 'timing/model': 0.6316475700587034, 'learning_rate/qwen_vl_interface': 9.858286924950714e-06, 'learning_rate/action_model': 9.858286924950713e-05, 'epoch': 0.26})
+[2026-07-28 11:49:10,068][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.03451910987496376, 'train/grad_norm_pre_clip': 1.5189027786254883, 'timing/data': 0.00038825348019599915, 'timing/model': 0.6305189644917846, 'learning_rate/qwen_vl_interface': 9.85735777082655e-06, 'learning_rate/action_model': 9.857357770826551e-05, 'epoch': 0.26})
+[2026-07-28 11:49:19,226][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.05760621279478073, 'train/grad_norm_pre_clip': 2.745300531387329, 'timing/data': 0.00034369900822639465, 'timing/model': 0.625710909254849, 'learning_rate/qwen_vl_interface': 9.856425627037951e-06, 'learning_rate/action_model': 9.856425627037952e-05, 'epoch': 0.26})
+[2026-07-28 11:49:28,343][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.042894355952739716, 'train/grad_norm_pre_clip': 1.6429734230041504, 'timing/data': 0.0003957115113735199, 'timing/model': 0.6234396174550056, 'learning_rate/qwen_vl_interface': 9.855490494189777e-06, 'learning_rate/action_model': 9.855490494189776e-05, 'epoch': 0.26})
+[2026-07-28 11:49:37,742][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.05737244337797165, 'train/grad_norm_pre_clip': 1.8071848154067993, 'timing/data': 0.0004348643124103546, 'timing/model': 0.6216402240097523, 'learning_rate/qwen_vl_interface': 9.854552372888819e-06, 'learning_rate/action_model': 9.854552372888819e-05, 'epoch': 0.26})
+[2026-07-28 11:49:46,859][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.09278860688209534, 'train/grad_norm_pre_clip': 1.5648231506347656, 'timing/data': 0.0004982789978384972, 'timing/model': 0.6187358684837818, 'learning_rate/qwen_vl_interface': 9.853611263743817e-06, 'learning_rate/action_model': 9.853611263743817e-05, 'epoch': 0.26})
+[2026-07-28 11:49:56,240][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.047445185482501984, 'train/grad_norm_pre_clip': 0.99732506275177, 'timing/data': 0.00032582320272922516, 'timing/model': 0.8585599455982447, 'learning_rate/qwen_vl_interface': 9.85266716736545e-06, 'learning_rate/action_model': 9.852667167365449e-05, 'epoch': 0.26})
+[2026-07-28 11:50:05,391][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.05729798972606659, 'train/grad_norm_pre_clip': 1.0476950407028198, 'timing/data': 0.0003844061866402626, 'timing/model': 0.6186133930459619, 'learning_rate/qwen_vl_interface': 9.851720084366324e-06, 'learning_rate/action_model': 9.851720084366324e-05, 'epoch': 0.26})
+[2026-07-28 11:50:14,738][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.04363346844911575, 'train/grad_norm_pre_clip': 1.1422706842422485, 'timing/data': 0.0003798315301537514, 'timing/model': 0.6195746148005128, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.26})
+[2026-07-28 11:50:23,968][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.07376321405172348, 'train/grad_norm_pre_clip': 2.0404648780822754, 'timing/data': 0.00038299523293972015, 'timing/model': 0.6333966637030244, 'learning_rate/qwen_vl_interface': 9.849816960965959e-06, 'learning_rate/action_model': 9.849816960965959e-05, 'epoch': 0.26})
+[2026-07-28 11:50:33,073][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.05427084118127823, 'train/grad_norm_pre_clip': 0.9169588685035706, 'timing/data': 0.0003282669931650162, 'timing/model': 0.6166048031300306, 'learning_rate/qwen_vl_interface': 9.848860921799633e-06, 'learning_rate/action_model': 9.848860921799632e-05, 'epoch': 0.26})
+[2026-07-28 11:50:42,483][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.05443394184112549, 'train/grad_norm_pre_clip': 2.0851292610168457, 'timing/data': 0.0003955848515033722, 'timing/model': 0.6214876640588045, 'learning_rate/qwen_vl_interface': 9.847901898482385e-06, 'learning_rate/action_model': 9.847901898482385e-05, 'epoch': 0.26})
+[2026-07-28 11:50:51,775][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.10154780000448227, 'train/grad_norm_pre_clip': 1.625733733177185, 'timing/data': 0.00040936656296253204, 'timing/model': 0.6217443346977234, 'learning_rate/qwen_vl_interface': 9.846939891636512e-06, 'learning_rate/action_model': 9.846939891636512e-05, 'epoch': 0.26})
+[2026-07-28 11:51:01,219][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.10987477749586105, 'train/grad_norm_pre_clip': 2.7718019485473633, 'timing/data': 0.00035101640969514847, 'timing/model': 0.6230125175788999, 'learning_rate/qwen_vl_interface': 9.845974901886253e-06, 'learning_rate/action_model': 9.845974901886252e-05, 'epoch': 0.26})
+[2026-07-28 11:51:10,627][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.08533310145139694, 'train/grad_norm_pre_clip': 1.917118787765503, 'timing/data': 0.0004205303266644478, 'timing/model': 0.6208341373130679, 'learning_rate/qwen_vl_interface': 9.845006929857777e-06, 'learning_rate/action_model': 9.845006929857776e-05, 'epoch': 0.26})
+[2026-07-28 11:51:19,773][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.07393132895231247, 'train/grad_norm_pre_clip': 1.6020188331604004, 'timing/data': 0.00032963603734970093, 'timing/model': 0.6203678315505385, 'learning_rate/qwen_vl_interface': 9.844035976179193e-06, 'learning_rate/action_model': 9.844035976179192e-05, 'epoch': 0.26})
+[2026-07-28 11:51:28,895][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.047483254224061966, 'train/grad_norm_pre_clip': 0.807950496673584, 'timing/data': 0.00032982882112264633, 'timing/model': 0.6182971810922027, 'learning_rate/qwen_vl_interface': 9.843062041480538e-06, 'learning_rate/action_model': 9.843062041480539e-05, 'epoch': 0.26})
+[2026-07-28 11:51:38,281][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.08495907485485077, 'train/grad_norm_pre_clip': 1.5810171365737915, 'timing/data': 0.0003883708268404007, 'timing/model': 0.622239463031292, 'learning_rate/qwen_vl_interface': 9.842085126393793e-06, 'learning_rate/action_model': 9.842085126393792e-05, 'epoch': 0.26})
+[2026-07-28 11:51:47,631][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.043592773377895355, 'train/grad_norm_pre_clip': 1.2205697298049927, 'timing/data': 0.0006085503846406937, 'timing/model': 0.6192063791677356, 'learning_rate/qwen_vl_interface': 9.841105231552863e-06, 'learning_rate/action_model': 9.841105231552864e-05, 'epoch': 0.26})
+[2026-07-28 11:51:56,517][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.058625273406505585, 'train/grad_norm_pre_clip': 1.2594293355941772, 'timing/data': 0.0003386782482266426, 'timing/model': 0.6184517750516534, 'learning_rate/qwen_vl_interface': 9.840122357593594e-06, 'learning_rate/action_model': 9.840122357593594e-05, 'epoch': 0.26})
+[2026-07-28 11:52:05,626][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.05401761084794998, 'train/grad_norm_pre_clip': 1.4947502613067627, 'timing/data': 0.00034460611641407013, 'timing/model': 0.6181083526462317, 'learning_rate/qwen_vl_interface': 9.839136505153762e-06, 'learning_rate/action_model': 9.839136505153762e-05, 'epoch': 0.26})
+[2026-07-28 11:52:14,522][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.05576746538281441, 'train/grad_norm_pre_clip': 2.898845911026001, 'timing/data': 0.0004324018955230713, 'timing/model': 0.6192723847925663, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.26})
+[2026-07-28 11:52:24,128][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.05690343305468559, 'train/grad_norm_pre_clip': 1.2555952072143555, 'timing/data': 0.0004237005487084389, 'timing/model': 0.6190564101561904, 'learning_rate/qwen_vl_interface': 9.837155867393176e-06, 'learning_rate/action_model': 9.837155867393175e-05, 'epoch': 0.26})
+[2026-07-28 11:52:33,021][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.08748864382505417, 'train/grad_norm_pre_clip': 1.4125479459762573, 'timing/data': 0.0003287205472588539, 'timing/model': 0.6182048907503486, 'learning_rate/qwen_vl_interface': 9.836161083357635e-06, 'learning_rate/action_model': 9.836161083357635e-05, 'epoch': 0.26})
+[2026-07-28 11:52:42,129][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.0867960974574089, 'train/grad_norm_pre_clip': 2.0125954151153564, 'timing/data': 0.00034248922020196915, 'timing/model': 0.6175736645236611, 'learning_rate/qwen_vl_interface': 9.835163323411957e-06, 'learning_rate/action_model': 9.835163323411958e-05, 'epoch': 0.26})
+[2026-07-28 11:52:51,023][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.06228318065404892, 'train/grad_norm_pre_clip': 1.7801820039749146, 'timing/data': 0.0003257188946008682, 'timing/model': 0.618979305960238, 'learning_rate/qwen_vl_interface': 9.834162588203581e-06, 'learning_rate/action_model': 9.83416258820358e-05, 'epoch': 0.26})
+[2026-07-28 11:53:00,403][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.04915327578783035, 'train/grad_norm_pre_clip': 1.1409170627593994, 'timing/data': 0.0003197193145751953, 'timing/model': 0.8320261966437101, 'learning_rate/qwen_vl_interface': 9.833158878381872e-06, 'learning_rate/action_model': 9.83315887838187e-05, 'epoch': 0.26})
+[2026-07-28 11:53:09,550][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.03416972979903221, 'train/grad_norm_pre_clip': 1.8714038133621216, 'timing/data': 0.00045652594417333603, 'timing/model': 0.6183396652340889, 'learning_rate/qwen_vl_interface': 9.832152194598123e-06, 'learning_rate/action_model': 9.832152194598122e-05, 'epoch': 0.26})
+[2026-07-28 11:53:18,712][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.05862712487578392, 'train/grad_norm_pre_clip': 1.6313539743423462, 'timing/data': 0.00037679821252822876, 'timing/model': 0.6170735312625766, 'learning_rate/qwen_vl_interface': 9.831142537505563e-06, 'learning_rate/action_model': 9.831142537505562e-05, 'epoch': 0.26})
+[2026-07-28 11:53:27,830][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.06405006349086761, 'train/grad_norm_pre_clip': 1.4283099174499512, 'timing/data': 0.00039727985858917236, 'timing/model': 0.6167549369856715, 'learning_rate/qwen_vl_interface': 9.830129907759347e-06, 'learning_rate/action_model': 9.830129907759347e-05, 'epoch': 0.26})
+[2026-07-28 11:53:36,958][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.05513487756252289, 'train/grad_norm_pre_clip': 2.0418765544891357, 'timing/data': 0.00041223596781492233, 'timing/model': 0.6186005817726254, 'learning_rate/qwen_vl_interface': 9.82911430601656e-06, 'learning_rate/action_model': 9.82911430601656e-05, 'epoch': 0.26})
+[2026-07-28 11:53:46,306][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.059577979147434235, 'train/grad_norm_pre_clip': 2.3409974575042725, 'timing/data': 0.0003320695832371712, 'timing/model': 0.6182426763698459, 'learning_rate/qwen_vl_interface': 9.828095732936216e-06, 'learning_rate/action_model': 9.828095732936216e-05, 'epoch': 0.26})
+[2026-07-28 11:53:55,652][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.041278522461652756, 'train/grad_norm_pre_clip': 1.2680896520614624, 'timing/data': 0.0004974529147148132, 'timing/model': 0.6190336439758539, 'learning_rate/qwen_vl_interface': 9.827074189179254e-06, 'learning_rate/action_model': 9.827074189179253e-05, 'epoch': 0.26})
+[2026-07-28 11:54:04,762][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.08952264487743378, 'train/grad_norm_pre_clip': 2.928553581237793, 'timing/data': 0.00033182743936777115, 'timing/model': 0.6175263123586774, 'learning_rate/qwen_vl_interface': 9.82604967540854e-06, 'learning_rate/action_model': 9.82604967540854e-05, 'epoch': 0.26})
+[2026-07-28 11:54:13,859][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.061432041227817535, 'train/grad_norm_pre_clip': 1.4525094032287598, 'timing/data': 0.00040939729660749435, 'timing/model': 0.6188367428258061, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.26})
+[2026-07-28 11:54:22,955][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.06186794117093086, 'train/grad_norm_pre_clip': 2.301388740539551, 'timing/data': 0.00040884315967559814, 'timing/model': 0.6168517107143998, 'learning_rate/qwen_vl_interface': 9.823991740486982e-06, 'learning_rate/action_model': 9.823991740486982e-05, 'epoch': 0.26})
+[2026-07-28 11:54:32,508][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.03565666079521179, 'train/grad_norm_pre_clip': 1.0507124662399292, 'timing/data': 0.0004159463569521904, 'timing/model': 0.8385583190247416, 'learning_rate/qwen_vl_interface': 9.822958320671506e-06, 'learning_rate/action_model': 9.822958320671506e-05, 'epoch': 0.26})
+[2026-07-28 11:54:41,578][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.06691502779722214, 'train/grad_norm_pre_clip': 0.9108678698539734, 'timing/data': 0.0003245873376727104, 'timing/model': 0.624052288942039, 'learning_rate/qwen_vl_interface': 9.821921933513023e-06, 'learning_rate/action_model': 9.821921933513023e-05, 'epoch': 0.26})
+[2026-07-28 11:54:50,462][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.04107268899679184, 'train/grad_norm_pre_clip': 1.2434804439544678, 'timing/data': 0.00033804867416620255, 'timing/model': 0.6178719718009233, 'learning_rate/qwen_vl_interface': 9.820882579684037e-06, 'learning_rate/action_model': 9.820882579684037e-05, 'epoch': 0.26})
+[2026-07-28 11:54:59,539][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.054633043706417084, 'train/grad_norm_pre_clip': 1.6003955602645874, 'timing/data': 0.0003086002543568611, 'timing/model': 0.6172718107700348, 'learning_rate/qwen_vl_interface': 9.819840259858968e-06, 'learning_rate/action_model': 9.819840259858968e-05, 'epoch': 0.27})
+[2026-07-28 11:55:08,870][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.05752424895763397, 'train/grad_norm_pre_clip': 1.3748767375946045, 'timing/data': 0.00033806636929512024, 'timing/model': 0.6184589797630906, 'learning_rate/qwen_vl_interface': 9.818794974714167e-06, 'learning_rate/action_model': 9.818794974714168e-05, 'epoch': 0.27})
+[2026-07-28 11:55:18,202][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.05747988075017929, 'train/grad_norm_pre_clip': 2.5101888179779053, 'timing/data': 0.00039208121597766876, 'timing/model': 0.6173321604728699, 'learning_rate/qwen_vl_interface': 9.817746724927912e-06, 'learning_rate/action_model': 9.817746724927912e-05, 'epoch': 0.27})
+[2026-07-28 11:55:27,313][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.08808249235153198, 'train/grad_norm_pre_clip': 2.1160740852355957, 'timing/data': 0.0003597624599933624, 'timing/model': 0.6166200479492545, 'learning_rate/qwen_vl_interface': 9.8166955111804e-06, 'learning_rate/action_model': 9.816695511180399e-05, 'epoch': 0.27})
+[2026-07-28 11:55:36,395][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.0744638741016388, 'train/grad_norm_pre_clip': 2.521695137023926, 'timing/data': 0.0004067877307534218, 'timing/model': 0.619106532074511, 'learning_rate/qwen_vl_interface': 9.815641334153746e-06, 'learning_rate/action_model': 9.815641334153746e-05, 'epoch': 0.27})
+[2026-07-28 11:55:45,778][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.02629910409450531, 'train/grad_norm_pre_clip': 0.6114101409912109, 'timing/data': 0.0004898020997643471, 'timing/model': 0.8608534764498472, 'learning_rate/qwen_vl_interface': 9.814584194532001e-06, 'learning_rate/action_model': 9.814584194532001e-05, 'epoch': 0.27})
+[2026-07-28 11:55:54,879][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.04286421835422516, 'train/grad_norm_pre_clip': 1.8658125400543213, 'timing/data': 0.0004060380160808563, 'timing/model': 0.6187164392322302, 'learning_rate/qwen_vl_interface': 9.813524093001128e-06, 'learning_rate/action_model': 9.813524093001129e-05, 'epoch': 0.27})
+[2026-07-28 11:56:04,208][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.05647550895810127, 'train/grad_norm_pre_clip': 1.054011583328247, 'timing/data': 0.0003176843747496605, 'timing/model': 0.6172758294269443, 'learning_rate/qwen_vl_interface': 9.81246103024902e-06, 'learning_rate/action_model': 9.81246103024902e-05, 'epoch': 0.27})
+[2026-07-28 11:56:13,313][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.053729355335235596, 'train/grad_norm_pre_clip': 1.2177578210830688, 'timing/data': 0.00034601613879203796, 'timing/model': 0.6185673791915178, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.27})
+[2026-07-28 11:56:22,406][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.08941921591758728, 'train/grad_norm_pre_clip': 1.9909969568252563, 'timing/data': 0.000321132130920887, 'timing/model': 0.6182209346443415, 'learning_rate/qwen_vl_interface': 9.810326023842245e-06, 'learning_rate/action_model': 9.810326023842245e-05, 'epoch': 0.27})
+[2026-07-28 11:56:31,487][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.057967737317085266, 'train/grad_norm_pre_clip': 2.6756021976470947, 'timing/data': 0.0003709150478243828, 'timing/model': 0.6175690442323685, 'learning_rate/qwen_vl_interface': 9.809254081572963e-06, 'learning_rate/action_model': 9.809254081572961e-05, 'epoch': 0.27})
+[2026-07-28 11:56:40,797][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.04967048019170761, 'train/grad_norm_pre_clip': 1.1573565006256104, 'timing/data': 0.0003114528954029083, 'timing/model': 0.6189596531912684, 'learning_rate/qwen_vl_interface': 9.808179180853204e-06, 'learning_rate/action_model': 9.808179180853204e-05, 'epoch': 0.27})
+[2026-07-28 11:56:49,875][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.05471979081630707, 'train/grad_norm_pre_clip': 2.068808078765869, 'timing/data': 0.0003336425870656967, 'timing/model': 0.6171976905316114, 'learning_rate/qwen_vl_interface': 9.807101322380464e-06, 'learning_rate/action_model': 9.807101322380464e-05, 'epoch': 0.27})
+[2026-07-28 11:56:58,962][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.10477464646100998, 'train/grad_norm_pre_clip': 4.521114349365234, 'timing/data': 0.00033469218760728836, 'timing/model': 0.6201538993045688, 'learning_rate/qwen_vl_interface': 9.806020506854152e-06, 'learning_rate/action_model': 9.806020506854153e-05, 'epoch': 0.27})
+[2026-07-28 11:57:08,219][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.06322140246629715, 'train/grad_norm_pre_clip': 1.5434904098510742, 'timing/data': 0.0003267880529165268, 'timing/model': 0.7852752460166812, 'learning_rate/qwen_vl_interface': 9.804936734975596e-06, 'learning_rate/action_model': 9.804936734975596e-05, 'epoch': 0.27})
+[2026-07-28 11:57:17,529][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.06677164137363434, 'train/grad_norm_pre_clip': 1.5504518747329712, 'timing/data': 0.0003757784143090248, 'timing/model': 0.8500159047544003, 'learning_rate/qwen_vl_interface': 9.803850007448046e-06, 'learning_rate/action_model': 9.803850007448045e-05, 'epoch': 0.27})
+[2026-07-28 11:57:26,603][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.07396278530359268, 'train/grad_norm_pre_clip': 1.7183336019515991, 'timing/data': 0.0003363369032740593, 'timing/model': 0.616555699147284, 'learning_rate/qwen_vl_interface': 9.802760324976666e-06, 'learning_rate/action_model': 9.802760324976667e-05, 'epoch': 0.27})
+[2026-07-28 11:57:35,714][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.02295006439089775, 'train/grad_norm_pre_clip': 0.8141822218894958, 'timing/data': 0.00032338593155145645, 'timing/model': 0.6180931804701686, 'learning_rate/qwen_vl_interface': 9.80166768826854e-06, 'learning_rate/action_model': 9.80166768826854e-05, 'epoch': 0.27})
+[2026-07-28 11:57:45,059][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.058635782450437546, 'train/grad_norm_pre_clip': 1.4470341205596924, 'timing/data': 0.0004033641889691353, 'timing/model': 0.8517585843801498, 'learning_rate/qwen_vl_interface': 9.800572098032668e-06, 'learning_rate/action_model': 9.800572098032668e-05, 'epoch': 0.27})
+[2026-07-28 11:57:54,375][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.04923746734857559, 'train/grad_norm_pre_clip': 2.014716863632202, 'timing/data': 0.0003283172845840454, 'timing/model': 0.8509494541212916, 'learning_rate/qwen_vl_interface': 9.799473554979966e-06, 'learning_rate/action_model': 9.799473554979965e-05, 'epoch': 0.27})
+[2026-07-28 11:58:03,460][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.053781576454639435, 'train/grad_norm_pre_clip': 0.9262841939926147, 'timing/data': 0.0005119098350405693, 'timing/model': 0.6172876777127385, 'learning_rate/qwen_vl_interface': 9.798372059823268e-06, 'learning_rate/action_model': 9.798372059823268e-05, 'epoch': 0.27})
+[2026-07-28 11:58:12,537][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.04401208832859993, 'train/grad_norm_pre_clip': 0.896116316318512, 'timing/data': 0.0003122445195913315, 'timing/model': 0.619339787401259, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.27})
+[2026-07-28 11:58:21,846][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.10741116106510162, 'train/grad_norm_pre_clip': 1.7456024885177612, 'timing/data': 0.00032453984022140503, 'timing/model': 0.6175679601728916, 'learning_rate/qwen_vl_interface': 9.79616021605879e-06, 'learning_rate/action_model': 9.796160216058788e-05, 'epoch': 0.27})
+[2026-07-28 11:58:31,156][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.04702123999595642, 'train/grad_norm_pre_clip': 2.4672837257385254, 'timing/data': 0.00034692976623773575, 'timing/model': 0.6178290024399757, 'learning_rate/qwen_vl_interface': 9.795049868886247e-06, 'learning_rate/action_model': 9.795049868886247e-05, 'epoch': 0.27})
+[2026-07-28 11:58:40,223][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.04379408061504364, 'train/grad_norm_pre_clip': 1.4591131210327148, 'timing/data': 0.0003580637276172638, 'timing/model': 0.614998578093946, 'learning_rate/qwen_vl_interface': 9.793936572480192e-06, 'learning_rate/action_model': 9.793936572480192e-05, 'epoch': 0.27})
+[2026-07-28 11:58:49,294][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.047629598528146744, 'train/grad_norm_pre_clip': 0.7591003179550171, 'timing/data': 0.0003271438181400299, 'timing/model': 0.616265487857163, 'learning_rate/qwen_vl_interface': 9.792820327563028e-06, 'learning_rate/action_model': 9.792820327563028e-05, 'epoch': 0.27})
+[2026-07-28 11:58:58,839][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.04458492621779442, 'train/grad_norm_pre_clip': 0.8675631880760193, 'timing/data': 0.00031461380422115326, 'timing/model': 0.8440602933987975, 'learning_rate/qwen_vl_interface': 9.791701134859076e-06, 'learning_rate/action_model': 9.791701134859075e-05, 'epoch': 0.27})
+[2026-07-28 11:59:07,921][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.05348639562726021, 'train/grad_norm_pre_clip': 1.6664025783538818, 'timing/data': 0.000328943133354187, 'timing/model': 0.617836382240057, 'learning_rate/qwen_vl_interface': 9.790578995094565e-06, 'learning_rate/action_model': 9.790578995094563e-05, 'epoch': 0.27})
+[2026-07-28 11:59:16,980][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.04226082190871239, 'train/grad_norm_pre_clip': 1.171057105064392, 'timing/data': 0.0003369022160768509, 'timing/model': 0.6176764527335763, 'learning_rate/qwen_vl_interface': 9.789453908997639e-06, 'learning_rate/action_model': 9.789453908997638e-05, 'epoch': 0.27})
+[2026-07-28 11:59:26,287][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.05567622557282448, 'train/grad_norm_pre_clip': 1.0476282835006714, 'timing/data': 0.0003117574378848076, 'timing/model': 0.8459750013425946, 'learning_rate/qwen_vl_interface': 9.788325877298355e-06, 'learning_rate/action_model': 9.788325877298355e-05, 'epoch': 0.27})
+[2026-07-28 11:59:35,350][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.035114072263240814, 'train/grad_norm_pre_clip': 1.2355022430419922, 'timing/data': 0.0003232872113585472, 'timing/model': 0.617152389138937, 'learning_rate/qwen_vl_interface': 9.787194900728682e-06, 'learning_rate/action_model': 9.787194900728683e-05, 'epoch': 0.27})
+[2026-07-28 11:59:44,638][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.05279427021741867, 'train/grad_norm_pre_clip': 1.5999982357025146, 'timing/data': 0.00031053368002176285, 'timing/model': 0.6144177243113518, 'learning_rate/qwen_vl_interface': 9.786060980022499e-06, 'learning_rate/action_model': 9.786060980022498e-05, 'epoch': 0.27})
+[2026-07-28 11:59:53,720][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.05068244785070419, 'train/grad_norm_pre_clip': 2.4290809631347656, 'timing/data': 0.0003201216459274292, 'timing/model': 0.616182160563767, 'learning_rate/qwen_vl_interface': 9.784924115915592e-06, 'learning_rate/action_model': 9.784924115915592e-05, 'epoch': 0.27})
+[2026-07-28 12:00:02,820][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.0787665918469429, 'train/grad_norm_pre_clip': 3.2053158283233643, 'timing/data': 0.00032483506947755814, 'timing/model': 0.6171495113521814, 'learning_rate/qwen_vl_interface': 9.783784309145663e-06, 'learning_rate/action_model': 9.783784309145661e-05, 'epoch': 0.27})
+[2026-07-28 12:00:12,136][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.032082416117191315, 'train/grad_norm_pre_clip': 1.625404715538025, 'timing/data': 0.0003369078040122986, 'timing/model': 0.6168311703950167, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.27})
+[2026-07-28 12:00:21,224][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.014192653819918633, 'train/grad_norm_pre_clip': 0.7803434729576111, 'timing/data': 0.00031800102442502975, 'timing/model': 0.6176828118041158, 'learning_rate/qwen_vl_interface': 9.78149587057707e-06, 'learning_rate/action_model': 9.78149587057707e-05, 'epoch': 0.27})
+[2026-07-28 12:00:30,529][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.12470995634794235, 'train/grad_norm_pre_clip': 2.126225233078003, 'timing/data': 0.00029650796204805374, 'timing/model': 0.8455030424520373, 'learning_rate/qwen_vl_interface': 9.780347240263354e-06, 'learning_rate/action_model': 9.780347240263353e-05, 'epoch': 0.27})
+[2026-07-28 12:00:39,606][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.05293286591768265, 'train/grad_norm_pre_clip': 2.58646559715271, 'timing/data': 0.0003333110362291336, 'timing/model': 0.6172094866633415, 'learning_rate/qwen_vl_interface': 9.779195670256498e-06, 'learning_rate/action_model': 9.779195670256497e-05, 'epoch': 0.27})
+[2026-07-28 12:00:48,934][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.06454693526029587, 'train/grad_norm_pre_clip': 1.0717827081680298, 'timing/data': 0.0003174440935254097, 'timing/model': 0.6177491564303637, 'learning_rate/qwen_vl_interface': 9.778041161303746e-06, 'learning_rate/action_model': 9.778041161303745e-05, 'epoch': 0.27})
+[2026-07-28 12:00:58,001][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.029936540871858597, 'train/grad_norm_pre_clip': 0.9171389937400818, 'timing/data': 0.0003700442612171173, 'timing/model': 0.6169313322752714, 'learning_rate/qwen_vl_interface': 9.776883714154242e-06, 'learning_rate/action_model': 9.776883714154241e-05, 'epoch': 0.27})
+[2026-07-28 12:01:07,091][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.040270011872053146, 'train/grad_norm_pre_clip': 1.1435261964797974, 'timing/data': 0.00030972063541412354, 'timing/model': 0.6176952365785837, 'learning_rate/qwen_vl_interface': 9.775723329559044e-06, 'learning_rate/action_model': 9.775723329559044e-05, 'epoch': 0.27})
+[2026-07-28 12:01:16,588][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.05986475571990013, 'train/grad_norm_pre_clip': 1.7502610683441162, 'timing/data': 0.0005453471094369888, 'timing/model': 0.7878308761864901, 'learning_rate/qwen_vl_interface': 9.774560008271114e-06, 'learning_rate/action_model': 9.774560008271114e-05, 'epoch': 0.27})
+[2026-07-28 12:01:25,697][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.04200323671102524, 'train/grad_norm_pre_clip': 1.5544949769973755, 'timing/data': 0.00036131124943494797, 'timing/model': 0.6160608716309071, 'learning_rate/qwen_vl_interface': 9.773393751045315e-06, 'learning_rate/action_model': 9.773393751045315e-05, 'epoch': 0.27})
+[2026-07-28 12:01:34,801][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.04814199358224869, 'train/grad_norm_pre_clip': 0.946046769618988, 'timing/data': 0.0003566788509488106, 'timing/model': 0.6246654791757464, 'learning_rate/qwen_vl_interface': 9.772224558638422e-06, 'learning_rate/action_model': 9.772224558638422e-05, 'epoch': 0.27})
+[2026-07-28 12:01:44,209][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.04558087885379791, 'train/grad_norm_pre_clip': 1.571349859237671, 'timing/data': 0.0003153681755065918, 'timing/model': 0.6231337292119861, 'learning_rate/qwen_vl_interface': 9.77105243180911e-06, 'learning_rate/action_model': 9.771052431809109e-05, 'epoch': 0.27})
+[2026-07-28 12:01:53,609][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.07660192996263504, 'train/grad_norm_pre_clip': 1.368944764137268, 'timing/data': 0.0003496231511235237, 'timing/model': 0.8493241239339113, 'learning_rate/qwen_vl_interface': 9.76987737131796e-06, 'learning_rate/action_model': 9.76987737131796e-05, 'epoch': 0.27})
+[2026-07-28 12:02:02,711][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.05171320214867592, 'train/grad_norm_pre_clip': 0.9754272103309631, 'timing/data': 0.00032266322523355484, 'timing/model': 0.6165213845670223, 'learning_rate/qwen_vl_interface': 9.768699377927454e-06, 'learning_rate/action_model': 9.768699377927453e-05, 'epoch': 0.27})
+[2026-07-28 12:02:11,842][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.03841588646173477, 'train/grad_norm_pre_clip': 1.6523005962371826, 'timing/data': 0.0004647374153137207, 'timing/model': 0.6187584139406681, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.27})
+[2026-07-28 12:02:21,247][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.04651257395744324, 'train/grad_norm_pre_clip': 0.8027244806289673, 'timing/data': 0.0003620665520429611, 'timing/model': 0.6198103362694383, 'learning_rate/qwen_vl_interface': 9.76633459550783e-06, 'learning_rate/action_model': 9.76633459550783e-05, 'epoch': 0.27})
+[2026-07-28 12:02:30,654][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.03941582888364792, 'train/grad_norm_pre_clip': 0.5971012115478516, 'timing/data': 0.0004414990544319153, 'timing/model': 0.6179055338725448, 'learning_rate/qwen_vl_interface': 9.765147808013193e-06, 'learning_rate/action_model': 9.765147808013192e-05, 'epoch': 0.27})
+[2026-07-28 12:02:39,758][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.09572401642799377, 'train/grad_norm_pre_clip': 3.2707881927490234, 'timing/data': 0.0003313245251774788, 'timing/model': 0.6166112087666988, 'learning_rate/qwen_vl_interface': 9.763958090688165e-06, 'learning_rate/action_model': 9.763958090688164e-05, 'epoch': 0.27})
+[2026-07-28 12:02:48,861][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.08661617338657379, 'train/grad_norm_pre_clip': 1.7803094387054443, 'timing/data': 0.00038852542638778687, 'timing/model': 0.6172694144770503, 'learning_rate/qwen_vl_interface': 9.762765444304737e-06, 'learning_rate/action_model': 9.762765444304737e-05, 'epoch': 0.27})
+[2026-07-28 12:02:58,181][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.06747306138277054, 'train/grad_norm_pre_clip': 1.1766228675842285, 'timing/data': 0.0003295261412858963, 'timing/model': 0.6179667608812451, 'learning_rate/qwen_vl_interface': 9.761569869636808e-06, 'learning_rate/action_model': 9.761569869636808e-05, 'epoch': 0.27})
+[2026-07-28 12:03:07,519][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.08215516805648804, 'train/grad_norm_pre_clip': 1.4893100261688232, 'timing/data': 0.0003736354410648346, 'timing/model': 0.6170828863978386, 'learning_rate/qwen_vl_interface': 9.760371367460172e-06, 'learning_rate/action_model': 9.760371367460171e-05, 'epoch': 0.27})
+[2026-07-28 12:03:16,600][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.043801773339509964, 'train/grad_norm_pre_clip': 0.8010393381118774, 'timing/data': 0.0003741588443517685, 'timing/model': 0.6169183338060975, 'learning_rate/qwen_vl_interface': 9.759169938552524e-06, 'learning_rate/action_model': 9.759169938552524e-05, 'epoch': 0.27})
+[2026-07-28 12:03:25,706][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.05164545029401779, 'train/grad_norm_pre_clip': 1.2914775609970093, 'timing/data': 0.0003551691770553589, 'timing/model': 0.6189100043848157, 'learning_rate/qwen_vl_interface': 9.757965583693458e-06, 'learning_rate/action_model': 9.757965583693457e-05, 'epoch': 0.28})
+[2026-07-28 12:03:35,168][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.0474548377096653, 'train/grad_norm_pre_clip': 0.8863987326622009, 'timing/data': 0.0003187805414199829, 'timing/model': 0.9961649533361197, 'learning_rate/qwen_vl_interface': 9.756758303664467e-06, 'learning_rate/action_model': 9.756758303664468e-05, 'epoch': 0.28})
+[2026-07-28 12:03:44,256][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1500
+[2026-07-28 12:03:54,690][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1500
+[2026-07-28 12:19:11,534][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500
+[2026-07-28 12:19:11,539][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.043373942375183105, 'train/grad_norm_pre_clip': 0.7147879004478455, 'eval/action_loss/seconds': 10.43071640189737, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.048856381326913834, 'eval/latency_3/loss': 0.048856381326913834, 'eval/flappy/loss': 0.048856381326913834, 'eval/flappy/latency_3/loss': 0.048856381326913834, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 332.3600039117038, 'rl_games_eval/mid_train/mean_length': 2695.55, 'rl_games_eval/mid_train/std_reward': 174.17145606979872, 'rl_games_eval/mid_train/std_length': 1403.4020263274526, 'rl_games_eval/mid_train/macro_mean_reward': 332.3600039117038, 'rl_games_eval/mid_train/macro_mean_length': 2695.55, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 332.3600039117038, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2695.55, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 174.17145606979872, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1403.4020263274526, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 332.3600039117038, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003870902583003044, 'timing/model': 0.617262882180512, 'learning_rate/qwen_vl_interface': 9.755548099248942e-06, 'learning_rate/action_model': 9.75554809924894e-05, 'epoch': 0.28})
+[2026-07-28 12:19:11,542][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1500_state
+[2026-07-28 12:19:11,542][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 12:19:49,469][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1500_state/pytorch_model
+[2026-07-28 12:19:49,470][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1500_state/sampler.bin
+[2026-07-28 12:19:49,470][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1500_state/sampler_1.bin
+[2026-07-28 12:19:49,475][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_1500_state/random_states_0.pkl
+[2026-07-28 12:19:57,581][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500
+[2026-07-28 12:19:57,582][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 12:19:57,603][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 12:20:07,190][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.05254390090703964, 'train/grad_norm_pre_clip': 1.1189053058624268, 'timing/data': 0.00039886217564344406, 'timing/model': 0.666709597222507, 'learning_rate/qwen_vl_interface': 9.75433497123217e-06, 'learning_rate/action_model': 9.754334971232171e-05, 'epoch': 0.28})
+[2026-07-28 12:20:16,686][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.08851814270019531, 'train/grad_norm_pre_clip': 1.8629565238952637, 'timing/data': 0.0003351988270878792, 'timing/model': 0.6188141535967588, 'learning_rate/qwen_vl_interface': 9.753118920401341e-06, 'learning_rate/action_model': 9.75311892040134e-05, 'epoch': 0.28})
+[2026-07-28 12:20:26,142][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.03811000660061836, 'train/grad_norm_pre_clip': 1.086137056350708, 'timing/data': 0.0003210296854376793, 'timing/model': 0.8423034269362688, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.28})
+[2026-07-28 12:20:35,413][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.020914822816848755, 'train/grad_norm_pre_clip': 0.4336574077606201, 'timing/data': 0.0003890702500939369, 'timing/model': 0.6173971854150295, 'learning_rate/qwen_vl_interface': 9.750678053455725e-06, 'learning_rate/action_model': 9.750678053455725e-05, 'epoch': 0.28})
+[2026-07-28 12:20:44,701][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.02837439998984337, 'train/grad_norm_pre_clip': 0.9872249364852905, 'timing/data': 0.00039142929017543793, 'timing/model': 0.842028190381825, 'learning_rate/qwen_vl_interface': 9.749453238924793e-06, 'learning_rate/action_model': 9.749453238924792e-05, 'epoch': 0.28})
+[2026-07-28 12:20:53,802][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.06924708932638168, 'train/grad_norm_pre_clip': 2.655020236968994, 'timing/data': 0.0003770105540752411, 'timing/model': 0.6166241364553571, 'learning_rate/qwen_vl_interface': 9.7482255047475e-06, 'learning_rate/action_model': 9.7482255047475e-05, 'epoch': 0.28})
+[2026-07-28 12:21:02,897][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.061059240251779556, 'train/grad_norm_pre_clip': 1.2827541828155518, 'timing/data': 0.0003984197974205017, 'timing/model': 0.6163113014772534, 'learning_rate/qwen_vl_interface': 9.746994851720518e-06, 'learning_rate/action_model': 9.746994851720517e-05, 'epoch': 0.28})
+[2026-07-28 12:21:11,758][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.03806576877832413, 'train/grad_norm_pre_clip': 1.32001793384552, 'timing/data': 0.00038118381053209305, 'timing/model': 0.6167291905730963, 'learning_rate/qwen_vl_interface': 9.745761280642397e-06, 'learning_rate/action_model': 9.745761280642397e-05, 'epoch': 0.28})
+[2026-07-28 12:21:20,628][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.025809163227677345, 'train/grad_norm_pre_clip': 0.6371992826461792, 'timing/data': 0.0003748266026377678, 'timing/model': 0.6168179241940379, 'learning_rate/qwen_vl_interface': 9.744524792313591e-06, 'learning_rate/action_model': 9.74452479231359e-05, 'epoch': 0.28})
+[2026-07-28 12:21:29,735][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.06616085767745972, 'train/grad_norm_pre_clip': 3.440885305404663, 'timing/data': 0.0003793453797698021, 'timing/model': 0.6197065711021423, 'learning_rate/qwen_vl_interface': 9.743285387536445e-06, 'learning_rate/action_model': 9.743285387536444e-05, 'epoch': 0.28})
+[2026-07-28 12:21:38,596][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.04145408421754837, 'train/grad_norm_pre_clip': 1.0571023225784302, 'timing/data': 0.0003722384572029114, 'timing/model': 0.6154527626931667, 'learning_rate/qwen_vl_interface': 9.742043067115193e-06, 'learning_rate/action_model': 9.742043067115192e-05, 'epoch': 0.28})
+[2026-07-28 12:21:47,646][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.029113315045833588, 'train/grad_norm_pre_clip': 1.2051639556884766, 'timing/data': 0.00037253834307193756, 'timing/model': 0.6160646229982376, 'learning_rate/qwen_vl_interface': 9.740797831855965e-06, 'learning_rate/action_model': 9.740797831855965e-05, 'epoch': 0.28})
+[2026-07-28 12:21:56,722][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.05612380430102348, 'train/grad_norm_pre_clip': 1.6970014572143555, 'timing/data': 0.00033483467996120453, 'timing/model': 0.6179048325866461, 'learning_rate/qwen_vl_interface': 9.73954968256678e-06, 'learning_rate/action_model': 9.73954968256678e-05, 'epoch': 0.28})
+[2026-07-28 12:22:05,808][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.027712108567357063, 'train/grad_norm_pre_clip': 1.527653694152832, 'timing/data': 0.00036782026290893555, 'timing/model': 0.6177807850763202, 'learning_rate/qwen_vl_interface': 9.738298620057547e-06, 'learning_rate/action_model': 9.738298620057546e-05, 'epoch': 0.28})
+[2026-07-28 12:22:15,115][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.06562919914722443, 'train/grad_norm_pre_clip': 1.363441824913025, 'timing/data': 0.00037449225783348083, 'timing/model': 0.8470214484259486, 'learning_rate/qwen_vl_interface': 9.73704464514007e-06, 'learning_rate/action_model': 9.737044645140068e-05, 'epoch': 0.28})
+[2026-07-28 12:22:24,440][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.06343892961740494, 'train/grad_norm_pre_clip': 2.1397407054901123, 'timing/data': 0.0003783348947763443, 'timing/model': 0.6199872959405184, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.28})
+[2026-07-28 12:22:33,516][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.045507270842790604, 'train/grad_norm_pre_clip': 0.9583513736724854, 'timing/data': 0.0006176382303237915, 'timing/model': 0.6180042792111635, 'learning_rate/qwen_vl_interface': 9.734527961337028e-06, 'learning_rate/action_model': 9.734527961337028e-05, 'epoch': 0.28})
+[2026-07-28 12:22:42,593][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.05091805383563042, 'train/grad_norm_pre_clip': 1.6614618301391602, 'timing/data': 0.0004018489271402359, 'timing/model': 0.6154081020504236, 'learning_rate/qwen_vl_interface': 9.733265254084514e-06, 'learning_rate/action_model': 9.733265254084513e-05, 'epoch': 0.28})
+[2026-07-28 12:22:51,693][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.0627983883023262, 'train/grad_norm_pre_clip': 2.095351457595825, 'timing/data': 0.0003843745216727257, 'timing/model': 0.6169717069715261, 'learning_rate/qwen_vl_interface': 9.73199963768985e-06, 'learning_rate/action_model': 9.731999637689849e-05, 'epoch': 0.28})
+[2026-07-28 12:23:01,005][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.0657511055469513, 'train/grad_norm_pre_clip': 1.2669517993927002, 'timing/data': 0.00036170799285173416, 'timing/model': 0.615635285153985, 'learning_rate/qwen_vl_interface': 9.73073111297428e-06, 'learning_rate/action_model': 9.73073111297428e-05, 'epoch': 0.28})
+[2026-07-28 12:23:10,311][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.062462352216243744, 'train/grad_norm_pre_clip': 1.7390319108963013, 'timing/data': 0.00038281548768281937, 'timing/model': 0.615931436419487, 'learning_rate/qwen_vl_interface': 9.72945968076094e-06, 'learning_rate/action_model': 9.729459680760939e-05, 'epoch': 0.28})
+[2026-07-28 12:23:19,376][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.04458194226026535, 'train/grad_norm_pre_clip': 2.225229501724243, 'timing/data': 0.00038569699972867966, 'timing/model': 0.6166917644441128, 'learning_rate/qwen_vl_interface': 9.728185341874844e-06, 'learning_rate/action_model': 9.728185341874844e-05, 'epoch': 0.28})
+[2026-07-28 12:23:28,782][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.07306647300720215, 'train/grad_norm_pre_clip': 2.9699466228485107, 'timing/data': 0.0003638779744505882, 'timing/model': 0.6169858984649181, 'learning_rate/qwen_vl_interface': 9.7269080971429e-06, 'learning_rate/action_model': 9.726908097142899e-05, 'epoch': 0.28})
+[2026-07-28 12:23:37,848][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.14590102434158325, 'train/grad_norm_pre_clip': 5.010130882263184, 'timing/data': 0.0003753090277314186, 'timing/model': 0.6164481537416577, 'learning_rate/qwen_vl_interface': 9.725627947393896e-06, 'learning_rate/action_model': 9.725627947393896e-05, 'epoch': 0.28})
+[2026-07-28 12:23:47,165][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.059790223836898804, 'train/grad_norm_pre_clip': 2.213052749633789, 'timing/data': 0.00039896368980407715, 'timing/model': 0.6152648692950606, 'learning_rate/qwen_vl_interface': 9.72434489345851e-06, 'learning_rate/action_model': 9.72434489345851e-05, 'epoch': 0.28})
+[2026-07-28 12:23:56,248][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.05985284596681595, 'train/grad_norm_pre_clip': 2.2206718921661377, 'timing/data': 0.0003845691680908203, 'timing/model': 0.6161450622603297, 'learning_rate/qwen_vl_interface': 9.723058936169298e-06, 'learning_rate/action_model': 9.723058936169297e-05, 'epoch': 0.28})
+[2026-07-28 12:24:05,592][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.0638958141207695, 'train/grad_norm_pre_clip': 1.4434168338775635, 'timing/data': 0.00037156976759433746, 'timing/model': 0.6159748593345284, 'learning_rate/qwen_vl_interface': 9.721770076360707e-06, 'learning_rate/action_model': 9.721770076360706e-05, 'epoch': 0.28})
+[2026-07-28 12:24:14,845][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.040724977850914, 'train/grad_norm_pre_clip': 0.8064927458763123, 'timing/data': 0.0003949543461203575, 'timing/model': 0.786752556450665, 'learning_rate/qwen_vl_interface': 9.720478314869061e-06, 'learning_rate/action_model': 9.72047831486906e-05, 'epoch': 0.28})
+[2026-07-28 12:24:23,919][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.06207767874002457, 'train/grad_norm_pre_clip': 1.0109630823135376, 'timing/data': 0.00037438515573740005, 'timing/model': 0.615135820582509, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.28})
+[2026-07-28 12:24:33,004][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.08964381366968155, 'train/grad_norm_pre_clip': 1.655908226966858, 'timing/data': 0.0003833109512925148, 'timing/model': 0.6131660724058747, 'learning_rate/qwen_vl_interface': 9.717886090191334e-06, 'learning_rate/action_model': 9.717886090191333e-05, 'epoch': 0.28})
+[2026-07-28 12:24:42,328][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.04670323431491852, 'train/grad_norm_pre_clip': 1.5918577909469604, 'timing/data': 0.0003721090033650398, 'timing/model': 0.6157768303528428, 'learning_rate/qwen_vl_interface': 9.716585628687317e-06, 'learning_rate/action_model': 9.716585628687316e-05, 'epoch': 0.28})
+[2026-07-28 12:24:51,646][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.05388839170336723, 'train/grad_norm_pre_clip': 0.9109488725662231, 'timing/data': 0.00037514325231313705, 'timing/model': 0.8521578768268228, 'learning_rate/qwen_vl_interface': 9.715282268864379e-06, 'learning_rate/action_model': 9.715282268864378e-05, 'epoch': 0.28})
+[2026-07-28 12:25:00,706][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.07536476105451584, 'train/grad_norm_pre_clip': 1.9398154020309448, 'timing/data': 0.0003893766552209854, 'timing/model': 0.6173924803733826, 'learning_rate/qwen_vl_interface': 9.713976011568256e-06, 'learning_rate/action_model': 9.713976011568256e-05, 'epoch': 0.28})
+[2026-07-28 12:25:09,790][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.07066197693347931, 'train/grad_norm_pre_clip': 2.7309794425964355, 'timing/data': 0.00036522839218378067, 'timing/model': 0.6187364384531975, 'learning_rate/qwen_vl_interface': 9.712666857646565e-06, 'learning_rate/action_model': 9.712666857646564e-05, 'epoch': 0.28})
+[2026-07-28 12:25:19,097][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.06709924340248108, 'train/grad_norm_pre_clip': 2.925527334213257, 'timing/data': 0.00038523413240909576, 'timing/model': 0.8422984145581722, 'learning_rate/qwen_vl_interface': 9.711354807948798e-06, 'learning_rate/action_model': 9.711354807948798e-05, 'epoch': 0.28})
+[2026-07-28 12:25:28,432][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.031455256044864655, 'train/grad_norm_pre_clip': 0.5995340943336487, 'timing/data': 0.0003733299672603607, 'timing/model': 0.6190452165901661, 'learning_rate/qwen_vl_interface': 9.710039863326334e-06, 'learning_rate/action_model': 9.710039863326334e-05, 'epoch': 0.28})
+[2026-07-28 12:25:37,501][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.053454771637916565, 'train/grad_norm_pre_clip': 1.5038121938705444, 'timing/data': 0.0003773905336856842, 'timing/model': 0.6150926994159818, 'learning_rate/qwen_vl_interface': 9.708722024632425e-06, 'learning_rate/action_model': 9.708722024632424e-05, 'epoch': 0.28})
+[2026-07-28 12:25:46,367][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.03555150330066681, 'train/grad_norm_pre_clip': 1.3225480318069458, 'timing/data': 0.0003914590924978256, 'timing/model': 0.6162516418844461, 'learning_rate/qwen_vl_interface': 9.7074012927222e-06, 'learning_rate/action_model': 9.7074012927222e-05, 'epoch': 0.28})
+[2026-07-28 12:25:55,462][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.03134515881538391, 'train/grad_norm_pre_clip': 1.025529146194458, 'timing/data': 0.0003297850489616394, 'timing/model': 0.6169046750292182, 'learning_rate/qwen_vl_interface': 9.706077668452671e-06, 'learning_rate/action_model': 9.706077668452671e-05, 'epoch': 0.28})
+[2026-07-28 12:26:04,790][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.07832741737365723, 'train/grad_norm_pre_clip': 3.0192196369171143, 'timing/data': 0.0003740862011909485, 'timing/model': 0.6170216584578156, 'learning_rate/qwen_vl_interface': 9.704751152682722e-06, 'learning_rate/action_model': 9.704751152682721e-05, 'epoch': 0.28})
+[2026-07-28 12:26:14,112][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.058926817029714584, 'train/grad_norm_pre_clip': 2.7336008548736572, 'timing/data': 0.00032021570950746536, 'timing/model': 0.6177912745624781, 'learning_rate/qwen_vl_interface': 9.703421746273115e-06, 'learning_rate/action_model': 9.703421746273114e-05, 'epoch': 0.28})
+[2026-07-28 12:26:23,180][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.045495666563510895, 'train/grad_norm_pre_clip': 1.8184674978256226, 'timing/data': 0.0003496278077363968, 'timing/model': 0.6152974832803011, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.28})
+[2026-07-28 12:26:32,428][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.07752391695976257, 'train/grad_norm_pre_clip': 2.410597801208496, 'timing/data': 0.00032495055347681046, 'timing/model': 0.793630488216877, 'learning_rate/qwen_vl_interface': 9.700754264987347e-06, 'learning_rate/action_model': 9.700754264987347e-05, 'epoch': 0.28})
+[2026-07-28 12:26:41,481][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.07761389017105103, 'train/grad_norm_pre_clip': 2.3370978832244873, 'timing/data': 0.00044482573866844177, 'timing/model': 0.616192233748734, 'learning_rate/qwen_vl_interface': 9.69941619184209e-06, 'learning_rate/action_model': 9.699416191842089e-05, 'epoch': 0.28})
+[2026-07-28 12:26:50,757][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.08160628378391266, 'train/grad_norm_pre_clip': 1.9232420921325684, 'timing/data': 0.0003308309242129326, 'timing/model': 0.8393081072717905, 'learning_rate/qwen_vl_interface': 9.69807523151897e-06, 'learning_rate/action_model': 9.698075231518969e-05, 'epoch': 0.28})
+[2026-07-28 12:26:59,830][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.04490768909454346, 'train/grad_norm_pre_clip': 1.4911961555480957, 'timing/data': 0.0003161020576953888, 'timing/model': 0.61482107732445, 'learning_rate/qwen_vl_interface': 9.696731384888125e-06, 'learning_rate/action_model': 9.696731384888125e-05, 'epoch': 0.28})
+[2026-07-28 12:27:09,139][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.03836885839700699, 'train/grad_norm_pre_clip': 1.1840633153915405, 'timing/data': 0.00035027600824832916, 'timing/model': 0.8463769713416696, 'learning_rate/qwen_vl_interface': 9.695384652821562e-06, 'learning_rate/action_model': 9.695384652821561e-05, 'epoch': 0.28})
+[2026-07-28 12:27:18,181][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.05197111889719963, 'train/grad_norm_pre_clip': 1.4726877212524414, 'timing/data': 0.0003835922107100487, 'timing/model': 0.6149664642289281, 'learning_rate/qwen_vl_interface': 9.694035036193162e-06, 'learning_rate/action_model': 9.694035036193161e-05, 'epoch': 0.28})
+[2026-07-28 12:27:27,471][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.05794866383075714, 'train/grad_norm_pre_clip': 1.378104567527771, 'timing/data': 0.0003782017156481743, 'timing/model': 0.6160120870918036, 'learning_rate/qwen_vl_interface': 9.692682535878674e-06, 'learning_rate/action_model': 9.692682535878673e-05, 'epoch': 0.28})
+[2026-07-28 12:27:36,542][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.03879319876432419, 'train/grad_norm_pre_clip': 1.2398658990859985, 'timing/data': 0.00034740567207336426, 'timing/model': 0.6159630455076694, 'learning_rate/qwen_vl_interface': 9.691327152755721e-06, 'learning_rate/action_model': 9.691327152755722e-05, 'epoch': 0.28})
+[2026-07-28 12:27:45,645][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.04796861857175827, 'train/grad_norm_pre_clip': 1.6568951606750488, 'timing/data': 0.0003294004127383232, 'timing/model': 0.6161549175158143, 'learning_rate/qwen_vl_interface': 9.6899688877038e-06, 'learning_rate/action_model': 9.689968887703798e-05, 'epoch': 0.28})
+[2026-07-28 12:27:54,967][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.02687694877386093, 'train/grad_norm_pre_clip': 1.0510209798812866, 'timing/data': 0.00036090798676013947, 'timing/model': 0.8533246116712689, 'learning_rate/qwen_vl_interface': 9.688607741604269e-06, 'learning_rate/action_model': 9.688607741604269e-05, 'epoch': 0.29})
+[2026-07-28 12:28:04,022][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.04946527257561684, 'train/grad_norm_pre_clip': 1.512717843055725, 'timing/data': 0.000345524400472641, 'timing/model': 0.61578468978405, 'learning_rate/qwen_vl_interface': 9.687243715340365e-06, 'learning_rate/action_model': 9.687243715340364e-05, 'epoch': 0.29})
+[2026-07-28 12:28:13,075][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.03302500396966934, 'train/grad_norm_pre_clip': 1.2528740167617798, 'timing/data': 0.000316510908305645, 'timing/model': 0.6212039086967707, 'learning_rate/qwen_vl_interface': 9.685876809797188e-06, 'learning_rate/action_model': 9.685876809797188e-05, 'epoch': 0.29})
+[2026-07-28 12:28:22,176][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.05210211127996445, 'train/grad_norm_pre_clip': 2.2779970169067383, 'timing/data': 0.00034018978476524353, 'timing/model': 0.6170709300786257, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.29})
+[2026-07-28 12:28:31,276][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.04588446766138077, 'train/grad_norm_pre_clip': 1.3608030080795288, 'timing/data': 0.00036080460995435715, 'timing/model': 0.6160042202100158, 'learning_rate/qwen_vl_interface': 9.683134364422766e-06, 'learning_rate/action_model': 9.683134364422765e-05, 'epoch': 0.29})
+[2026-07-28 12:28:40,612][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.05380193889141083, 'train/grad_norm_pre_clip': 2.8327383995056152, 'timing/data': 0.000394442118704319, 'timing/model': 0.6198697788640857, 'learning_rate/qwen_vl_interface': 9.681758826371064e-06, 'learning_rate/action_model': 9.681758826371063e-05, 'epoch': 0.29})
+[2026-07-28 12:28:50,096][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.037057917565107346, 'train/grad_norm_pre_clip': 1.226065754890442, 'timing/data': 0.00032931286841630936, 'timing/model': 0.6192284412682056, 'learning_rate/qwen_vl_interface': 9.680380412599174e-06, 'learning_rate/action_model': 9.680380412599173e-05, 'epoch': 0.29})
+[2026-07-28 12:28:59,235][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.06243810057640076, 'train/grad_norm_pre_clip': 1.3920499086380005, 'timing/data': 0.0003442252054810524, 'timing/model': 0.6163704618811607, 'learning_rate/qwen_vl_interface': 9.678999124001533e-06, 'learning_rate/action_model': 9.678999124001533e-05, 'epoch': 0.29})
+[2026-07-28 12:29:08,404][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.10594233870506287, 'train/grad_norm_pre_clip': 1.9729759693145752, 'timing/data': 0.00033870991319417953, 'timing/model': 0.6154288407415152, 'learning_rate/qwen_vl_interface': 9.67761496147445e-06, 'learning_rate/action_model': 9.677614961474449e-05, 'epoch': 0.29})
+[2026-07-28 12:29:17,540][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.09622310101985931, 'train/grad_norm_pre_clip': 3.698198080062866, 'timing/data': 0.00033497903496026993, 'timing/model': 0.6162768350914121, 'learning_rate/qwen_vl_interface': 9.676227925916082e-06, 'learning_rate/action_model': 9.676227925916082e-05, 'epoch': 0.29})
+[2026-07-28 12:29:26,886][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.08571877330541611, 'train/grad_norm_pre_clip': 3.1489884853363037, 'timing/data': 0.0003976495936512947, 'timing/model': 0.8650525929406285, 'learning_rate/qwen_vl_interface': 9.674838018226472e-06, 'learning_rate/action_model': 9.674838018226471e-05, 'epoch': 0.29})
+[2026-07-28 12:29:35,949][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.09212302416563034, 'train/grad_norm_pre_clip': 1.9025295972824097, 'timing/data': 0.0004369094967842102, 'timing/model': 0.6176399672403932, 'learning_rate/qwen_vl_interface': 9.67344523930751e-06, 'learning_rate/action_model': 9.673445239307511e-05, 'epoch': 0.29})
+[2026-07-28 12:29:45,027][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.08530972898006439, 'train/grad_norm_pre_clip': 1.6888577938079834, 'timing/data': 0.00038214679807424545, 'timing/model': 0.6177658950909972, 'learning_rate/qwen_vl_interface': 9.672049590062959e-06, 'learning_rate/action_model': 9.672049590062958e-05, 'epoch': 0.29})
+[2026-07-28 12:29:53,896][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.0548834502696991, 'train/grad_norm_pre_clip': 1.1793948411941528, 'timing/data': 0.00034853070974349976, 'timing/model': 0.6153861125931144, 'learning_rate/qwen_vl_interface': 9.670651071398436e-06, 'learning_rate/action_model': 9.670651071398436e-05, 'epoch': 0.29})
+[2026-07-28 12:30:02,761][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.07873111963272095, 'train/grad_norm_pre_clip': 1.5122443437576294, 'timing/data': 0.0004157554358243942, 'timing/model': 0.6172955902293324, 'learning_rate/qwen_vl_interface': 9.669249684221427e-06, 'learning_rate/action_model': 9.669249684221428e-05, 'epoch': 0.29})
+[2026-07-28 12:30:12,281][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.023452727124094963, 'train/grad_norm_pre_clip': 0.8338281512260437, 'timing/data': 0.0003125471994280815, 'timing/model': 0.8398461705073714, 'learning_rate/qwen_vl_interface': 9.66784542944128e-06, 'learning_rate/action_model': 9.66784542944128e-05, 'epoch': 0.29})
+[2026-07-28 12:30:21,343][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.06410717964172363, 'train/grad_norm_pre_clip': 1.0433952808380127, 'timing/data': 0.0003331257030367851, 'timing/model': 0.6172949401661754, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.29})
+[2026-07-28 12:30:30,401][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.06574895977973938, 'train/grad_norm_pre_clip': 1.5728819370269775, 'timing/data': 0.00033015012741088867, 'timing/model': 0.6151403188705444, 'learning_rate/qwen_vl_interface': 9.66502832071824e-06, 'learning_rate/action_model': 9.66502832071824e-05, 'epoch': 0.29})
+[2026-07-28 12:30:39,689][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.07230843603610992, 'train/grad_norm_pre_clip': 1.4338085651397705, 'timing/data': 0.0004601338878273964, 'timing/model': 0.6173178711906075, 'learning_rate/qwen_vl_interface': 9.663615468603344e-06, 'learning_rate/action_model': 9.663615468603343e-05, 'epoch': 0.29})
+[2026-07-28 12:30:48,769][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.051767587661743164, 'train/grad_norm_pre_clip': 1.5968883037567139, 'timing/data': 0.0005654646083712578, 'timing/model': 0.6140395924448967, 'learning_rate/qwen_vl_interface': 9.662199752541287e-06, 'learning_rate/action_model': 9.662199752541287e-05, 'epoch': 0.29})
+[2026-07-28 12:30:57,856][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.04490356892347336, 'train/grad_norm_pre_clip': 1.4896115064620972, 'timing/data': 0.00032070185989141464, 'timing/model': 0.6158803487196565, 'learning_rate/qwen_vl_interface': 9.660781173450711e-06, 'learning_rate/action_model': 9.660781173450712e-05, 'epoch': 0.29})
+[2026-07-28 12:31:06,944][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.08042509853839874, 'train/grad_norm_pre_clip': 2.454216480255127, 'timing/data': 0.00033495016396045685, 'timing/model': 0.6173619637265801, 'learning_rate/qwen_vl_interface': 9.659359732252121e-06, 'learning_rate/action_model': 9.659359732252121e-05, 'epoch': 0.29})
+[2026-07-28 12:31:16,246][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.06079907715320587, 'train/grad_norm_pre_clip': 1.2265182733535767, 'timing/data': 0.00031492579728364944, 'timing/model': 0.6147199803963304, 'learning_rate/qwen_vl_interface': 9.657935429867872e-06, 'learning_rate/action_model': 9.657935429867873e-05, 'epoch': 0.29})
+[2026-07-28 12:31:25,303][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.065059594810009, 'train/grad_norm_pre_clip': 3.25417423248291, 'timing/data': 0.00039032381027936935, 'timing/model': 0.615097195841372, 'learning_rate/qwen_vl_interface': 9.65650826722218e-06, 'learning_rate/action_model': 9.656508267222179e-05, 'epoch': 0.29})
+[2026-07-28 12:31:34,346][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.05505109578371048, 'train/grad_norm_pre_clip': 2.372699737548828, 'timing/data': 0.0004082517698407173, 'timing/model': 0.6143811531364918, 'learning_rate/qwen_vl_interface': 9.655078245241114e-06, 'learning_rate/action_model': 9.655078245241112e-05, 'epoch': 0.29})
+[2026-07-28 12:31:43,415][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.059480760246515274, 'train/grad_norm_pre_clip': 2.2208898067474365, 'timing/data': 0.0003898805007338524, 'timing/model': 0.6144619323313236, 'learning_rate/qwen_vl_interface': 9.653645364852598e-06, 'learning_rate/action_model': 9.653645364852597e-05, 'epoch': 0.29})
+[2026-07-28 12:31:52,710][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.08664491772651672, 'train/grad_norm_pre_clip': 2.1641359329223633, 'timing/data': 0.00030286144465208054, 'timing/model': 0.6159536186605692, 'learning_rate/qwen_vl_interface': 9.652209626986414e-06, 'learning_rate/action_model': 9.652209626986413e-05, 'epoch': 0.29})
+[2026-07-28 12:32:02,012][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.017900459468364716, 'train/grad_norm_pre_clip': 0.4418535828590393, 'timing/data': 0.0003195274621248245, 'timing/model': 0.6144619369879365, 'learning_rate/qwen_vl_interface': 9.6507710325742e-06, 'learning_rate/action_model': 9.6507710325742e-05, 'epoch': 0.29})
+[2026-07-28 12:32:11,057][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.04585377871990204, 'train/grad_norm_pre_clip': 1.4358458518981934, 'timing/data': 0.0003120880573987961, 'timing/model': 0.616055597551167, 'learning_rate/qwen_vl_interface': 9.649329582549439e-06, 'learning_rate/action_model': 9.649329582549439e-05, 'epoch': 0.29})
+[2026-07-28 12:32:20,121][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.049515947699546814, 'train/grad_norm_pre_clip': 0.9091949462890625, 'timing/data': 0.0003329208120703697, 'timing/model': 0.6166291823610663, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.29})
+[2026-07-28 12:32:29,387][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.04872418940067291, 'train/grad_norm_pre_clip': 1.8025376796722412, 'timing/data': 0.00038680434226989746, 'timing/model': 0.8402328956872225, 'learning_rate/qwen_vl_interface': 9.646438119405503e-06, 'learning_rate/action_model': 9.646438119405503e-05, 'epoch': 0.29})
+[2026-07-28 12:32:38,465][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.036667097359895706, 'train/grad_norm_pre_clip': 1.0519965887069702, 'timing/data': 0.0003226539120078087, 'timing/model': 0.6150965299457312, 'learning_rate/qwen_vl_interface': 9.644988108162566e-06, 'learning_rate/action_model': 9.644988108162567e-05, 'epoch': 0.29})
+[2026-07-28 12:32:47,768][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.07867155969142914, 'train/grad_norm_pre_clip': 2.0580973625183105, 'timing/data': 0.000323709100484848, 'timing/model': 0.6173880454152822, 'learning_rate/qwen_vl_interface': 9.643535245059562e-06, 'learning_rate/action_model': 9.643535245059561e-05, 'epoch': 0.29})
+[2026-07-28 12:32:56,814][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.05692831426858902, 'train/grad_norm_pre_clip': 2.9722964763641357, 'timing/data': 0.0004039844498038292, 'timing/model': 0.6152966720983386, 'learning_rate/qwen_vl_interface': 9.642079531039235e-06, 'learning_rate/action_model': 9.642079531039235e-05, 'epoch': 0.29})
+[2026-07-28 12:33:05,898][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.02768673188984394, 'train/grad_norm_pre_clip': 1.487550139427185, 'timing/data': 0.00033252034336328506, 'timing/model': 0.6153779858723283, 'learning_rate/qwen_vl_interface': 9.640620967046186e-06, 'learning_rate/action_model': 9.640620967046186e-05, 'epoch': 0.29})
+[2026-07-28 12:33:14,955][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.011127525940537453, 'train/grad_norm_pre_clip': 0.49125897884368896, 'timing/data': 0.0004052063450217247, 'timing/model': 0.6165814939886332, 'learning_rate/qwen_vl_interface': 9.63915955402686e-06, 'learning_rate/action_model': 9.63915955402686e-05, 'epoch': 0.29})
+[2026-07-28 12:33:24,460][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.033249951899051666, 'train/grad_norm_pre_clip': 0.7482113838195801, 'timing/data': 0.0003166552633047104, 'timing/model': 0.8424880178645253, 'learning_rate/qwen_vl_interface': 9.637695292929552e-06, 'learning_rate/action_model': 9.637695292929552e-05, 'epoch': 0.29})
+[2026-07-28 12:33:33,504][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.020473193377256393, 'train/grad_norm_pre_clip': 1.0252726078033447, 'timing/data': 0.00039368029683828354, 'timing/model': 0.6139946766197681, 'learning_rate/qwen_vl_interface': 9.636228184704405e-06, 'learning_rate/action_model': 9.636228184704405e-05, 'epoch': 0.29})
+[2026-07-28 12:33:42,526][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.057946182787418365, 'train/grad_norm_pre_clip': 2.016037940979004, 'timing/data': 0.0003166431561112404, 'timing/model': 0.6155935414135456, 'learning_rate/qwen_vl_interface': 9.634758230303408e-06, 'learning_rate/action_model': 9.634758230303408e-05, 'epoch': 0.29})
+[2026-07-28 12:33:51,582][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.06646571308374405, 'train/grad_norm_pre_clip': 1.479650855064392, 'timing/data': 0.0004060734063386917, 'timing/model': 0.6158310128375888, 'learning_rate/qwen_vl_interface': 9.633285430680401e-06, 'learning_rate/action_model': 9.633285430680401e-05, 'epoch': 0.29})
+[2026-07-28 12:34:00,873][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.07139171659946442, 'train/grad_norm_pre_clip': 1.6169548034667969, 'timing/data': 0.00046592019498348236, 'timing/model': 0.8420909466221929, 'learning_rate/qwen_vl_interface': 9.631809786791066e-06, 'learning_rate/action_model': 9.631809786791066e-05, 'epoch': 0.29})
+[2026-07-28 12:34:10,180][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.06136765331029892, 'train/grad_norm_pre_clip': 3.149357795715332, 'timing/data': 0.0003154110163450241, 'timing/model': 0.6162258321419358, 'learning_rate/qwen_vl_interface': 9.630331299592932e-06, 'learning_rate/action_model': 9.630331299592933e-05, 'epoch': 0.29})
+[2026-07-28 12:34:19,264][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.04439621791243553, 'train/grad_norm_pre_clip': 1.858156442642212, 'timing/data': 0.00038658734411001205, 'timing/model': 0.6152940355241299, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.29})
+[2026-07-28 12:34:28,305][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.07270446419715881, 'train/grad_norm_pre_clip': 1.4524896144866943, 'timing/data': 0.0003282269462943077, 'timing/model': 0.6151824174448848, 'learning_rate/qwen_vl_interface': 9.627365799109617e-06, 'learning_rate/action_model': 9.627365799109617e-05, 'epoch': 0.29})
+[2026-07-28 12:34:37,578][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.07450395822525024, 'train/grad_norm_pre_clip': 2.304793357849121, 'timing/data': 0.0004137679934501648, 'timing/model': 0.6152013996616006, 'learning_rate/qwen_vl_interface': 9.625878787748712e-06, 'learning_rate/action_model': 9.625878787748713e-05, 'epoch': 0.29})
+[2026-07-28 12:34:46,859][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.05508442968130112, 'train/grad_norm_pre_clip': 1.024060606956482, 'timing/data': 0.00033561699092388153, 'timing/model': 0.8474009651690722, 'learning_rate/qwen_vl_interface': 9.624388936927572e-06, 'learning_rate/action_model': 9.624388936927571e-05, 'epoch': 0.29})
+[2026-07-28 12:34:55,930][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.03477600961923599, 'train/grad_norm_pre_clip': 1.7814607620239258, 'timing/data': 0.00036063604056835175, 'timing/model': 0.6156795676797628, 'learning_rate/qwen_vl_interface': 9.622896247612943e-06, 'learning_rate/action_model': 9.622896247612942e-05, 'epoch': 0.29})
+[2026-07-28 12:35:04,964][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.07084725797176361, 'train/grad_norm_pre_clip': 1.5787149667739868, 'timing/data': 0.000308942049741745, 'timing/model': 0.6148484973236918, 'learning_rate/qwen_vl_interface': 9.621400720773414e-06, 'learning_rate/action_model': 9.621400720773414e-05, 'epoch': 0.29})
+[2026-07-28 12:35:14,236][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.041573215276002884, 'train/grad_norm_pre_clip': 1.5025417804718018, 'timing/data': 0.00032813195139169693, 'timing/model': 0.839694912545383, 'learning_rate/qwen_vl_interface': 9.619902357379416e-06, 'learning_rate/action_model': 9.619902357379415e-05, 'epoch': 0.29})
+[2026-07-28 12:35:23,585][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.048123423010110855, 'train/grad_norm_pre_clip': 1.0981956720352173, 'timing/data': 0.0003863144665956497, 'timing/model': 0.8808144070208073, 'learning_rate/qwen_vl_interface': 9.618401158403222e-06, 'learning_rate/action_model': 9.618401158403222e-05, 'epoch': 0.29})
+[2026-07-28 12:35:32,785][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.04136085882782936, 'train/grad_norm_pre_clip': 0.9221651554107666, 'timing/data': 0.0003921482712030411, 'timing/model': 0.6695286436006427, 'learning_rate/qwen_vl_interface': 9.616897124818944e-06, 'learning_rate/action_model': 9.616897124818945e-05, 'epoch': 0.29})
+[2026-07-28 12:35:41,936][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.07018426060676575, 'train/grad_norm_pre_clip': 3.057920455932617, 'timing/data': 0.0003861133009195328, 'timing/model': 0.6170125203207135, 'learning_rate/qwen_vl_interface': 9.615390257602532e-06, 'learning_rate/action_model': 9.615390257602531e-05, 'epoch': 0.29})
+[2026-07-28 12:35:51,013][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.02253848873078823, 'train/grad_norm_pre_clip': 1.3128465414047241, 'timing/data': 0.00032828189432621, 'timing/model': 0.617830628529191, 'learning_rate/qwen_vl_interface': 9.613880557731776e-06, 'learning_rate/action_model': 9.613880557731776e-05, 'epoch': 0.29})
+[2026-07-28 12:36:00,331][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.043547987937927246, 'train/grad_norm_pre_clip': 2.594912052154541, 'timing/data': 0.00032269489020109177, 'timing/model': 0.8556745387613773, 'learning_rate/qwen_vl_interface': 9.612368026186304e-06, 'learning_rate/action_model': 9.612368026186304e-05, 'epoch': 0.29})
+[2026-07-28 12:36:09,599][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.05674474686384201, 'train/grad_norm_pre_clip': 1.223831295967102, 'timing/data': 0.0003281356766819954, 'timing/model': 0.6201259652152658, 'learning_rate/qwen_vl_interface': 9.610852663947582e-06, 'learning_rate/action_model': 9.610852663947583e-05, 'epoch': 0.29})
+[2026-07-28 12:36:18,711][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.12293820083141327, 'train/grad_norm_pre_clip': 5.199396133422852, 'timing/data': 0.0003810487687587738, 'timing/model': 0.6510690366849303, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.3})
+[2026-07-28 12:36:27,792][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.05495508760213852, 'train/grad_norm_pre_clip': 2.327759027481079, 'timing/data': 0.0003832681104540825, 'timing/model': 0.6153260320425034, 'learning_rate/qwen_vl_interface': 9.607813451325431e-06, 'learning_rate/action_model': 9.60781345132543e-05, 'epoch': 0.3})
+[2026-07-28 12:36:36,868][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.10556289553642273, 'train/grad_norm_pre_clip': 3.1901135444641113, 'timing/data': 0.0003239056095480919, 'timing/model': 0.6161367343738675, 'learning_rate/qwen_vl_interface': 9.606289602914113e-06, 'learning_rate/action_model': 9.606289602914112e-05, 'epoch': 0.3})
+[2026-07-28 12:36:46,154][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.09709563106298447, 'train/grad_norm_pre_clip': 1.9078991413116455, 'timing/data': 0.0003197472542524338, 'timing/model': 0.8446413036435843, 'learning_rate/qwen_vl_interface': 9.604762927753768e-06, 'learning_rate/action_model': 9.604762927753768e-05, 'epoch': 0.3})
+[2026-07-28 12:36:55,211][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.05642775446176529, 'train/grad_norm_pre_clip': 1.632668137550354, 'timing/data': 0.00032292958348989487, 'timing/model': 0.6153986901044846, 'learning_rate/qwen_vl_interface': 9.603233426835037e-06, 'learning_rate/action_model': 9.603233426835036e-05, 'epoch': 0.3})
+[2026-07-28 12:37:04,270][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.04805568605661392, 'train/grad_norm_pre_clip': 0.637952983379364, 'timing/data': 0.0004169521853327751, 'timing/model': 0.6163403419777751, 'learning_rate/qwen_vl_interface': 9.601701101150399e-06, 'learning_rate/action_model': 9.601701101150398e-05, 'epoch': 0.3})
+[2026-07-28 12:37:13,331][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.056835100054740906, 'train/grad_norm_pre_clip': 0.7559251189231873, 'timing/data': 0.00033746473491191864, 'timing/model': 0.6149891940876842, 'learning_rate/qwen_vl_interface': 9.600165951694162e-06, 'learning_rate/action_model': 9.60016595169416e-05, 'epoch': 0.3})
+[2026-07-28 12:37:22,413][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.0935511440038681, 'train/grad_norm_pre_clip': 2.3171956539154053, 'timing/data': 0.000314970500767231, 'timing/model': 0.6159764342010021, 'learning_rate/qwen_vl_interface': 9.598627979462466e-06, 'learning_rate/action_model': 9.598627979462467e-05, 'epoch': 0.3})
+[2026-07-28 12:37:31,702][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.029506221413612366, 'train/grad_norm_pre_clip': 0.5653751492500305, 'timing/data': 0.00032295379787683487, 'timing/model': 0.6178787853568792, 'learning_rate/qwen_vl_interface': 9.59708718545329e-06, 'learning_rate/action_model': 9.597087185453288e-05, 'epoch': 0.3})
+[2026-07-28 12:37:40,812][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.06570862978696823, 'train/grad_norm_pre_clip': 1.492497444152832, 'timing/data': 0.00039579812437295914, 'timing/model': 0.6161058945581317, 'learning_rate/qwen_vl_interface': 9.595543570666433e-06, 'learning_rate/action_model': 9.595543570666432e-05, 'epoch': 0.3})
+[2026-07-28 12:37:50,108][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.06248519569635391, 'train/grad_norm_pre_clip': 1.3511216640472412, 'timing/data': 0.0003909161314368248, 'timing/model': 0.6148088844493032, 'learning_rate/qwen_vl_interface': 9.593997136103533e-06, 'learning_rate/action_model': 9.593997136103533e-05, 'epoch': 0.3})
+[2026-07-28 12:37:59,373][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.07447271049022675, 'train/grad_norm_pre_clip': 1.322663426399231, 'timing/data': 0.0003278590738773346, 'timing/model': 0.8216382497921586, 'learning_rate/qwen_vl_interface': 9.592447882768053e-06, 'learning_rate/action_model': 9.592447882768052e-05, 'epoch': 0.3})
+[2026-07-28 12:38:08,421][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.032471753656864166, 'train/grad_norm_pre_clip': 0.5841726064682007, 'timing/data': 0.00033750850707292557, 'timing/model': 0.6172984037548304, 'learning_rate/qwen_vl_interface': 9.590895811665284e-06, 'learning_rate/action_model': 9.590895811665285e-05, 'epoch': 0.3})
+[2026-07-28 12:38:17,272][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.04942687228322029, 'train/grad_norm_pre_clip': 1.7319422960281372, 'timing/data': 0.00032044388353824615, 'timing/model': 0.6155702881515026, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.3})
+[2026-07-28 12:38:26,368][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.062430575489997864, 'train/grad_norm_pre_clip': 2.5296411514282227, 'timing/data': 0.000441807322204113, 'timing/model': 0.6193665852770209, 'learning_rate/qwen_vl_interface': 9.587783220188212e-06, 'learning_rate/action_model': 9.587783220188212e-05, 'epoch': 0.3})
+[2026-07-28 12:38:35,447][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.055658672004938126, 'train/grad_norm_pre_clip': 1.3005132675170898, 'timing/data': 0.00032186415046453476, 'timing/model': 0.6164072044193745, 'learning_rate/qwen_vl_interface': 9.586222701833631e-06, 'learning_rate/action_model': 9.58622270183363e-05, 'epoch': 0.3})
+[2026-07-28 12:38:44,981][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.05567243695259094, 'train/grad_norm_pre_clip': 2.5386574268341064, 'timing/data': 0.00033583957701921463, 'timing/model': 0.853126716800034, 'learning_rate/qwen_vl_interface': 9.584659369751216e-06, 'learning_rate/action_model': 9.584659369751215e-05, 'epoch': 0.3})
+[2026-07-28 12:38:53,833][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.04536832123994827, 'train/grad_norm_pre_clip': 1.4028363227844238, 'timing/data': 0.00031525641679763794, 'timing/model': 0.6167153501883149, 'learning_rate/qwen_vl_interface': 9.583093224955397e-06, 'learning_rate/action_model': 9.583093224955397e-05, 'epoch': 0.3})
+[2026-07-28 12:39:02,890][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.07039742171764374, 'train/grad_norm_pre_clip': 1.5513031482696533, 'timing/data': 0.0005299923941493034, 'timing/model': 0.6182993482798338, 'learning_rate/qwen_vl_interface': 9.58152426846243e-06, 'learning_rate/action_model': 9.581524268462429e-05, 'epoch': 0.3})
+[2026-07-28 12:39:11,742][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.051568325608968735, 'train/grad_norm_pre_clip': 1.403220534324646, 'timing/data': 0.0003247857093811035, 'timing/model': 0.6163909882307053, 'learning_rate/qwen_vl_interface': 9.579952501290388e-06, 'learning_rate/action_model': 9.579952501290387e-05, 'epoch': 0.3})
+[2026-07-28 12:39:20,823][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.048952434211969376, 'train/grad_norm_pre_clip': 2.213970422744751, 'timing/data': 0.0003382731229066849, 'timing/model': 0.6180405514314771, 'learning_rate/qwen_vl_interface': 9.57837792445918e-06, 'learning_rate/action_model': 9.578377924459181e-05, 'epoch': 0.3})
+[2026-07-28 12:39:29,950][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.04697515815496445, 'train/grad_norm_pre_clip': 0.7788762450218201, 'timing/data': 0.00031890999525785446, 'timing/model': 0.8562167286872864, 'learning_rate/qwen_vl_interface': 9.57680053899053e-06, 'learning_rate/action_model': 9.576800538990529e-05, 'epoch': 0.3})
+[2026-07-28 12:39:38,853][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.0758265033364296, 'train/grad_norm_pre_clip': 2.088461399078369, 'timing/data': 0.0003351690247654915, 'timing/model': 0.6169195789843798, 'learning_rate/qwen_vl_interface': 9.575220345907984e-06, 'learning_rate/action_model': 9.575220345907984e-05, 'epoch': 0.3})
+[2026-07-28 12:39:47,715][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.07008777558803558, 'train/grad_norm_pre_clip': 1.0948021411895752, 'timing/data': 0.00031702686101198196, 'timing/model': 0.6175643485039473, 'learning_rate/qwen_vl_interface': 9.573637346236915e-06, 'learning_rate/action_model': 9.573637346236915e-05, 'epoch': 0.3})
+[2026-07-28 12:39:56,788][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.037056125700473785, 'train/grad_norm_pre_clip': 1.0577341318130493, 'timing/data': 0.00032735802233219147, 'timing/model': 0.6179923741146922, 'learning_rate/qwen_vl_interface': 9.572051541004513e-06, 'learning_rate/action_model': 9.572051541004513e-05, 'epoch': 0.3})
+[2026-07-28 12:40:05,858][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.03019905835390091, 'train/grad_norm_pre_clip': 0.8459957242012024, 'timing/data': 0.00033177807927131653, 'timing/model': 0.6164170783013105, 'learning_rate/qwen_vl_interface': 9.57046293123979e-06, 'learning_rate/action_model': 9.570462931239789e-05, 'epoch': 0.3})
+[2026-07-28 12:40:15,398][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.016923271119594574, 'train/grad_norm_pre_clip': 0.4038127362728119, 'timing/data': 0.0004094736650586128, 'timing/model': 0.6199743542820215, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.3})
+[2026-07-28 12:40:24,491][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.03024972788989544, 'train/grad_norm_pre_clip': 0.7902376651763916, 'timing/data': 0.00032464321702718735, 'timing/model': 0.6160694258287549, 'learning_rate/qwen_vl_interface': 9.567277302238524e-06, 'learning_rate/action_model': 9.567277302238524e-05, 'epoch': 0.3})
+[2026-07-28 12:40:33,545][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.0356629453599453, 'train/grad_norm_pre_clip': 1.5034034252166748, 'timing/data': 0.00031792838126420975, 'timing/model': 0.6158350259065628, 'learning_rate/qwen_vl_interface': 9.565680285069104e-06, 'learning_rate/action_model': 9.565680285069103e-05, 'epoch': 0.3})
+[2026-07-28 12:40:42,833][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.03195062279701233, 'train/grad_norm_pre_clip': 1.3373563289642334, 'timing/data': 0.0003985166549682617, 'timing/model': 0.6185864508152008, 'learning_rate/qwen_vl_interface': 9.564080467501601e-06, 'learning_rate/action_model': 9.5640804675016e-05, 'epoch': 0.3})
+[2026-07-28 12:40:52,147][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.018209367990493774, 'train/grad_norm_pre_clip': 0.9772635698318481, 'timing/data': 0.00033054687082767487, 'timing/model': 0.8437192384153605, 'learning_rate/qwen_vl_interface': 9.562477850574118e-06, 'learning_rate/action_model': 9.562477850574119e-05, 'epoch': 0.3})
+[2026-07-28 12:41:01,229][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.07187115401029587, 'train/grad_norm_pre_clip': 1.9341728687286377, 'timing/data': 0.0003216788172721863, 'timing/model': 0.6169404834508896, 'learning_rate/qwen_vl_interface': 9.56087243532658e-06, 'learning_rate/action_model': 9.56087243532658e-05, 'epoch': 0.3})
+[2026-07-28 12:41:10,302][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.06291302293539047, 'train/grad_norm_pre_clip': 1.9648571014404297, 'timing/data': 0.00032866187393665314, 'timing/model': 0.6150023639202118, 'learning_rate/qwen_vl_interface': 9.55926422280072e-06, 'learning_rate/action_model': 9.55926422280072e-05, 'epoch': 0.3})
+[2026-07-28 12:41:19,587][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.08516816049814224, 'train/grad_norm_pre_clip': 2.7367281913757324, 'timing/data': 0.0004114285111427307, 'timing/model': 0.6148673621937633, 'learning_rate/qwen_vl_interface': 9.55765321404009e-06, 'learning_rate/action_model': 9.557653214040091e-05, 'epoch': 0.3})
+[2026-07-28 12:41:28,855][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.01008692104369402, 'train/grad_norm_pre_clip': 0.5989983677864075, 'timing/data': 0.00033123791217803955, 'timing/model': 0.8016160475090146, 'learning_rate/qwen_vl_interface': 9.556039410090061e-06, 'learning_rate/action_model': 9.556039410090061e-05, 'epoch': 0.3})
+[2026-07-28 12:41:37,968][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.057226501405239105, 'train/grad_norm_pre_clip': 2.194044589996338, 'timing/data': 0.00033121462911367416, 'timing/model': 0.6151039283722639, 'learning_rate/qwen_vl_interface': 9.554422811997804e-06, 'learning_rate/action_model': 9.554422811997805e-05, 'epoch': 0.3})
+[2026-07-28 12:41:47,103][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.07431761175394058, 'train/grad_norm_pre_clip': 1.8049627542495728, 'timing/data': 0.0003205547109246254, 'timing/model': 0.6145321065559983, 'learning_rate/qwen_vl_interface': 9.55280342081232e-06, 'learning_rate/action_model': 9.55280342081232e-05, 'epoch': 0.3})
+[2026-07-28 12:41:56,386][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.034905269742012024, 'train/grad_norm_pre_clip': 1.3540031909942627, 'timing/data': 0.000477813184261322, 'timing/model': 0.6164735611528158, 'learning_rate/qwen_vl_interface': 9.55118123758441e-06, 'learning_rate/action_model': 9.551181237584409e-05, 'epoch': 0.3})
+[2026-07-28 12:42:05,696][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.029417965561151505, 'train/grad_norm_pre_clip': 1.5319017171859741, 'timing/data': 0.00041920319199562073, 'timing/model': 0.8500476842746139, 'learning_rate/qwen_vl_interface': 9.549556263366691e-06, 'learning_rate/action_model': 9.549556263366692e-05, 'epoch': 0.3})
+[2026-07-28 12:42:14,758][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.062376976013183594, 'train/grad_norm_pre_clip': 1.598446249961853, 'timing/data': 0.0003134924918413162, 'timing/model': 0.6152687314897776, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.3})
+[2026-07-28 12:42:23,826][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.033576760441064835, 'train/grad_norm_pre_clip': 1.0055941343307495, 'timing/data': 0.0003959452733397484, 'timing/model': 0.6166541799902916, 'learning_rate/qwen_vl_interface': 9.546297946181354e-06, 'learning_rate/action_model': 9.546297946181354e-05, 'epoch': 0.3})
+[2026-07-28 12:42:33,166][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.03711676970124245, 'train/grad_norm_pre_clip': 1.2644013166427612, 'timing/data': 0.00038912519812583923, 'timing/model': 0.835489122197032, 'learning_rate/qwen_vl_interface': 9.544664605328022e-06, 'learning_rate/action_model': 9.544664605328022e-05, 'epoch': 0.3})
+[2026-07-28 12:42:42,549][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.053219325840473175, 'train/grad_norm_pre_clip': 1.62215256690979, 'timing/data': 0.00038674380630254745, 'timing/model': 0.6160495411604643, 'learning_rate/qwen_vl_interface': 9.543028477713455e-06, 'learning_rate/action_model': 9.543028477713453e-05, 'epoch': 0.3})
+[2026-07-28 12:42:51,604][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.03174711391329765, 'train/grad_norm_pre_clip': 1.0314109325408936, 'timing/data': 0.00039375200867652893, 'timing/model': 0.6140943542122841, 'learning_rate/qwen_vl_interface': 9.541389564399316e-06, 'learning_rate/action_model': 9.541389564399315e-05, 'epoch': 0.3})
+[2026-07-28 12:43:00,469][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.04165930300951004, 'train/grad_norm_pre_clip': 2.6141197681427, 'timing/data': 0.0003239838406443596, 'timing/model': 0.6152048911899328, 'learning_rate/qwen_vl_interface': 9.53974786644908e-06, 'learning_rate/action_model': 9.53974786644908e-05, 'epoch': 0.3})
+[2026-07-28 12:43:09,557][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.044522564858198166, 'train/grad_norm_pre_clip': 1.8989243507385254, 'timing/data': 0.0008217357099056244, 'timing/model': 0.6173175079748034, 'learning_rate/qwen_vl_interface': 9.538103384928029e-06, 'learning_rate/action_model': 9.538103384928027e-05, 'epoch': 0.3})
+[2026-07-28 12:43:18,895][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.09126147627830505, 'train/grad_norm_pre_clip': 2.6192779541015625, 'timing/data': 0.0003362763673067093, 'timing/model': 0.6167558711022139, 'learning_rate/qwen_vl_interface': 9.536456120903244e-06, 'learning_rate/action_model': 9.536456120903244e-05, 'epoch': 0.3})
+[2026-07-28 12:43:28,180][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.09017948806285858, 'train/grad_norm_pre_clip': 3.328372001647949, 'timing/data': 0.00031818728893995285, 'timing/model': 0.6160981077700853, 'learning_rate/qwen_vl_interface': 9.534806075443619e-06, 'learning_rate/action_model': 9.534806075443619e-05, 'epoch': 0.3})
+[2026-07-28 12:43:37,273][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.02689925953745842, 'train/grad_norm_pre_clip': 1.0142643451690674, 'timing/data': 0.0004296721890568733, 'timing/model': 0.6172497496008873, 'learning_rate/qwen_vl_interface': 9.533153249619853e-06, 'learning_rate/action_model': 9.533153249619853e-05, 'epoch': 0.3})
+[2026-07-28 12:43:46,546][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.07026790827512741, 'train/grad_norm_pre_clip': 1.4028198719024658, 'timing/data': 0.0003256397321820259, 'timing/model': 0.7998294113203883, 'learning_rate/qwen_vl_interface': 9.531497644504444e-06, 'learning_rate/action_model': 9.531497644504445e-05, 'epoch': 0.3})
+[2026-07-28 12:43:55,628][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.04460112005472183, 'train/grad_norm_pre_clip': 2.607175588607788, 'timing/data': 0.00039567891508340836, 'timing/model': 0.6173359025269747, 'learning_rate/qwen_vl_interface': 9.529839261171698e-06, 'learning_rate/action_model': 9.529839261171696e-05, 'epoch': 0.3})
+[2026-07-28 12:44:04,913][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.07843643426895142, 'train/grad_norm_pre_clip': 2.5137839317321777, 'timing/data': 0.00031095556914806366, 'timing/model': 0.8312276042997837, 'learning_rate/qwen_vl_interface': 9.52817810069772e-06, 'learning_rate/action_model': 9.52817810069772e-05, 'epoch': 0.3})
+[2026-07-28 12:44:13,988][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.1048787385225296, 'train/grad_norm_pre_clip': 2.866466999053955, 'timing/data': 0.00042082276195287704, 'timing/model': 0.6163531057536602, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.3})
+[2026-07-28 12:44:23,325][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.038445036858320236, 'train/grad_norm_pre_clip': 0.9856283068656921, 'timing/data': 0.000334395095705986, 'timing/model': 0.849920280277729, 'learning_rate/qwen_vl_interface': 9.524847452639509e-06, 'learning_rate/action_model': 9.524847452639509e-05, 'epoch': 0.3})
+[2026-07-28 12:44:32,389][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.07380171865224838, 'train/grad_norm_pre_clip': 2.2217764854431152, 'timing/data': 0.0003157183527946472, 'timing/model': 0.6153538329526782, 'learning_rate/qwen_vl_interface': 9.523177967216496e-06, 'learning_rate/action_model': 9.523177967216496e-05, 'epoch': 0.31})
+[2026-07-28 12:44:41,666][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.07148037850856781, 'train/grad_norm_pre_clip': 1.4318963289260864, 'timing/data': 0.0003205714747309685, 'timing/model': 0.6162791177630424, 'learning_rate/qwen_vl_interface': 9.521505708974694e-06, 'learning_rate/action_model': 9.521505708974694e-05, 'epoch': 0.31})
+[2026-07-28 12:44:50,737][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.04415082931518555, 'train/grad_norm_pre_clip': 0.9449865818023682, 'timing/data': 0.00034465640783309937, 'timing/model': 0.6164515819400549, 'learning_rate/qwen_vl_interface': 9.519830678999214e-06, 'learning_rate/action_model': 9.519830678999213e-05, 'epoch': 0.31})
+[2026-07-28 12:45:00,120][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.04844088852405548, 'train/grad_norm_pre_clip': 1.3095438480377197, 'timing/data': 0.0003364328294992447, 'timing/model': 0.6572241075336933, 'learning_rate/qwen_vl_interface': 9.518152878376961e-06, 'learning_rate/action_model': 9.51815287837696e-05, 'epoch': 0.31})
+[2026-07-28 12:45:09,444][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.03196113184094429, 'train/grad_norm_pre_clip': 1.4187169075012207, 'timing/data': 0.0003878790885210037, 'timing/model': 0.8415875788778067, 'learning_rate/qwen_vl_interface': 9.516472308196646e-06, 'learning_rate/action_model': 9.516472308196645e-05, 'epoch': 0.31})
+[2026-07-28 12:45:18,514][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.06380761414766312, 'train/grad_norm_pre_clip': 2.3027544021606445, 'timing/data': 0.0003878604620695114, 'timing/model': 0.6161290882155299, 'learning_rate/qwen_vl_interface': 9.51478896954877e-06, 'learning_rate/action_model': 9.51478896954877e-05, 'epoch': 0.31})
+[2026-07-28 12:45:27,737][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.0583239421248436, 'train/grad_norm_pre_clip': 2.806989908218384, 'timing/data': 0.00031256116926670074, 'timing/model': 0.694453839212656, 'learning_rate/qwen_vl_interface': 9.51310286352563e-06, 'learning_rate/action_model': 9.51310286352563e-05, 'epoch': 0.31})
+[2026-07-28 12:45:37,046][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.05670284852385521, 'train/grad_norm_pre_clip': 1.4085618257522583, 'timing/data': 0.00041679292917251587, 'timing/model': 0.6182754626497626, 'learning_rate/qwen_vl_interface': 9.51141399122133e-06, 'learning_rate/action_model': 9.511413991221331e-05, 'epoch': 0.31})
+[2026-07-28 12:45:46,143][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.052554208785295486, 'train/grad_norm_pre_clip': 1.5892146825790405, 'timing/data': 0.00032338500022888184, 'timing/model': 0.6158493300899863, 'learning_rate/qwen_vl_interface': 9.509722353731757e-06, 'learning_rate/action_model': 9.509722353731756e-05, 'epoch': 0.31})
+[2026-07-28 12:45:55,434][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.022254448384046555, 'train/grad_norm_pre_clip': 0.7563071250915527, 'timing/data': 0.00032520759850740433, 'timing/model': 0.6172427320852876, 'learning_rate/qwen_vl_interface': 9.508027952154595e-06, 'learning_rate/action_model': 9.508027952154593e-05, 'epoch': 0.31})
+[2026-07-28 12:46:04,747][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.03384466841816902, 'train/grad_norm_pre_clip': 0.7957467436790466, 'timing/data': 0.00031561218202114105, 'timing/model': 0.61855423822999, 'learning_rate/qwen_vl_interface': 9.506330787589321e-06, 'learning_rate/action_model': 9.506330787589322e-05, 'epoch': 0.31})
+[2026-07-28 12:46:13,819][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.07073605805635452, 'train/grad_norm_pre_clip': 1.135028600692749, 'timing/data': 0.0003190618008375168, 'timing/model': 0.6160294450819492, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.31})
+[2026-07-28 12:46:22,897][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.048069827258586884, 'train/grad_norm_pre_clip': 1.5066200494766235, 'timing/data': 0.0004204818978905678, 'timing/model': 0.6178584648296237, 'learning_rate/qwen_vl_interface': 9.50292817390133e-06, 'learning_rate/action_model': 9.502928173901329e-05, 'epoch': 0.31})
+[2026-07-28 12:46:31,972][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.028058700263500214, 'train/grad_norm_pre_clip': 0.7659328579902649, 'timing/data': 0.0004054997116327286, 'timing/model': 0.6159298196434975, 'learning_rate/qwen_vl_interface': 9.501222726986525e-06, 'learning_rate/action_model': 9.501222726986524e-05, 'epoch': 0.31})
+[2026-07-28 12:46:41,293][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.03448515385389328, 'train/grad_norm_pre_clip': 1.7367229461669922, 'timing/data': 0.0003744130954146385, 'timing/model': 0.849592206068337, 'learning_rate/qwen_vl_interface': 9.49951452149945e-06, 'learning_rate/action_model': 9.499514521499451e-05, 'epoch': 0.31})
+[2026-07-28 12:46:50,354][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.05791658163070679, 'train/grad_norm_pre_clip': 0.9688555002212524, 'timing/data': 0.00037682708352804184, 'timing/model': 0.6172806117683649, 'learning_rate/qwen_vl_interface': 9.497803558548539e-06, 'learning_rate/action_model': 9.497803558548539e-05, 'epoch': 0.31})
+[2026-07-28 12:46:59,441][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.05577690899372101, 'train/grad_norm_pre_clip': 1.1649011373519897, 'timing/data': 0.0003786049783229828, 'timing/model': 0.631024700589478, 'learning_rate/qwen_vl_interface': 9.496089839244015e-06, 'learning_rate/action_model': 9.496089839244015e-05, 'epoch': 0.31})
+[2026-07-28 12:47:08,519][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.06203765422105789, 'train/grad_norm_pre_clip': 1.3793153762817383, 'timing/data': 0.00034356489777565, 'timing/model': 0.6150632947683334, 'learning_rate/qwen_vl_interface': 9.494373364697895e-06, 'learning_rate/action_model': 9.494373364697894e-05, 'epoch': 0.31})
+[2026-07-28 12:47:17,395][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.037957850843667984, 'train/grad_norm_pre_clip': 3.8209125995635986, 'timing/data': 0.0005386881530284882, 'timing/model': 0.6160690877586603, 'learning_rate/qwen_vl_interface': 9.492654136023978e-06, 'learning_rate/action_model': 9.492654136023979e-05, 'epoch': 0.31})
+[2026-07-28 12:47:27,022][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.033349402248859406, 'train/grad_norm_pre_clip': 1.1933789253234863, 'timing/data': 0.000331302173435688, 'timing/model': 0.8598242243751884, 'learning_rate/qwen_vl_interface': 9.490932154337855e-06, 'learning_rate/action_model': 9.490932154337855e-05, 'epoch': 0.31})
+[2026-07-28 12:47:36,085][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.01887226104736328, 'train/grad_norm_pre_clip': 0.6771377921104431, 'timing/data': 0.00031981058418750763, 'timing/model': 0.616836991161108, 'learning_rate/qwen_vl_interface': 9.4892074207569e-06, 'learning_rate/action_model': 9.489207420756899e-05, 'epoch': 0.31})
+[2026-07-28 12:47:45,133][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.044778358191251755, 'train/grad_norm_pre_clip': 2.5685601234436035, 'timing/data': 0.00037505291402339935, 'timing/model': 0.6169090392068028, 'learning_rate/qwen_vl_interface': 9.487479936400274e-06, 'learning_rate/action_model': 9.487479936400272e-05, 'epoch': 0.31})
+[2026-07-28 12:47:54,408][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.04090586677193642, 'train/grad_norm_pre_clip': 1.0521200895309448, 'timing/data': 0.00030312221497297287, 'timing/model': 0.6150411842390895, 'learning_rate/qwen_vl_interface': 9.485749702388923e-06, 'learning_rate/action_model': 9.485749702388923e-05, 'epoch': 0.31})
+[2026-07-28 12:48:03,452][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.03359244391322136, 'train/grad_norm_pre_clip': 1.4338326454162598, 'timing/data': 0.0003206683322787285, 'timing/model': 0.6160488734021783, 'learning_rate/qwen_vl_interface': 9.484016719845576e-06, 'learning_rate/action_model': 9.484016719845576e-05, 'epoch': 0.31})
+[2026-07-28 12:48:12,554][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.043355949223041534, 'train/grad_norm_pre_clip': 1.0355600118637085, 'timing/data': 0.00032036192715168, 'timing/model': 0.616699593141675, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.31})
+[2026-07-28 12:48:21,398][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.05462122708559036, 'train/grad_norm_pre_clip': 1.8517884016036987, 'timing/data': 0.00033217109739780426, 'timing/model': 0.6144553879275918, 'learning_rate/qwen_vl_interface': 9.480542513662734e-06, 'learning_rate/action_model': 9.480542513662734e-05, 'epoch': 0.31})
+[2026-07-28 12:48:30,909][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.03092029318213463, 'train/grad_norm_pre_clip': 1.5352632999420166, 'timing/data': 0.0003120293840765953, 'timing/model': 0.6164630027487874, 'learning_rate/qwen_vl_interface': 9.478801292277617e-06, 'learning_rate/action_model': 9.478801292277616e-05, 'epoch': 0.31})
+[2026-07-28 12:48:39,967][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.07878734171390533, 'train/grad_norm_pre_clip': 2.2795252799987793, 'timing/data': 0.0002996092662215233, 'timing/model': 0.6167506910860538, 'learning_rate/qwen_vl_interface': 9.477057326869249e-06, 'learning_rate/action_model': 9.477057326869249e-05, 'epoch': 0.31})
+[2026-07-28 12:48:49,024][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.05391133949160576, 'train/grad_norm_pre_clip': 1.6248602867126465, 'timing/data': 0.0003390219062566757, 'timing/model': 0.6172195384278893, 'learning_rate/qwen_vl_interface': 9.475310618569274e-06, 'learning_rate/action_model': 9.475310618569274e-05, 'epoch': 0.31})
+[2026-07-28 12:48:58,084][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.09427376091480255, 'train/grad_norm_pre_clip': 3.084352970123291, 'timing/data': 0.00038121454417705536, 'timing/model': 0.6155047873035073, 'learning_rate/qwen_vl_interface': 9.473561168511115e-06, 'learning_rate/action_model': 9.473561168511115e-05, 'epoch': 0.31})
+[2026-07-28 12:49:07,383][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.06348267942667007, 'train/grad_norm_pre_clip': 1.9524965286254883, 'timing/data': 0.0003268830478191376, 'timing/model': 0.6173107754439116, 'learning_rate/qwen_vl_interface': 9.471808977829967e-06, 'learning_rate/action_model': 9.471808977829967e-05, 'epoch': 0.31})
+[2026-07-28 12:49:16,704][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.0765572190284729, 'train/grad_norm_pre_clip': 1.5850636959075928, 'timing/data': 0.00031247083097696304, 'timing/model': 0.6147311478853226, 'learning_rate/qwen_vl_interface': 9.470054047662808e-06, 'learning_rate/action_model': 9.470054047662808e-05, 'epoch': 0.31})
+[2026-07-28 12:49:25,753][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.09058026224374771, 'train/grad_norm_pre_clip': 2.3978474140167236, 'timing/data': 0.0003145551308989525, 'timing/model': 0.6166510609909892, 'learning_rate/qwen_vl_interface': 9.468296379148394e-06, 'learning_rate/action_model': 9.468296379148393e-05, 'epoch': 0.31})
+[2026-07-28 12:49:34,818][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.07379990816116333, 'train/grad_norm_pre_clip': 1.281391978263855, 'timing/data': 0.00032172538340091705, 'timing/model': 0.6156309153884649, 'learning_rate/qwen_vl_interface': 9.466535973427259e-06, 'learning_rate/action_model': 9.466535973427258e-05, 'epoch': 0.31})
+[2026-07-28 12:49:44,131][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.06502754986286163, 'train/grad_norm_pre_clip': 3.7265584468841553, 'timing/data': 0.0003198031336069107, 'timing/model': 0.8318838831037283, 'learning_rate/qwen_vl_interface': 9.464772831641706e-06, 'learning_rate/action_model': 9.464772831641707e-05, 'epoch': 0.31})
+[2026-07-28 12:49:53,210][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.06833085417747498, 'train/grad_norm_pre_clip': 1.1201481819152832, 'timing/data': 0.0003806697204709053, 'timing/model': 0.6175086796283722, 'learning_rate/qwen_vl_interface': 9.463006954935823e-06, 'learning_rate/action_model': 9.463006954935822e-05, 'epoch': 0.31})
+[2026-07-28 12:50:02,285][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.04148908704519272, 'train/grad_norm_pre_clip': 0.8870294690132141, 'timing/data': 0.0003047548234462738, 'timing/model': 0.6160605270415545, 'learning_rate/qwen_vl_interface': 9.461238344455468e-06, 'learning_rate/action_model': 9.461238344455468e-05, 'epoch': 0.31})
+[2026-07-28 12:50:11,352][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.05799327790737152, 'train/grad_norm_pre_clip': 1.5602291822433472, 'timing/data': 0.0003031119704246521, 'timing/model': 0.6172185456380248, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.31})
+[2026-07-28 12:50:20,736][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.05030378699302673, 'train/grad_norm_pre_clip': 0.874940812587738, 'timing/data': 0.00030957628041505814, 'timing/model': 0.7825508285313845, 'learning_rate/qwen_vl_interface': 9.457692926763638e-06, 'learning_rate/action_model': 9.457692926763638e-05, 'epoch': 0.31})
+[2026-07-28 12:50:29,952][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.044103629887104034, 'train/grad_norm_pre_clip': 2.2635462284088135, 'timing/data': 0.0003306828439235687, 'timing/model': 0.6183241866528988, 'learning_rate/qwen_vl_interface': 9.455916121852748e-06, 'learning_rate/action_model': 9.455916121852747e-05, 'epoch': 0.31})
+[2026-07-28 12:50:39,283][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.057393722236156464, 'train/grad_norm_pre_clip': 1.6054006814956665, 'timing/data': 0.0003163497895002365, 'timing/model': 0.6208205157890916, 'learning_rate/qwen_vl_interface': 9.454136587768549e-06, 'learning_rate/action_model': 9.454136587768548e-05, 'epoch': 0.31})
+[2026-07-28 12:50:48,379][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.14816203713417053, 'train/grad_norm_pre_clip': 10.79772663116455, 'timing/data': 0.0003184722736477852, 'timing/model': 0.6178956497460604, 'learning_rate/qwen_vl_interface': 9.452354325665763e-06, 'learning_rate/action_model': 9.452354325665763e-05, 'epoch': 0.31})
+[2026-07-28 12:50:57,724][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.038208018988370895, 'train/grad_norm_pre_clip': 1.678022027015686, 'timing/data': 0.0003495030105113983, 'timing/model': 0.8488922165706754, 'learning_rate/qwen_vl_interface': 9.450569336700878e-06, 'learning_rate/action_model': 9.450569336700878e-05, 'epoch': 0.31})
+[2026-07-28 12:51:06,845][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.06183268129825592, 'train/grad_norm_pre_clip': 1.5800061225891113, 'timing/data': 0.000328991562128067, 'timing/model': 0.6167041826993227, 'learning_rate/qwen_vl_interface': 9.448781622032155e-06, 'learning_rate/action_model': 9.448781622032155e-05, 'epoch': 0.31})
+[2026-07-28 12:51:16,217][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.03654786944389343, 'train/grad_norm_pre_clip': 1.2689826488494873, 'timing/data': 0.00032428745180368423, 'timing/model': 0.8359453566372395, 'learning_rate/qwen_vl_interface': 9.446991182819624e-06, 'learning_rate/action_model': 9.446991182819623e-05, 'epoch': 0.31})
+[2026-07-28 12:51:25,362][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.04638943821191788, 'train/grad_norm_pre_clip': 1.4008022546768188, 'timing/data': 0.0003409748896956444, 'timing/model': 0.6170281963422894, 'learning_rate/qwen_vl_interface': 9.44519802022508e-06, 'learning_rate/action_model': 9.44519802022508e-05, 'epoch': 0.31})
+[2026-07-28 12:51:34,497][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.02103659138083458, 'train/grad_norm_pre_clip': 0.8106557130813599, 'timing/data': 0.0003326795995235443, 'timing/model': 0.6167609579861164, 'learning_rate/qwen_vl_interface': 9.443402135412088e-06, 'learning_rate/action_model': 9.443402135412088e-05, 'epoch': 0.31})
+[2026-07-28 12:51:43,789][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.039121102541685104, 'train/grad_norm_pre_clip': 1.4590204954147339, 'timing/data': 0.0003208508715033531, 'timing/model': 0.8429938796907663, 'learning_rate/qwen_vl_interface': 9.441603529545977e-06, 'learning_rate/action_model': 9.441603529545976e-05, 'epoch': 0.31})
+[2026-07-28 12:51:52,869][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.06227978318929672, 'train/grad_norm_pre_clip': 1.406995415687561, 'timing/data': 0.00031536538153886795, 'timing/model': 0.6182607794180512, 'learning_rate/qwen_vl_interface': 9.439802203793843e-06, 'learning_rate/action_model': 9.439802203793842e-05, 'epoch': 0.31})
+[2026-07-28 12:52:01,944][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.03925882652401924, 'train/grad_norm_pre_clip': 1.0462936162948608, 'timing/data': 0.0003168713301420212, 'timing/model': 0.6167883900925517, 'learning_rate/qwen_vl_interface': 9.437998159324547e-06, 'learning_rate/action_model': 9.437998159324547e-05, 'epoch': 0.31})
+[2026-07-28 12:52:11,032][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.05513637140393257, 'train/grad_norm_pre_clip': 1.4565809965133667, 'timing/data': 0.0003105662763118744, 'timing/model': 0.6173315783962607, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.31})
+[2026-07-28 12:52:20,122][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.03449762612581253, 'train/grad_norm_pre_clip': 1.2285182476043701, 'timing/data': 0.0003470703959465027, 'timing/model': 0.6186154065653682, 'learning_rate/qwen_vl_interface': 9.434381918918731e-06, 'learning_rate/action_model': 9.434381918918731e-05, 'epoch': 0.31})
+[2026-07-28 12:52:29,630][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.062240276485681534, 'train/grad_norm_pre_clip': 1.341011643409729, 'timing/data': 0.0004162602126598358, 'timing/model': 0.8575585959479213, 'learning_rate/qwen_vl_interface': 9.432569725328753e-06, 'learning_rate/action_model': 9.432569725328753e-05, 'epoch': 0.31})
+[2026-07-28 12:52:38,725][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.024433191865682602, 'train/grad_norm_pre_clip': 0.9133361577987671, 'timing/data': 0.0004162481054663658, 'timing/model': 0.6170591590926051, 'learning_rate/qwen_vl_interface': 9.430754817714686e-06, 'learning_rate/action_model': 9.430754817714687e-05, 'epoch': 0.31})
+[2026-07-28 12:52:48,022][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.08924686908721924, 'train/grad_norm_pre_clip': 3.221424102783203, 'timing/data': 0.0003639655187726021, 'timing/model': 0.6171337179839611, 'learning_rate/qwen_vl_interface': 9.42893719725421e-06, 'learning_rate/action_model': 9.42893719725421e-05, 'epoch': 0.32})
+[2026-07-28 12:52:57,114][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.07363779842853546, 'train/grad_norm_pre_clip': 2.403714179992676, 'timing/data': 0.0003697676584124565, 'timing/model': 0.6154427425935864, 'learning_rate/qwen_vl_interface': 9.427116865126757e-06, 'learning_rate/action_model': 9.427116865126756e-05, 'epoch': 0.32})
+[2026-07-28 12:53:06,156][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.043498359620571136, 'train/grad_norm_pre_clip': 2.9370524883270264, 'timing/data': 0.0003931252285838127, 'timing/model': 0.6165846539661288, 'learning_rate/qwen_vl_interface': 9.42529382251352e-06, 'learning_rate/action_model': 9.42529382251352e-05, 'epoch': 0.32})
+[2026-07-28 12:53:15,212][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.06624719500541687, 'train/grad_norm_pre_clip': 2.074680805206299, 'timing/data': 0.00038277730345726013, 'timing/model': 0.6164178280159831, 'learning_rate/qwen_vl_interface': 9.423468070597453e-06, 'learning_rate/action_model': 9.423468070597452e-05, 'epoch': 0.32})
+[2026-07-28 12:53:24,496][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.05209794640541077, 'train/grad_norm_pre_clip': 1.448279857635498, 'timing/data': 0.00037292297929525375, 'timing/model': 0.6187882237136364, 'learning_rate/qwen_vl_interface': 9.421639610563264e-06, 'learning_rate/action_model': 9.421639610563264e-05, 'epoch': 0.32})
+[2026-07-28 12:53:33,561][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.09316451847553253, 'train/grad_norm_pre_clip': 2.3450071811676025, 'timing/data': 0.00030800048261880875, 'timing/model': 0.6162492362782359, 'learning_rate/qwen_vl_interface': 9.419808443597422e-06, 'learning_rate/action_model': 9.419808443597423e-05, 'epoch': 0.32})
+[2026-07-28 12:53:42,888][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.09895341843366623, 'train/grad_norm_pre_clip': 2.0844528675079346, 'timing/data': 0.0003073764964938164, 'timing/model': 0.6184644484892488, 'learning_rate/qwen_vl_interface': 9.417974570888152e-06, 'learning_rate/action_model': 9.417974570888153e-05, 'epoch': 0.32})
+[2026-07-28 12:53:51,951][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.062037259340286255, 'train/grad_norm_pre_clip': 2.292407512664795, 'timing/data': 0.00032448582351207733, 'timing/model': 0.6149547677487135, 'learning_rate/qwen_vl_interface': 9.416137993625437e-06, 'learning_rate/action_model': 9.416137993625437e-05, 'epoch': 0.32})
+[2026-07-28 12:54:01,030][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.06750202178955078, 'train/grad_norm_pre_clip': 3.0475118160247803, 'timing/data': 0.00033374596387147903, 'timing/model': 0.6152657857164741, 'learning_rate/qwen_vl_interface': 9.414298713001007e-06, 'learning_rate/action_model': 9.414298713001006e-05, 'epoch': 0.32})
+[2026-07-28 12:54:10,313][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.05491974204778671, 'train/grad_norm_pre_clip': 2.3009307384490967, 'timing/data': 0.0003317529335618019, 'timing/model': 0.6158363595604897, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.32})
+[2026-07-28 12:54:19,365][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.07292318344116211, 'train/grad_norm_pre_clip': 2.5676941871643066, 'timing/data': 0.00040477048605680466, 'timing/model': 0.6144660878926516, 'learning_rate/qwen_vl_interface': 9.410612046442719e-06, 'learning_rate/action_model': 9.410612046442718e-05, 'epoch': 0.32})
+[2026-07-28 12:54:28,660][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.07874247431755066, 'train/grad_norm_pre_clip': 2.418222427368164, 'timing/data': 0.0003309529274702072, 'timing/model': 0.6158696254715323, 'learning_rate/qwen_vl_interface': 9.408764662901097e-06, 'learning_rate/action_model': 9.408764662901098e-05, 'epoch': 0.32})
+[2026-07-28 12:54:37,715][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.05476182699203491, 'train/grad_norm_pre_clip': 1.1220390796661377, 'timing/data': 0.000383010134100914, 'timing/model': 0.6166117256507277, 'learning_rate/qwen_vl_interface': 9.406914580782238e-06, 'learning_rate/action_model': 9.406914580782238e-05, 'epoch': 0.32})
+[2026-07-28 12:54:46,988][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.07842597365379333, 'train/grad_norm_pre_clip': 3.175079345703125, 'timing/data': 0.0003893645480275154, 'timing/model': 0.6141559574753046, 'learning_rate/qwen_vl_interface': 9.405061801286638e-06, 'learning_rate/action_model': 9.405061801286638e-05, 'epoch': 0.32})
+[2026-07-28 12:54:56,054][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.08901762962341309, 'train/grad_norm_pre_clip': 1.4560869932174683, 'timing/data': 0.0003752708435058594, 'timing/model': 0.619897648692131, 'learning_rate/qwen_vl_interface': 9.403206325616546e-06, 'learning_rate/action_model': 9.403206325616545e-05, 'epoch': 0.32})
+[2026-07-28 12:55:05,356][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.026422180235385895, 'train/grad_norm_pre_clip': 0.7161954641342163, 'timing/data': 0.0003118785098195076, 'timing/model': 0.6160037955269217, 'learning_rate/qwen_vl_interface': 9.40134815497596e-06, 'learning_rate/action_model': 9.40134815497596e-05, 'epoch': 0.32})
+[2026-07-28 12:55:14,608][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.10508301854133606, 'train/grad_norm_pre_clip': 8.3836669921875, 'timing/data': 0.00032295193523168564, 'timing/model': 0.8244340512901545, 'learning_rate/qwen_vl_interface': 9.399487290570629e-06, 'learning_rate/action_model': 9.399487290570629e-05, 'epoch': 0.32})
+[2026-07-28 12:55:23,665][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.10577614605426788, 'train/grad_norm_pre_clip': 3.8342649936676025, 'timing/data': 0.0003997059538960457, 'timing/model': 0.6154524888843298, 'learning_rate/qwen_vl_interface': 9.397623733608042e-06, 'learning_rate/action_model': 9.397623733608042e-05, 'epoch': 0.32})
+[2026-07-28 12:55:32,914][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.06104663759469986, 'train/grad_norm_pre_clip': 2.173483371734619, 'timing/data': 0.00048226211220026016, 'timing/model': 0.7846151292324066, 'learning_rate/qwen_vl_interface': 9.395757485297447e-06, 'learning_rate/action_model': 9.395757485297446e-05, 'epoch': 0.32})
+[2026-07-28 12:55:42,054][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.08193562179803848, 'train/grad_norm_pre_clip': 2.428971290588379, 'timing/data': 0.00041375309228897095, 'timing/model': 0.6150796273723245, 'learning_rate/qwen_vl_interface': 9.393888546849828e-06, 'learning_rate/action_model': 9.393888546849828e-05, 'epoch': 0.32})
+[2026-07-28 12:55:51,136][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.08596666157245636, 'train/grad_norm_pre_clip': 3.229907512664795, 'timing/data': 0.00034102797508239746, 'timing/model': 0.6154414052143693, 'learning_rate/qwen_vl_interface': 9.392016919477922e-06, 'learning_rate/action_model': 9.392016919477921e-05, 'epoch': 0.32})
+[2026-07-28 12:56:00,458][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.048123396933078766, 'train/grad_norm_pre_clip': 1.8617571592330933, 'timing/data': 0.0003147199749946594, 'timing/model': 0.6171478452160954, 'learning_rate/qwen_vl_interface': 9.390142604396206e-06, 'learning_rate/action_model': 9.390142604396205e-05, 'epoch': 0.32})
+[2026-07-28 12:56:09,760][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.09657548367977142, 'train/grad_norm_pre_clip': 4.667242050170898, 'timing/data': 0.00037974491715431213, 'timing/model': 0.8479215241968632, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.32})
+[2026-07-28 12:56:18,824][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.046051397919654846, 'train/grad_norm_pre_clip': 1.4402976036071777, 'timing/data': 0.0003157006576657295, 'timing/model': 0.6174882613122463, 'learning_rate/qwen_vl_interface': 9.386385915969975e-06, 'learning_rate/action_model': 9.386385915969976e-05, 'epoch': 0.32})
+[2026-07-28 12:56:27,897][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.059151604771614075, 'train/grad_norm_pre_clip': 1.3257752656936646, 'timing/data': 0.00038039032369852066, 'timing/model': 0.6160544417798519, 'learning_rate/qwen_vl_interface': 9.384503545063138e-06, 'learning_rate/action_model': 9.384503545063137e-05, 'epoch': 0.32})
+[2026-07-28 12:56:36,980][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.036975909024477005, 'train/grad_norm_pre_clip': 0.7844786643981934, 'timing/data': 0.0003871731460094452, 'timing/model': 0.6138058435171843, 'learning_rate/qwen_vl_interface': 9.382618491321837e-06, 'learning_rate/action_model': 9.382618491321837e-05, 'epoch': 0.32})
+[2026-07-28 12:56:46,502][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.08554744720458984, 'train/grad_norm_pre_clip': 0.9376795887947083, 'timing/data': 0.00038708001375198364, 'timing/model': 0.8343479856848717, 'learning_rate/qwen_vl_interface': 9.380730755969265e-06, 'learning_rate/action_model': 9.380730755969264e-05, 'epoch': 0.32})
+[2026-07-28 12:56:55,594][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.12282659113407135, 'train/grad_norm_pre_clip': 5.337294578552246, 'timing/data': 0.0003082519397139549, 'timing/model': 0.6163585688918829, 'learning_rate/qwen_vl_interface': 9.378840340230353e-06, 'learning_rate/action_model': 9.378840340230352e-05, 'epoch': 0.32})
+[2026-07-28 12:57:04,653][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.09710971266031265, 'train/grad_norm_pre_clip': 2.655055522918701, 'timing/data': 0.0003250306472182274, 'timing/model': 0.6150607410818338, 'learning_rate/qwen_vl_interface': 9.376947245331769e-06, 'learning_rate/action_model': 9.376947245331767e-05, 'epoch': 0.32})
+[2026-07-28 12:57:13,703][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.0548500120639801, 'train/grad_norm_pre_clip': 2.5092978477478027, 'timing/data': 0.0003216918557882309, 'timing/model': 0.6141506293788552, 'learning_rate/qwen_vl_interface': 9.375051472501922e-06, 'learning_rate/action_model': 9.375051472501923e-05, 'epoch': 0.32})
+[2026-07-28 12:57:23,024][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.0985662117600441, 'train/grad_norm_pre_clip': 2.345926284790039, 'timing/data': 0.0003147535026073456, 'timing/model': 0.6169069716706872, 'learning_rate/qwen_vl_interface': 9.373153022970961e-06, 'learning_rate/action_model': 9.373153022970961e-05, 'epoch': 0.32})
+[2026-07-28 12:57:32,290][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.07539147138595581, 'train/grad_norm_pre_clip': 1.655791997909546, 'timing/data': 0.00033315736800432205, 'timing/model': 0.6155209988355637, 'learning_rate/qwen_vl_interface': 9.371251897970768e-06, 'learning_rate/action_model': 9.371251897970768e-05, 'epoch': 0.32})
+[2026-07-28 12:57:41,358][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.041301481425762177, 'train/grad_norm_pre_clip': 1.1444153785705566, 'timing/data': 0.00031937938183546066, 'timing/model': 0.6157658733427525, 'learning_rate/qwen_vl_interface': 9.369348098734961e-06, 'learning_rate/action_model': 9.369348098734961e-05, 'epoch': 0.32})
+[2026-07-28 12:57:50,417][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.0221819169819355, 'train/grad_norm_pre_clip': 0.9983690977096558, 'timing/data': 0.000314248725771904, 'timing/model': 0.6151898959651589, 'learning_rate/qwen_vl_interface': 9.367441626498897e-06, 'learning_rate/action_model': 9.367441626498896e-05, 'epoch': 0.32})
+[2026-07-28 12:57:59,508][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.03760465979576111, 'train/grad_norm_pre_clip': 1.1250548362731934, 'timing/data': 0.00031899940222501755, 'timing/model': 0.6154309585690498, 'learning_rate/qwen_vl_interface': 9.365532482499663e-06, 'learning_rate/action_model': 9.365532482499662e-05, 'epoch': 0.32})
+[2026-07-28 12:58:08,762][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1750
+[2026-07-28 12:58:19,077][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1750
+[2026-07-28 13:09:34,862][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1750
+[2026-07-28 13:09:34,863][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.05471404641866684, 'train/grad_norm_pre_clip': 1.5324064493179321, 'eval/action_loss/seconds': 10.309889434836805, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.05072486028075218, 'eval/latency_3/loss': 0.05072486028075218, 'eval/flappy/loss': 0.05072486028075218, 'eval/flappy/latency_3/loss': 0.05072486028075218, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 262.9750031013042, 'rl_games_eval/mid_train/mean_length': 2137.6, 'rl_games_eval/mid_train/std_reward': 133.6748079966357, 'rl_games_eval/mid_train/std_length': 1077.1758166613286, 'rl_games_eval/mid_train/macro_mean_reward': 262.9750031013042, 'rl_games_eval/mid_train/macro_mean_length': 2137.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 262.9750031013042, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2137.6, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 133.6748079966357, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1077.1758166613286, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 262.9750031013042, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003056982532143593, 'timing/model': 0.6154752932488918, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.32})
+[2026-07-28 13:09:43,936][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.06399897485971451, 'train/grad_norm_pre_clip': 1.2480696439743042, 'timing/data': 0.00033553503453731537, 'timing/model': 0.6163394367322326, 'learning_rate/qwen_vl_interface': 9.361706184168714e-06, 'learning_rate/action_model': 9.361706184168713e-05, 'epoch': 0.32})
+[2026-07-28 13:09:52,981][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.07600198686122894, 'train/grad_norm_pre_clip': 1.7003366947174072, 'timing/data': 0.0003286413848400116, 'timing/model': 0.6165740555152297, 'learning_rate/qwen_vl_interface': 9.359789032319843e-06, 'learning_rate/action_model': 9.359789032319842e-05, 'epoch': 0.32})
+[2026-07-28 13:10:02,024][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.028921430930495262, 'train/grad_norm_pre_clip': 1.1957072019577026, 'timing/data': 0.00032729096710681915, 'timing/model': 0.6158297192305326, 'learning_rate/qwen_vl_interface': 9.357869213673484e-06, 'learning_rate/action_model': 9.357869213673485e-05, 'epoch': 0.32})
+[2026-07-28 13:10:11,304][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.0960155725479126, 'train/grad_norm_pre_clip': 1.721567988395691, 'timing/data': 0.00031267665326595306, 'timing/model': 0.6156237116083503, 'learning_rate/qwen_vl_interface': 9.355946729475393e-06, 'learning_rate/action_model': 9.355946729475393e-05, 'epoch': 0.32})
+[2026-07-28 13:10:20,348][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.030611863359808922, 'train/grad_norm_pre_clip': 0.8952792286872864, 'timing/data': 0.00033709220588207245, 'timing/model': 0.6148878689855337, 'learning_rate/qwen_vl_interface': 9.354021580973049e-06, 'learning_rate/action_model': 9.354021580973048e-05, 'epoch': 0.32})
+[2026-07-28 13:10:29,406][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.03138512745499611, 'train/grad_norm_pre_clip': 0.803475558757782, 'timing/data': 0.0004099356010556221, 'timing/model': 0.615182189270854, 'learning_rate/qwen_vl_interface': 9.352093769415655e-06, 'learning_rate/action_model': 9.352093769415655e-05, 'epoch': 0.32})
+[2026-07-28 13:10:38,642][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.03697649762034416, 'train/grad_norm_pre_clip': 0.8480358719825745, 'timing/data': 0.000320962630212307, 'timing/model': 0.7879901751875877, 'learning_rate/qwen_vl_interface': 9.35016329605415e-06, 'learning_rate/action_model': 9.350163296054151e-05, 'epoch': 0.32})
+[2026-07-28 13:10:47,910][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.01998400315642357, 'train/grad_norm_pre_clip': 0.5350545048713684, 'timing/data': 0.0003163600340485573, 'timing/model': 0.6171236289665103, 'learning_rate/qwen_vl_interface': 9.348230162141198e-06, 'learning_rate/action_model': 9.348230162141198e-05, 'epoch': 0.32})
+[2026-07-28 13:10:56,964][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.022186072543263435, 'train/grad_norm_pre_clip': 0.647121012210846, 'timing/data': 0.0003208909183740616, 'timing/model': 0.6156535651534796, 'learning_rate/qwen_vl_interface': 9.346294368931188e-06, 'learning_rate/action_model': 9.346294368931188e-05, 'epoch': 0.32})
+[2026-07-28 13:11:06,023][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.048413973301649094, 'train/grad_norm_pre_clip': 3.5255205631256104, 'timing/data': 0.0003557885065674782, 'timing/model': 0.6205043401569128, 'learning_rate/qwen_vl_interface': 9.344355917680235e-06, 'learning_rate/action_model': 9.344355917680235e-05, 'epoch': 0.32})
+[2026-07-28 13:11:15,301][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.10306822508573532, 'train/grad_norm_pre_clip': 3.5508759021759033, 'timing/data': 0.0003079129382967949, 'timing/model': 0.846549735404551, 'learning_rate/qwen_vl_interface': 9.342414809646179e-06, 'learning_rate/action_model': 9.342414809646179e-05, 'epoch': 0.32})
+[2026-07-28 13:11:24,142][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.0916609987616539, 'train/grad_norm_pre_clip': 2.650338888168335, 'timing/data': 0.00030981190502643585, 'timing/model': 0.615701911970973, 'learning_rate/qwen_vl_interface': 9.340471046088581e-06, 'learning_rate/action_model': 9.34047104608858e-05, 'epoch': 0.32})
+[2026-07-28 13:11:33,180][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.06690241396427155, 'train/grad_norm_pre_clip': 2.218992233276367, 'timing/data': 0.0003134738653898239, 'timing/model': 0.6158928163349628, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.32})
+[2026-07-28 13:11:42,212][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.09622084349393845, 'train/grad_norm_pre_clip': 3.057757616043091, 'timing/data': 0.00032167043536901474, 'timing/model': 0.614805887453258, 'learning_rate/qwen_vl_interface': 9.336575557449641e-06, 'learning_rate/action_model': 9.336575557449641e-05, 'epoch': 0.32})
+[2026-07-28 13:11:51,475][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.03327133506536484, 'train/grad_norm_pre_clip': 0.90949946641922, 'timing/data': 0.0004041474312543869, 'timing/model': 0.6133230542764068, 'learning_rate/qwen_vl_interface': 9.334623834896039e-06, 'learning_rate/action_model': 9.334623834896038e-05, 'epoch': 0.32})
+[2026-07-28 13:12:00,762][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.07451227307319641, 'train/grad_norm_pre_clip': 2.688741683959961, 'timing/data': 0.0004051933065056801, 'timing/model': 0.6164910914376378, 'learning_rate/qwen_vl_interface': 9.332669461874376e-06, 'learning_rate/action_model': 9.332669461874374e-05, 'epoch': 0.32})
+[2026-07-28 13:12:09,817][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.09381558001041412, 'train/grad_norm_pre_clip': 1.5000349283218384, 'timing/data': 0.0003090575337409973, 'timing/model': 0.6153947915881872, 'learning_rate/qwen_vl_interface': 9.330712439652824e-06, 'learning_rate/action_model': 9.330712439652823e-05, 'epoch': 0.32})
+[2026-07-28 13:12:18,867][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.0792386531829834, 'train/grad_norm_pre_clip': 2.2755274772644043, 'timing/data': 0.00031980592757463455, 'timing/model': 0.616567824035883, 'learning_rate/qwen_vl_interface': 9.328752769501272e-06, 'learning_rate/action_model': 9.328752769501272e-05, 'epoch': 0.32})
+[2026-07-28 13:12:28,174][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.08843915164470673, 'train/grad_norm_pre_clip': 3.659412145614624, 'timing/data': 0.00037901289761066437, 'timing/model': 0.6165302321314812, 'learning_rate/qwen_vl_interface': 9.326790452691332e-06, 'learning_rate/action_model': 9.326790452691332e-05, 'epoch': 0.32})
+[2026-07-28 13:12:37,240][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.07315906882286072, 'train/grad_norm_pre_clip': 2.4276788234710693, 'timing/data': 0.00030483491718769073, 'timing/model': 0.6165283750742674, 'learning_rate/qwen_vl_interface': 9.324825490496327e-06, 'learning_rate/action_model': 9.324825490496327e-05, 'epoch': 0.33})
+[2026-07-28 13:12:46,565][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.048831626772880554, 'train/grad_norm_pre_clip': 1.3228685855865479, 'timing/data': 0.00031855423003435135, 'timing/model': 0.8369265682995319, 'learning_rate/qwen_vl_interface': 9.322857884191303e-06, 'learning_rate/action_model': 9.322857884191302e-05, 'epoch': 0.33})
+[2026-07-28 13:12:55,869][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.048932552337646484, 'train/grad_norm_pre_clip': 1.3131495714187622, 'timing/data': 0.0003189649432897568, 'timing/model': 0.8478001691401005, 'learning_rate/qwen_vl_interface': 9.320887635053013e-06, 'learning_rate/action_model': 9.320887635053014e-05, 'epoch': 0.33})
+[2026-07-28 13:13:04,942][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.07914255559444427, 'train/grad_norm_pre_clip': 0.9479764699935913, 'timing/data': 0.00031368061900138855, 'timing/model': 0.6160444924607873, 'learning_rate/qwen_vl_interface': 9.318914744359936e-06, 'learning_rate/action_model': 9.318914744359937e-05, 'epoch': 0.33})
+[2026-07-28 13:13:14,012][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.05053345486521721, 'train/grad_norm_pre_clip': 0.5284456014633179, 'timing/data': 0.00032717175781726837, 'timing/model': 0.6169195920228958, 'learning_rate/qwen_vl_interface': 9.316939213392257e-06, 'learning_rate/action_model': 9.316939213392257e-05, 'epoch': 0.33})
+[2026-07-28 13:13:23,076][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.06121039390563965, 'train/grad_norm_pre_clip': 2.5399763584136963, 'timing/data': 0.00033116620033979416, 'timing/model': 0.6157543566077948, 'learning_rate/qwen_vl_interface': 9.314961043431877e-06, 'learning_rate/action_model': 9.314961043431876e-05, 'epoch': 0.33})
+[2026-07-28 13:13:32,394][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.036750804632902145, 'train/grad_norm_pre_clip': 1.48134446144104, 'timing/data': 0.0003498764708638191, 'timing/model': 0.6272450163960457, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.33})
+[2026-07-28 13:13:41,718][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.048503175377845764, 'train/grad_norm_pre_clip': 2.007634162902832, 'timing/data': 0.0003260672092437744, 'timing/model': 0.619311292655766, 'learning_rate/qwen_vl_interface': 9.310996791669175e-06, 'learning_rate/action_model': 9.310996791669175e-05, 'epoch': 0.33})
+[2026-07-28 13:13:50,768][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.047316230833530426, 'train/grad_norm_pre_clip': 1.7902227640151978, 'timing/data': 0.0003288509324193001, 'timing/model': 0.623156851157546, 'learning_rate/qwen_vl_interface': 9.309010712439213e-06, 'learning_rate/action_model': 9.309010712439213e-05, 'epoch': 0.33})
+[2026-07-28 13:14:00,065][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.061654768884181976, 'train/grad_norm_pre_clip': 2.078469753265381, 'timing/data': 0.0003211935982108116, 'timing/model': 0.6230320176109672, 'learning_rate/qwen_vl_interface': 9.307021999361267e-06, 'learning_rate/action_model': 9.307021999361267e-05, 'epoch': 0.33})
+[2026-07-28 13:14:09,130][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.051418378949165344, 'train/grad_norm_pre_clip': 1.1780884265899658, 'timing/data': 0.00032360944896936417, 'timing/model': 0.6209189891815186, 'learning_rate/qwen_vl_interface': 9.305030653725794e-06, 'learning_rate/action_model': 9.305030653725794e-05, 'epoch': 0.33})
+[2026-07-28 13:14:18,443][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.02257811836898327, 'train/grad_norm_pre_clip': 0.7196136116981506, 'timing/data': 0.0003119567409157753, 'timing/model': 0.6196039523929358, 'learning_rate/qwen_vl_interface': 9.303036676824954e-06, 'learning_rate/action_model': 9.303036676824953e-05, 'epoch': 0.33})
+[2026-07-28 13:14:27,513][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.04345041513442993, 'train/grad_norm_pre_clip': 1.1935632228851318, 'timing/data': 0.00032748375087976456, 'timing/model': 0.6203424334526062, 'learning_rate/qwen_vl_interface': 9.301040069952616e-06, 'learning_rate/action_model': 9.301040069952616e-05, 'epoch': 0.33})
+[2026-07-28 13:14:36,824][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.017141133546829224, 'train/grad_norm_pre_clip': 0.9654221534729004, 'timing/data': 0.00038147345185279846, 'timing/model': 0.619703134521842, 'learning_rate/qwen_vl_interface': 9.299040834404362e-06, 'learning_rate/action_model': 9.299040834404361e-05, 'epoch': 0.33})
+[2026-07-28 13:14:46,063][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.07438215613365173, 'train/grad_norm_pre_clip': 2.8342647552490234, 'timing/data': 0.0004125963896512985, 'timing/model': 0.7910177577286959, 'learning_rate/qwen_vl_interface': 9.297038971477468e-06, 'learning_rate/action_model': 9.297038971477466e-05, 'epoch': 0.33})
+[2026-07-28 13:14:55,148][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.04999527707695961, 'train/grad_norm_pre_clip': 2.8907675743103027, 'timing/data': 0.0003203507512807846, 'timing/model': 0.616409876383841, 'learning_rate/qwen_vl_interface': 9.295034482470922e-06, 'learning_rate/action_model': 9.295034482470923e-05, 'epoch': 0.33})
+[2026-07-28 13:15:04,220][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.018659822642803192, 'train/grad_norm_pre_clip': 1.2768055200576782, 'timing/data': 0.00034119561314582825, 'timing/model': 0.615920721553266, 'learning_rate/qwen_vl_interface': 9.293027368685418e-06, 'learning_rate/action_model': 9.293027368685419e-05, 'epoch': 0.33})
+[2026-07-28 13:15:13,505][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.05370977148413658, 'train/grad_norm_pre_clip': 2.373835325241089, 'timing/data': 0.00033011194318532944, 'timing/model': 0.6160555789247155, 'learning_rate/qwen_vl_interface': 9.29101763142335e-06, 'learning_rate/action_model': 9.29101763142335e-05, 'epoch': 0.33})
+[2026-07-28 13:15:22,805][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.03504064306616783, 'train/grad_norm_pre_clip': 1.7038105726242065, 'timing/data': 0.0003284430131316185, 'timing/model': 0.8488038573414087, 'learning_rate/qwen_vl_interface': 9.28900527198881e-06, 'learning_rate/action_model': 9.289005271988811e-05, 'epoch': 0.33})
+[2026-07-28 13:15:31,846][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.08306895196437836, 'train/grad_norm_pre_clip': 2.395364284515381, 'timing/data': 0.00032481923699378967, 'timing/model': 0.6166731789708138, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.33})
+[2026-07-28 13:15:40,898][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.10112844407558441, 'train/grad_norm_pre_clip': 3.3108701705932617, 'timing/data': 0.00032788515090942383, 'timing/model': 0.615743768401444, 'learning_rate/qwen_vl_interface': 9.284972691827226e-06, 'learning_rate/action_model': 9.284972691827224e-05, 'epoch': 0.33})
+[2026-07-28 13:15:50,166][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.04819967597723007, 'train/grad_norm_pre_clip': 1.9116169214248657, 'timing/data': 0.000407341867685318, 'timing/model': 0.8307465808466077, 'learning_rate/qwen_vl_interface': 9.282952473716873e-06, 'learning_rate/action_model': 9.282952473716874e-05, 'epoch': 0.33})
+[2026-07-28 13:15:59,484][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.07186770439147949, 'train/grad_norm_pre_clip': 1.4699581861495972, 'timing/data': 0.0003237994387745857, 'timing/model': 0.6223307568579912, 'learning_rate/qwen_vl_interface': 9.280929638667447e-06, 'learning_rate/action_model': 9.280929638667447e-05, 'epoch': 0.33})
+[2026-07-28 13:16:08,584][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.08663706481456757, 'train/grad_norm_pre_clip': 3.676117181777954, 'timing/data': 0.00032389629632234573, 'timing/model': 0.626991861499846, 'learning_rate/qwen_vl_interface': 9.278904187991542e-06, 'learning_rate/action_model': 9.278904187991541e-05, 'epoch': 0.33})
+[2026-07-28 13:16:17,453][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.03758596256375313, 'train/grad_norm_pre_clip': 1.143620491027832, 'timing/data': 0.0003424631431698799, 'timing/model': 0.6207404201850295, 'learning_rate/qwen_vl_interface': 9.27687612300345e-06, 'learning_rate/action_model': 9.27687612300345e-05, 'epoch': 0.33})
+[2026-07-28 13:16:26,552][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.11910539865493774, 'train/grad_norm_pre_clip': 3.761587619781494, 'timing/data': 0.0004035923629999161, 'timing/model': 0.6237995978444815, 'learning_rate/qwen_vl_interface': 9.27484544501916e-06, 'learning_rate/action_model': 9.274845445019159e-05, 'epoch': 0.33})
+[2026-07-28 13:16:35,897][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.0842384546995163, 'train/grad_norm_pre_clip': 4.245060920715332, 'timing/data': 0.00032476987689733505, 'timing/model': 0.621121434494853, 'learning_rate/qwen_vl_interface': 9.27281215535636e-06, 'learning_rate/action_model': 9.272812155356358e-05, 'epoch': 0.33})
+[2026-07-28 13:16:45,248][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.08262788504362106, 'train/grad_norm_pre_clip': 1.8909127712249756, 'timing/data': 0.00031697098165750504, 'timing/model': 0.6226501455530524, 'learning_rate/qwen_vl_interface': 9.270776255334425e-06, 'learning_rate/action_model': 9.270776255334425e-05, 'epoch': 0.33})
+[2026-07-28 13:16:54,317][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.08288724720478058, 'train/grad_norm_pre_clip': 2.876084566116333, 'timing/data': 0.0003173183649778366, 'timing/model': 0.6203088741749525, 'learning_rate/qwen_vl_interface': 9.268737746274432e-06, 'learning_rate/action_model': 9.268737746274432e-05, 'epoch': 0.33})
+[2026-07-28 13:17:03,558][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.05520111322402954, 'train/grad_norm_pre_clip': 1.0582928657531738, 'timing/data': 0.00032713450491428375, 'timing/model': 0.7927983524277806, 'learning_rate/qwen_vl_interface': 9.266696629499146e-06, 'learning_rate/action_model': 9.266696629499145e-05, 'epoch': 0.33})
+[2026-07-28 13:17:12,654][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.057583801448345184, 'train/grad_norm_pre_clip': 0.820588231086731, 'timing/data': 0.00032156798988580704, 'timing/model': 0.631987027823925, 'learning_rate/qwen_vl_interface': 9.264652906333026e-06, 'learning_rate/action_model': 9.264652906333026e-05, 'epoch': 0.33})
+[2026-07-28 13:17:21,984][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.03738800063729286, 'train/grad_norm_pre_clip': 0.9810428023338318, 'timing/data': 0.00031555164605379105, 'timing/model': 0.8545441962778568, 'learning_rate/qwen_vl_interface': 9.26260657810222e-06, 'learning_rate/action_model': 9.262606578102221e-05, 'epoch': 0.33})
+[2026-07-28 13:17:31,071][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.03504973277449608, 'train/grad_norm_pre_clip': 1.3252915143966675, 'timing/data': 0.0003268793225288391, 'timing/model': 0.6226358404383063, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.33})
+[2026-07-28 13:17:40,407][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.04258698970079422, 'train/grad_norm_pre_clip': 1.3502168655395508, 'timing/data': 0.00034115929156541824, 'timing/model': 0.8489743629470468, 'learning_rate/qwen_vl_interface': 9.258506111759608e-06, 'learning_rate/action_model': 9.258506111759609e-05, 'epoch': 0.33})
+[2026-07-28 13:17:49,491][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.039773911237716675, 'train/grad_norm_pre_clip': 2.225466251373291, 'timing/data': 0.00034523941576480865, 'timing/model': 0.6190886301919818, 'learning_rate/qwen_vl_interface': 9.256451976308548e-06, 'learning_rate/action_model': 9.25645197630855e-05, 'epoch': 0.33})
+[2026-07-28 13:17:58,789][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.06535898149013519, 'train/grad_norm_pre_clip': 2.4124467372894287, 'timing/data': 0.0003397725522518158, 'timing/model': 0.6205032356083393, 'learning_rate/qwen_vl_interface': 9.2543952411143e-06, 'learning_rate/action_model': 9.254395241114299e-05, 'epoch': 0.33})
+[2026-07-28 13:18:07,879][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.07278856635093689, 'train/grad_norm_pre_clip': 1.3829728364944458, 'timing/data': 0.00033915601670742035, 'timing/model': 0.6214113878086209, 'learning_rate/qwen_vl_interface': 9.252335907511452e-06, 'learning_rate/action_model': 9.252335907511453e-05, 'epoch': 0.33})
+[2026-07-28 13:18:17,214][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.037187863141298294, 'train/grad_norm_pre_clip': 1.2532024383544922, 'timing/data': 0.0003148224204778671, 'timing/model': 0.6222093440592289, 'learning_rate/qwen_vl_interface': 9.250273976836289e-06, 'learning_rate/action_model': 9.250273976836289e-05, 'epoch': 0.33})
+[2026-07-28 13:18:26,555][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.04941374808549881, 'train/grad_norm_pre_clip': 1.235695719718933, 'timing/data': 0.0003258129581809044, 'timing/model': 0.8567169029265642, 'learning_rate/qwen_vl_interface': 9.248209450426772e-06, 'learning_rate/action_model': 9.248209450426772e-05, 'epoch': 0.33})
+[2026-07-28 13:18:35,665][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.0467044971883297, 'train/grad_norm_pre_clip': 1.2762776613235474, 'timing/data': 0.0003193626180291176, 'timing/model': 0.6201881663873792, 'learning_rate/qwen_vl_interface': 9.24614232962255e-06, 'learning_rate/action_model': 9.24614232962255e-05, 'epoch': 0.33})
+[2026-07-28 13:18:44,738][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.0467648021876812, 'train/grad_norm_pre_clip': 1.517538070678711, 'timing/data': 0.00032404717057943344, 'timing/model': 0.6174324722960591, 'learning_rate/qwen_vl_interface': 9.244072615764957e-06, 'learning_rate/action_model': 9.244072615764956e-05, 'epoch': 0.33})
+[2026-07-28 13:18:53,834][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.03601507097482681, 'train/grad_norm_pre_clip': 0.7592388391494751, 'timing/data': 0.0003196028992533684, 'timing/model': 0.6173376282677054, 'learning_rate/qwen_vl_interface': 9.242000310197005e-06, 'learning_rate/action_model': 9.242000310197004e-05, 'epoch': 0.33})
+[2026-07-28 13:19:03,326][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.026209959760308266, 'train/grad_norm_pre_clip': 1.0654358863830566, 'timing/data': 0.00032513774931430817, 'timing/model': 0.8360244687646627, 'learning_rate/qwen_vl_interface': 9.239925414263392e-06, 'learning_rate/action_model': 9.239925414263392e-05, 'epoch': 0.33})
+[2026-07-28 13:19:12,402][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.05295005440711975, 'train/grad_norm_pre_clip': 2.258395195007324, 'timing/data': 0.0003173230215907097, 'timing/model': 0.6156539795920253, 'learning_rate/qwen_vl_interface': 9.237847929310497e-06, 'learning_rate/action_model': 9.237847929310496e-05, 'epoch': 0.33})
+[2026-07-28 13:19:21,480][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.03943771868944168, 'train/grad_norm_pre_clip': 1.5217547416687012, 'timing/data': 0.00040265172719955444, 'timing/model': 0.6151470737531781, 'learning_rate/qwen_vl_interface': 9.235767856686375e-06, 'learning_rate/action_model': 9.235767856686375e-05, 'epoch': 0.33})
+[2026-07-28 13:19:30,781][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.03248678147792816, 'train/grad_norm_pre_clip': 1.2181447744369507, 'timing/data': 0.0004461640492081642, 'timing/model': 0.6167426817119122, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.33})
+[2026-07-28 13:19:39,868][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.03962486982345581, 'train/grad_norm_pre_clip': 1.1291003227233887, 'timing/data': 0.000329604372382164, 'timing/model': 0.6166226277127862, 'learning_rate/qwen_vl_interface': 9.23159995382508e-06, 'learning_rate/action_model': 9.231599953825079e-05, 'epoch': 0.33})
+[2026-07-28 13:19:48,973][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.042552947998046875, 'train/grad_norm_pre_clip': 1.0684500932693481, 'timing/data': 0.0003345254808664322, 'timing/model': 0.6210357798263431, 'learning_rate/qwen_vl_interface': 9.22951212629241e-06, 'learning_rate/action_model': 9.229512126292411e-05, 'epoch': 0.33})
+[2026-07-28 13:19:57,837][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.07052649557590485, 'train/grad_norm_pre_clip': 1.5534615516662598, 'timing/data': 0.00031939800828695297, 'timing/model': 0.6169696822762489, 'learning_rate/qwen_vl_interface': 9.227421716497531e-06, 'learning_rate/action_model': 9.227421716497531e-05, 'epoch': 0.33})
+[2026-07-28 13:20:06,936][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.0498092956840992, 'train/grad_norm_pre_clip': 1.1711478233337402, 'timing/data': 0.00034014880657196045, 'timing/model': 0.8519174419343472, 'learning_rate/qwen_vl_interface': 9.225328725796879e-06, 'learning_rate/action_model': 9.22532872579688e-05, 'epoch': 0.33})
+[2026-07-28 13:20:16,005][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.039668142795562744, 'train/grad_norm_pre_clip': 0.9554173350334167, 'timing/data': 0.000400482676923275, 'timing/model': 0.6176334386691451, 'learning_rate/qwen_vl_interface': 9.223233155548579e-06, 'learning_rate/action_model': 9.223233155548578e-05, 'epoch': 0.33})
+[2026-07-28 13:20:25,072][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.0640462264418602, 'train/grad_norm_pre_clip': 1.4487475156784058, 'timing/data': 0.00033134594559669495, 'timing/model': 0.6164077846333385, 'learning_rate/qwen_vl_interface': 9.22113500711242e-06, 'learning_rate/action_model': 9.221135007112419e-05, 'epoch': 0.33})
+[2026-07-28 13:20:33,935][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.08266814053058624, 'train/grad_norm_pre_clip': 2.717329263687134, 'timing/data': 0.0003283582627773285, 'timing/model': 0.6164294555783272, 'learning_rate/qwen_vl_interface': 9.219034281849867e-06, 'learning_rate/action_model': 9.219034281849867e-05, 'epoch': 0.33})
+[2026-07-28 13:20:42,780][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.032816171646118164, 'train/grad_norm_pre_clip': 2.079935312271118, 'timing/data': 0.0003145560622215271, 'timing/model': 0.617529715411365, 'learning_rate/qwen_vl_interface': 9.216930981124062e-06, 'learning_rate/action_model': 9.216930981124062e-05, 'epoch': 0.33})
+[2026-07-28 13:20:51,622][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.04926285147666931, 'train/grad_norm_pre_clip': 1.1871455907821655, 'timing/data': 0.00031962525099515915, 'timing/model': 0.616584069095552, 'learning_rate/qwen_vl_interface': 9.214825106299808e-06, 'learning_rate/action_model': 9.214825106299808e-05, 'epoch': 0.34})
+[2026-07-28 13:21:00,929][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.06996599584817886, 'train/grad_norm_pre_clip': 1.774760365486145, 'timing/data': 0.00034565944224596024, 'timing/model': 0.6172488406300545, 'learning_rate/qwen_vl_interface': 9.212716658743592e-06, 'learning_rate/action_model': 9.212716658743592e-05, 'epoch': 0.34})
+[2026-07-28 13:21:10,033][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.06201887130737305, 'train/grad_norm_pre_clip': 2.173429250717163, 'timing/data': 0.00034657400101423264, 'timing/model': 0.6172909140586853, 'learning_rate/qwen_vl_interface': 9.210605639823558e-06, 'learning_rate/action_model': 9.210605639823557e-05, 'epoch': 0.34})
+[2026-07-28 13:21:19,142][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.07096562534570694, 'train/grad_norm_pre_clip': 2.101698875427246, 'timing/data': 0.0003126859664916992, 'timing/model': 0.6160033084452152, 'learning_rate/qwen_vl_interface': 9.208492050909523e-06, 'learning_rate/action_model': 9.208492050909523e-05, 'epoch': 0.34})
+[2026-07-28 13:21:28,511][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.0386631153523922, 'train/grad_norm_pre_clip': 1.3278998136520386, 'timing/data': 0.0003218790516257286, 'timing/model': 0.6167689459398389, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.34})
+[2026-07-28 13:21:37,841][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.040848225355148315, 'train/grad_norm_pre_clip': 0.9072228670120239, 'timing/data': 0.00034478306770324707, 'timing/model': 0.8518840381875634, 'learning_rate/qwen_vl_interface': 9.204257168587061e-06, 'learning_rate/action_model': 9.204257168587061e-05, 'epoch': 0.34})
+[2026-07-28 13:21:46,927][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.05441912263631821, 'train/grad_norm_pre_clip': 0.9674890041351318, 'timing/data': 0.00031365081667900085, 'timing/model': 0.6155678099021316, 'learning_rate/qwen_vl_interface': 9.2021358779266e-06, 'learning_rate/action_model': 9.202135877926601e-05, 'epoch': 0.34})
+[2026-07-28 13:21:56,198][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.05170441418886185, 'train/grad_norm_pre_clip': 2.154231071472168, 'timing/data': 0.00032208580523729324, 'timing/model': 0.7799554960802197, 'learning_rate/qwen_vl_interface': 9.200012022768077e-06, 'learning_rate/action_model': 9.200012022768077e-05, 'epoch': 0.34})
+[2026-07-28 13:22:05,303][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.08535534143447876, 'train/grad_norm_pre_clip': 1.8405327796936035, 'timing/data': 0.0003307946026325226, 'timing/model': 0.6170778619125485, 'learning_rate/qwen_vl_interface': 9.19788560448964e-06, 'learning_rate/action_model': 9.197885604489639e-05, 'epoch': 0.34})
+[2026-07-28 13:22:14,647][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.06483928859233856, 'train/grad_norm_pre_clip': 1.2564995288848877, 'timing/data': 0.0003237267956137657, 'timing/model': 0.8367234244942665, 'learning_rate/qwen_vl_interface': 9.195756624471094e-06, 'learning_rate/action_model': 9.195756624471094e-05, 'epoch': 0.34})
+[2026-07-28 13:22:23,775][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.08845823258161545, 'train/grad_norm_pre_clip': 2.8738491535186768, 'timing/data': 0.00031480100005865097, 'timing/model': 0.6154255205765367, 'learning_rate/qwen_vl_interface': 9.193625084093912e-06, 'learning_rate/action_model': 9.193625084093911e-05, 'epoch': 0.34})
+[2026-07-28 13:22:33,121][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.058583617210388184, 'train/grad_norm_pre_clip': 1.9278932809829712, 'timing/data': 0.0003235405310988426, 'timing/model': 0.6156214699149132, 'learning_rate/qwen_vl_interface': 9.191490984741227e-06, 'learning_rate/action_model': 9.191490984741227e-05, 'epoch': 0.34})
+[2026-07-28 13:22:42,233][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.03357904776930809, 'train/grad_norm_pre_clip': 0.7062719464302063, 'timing/data': 0.00041250884532928467, 'timing/model': 0.6144180009141564, 'learning_rate/qwen_vl_interface': 9.189354327797835e-06, 'learning_rate/action_model': 9.189354327797836e-05, 'epoch': 0.34})
+[2026-07-28 13:22:51,516][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.04893577843904495, 'train/grad_norm_pre_clip': 1.5223782062530518, 'timing/data': 0.0004090499132871628, 'timing/model': 0.6163394115865231, 'learning_rate/qwen_vl_interface': 9.18721511465019e-06, 'learning_rate/action_model': 9.187215114650191e-05, 'epoch': 0.34})
+[2026-07-28 13:23:00,845][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.041894637048244476, 'train/grad_norm_pre_clip': 1.3324722051620483, 'timing/data': 0.00032331235706806183, 'timing/model': 0.8456026175990701, 'learning_rate/qwen_vl_interface': 9.185073346686402e-06, 'learning_rate/action_model': 9.185073346686401e-05, 'epoch': 0.34})
+[2026-07-28 13:23:10,055][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.05903613567352295, 'train/grad_norm_pre_clip': 2.890065908432007, 'timing/data': 0.00032103806734085083, 'timing/model': 0.6195230884477496, 'learning_rate/qwen_vl_interface': 9.182929025296242e-06, 'learning_rate/action_model': 9.182929025296242e-05, 'epoch': 0.34})
+[2026-07-28 13:23:19,155][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.06520257890224457, 'train/grad_norm_pre_clip': 1.0470095872879028, 'timing/data': 0.00032026879489421844, 'timing/model': 0.6229463536292315, 'learning_rate/qwen_vl_interface': 9.180782151871138e-06, 'learning_rate/action_model': 9.180782151871137e-05, 'epoch': 0.34})
+[2026-07-28 13:23:28,263][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.06495112925767899, 'train/grad_norm_pre_clip': 1.0271666049957275, 'timing/data': 0.0005292808637022972, 'timing/model': 0.6233477611094713, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.34})
+[2026-07-28 13:23:37,339][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.06376759707927704, 'train/grad_norm_pre_clip': 1.4011651277542114, 'timing/data': 0.0003189835697412491, 'timing/model': 0.6143541019409895, 'learning_rate/qwen_vl_interface': 9.176480754490078e-06, 'learning_rate/action_model': 9.176480754490078e-05, 'epoch': 0.34})
+[2026-07-28 13:23:46,627][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.020221170037984848, 'train/grad_norm_pre_clip': 0.7997798919677734, 'timing/data': 0.00031652115285396576, 'timing/model': 0.6166828023269773, 'learning_rate/qwen_vl_interface': 9.174326233325253e-06, 'learning_rate/action_model': 9.174326233325253e-05, 'epoch': 0.34})
+[2026-07-28 13:23:55,709][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.030141102150082588, 'train/grad_norm_pre_clip': 1.015028715133667, 'timing/data': 0.00032217055559158325, 'timing/model': 0.6167120169848204, 'learning_rate/qwen_vl_interface': 9.17216916570774e-06, 'learning_rate/action_model': 9.172169165707741e-05, 'epoch': 0.34})
+[2026-07-28 13:24:05,050][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.06844840198755264, 'train/grad_norm_pre_clip': 1.9460316896438599, 'timing/data': 0.00031582824885845184, 'timing/model': 0.6148518016561866, 'learning_rate/qwen_vl_interface': 9.170009553037239e-06, 'learning_rate/action_model': 9.170009553037238e-05, 'epoch': 0.34})
+[2026-07-28 13:24:14,109][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.07614724338054657, 'train/grad_norm_pre_clip': 1.5223511457443237, 'timing/data': 0.00040081702172756195, 'timing/model': 0.6174169881269336, 'learning_rate/qwen_vl_interface': 9.167847396715098e-06, 'learning_rate/action_model': 9.167847396715098e-05, 'epoch': 0.34})
+[2026-07-28 13:24:23,211][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.03695093095302582, 'train/grad_norm_pre_clip': 0.8805248737335205, 'timing/data': 0.0003235936164855957, 'timing/model': 0.6168485563248396, 'learning_rate/qwen_vl_interface': 9.165682698144317e-06, 'learning_rate/action_model': 9.165682698144317e-05, 'epoch': 0.34})
+[2026-07-28 13:24:32,519][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.04522966593503952, 'train/grad_norm_pre_clip': 1.6130495071411133, 'timing/data': 0.0003397176042199135, 'timing/model': 0.8402069453150034, 'learning_rate/qwen_vl_interface': 9.163515458729545e-06, 'learning_rate/action_model': 9.163515458729545e-05, 'epoch': 0.34})
+[2026-07-28 13:24:41,629][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.06961153447628021, 'train/grad_norm_pre_clip': 1.6479640007019043, 'timing/data': 0.0003186659887433052, 'timing/model': 0.6245519481599331, 'learning_rate/qwen_vl_interface': 9.16134567987708e-06, 'learning_rate/action_model': 9.16134567987708e-05, 'epoch': 0.34})
+[2026-07-28 13:24:50,700][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.04865982383489609, 'train/grad_norm_pre_clip': 1.633686900138855, 'timing/data': 0.00031619705259799957, 'timing/model': 0.6201393231749535, 'learning_rate/qwen_vl_interface': 9.159173362994868e-06, 'learning_rate/action_model': 9.159173362994869e-05, 'epoch': 0.34})
+[2026-07-28 13:25:00,025][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.0340832956135273, 'train/grad_norm_pre_clip': 0.7186917662620544, 'timing/data': 0.00033342186361551285, 'timing/model': 0.8372515393421054, 'learning_rate/qwen_vl_interface': 9.156998509492503e-06, 'learning_rate/action_model': 9.156998509492503e-05, 'epoch': 0.34})
+[2026-07-28 13:25:09,134][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.05170251801609993, 'train/grad_norm_pre_clip': 2.118300199508667, 'timing/data': 0.0003184806555509567, 'timing/model': 0.6192946964874864, 'learning_rate/qwen_vl_interface': 9.154821120781221e-06, 'learning_rate/action_model': 9.154821120781222e-05, 'epoch': 0.34})
+[2026-07-28 13:25:18,233][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.04607362672686577, 'train/grad_norm_pre_clip': 2.0377187728881836, 'timing/data': 0.00032061804085969925, 'timing/model': 0.6229524714872241, 'learning_rate/qwen_vl_interface': 9.152641198273911e-06, 'learning_rate/action_model': 9.15264119827391e-05, 'epoch': 0.34})
+[2026-07-28 13:25:27,311][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.031353890895843506, 'train/grad_norm_pre_clip': 0.7537551522254944, 'timing/data': 0.00031262729316949844, 'timing/model': 0.6216066824272275, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.34})
+[2026-07-28 13:25:36,392][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.04644925519824028, 'train/grad_norm_pre_clip': 0.9062612652778625, 'timing/data': 0.0003548627719283104, 'timing/model': 0.6243370631709695, 'learning_rate/qwen_vl_interface': 9.148273757530948e-06, 'learning_rate/action_model': 9.148273757530948e-05, 'epoch': 0.34})
+[2026-07-28 13:25:45,733][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.05798882991075516, 'train/grad_norm_pre_clip': 2.048452138900757, 'timing/data': 0.00045422371476888657, 'timing/model': 0.6252456465736032, 'learning_rate/qwen_vl_interface': 9.146086242129283e-06, 'learning_rate/action_model': 9.146086242129282e-05, 'epoch': 0.34})
+[2026-07-28 13:25:54,925][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.018680695444345474, 'train/grad_norm_pre_clip': 0.5060960650444031, 'timing/data': 0.0004206579178571701, 'timing/model': 0.6235284069553018, 'learning_rate/qwen_vl_interface': 9.143896198599553e-06, 'learning_rate/action_model': 9.143896198599553e-05, 'epoch': 0.34})
+[2026-07-28 13:26:04,341][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.022878006100654602, 'train/grad_norm_pre_clip': 1.0449942350387573, 'timing/data': 0.00032826606184244156, 'timing/model': 0.8594523034989834, 'learning_rate/qwen_vl_interface': 9.141703628362858e-06, 'learning_rate/action_model': 9.141703628362858e-05, 'epoch': 0.34})
+[2026-07-28 13:26:13,491][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.029155559837818146, 'train/grad_norm_pre_clip': 1.8157280683517456, 'timing/data': 0.0003194846212863922, 'timing/model': 0.6205919682979584, 'learning_rate/qwen_vl_interface': 9.139508532841928e-06, 'learning_rate/action_model': 9.139508532841927e-05, 'epoch': 0.34})
+[2026-07-28 13:26:22,845][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.05873233079910278, 'train/grad_norm_pre_clip': 2.3894705772399902, 'timing/data': 0.000333382748067379, 'timing/model': 0.6199675016105175, 'learning_rate/qwen_vl_interface': 9.137310913461139e-06, 'learning_rate/action_model': 9.137310913461139e-05, 'epoch': 0.34})
+[2026-07-28 13:26:31,946][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.07389657944440842, 'train/grad_norm_pre_clip': 1.8027616739273071, 'timing/data': 0.00032275542616844177, 'timing/model': 0.6215737871825695, 'learning_rate/qwen_vl_interface': 9.135110771646502e-06, 'learning_rate/action_model': 9.135110771646502e-05, 'epoch': 0.34})
+[2026-07-28 13:26:41,037][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.01931900903582573, 'train/grad_norm_pre_clip': 1.2036579847335815, 'timing/data': 0.0003442969173192978, 'timing/model': 0.6221367409452796, 'learning_rate/qwen_vl_interface': 9.132908108825667e-06, 'learning_rate/action_model': 9.132908108825667e-05, 'epoch': 0.34})
+[2026-07-28 13:26:50,372][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.03117474541068077, 'train/grad_norm_pre_clip': 1.2202589511871338, 'timing/data': 0.000326329842209816, 'timing/model': 0.6179044609889388, 'learning_rate/qwen_vl_interface': 9.130702926427915e-06, 'learning_rate/action_model': 9.130702926427914e-05, 'epoch': 0.34})
+[2026-07-28 13:26:59,697][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.05205806344747543, 'train/grad_norm_pre_clip': 2.8569226264953613, 'timing/data': 0.00031756050884723663, 'timing/model': 0.6219244636595249, 'learning_rate/qwen_vl_interface': 9.128495225884167e-06, 'learning_rate/action_model': 9.128495225884166e-05, 'epoch': 0.34})
+[2026-07-28 13:27:08,780][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.0756741464138031, 'train/grad_norm_pre_clip': 3.8018765449523926, 'timing/data': 0.0003238748759031296, 'timing/model': 0.6226226566359401, 'learning_rate/qwen_vl_interface': 9.126285008626975e-06, 'learning_rate/action_model': 9.126285008626974e-05, 'epoch': 0.34})
+[2026-07-28 13:27:17,892][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.11717469990253448, 'train/grad_norm_pre_clip': 3.0129659175872803, 'timing/data': 0.000326281413435936, 'timing/model': 0.6235850313678384, 'learning_rate/qwen_vl_interface': 9.124072276090524e-06, 'learning_rate/action_model': 9.124072276090524e-05, 'epoch': 0.34})
+[2026-07-28 13:27:27,247][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.07970364391803741, 'train/grad_norm_pre_clip': 2.231261730194092, 'timing/data': 0.0004069888964295387, 'timing/model': 0.6225783824920654, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.34})
+[2026-07-28 13:27:36,596][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.030300404876470566, 'train/grad_norm_pre_clip': 1.0151057243347168, 'timing/data': 0.0004201531410217285, 'timing/model': 0.8553774962201715, 'learning_rate/qwen_vl_interface': 9.119639270924754e-06, 'learning_rate/action_model': 9.119639270924754e-05, 'epoch': 0.34})
+[2026-07-28 13:27:45,692][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.0703587606549263, 'train/grad_norm_pre_clip': 2.9504551887512207, 'timing/data': 0.00031874142587184906, 'timing/model': 0.6219889158383012, 'learning_rate/qwen_vl_interface': 9.11741900117196e-06, 'learning_rate/action_model': 9.11741900117196e-05, 'epoch': 0.34})
+[2026-07-28 13:27:54,776][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.0895586907863617, 'train/grad_norm_pre_clip': 2.4157721996307373, 'timing/data': 0.00032485276460647583, 'timing/model': 0.629021312110126, 'learning_rate/qwen_vl_interface': 9.115196221892969e-06, 'learning_rate/action_model': 9.115196221892969e-05, 'epoch': 0.34})
+[2026-07-28 13:28:04,197][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.023635799065232277, 'train/grad_norm_pre_clip': 0.7351512312889099, 'timing/data': 0.0003255680203437805, 'timing/model': 0.620654035359621, 'learning_rate/qwen_vl_interface': 9.112970934530111e-06, 'learning_rate/action_model': 9.112970934530111e-05, 'epoch': 0.34})
+[2026-07-28 13:28:13,435][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.05260702222585678, 'train/grad_norm_pre_clip': 1.9035998582839966, 'timing/data': 0.00032253749668598175, 'timing/model': 0.6236184453591704, 'learning_rate/qwen_vl_interface': 9.110743140527353e-06, 'learning_rate/action_model': 9.110743140527352e-05, 'epoch': 0.34})
+[2026-07-28 13:28:22,677][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.050778113305568695, 'train/grad_norm_pre_clip': 1.7146382331848145, 'timing/data': 0.0003208080306649208, 'timing/model': 0.6257734391838312, 'learning_rate/qwen_vl_interface': 9.108512841330288e-06, 'learning_rate/action_model': 9.108512841330287e-05, 'epoch': 0.34})
+[2026-07-28 13:28:32,018][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.03990200161933899, 'train/grad_norm_pre_clip': 0.7402605414390564, 'timing/data': 0.0003227079287171364, 'timing/model': 0.6226047547534108, 'learning_rate/qwen_vl_interface': 9.106280038386129e-06, 'learning_rate/action_model': 9.10628003838613e-05, 'epoch': 0.34})
+[2026-07-28 13:28:41,110][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.0549483597278595, 'train/grad_norm_pre_clip': 1.847419023513794, 'timing/data': 0.0003201998770236969, 'timing/model': 0.6330932658165693, 'learning_rate/qwen_vl_interface': 9.104044733143723e-06, 'learning_rate/action_model': 9.104044733143723e-05, 'epoch': 0.34})
+[2026-07-28 13:28:50,416][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.03530310094356537, 'train/grad_norm_pre_clip': 1.1117907762527466, 'timing/data': 0.0003362540155649185, 'timing/model': 0.8505871007218957, 'learning_rate/qwen_vl_interface': 9.10180692705353e-06, 'learning_rate/action_model': 9.10180692705353e-05, 'epoch': 0.34})
+[2026-07-28 13:28:59,506][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.055634111166000366, 'train/grad_norm_pre_clip': 2.274012804031372, 'timing/data': 0.0003359271213412285, 'timing/model': 0.6178310932591558, 'learning_rate/qwen_vl_interface': 9.099566621567643e-06, 'learning_rate/action_model': 9.099566621567642e-05, 'epoch': 0.34})
+[2026-07-28 13:29:08,590][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.07906083762645721, 'train/grad_norm_pre_clip': 1.33942449092865, 'timing/data': 0.00039479508996009827, 'timing/model': 0.6211624350398779, 'learning_rate/qwen_vl_interface': 9.097323818139767e-06, 'learning_rate/action_model': 9.097323818139767e-05, 'epoch': 0.34})
+[2026-07-28 13:29:17,697][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.050502847880125046, 'train/grad_norm_pre_clip': 0.8918206691741943, 'timing/data': 0.0003359336405992508, 'timing/model': 0.6158528029918671, 'learning_rate/qwen_vl_interface': 9.095078518225237e-06, 'learning_rate/action_model': 9.095078518225236e-05, 'epoch': 0.35})
+[2026-07-28 13:29:26,787][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.05776352807879448, 'train/grad_norm_pre_clip': 1.6232763528823853, 'timing/data': 0.00031947996467351913, 'timing/model': 0.6163480943068862, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.35})
+[2026-07-28 13:29:36,284][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.03134310990571976, 'train/grad_norm_pre_clip': 0.8540975451469421, 'timing/data': 0.00032569095492362976, 'timing/model': 0.8499309355393052, 'learning_rate/qwen_vl_interface': 9.09058043476563e-06, 'learning_rate/action_model': 9.09058043476563e-05, 'epoch': 0.35})
+[2026-07-28 13:29:45,382][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.045282065868377686, 'train/grad_norm_pre_clip': 1.3150925636291504, 'timing/data': 0.0003267629072070122, 'timing/model': 0.6178161259740591, 'learning_rate/qwen_vl_interface': 9.088327654139311e-06, 'learning_rate/action_model': 9.08832765413931e-05, 'epoch': 0.35})
+[2026-07-28 13:29:54,719][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.03100917488336563, 'train/grad_norm_pre_clip': 0.8804231286048889, 'timing/data': 0.00033761654049158096, 'timing/model': 0.6193646937608719, 'learning_rate/qwen_vl_interface': 9.08607238286385e-06, 'learning_rate/action_model': 9.08607238286385e-05, 'epoch': 0.35})
+[2026-07-28 13:30:03,813][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.039971496909856796, 'train/grad_norm_pre_clip': 0.608465313911438, 'timing/data': 0.0003411760553717613, 'timing/model': 0.6171297878026962, 'learning_rate/qwen_vl_interface': 9.083814622402667e-06, 'learning_rate/action_model': 9.083814622402666e-05, 'epoch': 0.35})
+[2026-07-28 13:30:12,960][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.045469291508197784, 'train/grad_norm_pre_clip': 0.9157139658927917, 'timing/data': 0.00035502389073371887, 'timing/model': 0.6155189415439963, 'learning_rate/qwen_vl_interface': 9.0815543742208e-06, 'learning_rate/action_model': 9.081554374220799e-05, 'epoch': 0.35})
+[2026-07-28 13:30:22,240][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.0721607580780983, 'train/grad_norm_pre_clip': 1.2876973152160645, 'timing/data': 0.00032197684049606323, 'timing/model': 0.8301478065550327, 'learning_rate/qwen_vl_interface': 9.079291639784897e-06, 'learning_rate/action_model': 9.079291639784896e-05, 'epoch': 0.35})
+[2026-07-28 13:30:31,557][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.026259638369083405, 'train/grad_norm_pre_clip': 1.1500179767608643, 'timing/data': 0.0003257105126976967, 'timing/model': 0.8507406823337078, 'learning_rate/qwen_vl_interface': 9.077026420563223e-06, 'learning_rate/action_model': 9.077026420563223e-05, 'epoch': 0.35})
+[2026-07-28 13:30:40,638][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.09221558272838593, 'train/grad_norm_pre_clip': 2.588320016860962, 'timing/data': 0.00032060034573078156, 'timing/model': 0.6158505864441395, 'learning_rate/qwen_vl_interface': 9.074758718025652e-06, 'learning_rate/action_model': 9.074758718025652e-05, 'epoch': 0.35})
+[2026-07-28 13:30:49,720][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.03163160756230354, 'train/grad_norm_pre_clip': 0.8667801022529602, 'timing/data': 0.0003224071115255356, 'timing/model': 0.6153712617233396, 'learning_rate/qwen_vl_interface': 9.072488533643678e-06, 'learning_rate/action_model': 9.072488533643677e-05, 'epoch': 0.35})
+[2026-07-28 13:30:58,791][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.022686496376991272, 'train/grad_norm_pre_clip': 0.5380873084068298, 'timing/data': 0.00031598377972841263, 'timing/model': 0.6168198753148317, 'learning_rate/qwen_vl_interface': 9.070215868890391e-06, 'learning_rate/action_model': 9.070215868890391e-05, 'epoch': 0.35})
+[2026-07-28 13:31:07,893][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.03387252613902092, 'train/grad_norm_pre_clip': 1.1003648042678833, 'timing/data': 0.00032022688537836075, 'timing/model': 0.6160854818299413, 'learning_rate/qwen_vl_interface': 9.067940725240503e-06, 'learning_rate/action_model': 9.067940725240503e-05, 'epoch': 0.35})
+[2026-07-28 13:31:17,213][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.06332562863826752, 'train/grad_norm_pre_clip': 1.2375133037567139, 'timing/data': 0.0003615794703364372, 'timing/model': 0.6200981577858329, 'learning_rate/qwen_vl_interface': 9.06566310417033e-06, 'learning_rate/action_model': 9.06566310417033e-05, 'epoch': 0.35})
+[2026-07-28 13:31:26,098][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.06359197944402695, 'train/grad_norm_pre_clip': 2.0442869663238525, 'timing/data': 0.00033217575401067734, 'timing/model': 0.6218605171889067, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.35})
+[2026-07-28 13:31:35,178][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.04324354976415634, 'train/grad_norm_pre_clip': 1.2745037078857422, 'timing/data': 0.0003220876678824425, 'timing/model': 0.6180430175736547, 'learning_rate/qwen_vl_interface': 9.061100435682429e-06, 'learning_rate/action_model': 9.061100435682429e-05, 'epoch': 0.35})
+[2026-07-28 13:31:44,290][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.03249165788292885, 'train/grad_norm_pre_clip': 0.9660212993621826, 'timing/data': 0.0007990105077624321, 'timing/model': 0.8521614810451865, 'learning_rate/qwen_vl_interface': 9.058815391225363e-06, 'learning_rate/action_model': 9.058815391225362e-05, 'epoch': 0.35})
+[2026-07-28 13:31:53,590][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.043106067925691605, 'train/grad_norm_pre_clip': 2.1145591735839844, 'timing/data': 0.00040060747414827347, 'timing/model': 0.6156895877793431, 'learning_rate/qwen_vl_interface': 9.056527875269344e-06, 'learning_rate/action_model': 9.056527875269344e-05, 'epoch': 0.35})
+[2026-07-28 13:32:02,681][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.03092648833990097, 'train/grad_norm_pre_clip': 0.9838433861732483, 'timing/data': 0.00031117163598537445, 'timing/model': 0.616529680788517, 'learning_rate/qwen_vl_interface': 9.054237889298714e-06, 'learning_rate/action_model': 9.054237889298714e-05, 'epoch': 0.35})
+[2026-07-28 13:32:11,787][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.04019748046994209, 'train/grad_norm_pre_clip': 0.9227521419525146, 'timing/data': 0.00033001694828271866, 'timing/model': 0.6166262933984399, 'learning_rate/qwen_vl_interface': 9.05194543479942e-06, 'learning_rate/action_model': 9.05194543479942e-05, 'epoch': 0.35})
+[2026-07-28 13:32:20,860][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.026578355580568314, 'train/grad_norm_pre_clip': 1.181454062461853, 'timing/data': 0.00032768305391073227, 'timing/model': 0.6185560869053006, 'learning_rate/qwen_vl_interface': 9.049650513259007e-06, 'learning_rate/action_model': 9.049650513259008e-05, 'epoch': 0.35})
+[2026-07-28 13:32:30,187][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.06357277929782867, 'train/grad_norm_pre_clip': 1.3685883283615112, 'timing/data': 0.0003422684967517853, 'timing/model': 0.8447804842144251, 'learning_rate/qwen_vl_interface': 9.047353126166632e-06, 'learning_rate/action_model': 9.047353126166632e-05, 'epoch': 0.35})
+[2026-07-28 13:32:39,504][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.04147753119468689, 'train/grad_norm_pre_clip': 1.2822611331939697, 'timing/data': 0.0003367653116583824, 'timing/model': 0.8659473843872547, 'learning_rate/qwen_vl_interface': 9.045053275013038e-06, 'learning_rate/action_model': 9.045053275013038e-05, 'epoch': 0.35})
+[2026-07-28 13:32:48,575][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.06587927788496017, 'train/grad_norm_pre_clip': 1.446110725402832, 'timing/data': 0.00034014880657196045, 'timing/model': 0.6175219649448991, 'learning_rate/qwen_vl_interface': 9.042750961290578e-06, 'learning_rate/action_model': 9.042750961290578e-05, 'epoch': 0.35})
+[2026-07-28 13:32:57,465][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.02934369072318077, 'train/grad_norm_pre_clip': 1.3220831155776978, 'timing/data': 0.0003344295546412468, 'timing/model': 0.6248837951570749, 'learning_rate/qwen_vl_interface': 9.040446186493195e-06, 'learning_rate/action_model': 9.040446186493194e-05, 'epoch': 0.35})
+[2026-07-28 13:33:06,564][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.04341235011816025, 'train/grad_norm_pre_clip': 1.3405537605285645, 'timing/data': 0.00031504034996032715, 'timing/model': 0.6231667576357722, 'learning_rate/qwen_vl_interface': 9.038138952116435e-06, 'learning_rate/action_model': 9.038138952116435e-05, 'epoch': 0.35})
+[2026-07-28 13:33:15,642][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.034525323659181595, 'train/grad_norm_pre_clip': 1.1541171073913574, 'timing/data': 0.0003221062943339348, 'timing/model': 0.6215416332706809, 'learning_rate/qwen_vl_interface': 9.035829259657437e-06, 'learning_rate/action_model': 9.035829259657436e-05, 'epoch': 0.35})
+[2026-07-28 13:33:24,937][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.017951197922229767, 'train/grad_norm_pre_clip': 0.9003872871398926, 'timing/data': 0.0003168918192386627, 'timing/model': 0.6245985962450504, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.35})
+[2026-07-28 13:33:34,249][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.05616048723459244, 'train/grad_norm_pre_clip': 1.8567261695861816, 'timing/data': 0.00034898892045021057, 'timing/model': 0.6208688793703914, 'learning_rate/qwen_vl_interface': 9.031202506489254e-06, 'learning_rate/action_model': 9.031202506489253e-05, 'epoch': 0.35})
+[2026-07-28 13:33:43,338][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.03595249354839325, 'train/grad_norm_pre_clip': 0.8832582235336304, 'timing/data': 0.0003274036571383476, 'timing/model': 0.6217773323878646, 'learning_rate/qwen_vl_interface': 9.02888544878232e-06, 'learning_rate/action_model': 9.028885448782319e-05, 'epoch': 0.35})
+[2026-07-28 13:33:52,428][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.08198113739490509, 'train/grad_norm_pre_clip': 1.802194595336914, 'timing/data': 0.00032060686498880386, 'timing/model': 0.6245911084115505, 'learning_rate/qwen_vl_interface': 9.026565938997644e-06, 'learning_rate/action_model': 9.026565938997645e-05, 'epoch': 0.35})
+[2026-07-28 13:34:01,542][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.061236605048179626, 'train/grad_norm_pre_clip': 2.825389862060547, 'timing/data': 0.000333552248775959, 'timing/model': 0.6169015187770128, 'learning_rate/qwen_vl_interface': 9.024243978640334e-06, 'learning_rate/action_model': 9.024243978640333e-05, 'epoch': 0.35})
+[2026-07-28 13:34:10,858][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.061575863510370255, 'train/grad_norm_pre_clip': 1.6803393363952637, 'timing/data': 0.0003426112234592438, 'timing/model': 0.6181056443601847, 'learning_rate/qwen_vl_interface': 9.02191956921708e-06, 'learning_rate/action_model': 9.02191956921708e-05, 'epoch': 0.35})
+[2026-07-28 13:34:20,197][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.03509629890322685, 'train/grad_norm_pre_clip': 0.9325854778289795, 'timing/data': 0.0003209281712770462, 'timing/model': 0.6165737845003605, 'learning_rate/qwen_vl_interface': 9.019592712236172e-06, 'learning_rate/action_model': 9.019592712236171e-05, 'epoch': 0.35})
+[2026-07-28 13:34:29,297][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.04527011513710022, 'train/grad_norm_pre_clip': 1.2380154132843018, 'timing/data': 0.00030753109604120255, 'timing/model': 0.6195291923359036, 'learning_rate/qwen_vl_interface': 9.017263409207478e-06, 'learning_rate/action_model': 9.017263409207478e-05, 'epoch': 0.35})
+[2026-07-28 13:34:38,390][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.05113225430250168, 'train/grad_norm_pre_clip': 1.1024473905563354, 'timing/data': 0.00033209938555955887, 'timing/model': 0.6173534383997321, 'learning_rate/qwen_vl_interface': 9.014931661642455e-06, 'learning_rate/action_model': 9.014931661642455e-05, 'epoch': 0.35})
+[2026-07-28 13:34:47,687][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.04237523674964905, 'train/grad_norm_pre_clip': 0.8950833082199097, 'timing/data': 0.0003112703561782837, 'timing/model': 0.6159469680860639, 'learning_rate/qwen_vl_interface': 9.012597471054154e-06, 'learning_rate/action_model': 9.012597471054154e-05, 'epoch': 0.35})
+[2026-07-28 13:34:57,050][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.04140527546405792, 'train/grad_norm_pre_clip': 1.2186639308929443, 'timing/data': 0.0003107115626335144, 'timing/model': 0.6187309361994267, 'learning_rate/qwen_vl_interface': 9.010260838957206e-06, 'learning_rate/action_model': 9.010260838957205e-05, 'epoch': 0.35})
+[2026-07-28 13:35:06,176][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.07364270836114883, 'train/grad_norm_pre_clip': 0.9418159127235413, 'timing/data': 0.0003102542832493782, 'timing/model': 0.6178173068910837, 'learning_rate/qwen_vl_interface': 9.007921766867823e-06, 'learning_rate/action_model': 9.007921766867822e-05, 'epoch': 0.35})
+[2026-07-28 13:35:15,467][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.04544823616743088, 'train/grad_norm_pre_clip': 1.9934358596801758, 'timing/data': 0.0003246292471885681, 'timing/model': 0.840838341973722, 'learning_rate/qwen_vl_interface': 9.005580256303804e-06, 'learning_rate/action_model': 9.005580256303803e-05, 'epoch': 0.35})
+[2026-07-28 13:35:24,552][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.03806551918387413, 'train/grad_norm_pre_clip': 0.820195734500885, 'timing/data': 0.0003292309120297432, 'timing/model': 0.615773132070899, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.35})
+[2026-07-28 13:35:33,674][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.03780040144920349, 'train/grad_norm_pre_clip': 0.9742571115493774, 'timing/data': 0.0004637250676751137, 'timing/model': 0.6176801444962621, 'learning_rate/qwen_vl_interface': 9.000889925830964e-06, 'learning_rate/action_model': 9.000889925830965e-05, 'epoch': 0.35})
+[2026-07-28 13:35:43,006][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.04612250253558159, 'train/grad_norm_pre_clip': 1.2673426866531372, 'timing/data': 0.0003232676535844803, 'timing/model': 0.6179131958633661, 'learning_rate/qwen_vl_interface': 8.998541108965648e-06, 'learning_rate/action_model': 8.998541108965648e-05, 'epoch': 0.35})
+[2026-07-28 13:35:52,075][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.026208728551864624, 'train/grad_norm_pre_clip': 0.971337080001831, 'timing/data': 0.00045981351286172867, 'timing/model': 0.6180132608860731, 'learning_rate/qwen_vl_interface': 8.996189859712704e-06, 'learning_rate/action_model': 8.996189859712703e-05, 'epoch': 0.35})
+[2026-07-28 13:36:01,415][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.05348537489771843, 'train/grad_norm_pre_clip': 1.2760708332061768, 'timing/data': 0.0003383392468094826, 'timing/model': 0.6182144144549966, 'learning_rate/qwen_vl_interface': 8.993836179597833e-06, 'learning_rate/action_model': 8.993836179597833e-05, 'epoch': 0.35})
+[2026-07-28 13:36:10,506][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.0399322435259819, 'train/grad_norm_pre_clip': 1.1774665117263794, 'timing/data': 0.00032993126660585403, 'timing/model': 0.6202189950272441, 'learning_rate/qwen_vl_interface': 8.99148007014831e-06, 'learning_rate/action_model': 8.99148007014831e-05, 'epoch': 0.35})
+[2026-07-28 13:36:19,579][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.026294568553566933, 'train/grad_norm_pre_clip': 0.7206201553344727, 'timing/data': 0.000320998951792717, 'timing/model': 0.6167859584093094, 'learning_rate/qwen_vl_interface': 8.989121532892989e-06, 'learning_rate/action_model': 8.989121532892988e-05, 'epoch': 0.35})
+[2026-07-28 13:36:28,887][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.01760033518075943, 'train/grad_norm_pre_clip': 0.49959826469421387, 'timing/data': 0.0003925105556845665, 'timing/model': 0.6172276642173529, 'learning_rate/qwen_vl_interface': 8.986760569362302e-06, 'learning_rate/action_model': 8.986760569362301e-05, 'epoch': 0.35})
+[2026-07-28 13:36:37,934][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.057797957211732864, 'train/grad_norm_pre_clip': 2.7979376316070557, 'timing/data': 0.00031993351876735687, 'timing/model': 0.614796930924058, 'learning_rate/qwen_vl_interface': 8.984397181088251e-06, 'learning_rate/action_model': 8.984397181088251e-05, 'epoch': 0.35})
+[2026-07-28 13:36:47,252][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.033989157527685165, 'train/grad_norm_pre_clip': 1.7709311246871948, 'timing/data': 0.0003293771296739578, 'timing/model': 0.8459270047023892, 'learning_rate/qwen_vl_interface': 8.982031369604412e-06, 'learning_rate/action_model': 8.982031369604411e-05, 'epoch': 0.35})
+[2026-07-28 13:36:56,358][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.02273937501013279, 'train/grad_norm_pre_clip': 1.1454237699508667, 'timing/data': 0.0003197910264134407, 'timing/model': 0.6340546011924744, 'learning_rate/qwen_vl_interface': 8.979663136445936e-06, 'learning_rate/action_model': 8.979663136445935e-05, 'epoch': 0.35})
+[2026-07-28 13:37:05,680][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.029589055106043816, 'train/grad_norm_pre_clip': 1.009171485900879, 'timing/data': 0.0003340495750308037, 'timing/model': 0.856941276229918, 'learning_rate/qwen_vl_interface': 8.977292483149542e-06, 'learning_rate/action_model': 8.977292483149542e-05, 'epoch': 0.35})
+[2026-07-28 13:37:14,779][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.044160328805446625, 'train/grad_norm_pre_clip': 1.7388304471969604, 'timing/data': 0.0003330167382955551, 'timing/model': 0.6158232120797038, 'learning_rate/qwen_vl_interface': 8.974919411253523e-06, 'learning_rate/action_model': 8.974919411253523e-05, 'epoch': 0.35})
+[2026-07-28 13:37:23,888][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.032640136778354645, 'train/grad_norm_pre_clip': 1.5375161170959473, 'timing/data': 0.00033357273787260056, 'timing/model': 0.6169961364939809, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.35})
+[2026-07-28 13:37:33,215][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.038900457322597504, 'train/grad_norm_pre_clip': 1.4539283514022827, 'timing/data': 0.00039921049028635025, 'timing/model': 0.6185533693060279, 'learning_rate/qwen_vl_interface': 8.970166017823621e-06, 'learning_rate/action_model': 8.97016601782362e-05, 'epoch': 0.36})
+[2026-07-28 13:37:42,304][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.016712870448827744, 'train/grad_norm_pre_clip': 0.6880984306335449, 'timing/data': 0.0003167390823364258, 'timing/model': 0.6146280597895384, 'learning_rate/qwen_vl_interface': 8.96778569937416e-06, 'learning_rate/action_model': 8.96778569937416e-05, 'epoch': 0.36})
+[2026-07-28 13:37:51,618][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.04715745151042938, 'train/grad_norm_pre_clip': 1.5130610466003418, 'timing/data': 0.00040522869676351547, 'timing/model': 0.6198730235919356, 'learning_rate/qwen_vl_interface': 8.965402968493922e-06, 'learning_rate/action_model': 8.965402968493922e-05, 'epoch': 0.36})
+[2026-07-28 13:38:00,913][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.07800108194351196, 'train/grad_norm_pre_clip': 2.446390151977539, 'timing/data': 0.00031931977719068527, 'timing/model': 0.615926525555551, 'learning_rate/qwen_vl_interface': 8.96301782672904e-06, 'learning_rate/action_model': 8.963017826729038e-05, 'epoch': 0.36})
+[2026-07-28 13:38:09,999][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.028859741985797882, 'train/grad_norm_pre_clip': 0.7323897480964661, 'timing/data': 0.00032267719507217407, 'timing/model': 0.6174290897324681, 'learning_rate/qwen_vl_interface': 8.960630275627197e-06, 'learning_rate/action_model': 8.960630275627197e-05, 'epoch': 0.36})
+[2026-07-28 13:38:19,101][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.03457747399806976, 'train/grad_norm_pre_clip': 1.4769397974014282, 'timing/data': 0.00032507162541151047, 'timing/model': 0.616988436318934, 'learning_rate/qwen_vl_interface': 8.958240316737655e-06, 'learning_rate/action_model': 8.958240316737656e-05, 'epoch': 0.36})
+[2026-07-28 13:38:28,434][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.061091870069503784, 'train/grad_norm_pre_clip': 1.769493818283081, 'timing/data': 0.00034279748797416687, 'timing/model': 0.8523440845310688, 'learning_rate/qwen_vl_interface': 8.955847951611233e-06, 'learning_rate/action_model': 8.955847951611232e-05, 'epoch': 0.36})
+[2026-07-28 13:38:37,741][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.06898234039545059, 'train/grad_norm_pre_clip': 1.6146589517593384, 'timing/data': 0.0003275657072663307, 'timing/model': 0.6167624806985259, 'learning_rate/qwen_vl_interface': 8.953453181800307e-06, 'learning_rate/action_model': 8.953453181800306e-05, 'epoch': 0.36})
+[2026-07-28 13:38:46,809][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.037169843912124634, 'train/grad_norm_pre_clip': 0.8699726462364197, 'timing/data': 0.0003929426893591881, 'timing/model': 0.6164139034226537, 'learning_rate/qwen_vl_interface': 8.951056008858819e-06, 'learning_rate/action_model': 8.95105600885882e-05, 'epoch': 0.36})
+[2026-07-28 13:38:55,940][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.020486576482653618, 'train/grad_norm_pre_clip': 1.1123688220977783, 'timing/data': 0.0003260774537920952, 'timing/model': 0.6170000778511167, 'learning_rate/qwen_vl_interface': 8.94865643434227e-06, 'learning_rate/action_model': 8.948656434342269e-05, 'epoch': 0.36})
+[2026-07-28 13:39:05,017][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.06405781209468842, 'train/grad_norm_pre_clip': 1.1704696416854858, 'timing/data': 0.00032457709312438965, 'timing/model': 0.61663568764925, 'learning_rate/qwen_vl_interface': 8.946254459807717e-06, 'learning_rate/action_model': 8.946254459807717e-05, 'epoch': 0.36})
+[2026-07-28 13:39:14,329][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.07102262228727341, 'train/grad_norm_pre_clip': 1.8092304468154907, 'timing/data': 0.0003294050693511963, 'timing/model': 0.8519266489893198, 'learning_rate/qwen_vl_interface': 8.943850086813773e-06, 'learning_rate/action_model': 8.943850086813773e-05, 'epoch': 0.36})
+[2026-07-28 13:39:23,630][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.07433998584747314, 'train/grad_norm_pre_clip': 0.996615469455719, 'timing/data': 0.00032904837280511856, 'timing/model': 0.6329173985868692, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.36})
+[2026-07-28 13:39:32,701][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.06534484773874283, 'train/grad_norm_pre_clip': 2.403557300567627, 'timing/data': 0.0003402670845389366, 'timing/model': 0.6176732070744038, 'learning_rate/qwen_vl_interface': 8.93903415168996e-06, 'learning_rate/action_model': 8.93903415168996e-05, 'epoch': 0.36})
+[2026-07-28 13:39:41,820][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.0320582315325737, 'train/grad_norm_pre_clip': 0.8537276983261108, 'timing/data': 0.00032936036586761475, 'timing/model': 0.6173063814640045, 'learning_rate/qwen_vl_interface': 8.9366225926851e-06, 'learning_rate/action_model': 8.936622592685098e-05, 'epoch': 0.36})
+[2026-07-28 13:39:50,908][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.0914105772972107, 'train/grad_norm_pre_clip': 1.4178516864776611, 'timing/data': 0.0003533335402607918, 'timing/model': 0.6168014919385314, 'learning_rate/qwen_vl_interface': 8.934208641470863e-06, 'learning_rate/action_model': 8.934208641470863e-05, 'epoch': 0.36})
+[2026-07-28 13:40:00,237][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.05136547610163689, 'train/grad_norm_pre_clip': 0.9441273212432861, 'timing/data': 0.0006778417155146599, 'timing/model': 0.6219086451455951, 'learning_rate/qwen_vl_interface': 8.931792299613637e-06, 'learning_rate/action_model': 8.931792299613636e-05, 'epoch': 0.36})
+[2026-07-28 13:40:09,561][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.03102109208703041, 'train/grad_norm_pre_clip': 1.7874096632003784, 'timing/data': 0.00032504182308912277, 'timing/model': 0.6184010598808527, 'learning_rate/qwen_vl_interface': 8.92937356868136e-06, 'learning_rate/action_model': 8.92937356868136e-05, 'epoch': 0.36})
+[2026-07-28 13:40:18,634][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.03500346094369888, 'train/grad_norm_pre_clip': 1.1536349058151245, 'timing/data': 0.00032450444996356964, 'timing/model': 0.6187466196715832, 'learning_rate/qwen_vl_interface': 8.926952450243522e-06, 'learning_rate/action_model': 8.926952450243522e-05, 'epoch': 0.36})
+[2026-07-28 13:40:27,732][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.045768171548843384, 'train/grad_norm_pre_clip': 0.9579029083251953, 'timing/data': 0.00032447557896375656, 'timing/model': 0.6183880977332592, 'learning_rate/qwen_vl_interface': 8.924528945871157e-06, 'learning_rate/action_model': 8.924528945871156e-05, 'epoch': 0.36})
+[2026-07-28 13:40:36,831][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.0472494512796402, 'train/grad_norm_pre_clip': 0.9450500011444092, 'timing/data': 0.00041606929153203964, 'timing/model': 0.6175870141014457, 'learning_rate/qwen_vl_interface': 8.922103057136857e-06, 'learning_rate/action_model': 8.922103057136856e-05, 'epoch': 0.36})
+[2026-07-28 13:40:46,152][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.03041756898164749, 'train/grad_norm_pre_clip': 0.8052560687065125, 'timing/data': 0.0003238227218389511, 'timing/model': 0.6211142484098673, 'learning_rate/qwen_vl_interface': 8.919674785614747e-06, 'learning_rate/action_model': 8.919674785614747e-05, 'epoch': 0.36})
+[2026-07-28 13:40:55,420][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.052525945007801056, 'train/grad_norm_pre_clip': 1.2803373336791992, 'timing/data': 0.0003211461007595062, 'timing/model': 0.8221837915480137, 'learning_rate/qwen_vl_interface': 8.91724413288051e-06, 'learning_rate/action_model': 8.91724413288051e-05, 'epoch': 0.36})
+[2026-07-28 13:41:04,517][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.05003604292869568, 'train/grad_norm_pre_clip': 1.7431766986846924, 'timing/data': 0.00032321643084287643, 'timing/model': 0.6172793032601476, 'learning_rate/qwen_vl_interface': 8.91481110051137e-06, 'learning_rate/action_model': 8.91481110051137e-05, 'epoch': 0.36})
+[2026-07-28 13:41:13,596][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.0195660088211298, 'train/grad_norm_pre_clip': 0.5812111496925354, 'timing/data': 0.0003598751500248909, 'timing/model': 0.6227564746513963, 'learning_rate/qwen_vl_interface': 8.912375690086097e-06, 'learning_rate/action_model': 8.912375690086096e-05, 'epoch': 0.36})
+[2026-07-28 13:41:22,704][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.039510153234004974, 'train/grad_norm_pre_clip': 1.7332122325897217, 'timing/data': 0.0003650262951850891, 'timing/model': 0.6227944986894727, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.36})
+[2026-07-28 13:41:32,043][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.034280672669410706, 'train/grad_norm_pre_clip': 1.5724906921386719, 'timing/data': 0.00038776081055402756, 'timing/model': 0.6326417019590735, 'learning_rate/qwen_vl_interface': 8.907497741389932e-06, 'learning_rate/action_model': 8.907497741389932e-05, 'epoch': 0.36})
+[2026-07-28 13:41:41,371][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.037399858236312866, 'train/grad_norm_pre_clip': 1.2113598585128784, 'timing/data': 0.00037211179733276367, 'timing/model': 0.6361636677756906, 'learning_rate/qwen_vl_interface': 8.905055206284289e-06, 'learning_rate/action_model': 8.905055206284288e-05, 'epoch': 0.36})
+[2026-07-28 13:41:50,469][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.029081575572490692, 'train/grad_norm_pre_clip': 1.1423810720443726, 'timing/data': 0.0003926018252968788, 'timing/model': 0.6164363250136375, 'learning_rate/qwen_vl_interface': 8.902610299453005e-06, 'learning_rate/action_model': 8.902610299453004e-05, 'epoch': 0.36})
+[2026-07-28 13:41:59,543][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.07204999774694443, 'train/grad_norm_pre_clip': 1.8250435590744019, 'timing/data': 0.0003802785649895668, 'timing/model': 0.617331625893712, 'learning_rate/qwen_vl_interface': 8.900163022482554e-06, 'learning_rate/action_model': 8.900163022482553e-05, 'epoch': 0.36})
+[2026-07-28 13:42:08,889][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.032913997769355774, 'train/grad_norm_pre_clip': 1.3678768873214722, 'timing/data': 0.0003876360133290291, 'timing/model': 0.878272145986557, 'learning_rate/qwen_vl_interface': 8.897713376960946e-06, 'learning_rate/action_model': 8.897713376960946e-05, 'epoch': 0.36})
+[2026-07-28 13:42:18,177][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.06229177862405777, 'train/grad_norm_pre_clip': 1.3954136371612549, 'timing/data': 0.0003769751638174057, 'timing/model': 0.8312528245151043, 'learning_rate/qwen_vl_interface': 8.895261364477731e-06, 'learning_rate/action_model': 8.895261364477731e-05, 'epoch': 0.36})
+[2026-07-28 13:42:27,279][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.05638020485639572, 'train/grad_norm_pre_clip': 2.557941198348999, 'timing/data': 0.00036834273487329483, 'timing/model': 0.6172109106555581, 'learning_rate/qwen_vl_interface': 8.892806986623993e-06, 'learning_rate/action_model': 8.892806986623992e-05, 'epoch': 0.36})
+[2026-07-28 13:42:36,365][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.08366359770298004, 'train/grad_norm_pre_clip': 2.3249940872192383, 'timing/data': 0.0003842059522867203, 'timing/model': 0.6157104447484016, 'learning_rate/qwen_vl_interface': 8.890350244992349e-06, 'learning_rate/action_model': 8.890350244992348e-05, 'epoch': 0.36})
+[2026-07-28 13:42:45,427][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.0305643267929554, 'train/grad_norm_pre_clip': 1.367679238319397, 'timing/data': 0.0003882395103573799, 'timing/model': 0.6173720303922892, 'learning_rate/qwen_vl_interface': 8.887891141176954e-06, 'learning_rate/action_model': 8.887891141176954e-05, 'epoch': 0.36})
+[2026-07-28 13:42:54,739][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.0637790635228157, 'train/grad_norm_pre_clip': 4.770643711090088, 'timing/data': 0.0003817034885287285, 'timing/model': 0.6169850956648588, 'learning_rate/qwen_vl_interface': 8.885429676773495e-06, 'learning_rate/action_model': 8.885429676773495e-05, 'epoch': 0.36})
+[2026-07-28 13:43:03,804][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.04143516719341278, 'train/grad_norm_pre_clip': 0.9393490552902222, 'timing/data': 0.000322781503200531, 'timing/model': 0.6167880650609732, 'learning_rate/qwen_vl_interface': 8.882965853379188e-06, 'learning_rate/action_model': 8.882965853379187e-05, 'epoch': 0.36})
+[2026-07-28 13:43:12,681][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.03932001069188118, 'train/grad_norm_pre_clip': 1.5571404695510864, 'timing/data': 0.0004171738401055336, 'timing/model': 0.617434992454946, 'learning_rate/qwen_vl_interface': 8.88049967259278e-06, 'learning_rate/action_model': 8.880499672592778e-05, 'epoch': 0.36})
+[2026-07-28 13:43:22,009][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.03175737336277962, 'train/grad_norm_pre_clip': 1.1470602750778198, 'timing/data': 0.0003284122794866562, 'timing/model': 0.8434741543605924, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.36})
+[2026-07-28 13:43:31,331][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.04616511985659599, 'train/grad_norm_pre_clip': 1.4775969982147217, 'timing/data': 0.0003936765715479851, 'timing/model': 0.8536782171577215, 'learning_rate/qwen_vl_interface': 8.875560245246307e-06, 'learning_rate/action_model': 8.875560245246307e-05, 'epoch': 0.36})
+[2026-07-28 13:43:40,426][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.0521186962723732, 'train/grad_norm_pre_clip': 0.9889653325080872, 'timing/data': 0.0003851214423775673, 'timing/model': 0.617436152882874, 'learning_rate/qwen_vl_interface': 8.873087001891381e-06, 'learning_rate/action_model': 8.87308700189138e-05, 'epoch': 0.36})
+[2026-07-28 13:43:49,502][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.05006113648414612, 'train/grad_norm_pre_clip': 2.159437656402588, 'timing/data': 0.0003954032436013222, 'timing/model': 0.6162291970103979, 'learning_rate/qwen_vl_interface': 8.870611407554635e-06, 'learning_rate/action_model': 8.870611407554636e-05, 'epoch': 0.36})
+[2026-07-28 13:43:58,584][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.06507733464241028, 'train/grad_norm_pre_clip': 1.8420641422271729, 'timing/data': 0.0003358125686645508, 'timing/model': 0.6176161505281925, 'learning_rate/qwen_vl_interface': 8.868133463842456e-06, 'learning_rate/action_model': 8.868133463842457e-05, 'epoch': 0.36})
+[2026-07-28 13:44:07,897][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.03631630539894104, 'train/grad_norm_pre_clip': 0.6857712864875793, 'timing/data': 0.0005688974633812904, 'timing/model': 0.616558350622654, 'learning_rate/qwen_vl_interface': 8.865653172362755e-06, 'learning_rate/action_model': 8.865653172362754e-05, 'epoch': 0.36})
+[2026-07-28 13:44:17,231][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.032435718923807144, 'train/grad_norm_pre_clip': 0.8860750198364258, 'timing/data': 0.00040988530963659286, 'timing/model': 0.6159818125888705, 'learning_rate/qwen_vl_interface': 8.863170534724962e-06, 'learning_rate/action_model': 8.863170534724961e-05, 'epoch': 0.36})
+[2026-07-28 13:44:26,305][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.05482558533549309, 'train/grad_norm_pre_clip': 1.0712143182754517, 'timing/data': 0.0003377804532647133, 'timing/model': 0.6185596901923418, 'learning_rate/qwen_vl_interface': 8.860685552540038e-06, 'learning_rate/action_model': 8.860685552540037e-05, 'epoch': 0.36})
+[2026-07-28 13:44:35,174][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.02668331190943718, 'train/grad_norm_pre_clip': 0.7165483832359314, 'timing/data': 0.0003249403089284897, 'timing/model': 0.6157616842538118, 'learning_rate/qwen_vl_interface': 8.858198227420461e-06, 'learning_rate/action_model': 8.85819822742046e-05, 'epoch': 0.36})
+[2026-07-28 13:44:44,263][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.05173656344413757, 'train/grad_norm_pre_clip': 0.9319248795509338, 'timing/data': 0.00032027531415224075, 'timing/model': 0.6331554651260376, 'learning_rate/qwen_vl_interface': 8.855708560980225e-06, 'learning_rate/action_model': 8.855708560980225e-05, 'epoch': 0.36})
+[2026-07-28 13:44:53,422][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.029863249510526657, 'train/grad_norm_pre_clip': 1.225673794746399, 'timing/data': 0.00032798293977975845, 'timing/model': 0.8832989456132054, 'learning_rate/qwen_vl_interface': 8.853216554834852e-06, 'learning_rate/action_model': 8.85321655483485e-05, 'epoch': 0.36})
+[2026-07-28 13:45:02,433][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.04056299105286598, 'train/grad_norm_pre_clip': 0.949761688709259, 'timing/data': 0.000401567667722702, 'timing/model': 0.6144252829253674, 'learning_rate/qwen_vl_interface': 8.850722210601376e-06, 'learning_rate/action_model': 8.850722210601377e-05, 'epoch': 0.36})
+[2026-07-28 13:45:11,651][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.04156433045864105, 'train/grad_norm_pre_clip': 0.7822466492652893, 'timing/data': 0.00032764114439487457, 'timing/model': 0.988401641137898, 'learning_rate/qwen_vl_interface': 8.848225529898351e-06, 'learning_rate/action_model': 8.84822552989835e-05, 'epoch': 0.36})
+[2026-07-28 13:45:20,865][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.03739568218588829, 'train/grad_norm_pre_clip': 1.2140426635742188, 'timing/data': 0.00035536661744117737, 'timing/model': 0.6307061463594437, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.36})
+[2026-07-28 13:45:30,121][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.01837325096130371, 'train/grad_norm_pre_clip': 1.0755128860473633, 'timing/data': 0.0004017464816570282, 'timing/model': 0.6188384993001819, 'learning_rate/qwen_vl_interface': 8.84322516556544e-06, 'learning_rate/action_model': 8.84322516556544e-05, 'epoch': 0.36})
+[2026-07-28 13:45:39,661][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.05127568542957306, 'train/grad_norm_pre_clip': 1.8450043201446533, 'timing/data': 0.00033478066325187683, 'timing/model': 0.6169400103390217, 'learning_rate/qwen_vl_interface': 8.840721485180237e-06, 'learning_rate/action_model': 8.840721485180236e-05, 'epoch': 0.36})
+[2026-07-28 13:45:48,747][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.05773169547319412, 'train/grad_norm_pre_clip': 2.3727669715881348, 'timing/data': 0.00032854173332452774, 'timing/model': 0.616304425522685, 'learning_rate/qwen_vl_interface': 8.838215474814848e-06, 'learning_rate/action_model': 8.838215474814847e-05, 'epoch': 0.36})
+[2026-07-28 13:45:57,859][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.05011668801307678, 'train/grad_norm_pre_clip': 1.760660171508789, 'timing/data': 0.0003922227770090103, 'timing/model': 0.6150147402659059, 'learning_rate/qwen_vl_interface': 8.835707136095392e-06, 'learning_rate/action_model': 8.835707136095391e-05, 'epoch': 0.37})
+[2026-07-28 13:46:07,176][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.04394857957959175, 'train/grad_norm_pre_clip': 1.7909162044525146, 'timing/data': 0.00031301379203796387, 'timing/model': 0.6186968544498086, 'learning_rate/qwen_vl_interface': 8.833196470649505e-06, 'learning_rate/action_model': 8.833196470649505e-05, 'epoch': 0.37})
+[2026-07-28 13:46:16,481][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.09875864535570145, 'train/grad_norm_pre_clip': 3.530860662460327, 'timing/data': 0.00039754994213581085, 'timing/model': 0.8475725837051868, 'learning_rate/qwen_vl_interface': 8.83068348010633e-06, 'learning_rate/action_model': 8.83068348010633e-05, 'epoch': 0.37})
+[2026-07-28 13:46:25,565][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.07254365086555481, 'train/grad_norm_pre_clip': 2.1204440593719482, 'timing/data': 0.00040502194315195084, 'timing/model': 0.6197163509204984, 'learning_rate/qwen_vl_interface': 8.82816816609652e-06, 'learning_rate/action_model': 8.828168166096521e-05, 'epoch': 0.37})
+[2026-07-28 13:46:34,644][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.06216660887002945, 'train/grad_norm_pre_clip': 1.5215562582015991, 'timing/data': 0.00039661210030317307, 'timing/model': 0.6147826937958598, 'learning_rate/qwen_vl_interface': 8.825650530252238e-06, 'learning_rate/action_model': 8.825650530252238e-05, 'epoch': 0.37})
+[2026-07-28 13:46:43,931][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.04733988270163536, 'train/grad_norm_pre_clip': 1.2800010442733765, 'timing/data': 0.0003209207206964493, 'timing/model': 0.6156240338459611, 'learning_rate/qwen_vl_interface': 8.823130574207144e-06, 'learning_rate/action_model': 8.823130574207143e-05, 'epoch': 0.37})
+[2026-07-28 13:46:53,000][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.07626649737358093, 'train/grad_norm_pre_clip': 2.0063624382019043, 'timing/data': 0.00034779682755470276, 'timing/model': 0.6153202336281538, 'learning_rate/qwen_vl_interface': 8.82060829959641e-06, 'learning_rate/action_model': 8.820608299596411e-05, 'epoch': 0.37})
+[2026-07-28 13:47:02,307][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.030319271609187126, 'train/grad_norm_pre_clip': 0.8438180685043335, 'timing/data': 0.0003069760277867317, 'timing/model': 0.617860802449286, 'learning_rate/qwen_vl_interface': 8.81808370805672e-06, 'learning_rate/action_model': 8.818083708056721e-05, 'epoch': 0.37})
+[2026-07-28 13:47:11,620][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.04671825096011162, 'train/grad_norm_pre_clip': 1.7050837278366089, 'timing/data': 0.00039588380604982376, 'timing/model': 0.8377083456143737, 'learning_rate/qwen_vl_interface': 8.815556801226251e-06, 'learning_rate/action_model': 8.815556801226251e-05, 'epoch': 0.37})
+[2026-07-28 13:47:20,735][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.02550564706325531, 'train/grad_norm_pre_clip': 1.5800144672393799, 'timing/data': 0.00039311032742261887, 'timing/model': 0.6150561841204762, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.37})
+[2026-07-28 13:47:29,804][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.0505886934697628, 'train/grad_norm_pre_clip': 1.3172824382781982, 'timing/data': 0.0003823796287178993, 'timing/model': 0.6248427759855986, 'learning_rate/qwen_vl_interface': 8.810496048253203e-06, 'learning_rate/action_model': 8.810496048253202e-05, 'epoch': 0.37})
+[2026-07-28 13:47:38,931][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.022836050018668175, 'train/grad_norm_pre_clip': 1.1009420156478882, 'timing/data': 0.0003115888684988022, 'timing/model': 0.6328096147626638, 'learning_rate/qwen_vl_interface': 8.80796220539449e-06, 'learning_rate/action_model': 8.80796220539449e-05, 'epoch': 0.37})
+[2026-07-28 13:47:47,783][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000
+[2026-07-28 13:47:58,181][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000
+[2026-07-28 13:58:37,057][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000
+[2026-07-28 13:58:37,059][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.061030685901641846, 'train/grad_norm_pre_clip': 2.4788341522216797, 'eval/action_loss/seconds': 10.395160723477602, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.04522417485713959, 'eval/latency_3/loss': 0.04522417485713959, 'eval/flappy/loss': 0.04522417485713959, 'eval/flappy/latency_3/loss': 0.04522417485713959, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 214.70500253476203, 'rl_games_eval/mid_train/mean_length': 1747.45, 'rl_games_eval/mid_train/std_reward': 161.25401726393875, 'rl_games_eval/mid_train/std_length': 1301.0273431023654, 'rl_games_eval/mid_train/macro_mean_reward': 214.70500253476203, 'rl_games_eval/mid_train/macro_mean_length': 1747.45, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 214.70500253476203, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 1747.45, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 161.25401726393875, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1301.0273431023654, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 214.70500253476203, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0004272507503628731, 'timing/model': 0.6169736478477716, 'learning_rate/qwen_vl_interface': 8.805426053812731e-06, 'learning_rate/action_model': 8.805426053812731e-05, 'epoch': 0.37})
+[2026-07-28 13:58:37,061][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2000_state
+[2026-07-28 13:58:37,061][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 13:59:14,297][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2000_state/pytorch_model
+[2026-07-28 13:59:14,298][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2000_state/sampler.bin
+[2026-07-28 13:59:14,298][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2000_state/sampler_1.bin
+[2026-07-28 13:59:14,300][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2000_state/random_states_0.pkl
+[2026-07-28 13:59:22,341][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000
+[2026-07-28 13:59:22,341][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 13:59:22,363][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 13:59:31,568][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.06864193081855774, 'train/grad_norm_pre_clip': 2.046590566635132, 'timing/data': 0.00040469132363796234, 'timing/model': 0.61530083976686, 'learning_rate/qwen_vl_interface': 8.802887595153607e-06, 'learning_rate/action_model': 8.802887595153606e-05, 'epoch': 0.37})
+[2026-07-28 13:59:40,867][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.03766249865293503, 'train/grad_norm_pre_clip': 1.9252655506134033, 'timing/data': 0.00040595512837171555, 'timing/model': 0.6158917136490345, 'learning_rate/qwen_vl_interface': 8.800346831064294e-06, 'learning_rate/action_model': 8.800346831064294e-05, 'epoch': 0.37})
+[2026-07-28 13:59:49,930][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.08938194811344147, 'train/grad_norm_pre_clip': 2.000363349914551, 'timing/data': 0.0007972028106451035, 'timing/model': 0.6146765742450953, 'learning_rate/qwen_vl_interface': 8.797803763193471e-06, 'learning_rate/action_model': 8.797803763193471e-05, 'epoch': 0.37})
+[2026-07-28 13:59:59,057][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.0390050932765007, 'train/grad_norm_pre_clip': 2.3287110328674316, 'timing/data': 0.00033026933670043945, 'timing/model': 0.6162609048187733, 'learning_rate/qwen_vl_interface': 8.795258393191302e-06, 'learning_rate/action_model': 8.795258393191301e-05, 'epoch': 0.37})
+[2026-07-28 14:00:08,412][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.0518568679690361, 'train/grad_norm_pre_clip': 1.3752548694610596, 'timing/data': 0.00032853521406650543, 'timing/model': 0.6145437797531486, 'learning_rate/qwen_vl_interface': 8.792710722709451e-06, 'learning_rate/action_model': 8.792710722709451e-05, 'epoch': 0.37})
+[2026-07-28 14:00:17,709][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.08170551061630249, 'train/grad_norm_pre_clip': 2.0958385467529297, 'timing/data': 0.00032745953649282455, 'timing/model': 0.6173415062949061, 'learning_rate/qwen_vl_interface': 8.790160753401076e-06, 'learning_rate/action_model': 8.790160753401077e-05, 'epoch': 0.37})
+[2026-07-28 14:00:26,797][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.046748846769332886, 'train/grad_norm_pre_clip': 1.0325204133987427, 'timing/data': 0.00033772923052310944, 'timing/model': 0.6162242293357849, 'learning_rate/qwen_vl_interface': 8.787608486920824e-06, 'learning_rate/action_model': 8.787608486920824e-05, 'epoch': 0.37})
+[2026-07-28 14:00:35,646][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.0700296014547348, 'train/grad_norm_pre_clip': 1.7677241563796997, 'timing/data': 0.00033387355506420135, 'timing/model': 0.6184927336871624, 'learning_rate/qwen_vl_interface': 8.78505392492483e-06, 'learning_rate/action_model': 8.78505392492483e-05, 'epoch': 0.37})
+[2026-07-28 14:00:44,491][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.02908051572740078, 'train/grad_norm_pre_clip': 0.82437664270401, 'timing/data': 0.00033505260944366455, 'timing/model': 0.6186279496178031, 'learning_rate/qwen_vl_interface': 8.782497069070725e-06, 'learning_rate/action_model': 8.782497069070725e-05, 'epoch': 0.37})
+[2026-07-28 14:00:53,994][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.03846253454685211, 'train/grad_norm_pre_clip': 1.0973186492919922, 'timing/data': 0.0004244251176714897, 'timing/model': 0.8328185239806771, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.37})
+[2026-07-28 14:01:03,057][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.07391060888767242, 'train/grad_norm_pre_clip': 1.2454293966293335, 'timing/data': 0.00034410785883665085, 'timing/model': 0.616247046738863, 'learning_rate/qwen_vl_interface': 8.777376482426129e-06, 'learning_rate/action_model': 8.777376482426128e-05, 'epoch': 0.37})
+[2026-07-28 14:01:12,102][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.06309646368026733, 'train/grad_norm_pre_clip': 0.997200608253479, 'timing/data': 0.00039628706872463226, 'timing/model': 0.6165360575541854, 'learning_rate/qwen_vl_interface': 8.774812754958332e-06, 'learning_rate/action_model': 8.774812754958331e-05, 'epoch': 0.37})
+[2026-07-28 14:01:21,139][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.04304405301809311, 'train/grad_norm_pre_clip': 1.2573158740997314, 'timing/data': 0.0003448091447353363, 'timing/model': 0.6157452799379826, 'learning_rate/qwen_vl_interface': 8.772246740277805e-06, 'learning_rate/action_model': 8.772246740277805e-05, 'epoch': 0.37})
+[2026-07-28 14:01:30,407][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.05157874897122383, 'train/grad_norm_pre_clip': 1.3459504842758179, 'timing/data': 0.0003096098080277443, 'timing/model': 0.6151208141818643, 'learning_rate/qwen_vl_interface': 8.769678440049608e-06, 'learning_rate/action_model': 8.769678440049608e-05, 'epoch': 0.37})
+[2026-07-28 14:01:39,692][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.06215780973434448, 'train/grad_norm_pre_clip': 1.431925654411316, 'timing/data': 0.0003413120284676552, 'timing/model': 0.614301573485136, 'learning_rate/qwen_vl_interface': 8.767107855940287e-06, 'learning_rate/action_model': 8.767107855940286e-05, 'epoch': 0.37})
+[2026-07-28 14:01:48,764][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.047293536365032196, 'train/grad_norm_pre_clip': 0.9544617533683777, 'timing/data': 0.0003944654017686844, 'timing/model': 0.614156155847013, 'learning_rate/qwen_vl_interface': 8.764534989617861e-06, 'learning_rate/action_model': 8.76453498961786e-05, 'epoch': 0.37})
+[2026-07-28 14:01:57,808][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.026990395039319992, 'train/grad_norm_pre_clip': 0.7484792470932007, 'timing/data': 0.00032181013375520706, 'timing/model': 0.6162392236292362, 'learning_rate/qwen_vl_interface': 8.761959842751837e-06, 'learning_rate/action_model': 8.761959842751837e-05, 'epoch': 0.37})
+[2026-07-28 14:02:07,298][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.05288656800985336, 'train/grad_norm_pre_clip': 1.3026491403579712, 'timing/data': 0.0003390209749341011, 'timing/model': 0.6144297160208225, 'learning_rate/qwen_vl_interface': 8.7593824170132e-06, 'learning_rate/action_model': 8.7593824170132e-05, 'epoch': 0.37})
+[2026-07-28 14:02:16,342][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.035882268100976944, 'train/grad_norm_pre_clip': 1.0574160814285278, 'timing/data': 0.000324079766869545, 'timing/model': 0.6156867928802967, 'learning_rate/qwen_vl_interface': 8.756802714074412e-06, 'learning_rate/action_model': 8.756802714074412e-05, 'epoch': 0.37})
+[2026-07-28 14:02:25,200][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.06085926294326782, 'train/grad_norm_pre_clip': 1.3096050024032593, 'timing/data': 0.0003611566498875618, 'timing/model': 0.6147995507344604, 'learning_rate/qwen_vl_interface': 8.75422073560942e-06, 'learning_rate/action_model': 8.754220735609419e-05, 'epoch': 0.37})
+[2026-07-28 14:02:34,436][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.02245641127228737, 'train/grad_norm_pre_clip': 1.0735942125320435, 'timing/data': 0.0003537200391292572, 'timing/model': 0.7964963261038065, 'learning_rate/qwen_vl_interface': 8.751636483293635e-06, 'learning_rate/action_model': 8.751636483293634e-05, 'epoch': 0.37})
+[2026-07-28 14:02:43,488][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.06092720851302147, 'train/grad_norm_pre_clip': 1.6349493265151978, 'timing/data': 0.00032310839742422104, 'timing/model': 0.6154612191021442, 'learning_rate/qwen_vl_interface': 8.749049958803952e-06, 'learning_rate/action_model': 8.749049958803951e-05, 'epoch': 0.37})
+[2026-07-28 14:02:52,531][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.04446333646774292, 'train/grad_norm_pre_clip': 2.3084585666656494, 'timing/data': 0.0003263913094997406, 'timing/model': 0.6163499681279063, 'learning_rate/qwen_vl_interface': 8.74646116381874e-06, 'learning_rate/action_model': 8.74646116381874e-05, 'epoch': 0.37})
+[2026-07-28 14:03:01,566][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.019858941435813904, 'train/grad_norm_pre_clip': 0.5958759784698486, 'timing/data': 0.00031271111220121384, 'timing/model': 0.6172327343374491, 'learning_rate/qwen_vl_interface': 8.74387010001784e-06, 'learning_rate/action_model': 8.743870100017838e-05, 'epoch': 0.37})
+[2026-07-28 14:03:10,933][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.06724436581134796, 'train/grad_norm_pre_clip': 2.2708756923675537, 'timing/data': 0.0004196772351861, 'timing/model': 0.8943207198753953, 'learning_rate/qwen_vl_interface': 8.741276769082562e-06, 'learning_rate/action_model': 8.741276769082562e-05, 'epoch': 0.37})
+[2026-07-28 14:03:20,212][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.04449842870235443, 'train/grad_norm_pre_clip': 0.8497711420059204, 'timing/data': 0.00034522730857133865, 'timing/model': 0.6168093169108033, 'learning_rate/qwen_vl_interface': 8.738681172695694e-06, 'learning_rate/action_model': 8.738681172695694e-05, 'epoch': 0.37})
+[2026-07-28 14:03:29,286][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.04408183693885803, 'train/grad_norm_pre_clip': 0.8226591348648071, 'timing/data': 0.0003370484337210655, 'timing/model': 0.6165727945044637, 'learning_rate/qwen_vl_interface': 8.73608331254149e-06, 'learning_rate/action_model': 8.736083312541489e-05, 'epoch': 0.37})
+[2026-07-28 14:03:38,165][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.02896120771765709, 'train/grad_norm_pre_clip': 0.5858485698699951, 'timing/data': 0.00042713433504104614, 'timing/model': 0.6152370302006602, 'learning_rate/qwen_vl_interface': 8.733483190305671e-06, 'learning_rate/action_model': 8.733483190305671e-05, 'epoch': 0.37})
+[2026-07-28 14:03:47,223][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.04374150186777115, 'train/grad_norm_pre_clip': 0.9153249263763428, 'timing/data': 0.0003380514681339264, 'timing/model': 0.6153882769867778, 'learning_rate/qwen_vl_interface': 8.730880807675431e-06, 'learning_rate/action_model': 8.73088080767543e-05, 'epoch': 0.37})
+[2026-07-28 14:03:56,309][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.025873344391584396, 'train/grad_norm_pre_clip': 0.7173786163330078, 'timing/data': 0.0006024409085512161, 'timing/model': 0.8538172403350472, 'learning_rate/qwen_vl_interface': 8.728276166339426e-06, 'learning_rate/action_model': 8.728276166339426e-05, 'epoch': 0.37})
+[2026-07-28 14:04:05,151][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.025796392932534218, 'train/grad_norm_pre_clip': 1.4631017446517944, 'timing/data': 0.0003148987889289856, 'timing/model': 0.6181679824367166, 'learning_rate/qwen_vl_interface': 8.725669267987779e-06, 'learning_rate/action_model': 8.725669267987778e-05, 'epoch': 0.37})
+[2026-07-28 14:04:13,989][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.036017291247844696, 'train/grad_norm_pre_clip': 1.0363520383834839, 'timing/data': 0.0003419620916247368, 'timing/model': 0.6225198842585087, 'learning_rate/qwen_vl_interface': 8.723060114312079e-06, 'learning_rate/action_model': 8.723060114312079e-05, 'epoch': 0.37})
+[2026-07-28 14:04:23,064][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.042185720056295395, 'train/grad_norm_pre_clip': 1.815225601196289, 'timing/data': 0.00035011768341064453, 'timing/model': 0.6175677673891187, 'learning_rate/qwen_vl_interface': 8.720448707005377e-06, 'learning_rate/action_model': 8.720448707005376e-05, 'epoch': 0.37})
+[2026-07-28 14:04:32,125][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.05165144056081772, 'train/grad_norm_pre_clip': 1.6972665786743164, 'timing/data': 0.0004622749984264374, 'timing/model': 0.6178479660302401, 'learning_rate/qwen_vl_interface': 8.717835047762188e-06, 'learning_rate/action_model': 8.717835047762187e-05, 'epoch': 0.37})
+[2026-07-28 14:04:41,199][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.04594530165195465, 'train/grad_norm_pre_clip': 1.3120747804641724, 'timing/data': 0.0004025287926197052, 'timing/model': 0.6190088307484984, 'learning_rate/qwen_vl_interface': 8.715219138278486e-06, 'learning_rate/action_model': 8.715219138278486e-05, 'epoch': 0.37})
+[2026-07-28 14:04:50,489][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.03178905323147774, 'train/grad_norm_pre_clip': 1.3500733375549316, 'timing/data': 0.0004110904410481453, 'timing/model': 0.6213012021034956, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.37})
+[2026-07-28 14:04:59,797][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.03605347499251366, 'train/grad_norm_pre_clip': 1.5341715812683105, 'timing/data': 0.0003886884078383446, 'timing/model': 0.6202861275523901, 'learning_rate/qwen_vl_interface': 8.709980575380741e-06, 'learning_rate/action_model': 8.709980575380741e-05, 'epoch': 0.37})
+[2026-07-28 14:05:08,860][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.024193383753299713, 'train/grad_norm_pre_clip': 1.5487339496612549, 'timing/data': 0.0003232313320040703, 'timing/model': 0.6144242202863097, 'learning_rate/qwen_vl_interface': 8.707357925365949e-06, 'learning_rate/action_model': 8.707357925365948e-05, 'epoch': 0.37})
+[2026-07-28 14:05:17,927][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.04178386554121971, 'train/grad_norm_pre_clip': 1.1266001462936401, 'timing/data': 0.00034573301672935486, 'timing/model': 0.6146438969299197, 'learning_rate/qwen_vl_interface': 8.704733031909134e-06, 'learning_rate/action_model': 8.704733031909133e-05, 'epoch': 0.37})
+[2026-07-28 14:05:27,198][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.07962404191493988, 'train/grad_norm_pre_clip': 1.9046542644500732, 'timing/data': 0.00031623896211385727, 'timing/model': 0.8317461144179106, 'learning_rate/qwen_vl_interface': 8.70210589671356e-06, 'learning_rate/action_model': 8.702105896713561e-05, 'epoch': 0.37})
+[2026-07-28 14:05:36,259][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.03872215002775192, 'train/grad_norm_pre_clip': 1.0541191101074219, 'timing/data': 0.0003305254504084587, 'timing/model': 0.6143876546993852, 'learning_rate/qwen_vl_interface': 8.699476521483952e-06, 'learning_rate/action_model': 8.699476521483952e-05, 'epoch': 0.37})
+[2026-07-28 14:05:45,317][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.07531075924634933, 'train/grad_norm_pre_clip': 2.6293516159057617, 'timing/data': 0.00043816398829221725, 'timing/model': 0.6139540020376444, 'learning_rate/qwen_vl_interface': 8.696844907926478e-06, 'learning_rate/action_model': 8.696844907926477e-05, 'epoch': 0.38})
+[2026-07-28 14:05:54,595][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.05221433937549591, 'train/grad_norm_pre_clip': 1.2030036449432373, 'timing/data': 0.0003848513588309288, 'timing/model': 0.831795671954751, 'learning_rate/qwen_vl_interface': 8.694211057748766e-06, 'learning_rate/action_model': 8.694211057748765e-05, 'epoch': 0.38})
+[2026-07-28 14:06:03,926][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.03606964275240898, 'train/grad_norm_pre_clip': 0.8431869149208069, 'timing/data': 0.0003174189478158951, 'timing/model': 0.6387701015919447, 'learning_rate/qwen_vl_interface': 8.69157497265989e-06, 'learning_rate/action_model': 8.69157497265989e-05, 'epoch': 0.38})
+[2026-07-28 14:06:13,005][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.10910530388355255, 'train/grad_norm_pre_clip': 1.3939776420593262, 'timing/data': 0.00033536460250616074, 'timing/model': 0.6179234087467194, 'learning_rate/qwen_vl_interface': 8.68893665437038e-06, 'learning_rate/action_model': 8.688936654370378e-05, 'epoch': 0.38})
+[2026-07-28 14:06:22,120][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.03040684014558792, 'train/grad_norm_pre_clip': 1.2721025943756104, 'timing/data': 0.0003300448879599571, 'timing/model': 0.613846099935472, 'learning_rate/qwen_vl_interface': 8.68629610459221e-06, 'learning_rate/action_model': 8.686296104592209e-05, 'epoch': 0.38})
+[2026-07-28 14:06:31,466][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.02977500855922699, 'train/grad_norm_pre_clip': 0.6551321148872375, 'timing/data': 0.0003282204270362854, 'timing/model': 0.8333563525229692, 'learning_rate/qwen_vl_interface': 8.683653325038805e-06, 'learning_rate/action_model': 8.683653325038804e-05, 'epoch': 0.38})
+[2026-07-28 14:06:40,544][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.07709822058677673, 'train/grad_norm_pre_clip': 2.1907641887664795, 'timing/data': 0.0003271922469139099, 'timing/model': 0.6172791514545679, 'learning_rate/qwen_vl_interface': 8.681008317425037e-06, 'learning_rate/action_model': 8.681008317425037e-05, 'epoch': 0.38})
+[2026-07-28 14:06:49,626][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.04849596321582794, 'train/grad_norm_pre_clip': 2.0061657428741455, 'timing/data': 0.0003392426297068596, 'timing/model': 0.6166107142344117, 'learning_rate/qwen_vl_interface': 8.678361083467221e-06, 'learning_rate/action_model': 8.67836108346722e-05, 'epoch': 0.38})
+[2026-07-28 14:06:58,966][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.041654109954833984, 'train/grad_norm_pre_clip': 0.7041492462158203, 'timing/data': 0.00034952443093061447, 'timing/model': 0.8594902837648988, 'learning_rate/qwen_vl_interface': 8.67571162488312e-06, 'learning_rate/action_model': 8.67571162488312e-05, 'epoch': 0.38})
+[2026-07-28 14:07:08,088][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.05311371386051178, 'train/grad_norm_pre_clip': 0.8180781602859497, 'timing/data': 0.00032879970967769623, 'timing/model': 0.6189648881554604, 'learning_rate/qwen_vl_interface': 8.673059943391937e-06, 'learning_rate/action_model': 8.673059943391936e-05, 'epoch': 0.38})
+[2026-07-28 14:07:17,555][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.04877627268433571, 'train/grad_norm_pre_clip': 0.8446621894836426, 'timing/data': 0.00032575055956840515, 'timing/model': 0.6673104390501976, 'learning_rate/qwen_vl_interface': 8.670406040714321e-06, 'learning_rate/action_model': 8.670406040714321e-05, 'epoch': 0.38})
+[2026-07-28 14:07:26,740][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.04022345691919327, 'train/grad_norm_pre_clip': 1.362566351890564, 'timing/data': 0.00034711696207523346, 'timing/model': 0.624856106005609, 'learning_rate/qwen_vl_interface': 8.667749918572361e-06, 'learning_rate/action_model': 8.667749918572361e-05, 'epoch': 0.38})
+[2026-07-28 14:07:35,830][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.038787975907325745, 'train/grad_norm_pre_clip': 1.053065299987793, 'timing/data': 0.00042632222175598145, 'timing/model': 0.6166404318064451, 'learning_rate/qwen_vl_interface': 8.665091578689584e-06, 'learning_rate/action_model': 8.665091578689584e-05, 'epoch': 0.38})
+[2026-07-28 14:07:44,894][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.06384703516960144, 'train/grad_norm_pre_clip': 1.5992059707641602, 'timing/data': 0.00041685067117214203, 'timing/model': 0.617823438718915, 'learning_rate/qwen_vl_interface': 8.662431022790958e-06, 'learning_rate/action_model': 8.662431022790958e-05, 'epoch': 0.38})
+[2026-07-28 14:07:54,187][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.02779475599527359, 'train/grad_norm_pre_clip': 0.48626428842544556, 'timing/data': 0.00039585866034030914, 'timing/model': 0.6146978521719575, 'learning_rate/qwen_vl_interface': 8.65976825260289e-06, 'learning_rate/action_model': 8.659768252602888e-05, 'epoch': 0.38})
+[2026-07-28 14:08:03,453][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.08193245530128479, 'train/grad_norm_pre_clip': 2.9084253311157227, 'timing/data': 0.0008563296869397163, 'timing/model': 0.8274399153888226, 'learning_rate/qwen_vl_interface': 8.657103269853222e-06, 'learning_rate/action_model': 8.657103269853222e-05, 'epoch': 0.38})
+[2026-07-28 14:08:12,505][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.04273471236228943, 'train/grad_norm_pre_clip': 0.6776876449584961, 'timing/data': 0.00032345671206712723, 'timing/model': 0.6153354570269585, 'learning_rate/qwen_vl_interface': 8.654436076271232e-06, 'learning_rate/action_model': 8.654436076271232e-05, 'epoch': 0.38})
+[2026-07-28 14:08:21,567][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.02483772486448288, 'train/grad_norm_pre_clip': 0.7904509902000427, 'timing/data': 0.000328700989484787, 'timing/model': 0.6137089161202312, 'learning_rate/qwen_vl_interface': 8.651766673587635e-06, 'learning_rate/action_model': 8.651766673587635e-05, 'epoch': 0.38})
+[2026-07-28 14:08:30,861][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.06542718410491943, 'train/grad_norm_pre_clip': 1.6971865892410278, 'timing/data': 0.00036064721643924713, 'timing/model': 0.8492727894335985, 'learning_rate/qwen_vl_interface': 8.649095063534576e-06, 'learning_rate/action_model': 8.649095063534576e-05, 'epoch': 0.38})
+[2026-07-28 14:08:39,941][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.045555420219898224, 'train/grad_norm_pre_clip': 1.649233102798462, 'timing/data': 0.00033673737198114395, 'timing/model': 0.6145106730982661, 'learning_rate/qwen_vl_interface': 8.646421247845633e-06, 'learning_rate/action_model': 8.646421247845633e-05, 'epoch': 0.38})
+[2026-07-28 14:08:49,212][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.03348039463162422, 'train/grad_norm_pre_clip': 1.177319049835205, 'timing/data': 0.0003235097974538803, 'timing/model': 0.6158737307414412, 'learning_rate/qwen_vl_interface': 8.643745228255816e-06, 'learning_rate/action_model': 8.643745228255817e-05, 'epoch': 0.38})
+[2026-07-28 14:08:58,280][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.039986856281757355, 'train/grad_norm_pre_clip': 2.0790183544158936, 'timing/data': 0.0003266669809818268, 'timing/model': 0.6148456744849682, 'learning_rate/qwen_vl_interface': 8.641067006501568e-06, 'learning_rate/action_model': 8.641067006501567e-05, 'epoch': 0.38})
+[2026-07-28 14:09:07,310][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.06363825500011444, 'train/grad_norm_pre_clip': 1.6165251731872559, 'timing/data': 0.00032644253224134445, 'timing/model': 0.6159684862941504, 'learning_rate/qwen_vl_interface': 8.638386584320753e-06, 'learning_rate/action_model': 8.638386584320752e-05, 'epoch': 0.38})
+[2026-07-28 14:09:16,588][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.04540581628680229, 'train/grad_norm_pre_clip': 1.6205949783325195, 'timing/data': 0.000328725203871727, 'timing/model': 0.6152253858745098, 'learning_rate/qwen_vl_interface': 8.635703963452673e-06, 'learning_rate/action_model': 8.635703963452673e-05, 'epoch': 0.38})
+[2026-07-28 14:09:25,624][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.038877781480550766, 'train/grad_norm_pre_clip': 0.7705371379852295, 'timing/data': 0.00033659953624010086, 'timing/model': 0.6167612001299858, 'learning_rate/qwen_vl_interface': 8.633019145638048e-06, 'learning_rate/action_model': 8.633019145638048e-05, 'epoch': 0.38})
+[2026-07-28 14:09:34,687][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.03656069189310074, 'train/grad_norm_pre_clip': 1.2417068481445312, 'timing/data': 0.00039020273834466934, 'timing/model': 0.6149812778457999, 'learning_rate/qwen_vl_interface': 8.630332132619029e-06, 'learning_rate/action_model': 8.630332132619028e-05, 'epoch': 0.38})
+[2026-07-28 14:09:43,745][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.060865432024002075, 'train/grad_norm_pre_clip': 1.8707637786865234, 'timing/data': 0.0004100445657968521, 'timing/model': 0.6142096957191825, 'learning_rate/qwen_vl_interface': 8.627642926139187e-06, 'learning_rate/action_model': 8.627642926139187e-05, 'epoch': 0.38})
+[2026-07-28 14:09:53,013][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.05442094802856445, 'train/grad_norm_pre_clip': 1.1140044927597046, 'timing/data': 0.0003246935084462166, 'timing/model': 0.614835768006742, 'learning_rate/qwen_vl_interface': 8.624951527943522e-06, 'learning_rate/action_model': 8.624951527943522e-05, 'epoch': 0.38})
+[2026-07-28 14:10:02,270][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.035159219056367874, 'train/grad_norm_pre_clip': 0.8234958648681641, 'timing/data': 0.00038690678775310516, 'timing/model': 0.6150294449180365, 'learning_rate/qwen_vl_interface': 8.622257939778453e-06, 'learning_rate/action_model': 8.622257939778453e-05, 'epoch': 0.38})
+[2026-07-28 14:10:11,333][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.021259645000100136, 'train/grad_norm_pre_clip': 0.7815336585044861, 'timing/data': 0.00039875321090221405, 'timing/model': 0.6152402013540268, 'learning_rate/qwen_vl_interface': 8.619562163391818e-06, 'learning_rate/action_model': 8.619562163391817e-05, 'epoch': 0.38})
+[2026-07-28 14:10:20,393][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.03925701975822449, 'train/grad_norm_pre_clip': 1.3213106393814087, 'timing/data': 0.00039496924728155136, 'timing/model': 0.6144246999174356, 'learning_rate/qwen_vl_interface': 8.616864200532877e-06, 'learning_rate/action_model': 8.616864200532877e-05, 'epoch': 0.38})
+[2026-07-28 14:10:29,677][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.03165218606591225, 'train/grad_norm_pre_clip': 0.9365904927253723, 'timing/data': 0.0004511019214987755, 'timing/model': 0.6156289111822844, 'learning_rate/qwen_vl_interface': 8.61416405295231e-06, 'learning_rate/action_model': 8.61416405295231e-05, 'epoch': 0.38})
+[2026-07-28 14:10:38,740][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.05531267076730728, 'train/grad_norm_pre_clip': 1.4511828422546387, 'timing/data': 0.00032498687505722046, 'timing/model': 0.613114426843822, 'learning_rate/qwen_vl_interface': 8.611461722402214e-06, 'learning_rate/action_model': 8.611461722402215e-05, 'epoch': 0.38})
+[2026-07-28 14:10:48,013][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.07325035333633423, 'train/grad_norm_pre_clip': 2.602703094482422, 'timing/data': 0.000390809029340744, 'timing/model': 0.6203553518280387, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.38})
+[2026-07-28 14:10:57,194][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.10386600345373154, 'train/grad_norm_pre_clip': 2.705825090408325, 'timing/data': 0.00032881367951631546, 'timing/model': 0.6174126490950584, 'learning_rate/qwen_vl_interface': 8.606050519408901e-06, 'learning_rate/action_model': 8.606050519408901e-05, 'epoch': 0.38})
+[2026-07-28 14:11:06,524][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.04933708533644676, 'train/grad_norm_pre_clip': 1.831207036972046, 'timing/data': 0.0003329077735543251, 'timing/model': 0.8319254405796528, 'learning_rate/qwen_vl_interface': 8.603341650476953e-06, 'learning_rate/action_model': 8.603341650476952e-05, 'epoch': 0.38})
+[2026-07-28 14:11:15,626][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.028283290565013885, 'train/grad_norm_pre_clip': 0.7319905757904053, 'timing/data': 0.00032695475965738297, 'timing/model': 0.6151438131928444, 'learning_rate/qwen_vl_interface': 8.600630605598016e-06, 'learning_rate/action_model': 8.600630605598015e-05, 'epoch': 0.38})
+[2026-07-28 14:11:24,960][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.08880801498889923, 'train/grad_norm_pre_clip': 1.946952223777771, 'timing/data': 0.0003472743555903435, 'timing/model': 0.6162922512739897, 'learning_rate/qwen_vl_interface': 8.597917386531255e-06, 'learning_rate/action_model': 8.597917386531254e-05, 'epoch': 0.38})
+[2026-07-28 14:11:34,016][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.06932877749204636, 'train/grad_norm_pre_clip': 2.3736650943756104, 'timing/data': 0.0003325510770082474, 'timing/model': 0.6146466936916113, 'learning_rate/qwen_vl_interface': 8.595201995037252e-06, 'learning_rate/action_model': 8.595201995037251e-05, 'epoch': 0.38})
+[2026-07-28 14:11:43,098][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.08049794286489487, 'train/grad_norm_pre_clip': 3.153468132019043, 'timing/data': 0.00032970868051052094, 'timing/model': 0.6167889833450317, 'learning_rate/qwen_vl_interface': 8.59248443287799e-06, 'learning_rate/action_model': 8.592484432877989e-05, 'epoch': 0.38})
+[2026-07-28 14:11:52,345][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.03155296668410301, 'train/grad_norm_pre_clip': 1.1944082975387573, 'timing/data': 0.0003338642418384552, 'timing/model': 0.7869496969506145, 'learning_rate/qwen_vl_interface': 8.589764701816868e-06, 'learning_rate/action_model': 8.589764701816868e-05, 'epoch': 0.38})
+[2026-07-28 14:12:01,614][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.05455116555094719, 'train/grad_norm_pre_clip': 2.2130446434020996, 'timing/data': 0.000338616780936718, 'timing/model': 0.614169311709702, 'learning_rate/qwen_vl_interface': 8.587042803618693e-06, 'learning_rate/action_model': 8.587042803618693e-05, 'epoch': 0.38})
+[2026-07-28 14:12:10,673][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.05603696033358574, 'train/grad_norm_pre_clip': 1.2655566930770874, 'timing/data': 0.0005879336968064308, 'timing/model': 0.615675107575953, 'learning_rate/qwen_vl_interface': 8.584318740049672e-06, 'learning_rate/action_model': 8.584318740049672e-05, 'epoch': 0.38})
+[2026-07-28 14:12:19,745][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.04261493682861328, 'train/grad_norm_pre_clip': 1.0910080671310425, 'timing/data': 0.0003210678696632385, 'timing/model': 0.6168298311531544, 'learning_rate/qwen_vl_interface': 8.581592512877419e-06, 'learning_rate/action_model': 8.581592512877418e-05, 'epoch': 0.38})
+[2026-07-28 14:12:29,183][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.03433474153280258, 'train/grad_norm_pre_clip': 1.1889078617095947, 'timing/data': 0.0003255484625697136, 'timing/model': 0.865786955691874, 'learning_rate/qwen_vl_interface': 8.578864123870955e-06, 'learning_rate/action_model': 8.578864123870955e-05, 'epoch': 0.38})
+[2026-07-28 14:12:38,281][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.016137486323714256, 'train/grad_norm_pre_clip': 0.39369046688079834, 'timing/data': 0.0003271438181400299, 'timing/model': 0.6195255862548947, 'learning_rate/qwen_vl_interface': 8.576133574800703e-06, 'learning_rate/action_model': 8.576133574800703e-05, 'epoch': 0.38})
+[2026-07-28 14:12:47,608][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.1028575748205185, 'train/grad_norm_pre_clip': 1.8991535902023315, 'timing/data': 0.0003131898120045662, 'timing/model': 0.6181865110993385, 'learning_rate/qwen_vl_interface': 8.573400867438486e-06, 'learning_rate/action_model': 8.573400867438486e-05, 'epoch': 0.38})
+[2026-07-28 14:12:56,707][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.06905369460582733, 'train/grad_norm_pre_clip': 1.492828130722046, 'timing/data': 0.00033851154148578644, 'timing/model': 0.6189412362873554, 'learning_rate/qwen_vl_interface': 8.570666003557528e-06, 'learning_rate/action_model': 8.570666003557528e-05, 'epoch': 0.38})
+[2026-07-28 14:13:06,058][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.03530194237828255, 'train/grad_norm_pre_clip': 1.2650007009506226, 'timing/data': 0.0003511998802423477, 'timing/model': 0.621623869985342, 'learning_rate/qwen_vl_interface': 8.567928984932452e-06, 'learning_rate/action_model': 8.567928984932452e-05, 'epoch': 0.38})
+[2026-07-28 14:13:15,161][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.05248687416315079, 'train/grad_norm_pre_clip': 1.2399336099624634, 'timing/data': 0.00043199025094509125, 'timing/model': 0.6179049946367741, 'learning_rate/qwen_vl_interface': 8.565189813339278e-06, 'learning_rate/action_model': 8.565189813339278e-05, 'epoch': 0.38})
+[2026-07-28 14:13:24,287][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.053953684866428375, 'train/grad_norm_pre_clip': 1.0887901782989502, 'timing/data': 0.00032203085720539093, 'timing/model': 0.6182890627533197, 'learning_rate/qwen_vl_interface': 8.562448490555429e-06, 'learning_rate/action_model': 8.562448490555429e-05, 'epoch': 0.38})
+[2026-07-28 14:13:33,590][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.05592074245214462, 'train/grad_norm_pre_clip': 1.899783730506897, 'timing/data': 0.00032367557287216187, 'timing/model': 0.8256932906806469, 'learning_rate/qwen_vl_interface': 8.559705018359716e-06, 'learning_rate/action_model': 8.559705018359716e-05, 'epoch': 0.38})
+[2026-07-28 14:13:42,697][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.04517350345849991, 'train/grad_norm_pre_clip': 1.0739715099334717, 'timing/data': 0.0003199279308319092, 'timing/model': 0.6153344912454486, 'learning_rate/qwen_vl_interface': 8.55695939853235e-06, 'learning_rate/action_model': 8.55695939853235e-05, 'epoch': 0.38})
+[2026-07-28 14:13:51,778][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.03888366371393204, 'train/grad_norm_pre_clip': 1.6282449960708618, 'timing/data': 0.0003342125564813614, 'timing/model': 0.6163336178287864, 'learning_rate/qwen_vl_interface': 8.554211632854932e-06, 'learning_rate/action_model': 8.554211632854932e-05, 'epoch': 0.38})
+[2026-07-28 14:14:00,658][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.02030961960554123, 'train/grad_norm_pre_clip': 0.9604418277740479, 'timing/data': 0.0003565363585948944, 'timing/model': 0.6180169424042106, 'learning_rate/qwen_vl_interface': 8.551461723110461e-06, 'learning_rate/action_model': 8.551461723110461e-05, 'epoch': 0.38})
+[2026-07-28 14:14:09,927][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.03489247336983681, 'train/grad_norm_pre_clip': 1.3225445747375488, 'timing/data': 0.0003252793103456497, 'timing/model': 0.7855968754738569, 'learning_rate/qwen_vl_interface': 8.548709671083321e-06, 'learning_rate/action_model': 8.54870967108332e-05, 'epoch': 0.39})
+[2026-07-28 14:14:19,243][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.07788588106632233, 'train/grad_norm_pre_clip': 1.7355520725250244, 'timing/data': 0.0003224853426218033, 'timing/model': 0.614757482893765, 'learning_rate/qwen_vl_interface': 8.545955478559288e-06, 'learning_rate/action_model': 8.545955478559288e-05, 'epoch': 0.39})
+[2026-07-28 14:14:28,131][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.04687761515378952, 'train/grad_norm_pre_clip': 2.2772722244262695, 'timing/data': 0.0003338688984513283, 'timing/model': 0.6151543231680989, 'learning_rate/qwen_vl_interface': 8.54319914732553e-06, 'learning_rate/action_model': 8.543199147325529e-05, 'epoch': 0.39})
+[2026-07-28 14:14:36,992][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.05091926082968712, 'train/grad_norm_pre_clip': 1.070121169090271, 'timing/data': 0.00031289272010326385, 'timing/model': 0.6170877674594522, 'learning_rate/qwen_vl_interface': 8.540440679170599e-06, 'learning_rate/action_model': 8.540440679170598e-05, 'epoch': 0.39})
+[2026-07-28 14:14:46,429][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.0357082299888134, 'train/grad_norm_pre_clip': 1.2109211683273315, 'timing/data': 0.000336294062435627, 'timing/model': 0.8812475567683578, 'learning_rate/qwen_vl_interface': 8.537680075884433e-06, 'learning_rate/action_model': 8.537680075884433e-05, 'epoch': 0.39})
+[2026-07-28 14:14:55,618][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.05298364534974098, 'train/grad_norm_pre_clip': 1.3497302532196045, 'timing/data': 0.00036874692887067795, 'timing/model': 0.6187449190765619, 'learning_rate/qwen_vl_interface': 8.53491733925836e-06, 'learning_rate/action_model': 8.53491733925836e-05, 'epoch': 0.39})
+[2026-07-28 14:15:04,735][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.035542719066143036, 'train/grad_norm_pre_clip': 1.5400383472442627, 'timing/data': 0.0003125164657831192, 'timing/model': 0.6162099875509739, 'learning_rate/qwen_vl_interface': 8.532152471085092e-06, 'learning_rate/action_model': 8.53215247108509e-05, 'epoch': 0.39})
+[2026-07-28 14:15:13,875][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.03960268944501877, 'train/grad_norm_pre_clip': 0.8878350853919983, 'timing/data': 0.00043939612805843353, 'timing/model': 0.6155226631090045, 'learning_rate/qwen_vl_interface': 8.529385473158717e-06, 'learning_rate/action_model': 8.529385473158717e-05, 'epoch': 0.39})
+[2026-07-28 14:15:23,237][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.03052348643541336, 'train/grad_norm_pre_clip': 0.8850259780883789, 'timing/data': 0.0003497013822197914, 'timing/model': 0.8351119970902801, 'learning_rate/qwen_vl_interface': 8.526616347274712e-06, 'learning_rate/action_model': 8.526616347274712e-05, 'epoch': 0.39})
+[2026-07-28 14:15:32,580][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.05833173543214798, 'train/grad_norm_pre_clip': 1.5215704441070557, 'timing/data': 0.0003438228741288185, 'timing/model': 0.853849402628839, 'learning_rate/qwen_vl_interface': 8.523845095229935e-06, 'learning_rate/action_model': 8.523845095229935e-05, 'epoch': 0.39})
+[2026-07-28 14:15:41,660][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.04804055392742157, 'train/grad_norm_pre_clip': 0.8861150741577148, 'timing/data': 0.00045076943933963776, 'timing/model': 0.6150524010881782, 'learning_rate/qwen_vl_interface': 8.521071718822616e-06, 'learning_rate/action_model': 8.521071718822616e-05, 'epoch': 0.39})
+[2026-07-28 14:15:50,786][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.01160433515906334, 'train/grad_norm_pre_clip': 0.5744754076004028, 'timing/data': 0.00032451190054416656, 'timing/model': 0.6170419482514262, 'learning_rate/qwen_vl_interface': 8.518296219852375e-06, 'learning_rate/action_model': 8.518296219852375e-05, 'epoch': 0.39})
+[2026-07-28 14:15:59,877][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.024709166958928108, 'train/grad_norm_pre_clip': 0.5701195001602173, 'timing/data': 0.0004172753542661667, 'timing/model': 0.6154389781877398, 'learning_rate/qwen_vl_interface': 8.515518600120201e-06, 'learning_rate/action_model': 8.5155186001202e-05, 'epoch': 0.39})
+[2026-07-28 14:16:08,984][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.02755512297153473, 'train/grad_norm_pre_clip': 1.0267385244369507, 'timing/data': 0.0003124848008155823, 'timing/model': 0.614963810890913, 'learning_rate/qwen_vl_interface': 8.512738861428458e-06, 'learning_rate/action_model': 8.512738861428458e-05, 'epoch': 0.39})
+[2026-07-28 14:16:18,302][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.020367519930005074, 'train/grad_norm_pre_clip': 1.1122455596923828, 'timing/data': 0.0005790228024125099, 'timing/model': 0.6169765135273337, 'learning_rate/qwen_vl_interface': 8.509957005580892e-06, 'learning_rate/action_model': 8.509957005580892e-05, 'epoch': 0.39})
+[2026-07-28 14:16:27,405][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.08074359595775604, 'train/grad_norm_pre_clip': 1.9830131530761719, 'timing/data': 0.0003468422219157219, 'timing/model': 0.617334839887917, 'learning_rate/qwen_vl_interface': 8.507173034382617e-06, 'learning_rate/action_model': 8.507173034382617e-05, 'epoch': 0.39})
+[2026-07-28 14:16:36,720][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.019049346446990967, 'train/grad_norm_pre_clip': 1.2628984451293945, 'timing/data': 0.00041981786489486694, 'timing/model': 0.6148236850276589, 'learning_rate/qwen_vl_interface': 8.504386949640124e-06, 'learning_rate/action_model': 8.504386949640124e-05, 'epoch': 0.39})
+[2026-07-28 14:16:46,004][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.032296810299158096, 'train/grad_norm_pre_clip': 1.6149656772613525, 'timing/data': 0.00034431926906108856, 'timing/model': 0.6165166478604078, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.39})
+[2026-07-28 14:16:55,057][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.054898813366889954, 'train/grad_norm_pre_clip': 2.1412529945373535, 'timing/data': 0.00031820405274629593, 'timing/model': 0.6165210055187345, 'learning_rate/qwen_vl_interface': 8.498808446755291e-06, 'learning_rate/action_model': 8.498808446755291e-05, 'epoch': 0.39})
+[2026-07-28 14:17:04,102][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.030070725828409195, 'train/grad_norm_pre_clip': 1.4134821891784668, 'timing/data': 0.0003432156518101692, 'timing/model': 0.6144177578389645, 'learning_rate/qwen_vl_interface': 8.49601603223278e-06, 'learning_rate/action_model': 8.49601603223278e-05, 'epoch': 0.39})
+[2026-07-28 14:17:13,339][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.043898746371269226, 'train/grad_norm_pre_clip': 1.3197139501571655, 'timing/data': 0.00043049734085798264, 'timing/model': 0.7768515637144446, 'learning_rate/qwen_vl_interface': 8.493221511405708e-06, 'learning_rate/action_model': 8.493221511405707e-05, 'epoch': 0.39})
+[2026-07-28 14:17:22,486][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.022077376022934914, 'train/grad_norm_pre_clip': 0.8456699848175049, 'timing/data': 0.00031861569732427597, 'timing/model': 0.6132756033912301, 'learning_rate/qwen_vl_interface': 8.490424886087407e-06, 'learning_rate/action_model': 8.490424886087407e-05, 'epoch': 0.39})
+[2026-07-28 14:17:31,765][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.056001804769039154, 'train/grad_norm_pre_clip': 2.6710293292999268, 'timing/data': 0.00032182689756155014, 'timing/model': 0.8326051477342844, 'learning_rate/qwen_vl_interface': 8.487626158092578e-06, 'learning_rate/action_model': 8.487626158092579e-05, 'epoch': 0.39})
+[2026-07-28 14:17:40,838][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.09598243236541748, 'train/grad_norm_pre_clip': 1.7732261419296265, 'timing/data': 0.00032460596412420273, 'timing/model': 0.6156796673312783, 'learning_rate/qwen_vl_interface': 8.484825329237287e-06, 'learning_rate/action_model': 8.484825329237287e-05, 'epoch': 0.39})
+[2026-07-28 14:17:50,139][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.053094785660505295, 'train/grad_norm_pre_clip': 2.016838788986206, 'timing/data': 0.00034896470606327057, 'timing/model': 0.6165024619549513, 'learning_rate/qwen_vl_interface': 8.48202240133896e-06, 'learning_rate/action_model': 8.482022401338959e-05, 'epoch': 0.39})
+[2026-07-28 14:17:59,008][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.09616755694150925, 'train/grad_norm_pre_clip': 3.6671266555786133, 'timing/data': 0.0003171432763338089, 'timing/model': 0.6147571438923478, 'learning_rate/qwen_vl_interface': 8.479217376216385e-06, 'learning_rate/action_model': 8.479217376216385e-05, 'epoch': 0.39})
+[2026-07-28 14:18:08,123][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.03249022737145424, 'train/grad_norm_pre_clip': 1.6230731010437012, 'timing/data': 0.0003407225012779236, 'timing/model': 0.6234062956646085, 'learning_rate/qwen_vl_interface': 8.476410255689714e-06, 'learning_rate/action_model': 8.476410255689713e-05, 'epoch': 0.39})
+[2026-07-28 14:18:17,430][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.06181038171052933, 'train/grad_norm_pre_clip': 1.7769798040390015, 'timing/data': 0.0003368854522705078, 'timing/model': 0.8374528512358665, 'learning_rate/qwen_vl_interface': 8.473601041580457e-06, 'learning_rate/action_model': 8.473601041580457e-05, 'epoch': 0.39})
+[2026-07-28 14:18:26,513][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.04574410617351532, 'train/grad_norm_pre_clip': 1.7507598400115967, 'timing/data': 0.0003190981224179268, 'timing/model': 0.6156683610752225, 'learning_rate/qwen_vl_interface': 8.470789735711483e-06, 'learning_rate/action_model': 8.470789735711483e-05, 'epoch': 0.39})
+[2026-07-28 14:18:35,796][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.051912471652030945, 'train/grad_norm_pre_clip': 1.200255274772644, 'timing/data': 0.0003262050449848175, 'timing/model': 0.8468971941620111, 'learning_rate/qwen_vl_interface': 8.467976339907015e-06, 'learning_rate/action_model': 8.467976339907016e-05, 'epoch': 0.39})
+[2026-07-28 14:18:44,868][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.03760145604610443, 'train/grad_norm_pre_clip': 0.9083719253540039, 'timing/data': 0.0004252959042787552, 'timing/model': 0.6154543487355113, 'learning_rate/qwen_vl_interface': 8.465160855992639e-06, 'learning_rate/action_model': 8.465160855992638e-05, 'epoch': 0.39})
+[2026-07-28 14:18:53,922][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.047383587807416916, 'train/grad_norm_pre_clip': 1.4054044485092163, 'timing/data': 0.00033493340015411377, 'timing/model': 0.6142845386639237, 'learning_rate/qwen_vl_interface': 8.462343285795289e-06, 'learning_rate/action_model': 8.462343285795288e-05, 'epoch': 0.39})
+[2026-07-28 14:19:03,214][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.035592176020145416, 'train/grad_norm_pre_clip': 2.041468381881714, 'timing/data': 0.0003225887194275856, 'timing/model': 0.616163152270019, 'learning_rate/qwen_vl_interface': 8.459523631143259e-06, 'learning_rate/action_model': 8.459523631143258e-05, 'epoch': 0.39})
+[2026-07-28 14:19:12,261][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.025099705904722214, 'train/grad_norm_pre_clip': 0.5474821925163269, 'timing/data': 0.00032594986259937286, 'timing/model': 0.6164272557944059, 'learning_rate/qwen_vl_interface': 8.456701893866188e-06, 'learning_rate/action_model': 8.456701893866187e-05, 'epoch': 0.39})
+[2026-07-28 14:19:21,321][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.018167240545153618, 'train/grad_norm_pre_clip': 0.7564563751220703, 'timing/data': 0.0004102969542145729, 'timing/model': 0.6176977436989546, 'learning_rate/qwen_vl_interface': 8.453878075795072e-06, 'learning_rate/action_model': 8.453878075795072e-05, 'epoch': 0.39})
+[2026-07-28 14:19:30,663][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.03600737079977989, 'train/grad_norm_pre_clip': 0.9244385957717896, 'timing/data': 0.0003698877990245819, 'timing/model': 0.6173143889755011, 'learning_rate/qwen_vl_interface': 8.451052178762258e-06, 'learning_rate/action_model': 8.451052178762258e-05, 'epoch': 0.39})
+[2026-07-28 14:19:39,712][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.04649447649717331, 'train/grad_norm_pre_clip': 1.2012041807174683, 'timing/data': 0.00037699565291404724, 'timing/model': 0.61724235676229, 'learning_rate/qwen_vl_interface': 8.448224204601439e-06, 'learning_rate/action_model': 8.448224204601438e-05, 'epoch': 0.39})
+[2026-07-28 14:19:49,025][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.07229653745889664, 'train/grad_norm_pre_clip': 1.3215606212615967, 'timing/data': 0.00036989524960517883, 'timing/model': 0.8475815085694194, 'learning_rate/qwen_vl_interface': 8.445394155147655e-06, 'learning_rate/action_model': 8.445394155147653e-05, 'epoch': 0.39})
+[2026-07-28 14:19:58,105][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.03979170322418213, 'train/grad_norm_pre_clip': 1.55159330368042, 'timing/data': 0.0003978544846177101, 'timing/model': 0.6181124318391085, 'learning_rate/qwen_vl_interface': 8.442562032237295e-06, 'learning_rate/action_model': 8.442562032237296e-05, 'epoch': 0.39})
+[2026-07-28 14:20:07,424][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.056665048003196716, 'train/grad_norm_pre_clip': 1.2143384218215942, 'timing/data': 0.00038013793528079987, 'timing/model': 0.8544848533347249, 'learning_rate/qwen_vl_interface': 8.439727837708094e-06, 'learning_rate/action_model': 8.439727837708095e-05, 'epoch': 0.39})
+[2026-07-28 14:20:16,496][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.05303862690925598, 'train/grad_norm_pre_clip': 0.9681268334388733, 'timing/data': 0.0003867950290441513, 'timing/model': 0.6148998802527785, 'learning_rate/qwen_vl_interface': 8.436891573399132e-06, 'learning_rate/action_model': 8.436891573399131e-05, 'epoch': 0.39})
+[2026-07-28 14:20:25,549][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.0259096622467041, 'train/grad_norm_pre_clip': 0.6010478138923645, 'timing/data': 0.00048783980309963226, 'timing/model': 0.617503079585731, 'learning_rate/qwen_vl_interface': 8.434053241150826e-06, 'learning_rate/action_model': 8.434053241150825e-05, 'epoch': 0.39})
+[2026-07-28 14:20:34,862][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.03644843399524689, 'train/grad_norm_pre_clip': 0.8113843202590942, 'timing/data': 0.00038227252662181854, 'timing/model': 0.615977636538446, 'learning_rate/qwen_vl_interface': 8.43121284280494e-06, 'learning_rate/action_model': 8.431212842804939e-05, 'epoch': 0.39})
+[2026-07-28 14:20:43,942][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.0674610286951065, 'train/grad_norm_pre_clip': 2.3300657272338867, 'timing/data': 0.0004150504246354103, 'timing/model': 0.6183592341840267, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.39})
+[2026-07-28 14:20:53,262][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.05322875827550888, 'train/grad_norm_pre_clip': 1.8910562992095947, 'timing/data': 0.0003904644399881363, 'timing/model': 0.8549396563321352, 'learning_rate/qwen_vl_interface': 8.425525855194184e-06, 'learning_rate/action_model': 8.425525855194183e-05, 'epoch': 0.39})
+[2026-07-28 14:21:02,562][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.03337957710027695, 'train/grad_norm_pre_clip': 0.9451018571853638, 'timing/data': 0.0003742808476090431, 'timing/model': 0.6162014920264482, 'learning_rate/qwen_vl_interface': 8.422679269619538e-06, 'learning_rate/action_model': 8.422679269619538e-05, 'epoch': 0.39})
+[2026-07-28 14:21:11,636][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.046017006039619446, 'train/grad_norm_pre_clip': 0.8018792867660522, 'timing/data': 0.000377030111849308, 'timing/model': 0.6169548751786351, 'learning_rate/qwen_vl_interface': 8.419830625327758e-06, 'learning_rate/action_model': 8.419830625327758e-05, 'epoch': 0.39})
+[2026-07-28 14:21:20,508][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.05563953518867493, 'train/grad_norm_pre_clip': 0.6316776275634766, 'timing/data': 0.00038202572613954544, 'timing/model': 0.6164348088204861, 'learning_rate/qwen_vl_interface': 8.416979924167303e-06, 'learning_rate/action_model': 8.416979924167303e-05, 'epoch': 0.39})
+[2026-07-28 14:21:29,376][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.05006016790866852, 'train/grad_norm_pre_clip': 1.00762939453125, 'timing/data': 0.00038038939237594604, 'timing/model': 0.6169723523780704, 'learning_rate/qwen_vl_interface': 8.414127167987957e-06, 'learning_rate/action_model': 8.414127167987957e-05, 'epoch': 0.39})
+[2026-07-28 14:21:38,909][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.0756923258304596, 'train/grad_norm_pre_clip': 3.132168769836426, 'timing/data': 0.00045031309127807617, 'timing/model': 0.6189919533208013, 'learning_rate/qwen_vl_interface': 8.411272358640843e-06, 'learning_rate/action_model': 8.411272358640844e-05, 'epoch': 0.39})
+[2026-07-28 14:21:47,986][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.0382843092083931, 'train/grad_norm_pre_clip': 1.285272479057312, 'timing/data': 0.00032385997474193573, 'timing/model': 0.6160216592252254, 'learning_rate/qwen_vl_interface': 8.408415497978418e-06, 'learning_rate/action_model': 8.408415497978418e-05, 'epoch': 0.39})
+[2026-07-28 14:21:57,053][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.03341872990131378, 'train/grad_norm_pre_clip': 0.9008494019508362, 'timing/data': 0.00032466091215610504, 'timing/model': 0.615763833746314, 'learning_rate/qwen_vl_interface': 8.405556587854468e-06, 'learning_rate/action_model': 8.405556587854467e-05, 'epoch': 0.39})
+[2026-07-28 14:22:06,151][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.02119305543601513, 'train/grad_norm_pre_clip': 0.7674170732498169, 'timing/data': 0.0003308430314064026, 'timing/model': 0.6166792241856456, 'learning_rate/qwen_vl_interface': 8.402695630124106e-06, 'learning_rate/action_model': 8.402695630124106e-05, 'epoch': 0.39})
+[2026-07-28 14:22:15,445][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.049326978623867035, 'train/grad_norm_pre_clip': 1.5451829433441162, 'timing/data': 0.000341121107339859, 'timing/model': 0.6153828604146838, 'learning_rate/qwen_vl_interface': 8.399832626643778e-06, 'learning_rate/action_model': 8.399832626643778e-05, 'epoch': 0.39})
+[2026-07-28 14:22:24,752][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.05884836986660957, 'train/grad_norm_pre_clip': 1.2191640138626099, 'timing/data': 0.0003447839990258217, 'timing/model': 0.6180258132517338, 'learning_rate/qwen_vl_interface': 8.396967579271256e-06, 'learning_rate/action_model': 8.396967579271257e-05, 'epoch': 0.4})
+[2026-07-28 14:22:33,821][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.055668532848358154, 'train/grad_norm_pre_clip': 1.3918770551681519, 'timing/data': 0.0003423783928155899, 'timing/model': 0.6169697623699903, 'learning_rate/qwen_vl_interface': 8.394100489865638e-06, 'learning_rate/action_model': 8.394100489865638e-05, 'epoch': 0.4})
+[2026-07-28 14:22:43,142][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.0448254831135273, 'train/grad_norm_pre_clip': 3.7541887760162354, 'timing/data': 0.00035621877759695053, 'timing/model': 0.6178388418629766, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.4})
+[2026-07-28 14:22:52,224][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.09006289392709732, 'train/grad_norm_pre_clip': 2.6082842350006104, 'timing/data': 0.0003361031413078308, 'timing/model': 0.6159373875707388, 'learning_rate/qwen_vl_interface': 8.388360192398131e-06, 'learning_rate/action_model': 8.38836019239813e-05, 'epoch': 0.4})
+[2026-07-28 14:23:01,546][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.057737186551094055, 'train/grad_norm_pre_clip': 1.6196842193603516, 'timing/data': 0.0003154641017317772, 'timing/model': 0.6215456034988165, 'learning_rate/qwen_vl_interface': 8.385486988061058e-06, 'learning_rate/action_model': 8.385486988061058e-05, 'epoch': 0.4})
+[2026-07-28 14:23:10,629][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.03159415349364281, 'train/grad_norm_pre_clip': 0.5203611254692078, 'timing/data': 0.0003258921205997467, 'timing/model': 0.6151031814515591, 'learning_rate/qwen_vl_interface': 8.38261174914052e-06, 'learning_rate/action_model': 8.38261174914052e-05, 'epoch': 0.4})
+[2026-07-28 14:23:19,892][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.04852934181690216, 'train/grad_norm_pre_clip': 1.9074959754943848, 'timing/data': 0.0003164708614349365, 'timing/model': 0.8300205953419209, 'learning_rate/qwen_vl_interface': 8.379734477502229e-06, 'learning_rate/action_model': 8.379734477502228e-05, 'epoch': 0.4})
+[2026-07-28 14:23:28,959][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.018895749002695084, 'train/grad_norm_pre_clip': 0.6684478521347046, 'timing/data': 0.0003246208652853966, 'timing/model': 0.6160795539617538, 'learning_rate/qwen_vl_interface': 8.376855175013214e-06, 'learning_rate/action_model': 8.376855175013213e-05, 'epoch': 0.4})
+[2026-07-28 14:23:38,283][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.06651771813631058, 'train/grad_norm_pre_clip': 1.4509382247924805, 'timing/data': 0.00039255712181329727, 'timing/model': 0.6208736794069409, 'learning_rate/qwen_vl_interface': 8.373973843541822e-06, 'learning_rate/action_model': 8.373973843541822e-05, 'epoch': 0.4})
+[2026-07-28 14:23:47,338][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.048162881284952164, 'train/grad_norm_pre_clip': 1.1968077421188354, 'timing/data': 0.00038159359246492386, 'timing/model': 0.6151369707658887, 'learning_rate/qwen_vl_interface': 8.371090484957722e-06, 'learning_rate/action_model': 8.371090484957722e-05, 'epoch': 0.4})
+[2026-07-28 14:23:56,439][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.03988123685121536, 'train/grad_norm_pre_clip': 1.1172157526016235, 'timing/data': 0.00037661753594875336, 'timing/model': 0.6172218378633261, 'learning_rate/qwen_vl_interface': 8.368205101131891e-06, 'learning_rate/action_model': 8.368205101131891e-05, 'epoch': 0.4})
+[2026-07-28 14:24:05,531][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.04936833307147026, 'train/grad_norm_pre_clip': 1.5121283531188965, 'timing/data': 0.0003789439797401428, 'timing/model': 0.6157366065308452, 'learning_rate/qwen_vl_interface': 8.365317693936624e-06, 'learning_rate/action_model': 8.365317693936623e-05, 'epoch': 0.4})
+[2026-07-28 14:24:14,634][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.026423394680023193, 'train/grad_norm_pre_clip': 0.7535083293914795, 'timing/data': 0.00037215743213891983, 'timing/model': 0.6201914418488741, 'learning_rate/qwen_vl_interface': 8.362428265245526e-06, 'learning_rate/action_model': 8.362428265245526e-05, 'epoch': 0.4})
+[2026-07-28 14:24:23,682][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.047567807137966156, 'train/grad_norm_pre_clip': 1.82973051071167, 'timing/data': 0.00037931278347969055, 'timing/model': 0.617284782230854, 'learning_rate/qwen_vl_interface': 8.359536816933518e-06, 'learning_rate/action_model': 8.359536816933518e-05, 'epoch': 0.4})
+[2026-07-28 14:24:32,550][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.0571131706237793, 'train/grad_norm_pre_clip': 1.7218652963638306, 'timing/data': 0.000630534254014492, 'timing/model': 0.6157959951087832, 'learning_rate/qwen_vl_interface': 8.356643350876832e-06, 'learning_rate/action_model': 8.356643350876832e-05, 'epoch': 0.4})
+[2026-07-28 14:24:41,607][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.0933201014995575, 'train/grad_norm_pre_clip': 2.6683783531188965, 'timing/data': 0.00033565983176231384, 'timing/model': 0.6150255529209971, 'learning_rate/qwen_vl_interface': 8.353747868953002e-06, 'learning_rate/action_model': 8.353747868953003e-05, 'epoch': 0.4})
+[2026-07-28 14:24:51,214][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.02579151839017868, 'train/grad_norm_pre_clip': 0.8737279772758484, 'timing/data': 0.00038738735020160675, 'timing/model': 0.8354089371860027, 'learning_rate/qwen_vl_interface': 8.350850373040876e-06, 'learning_rate/action_model': 8.350850373040876e-05, 'epoch': 0.4})
+[2026-07-28 14:25:00,342][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.04770924150943756, 'train/grad_norm_pre_clip': 1.256037712097168, 'timing/data': 0.0003836061805486679, 'timing/model': 0.6240529865026474, 'learning_rate/qwen_vl_interface': 8.34795086502061e-06, 'learning_rate/action_model': 8.347950865020609e-05, 'epoch': 0.4})
+[2026-07-28 14:25:09,473][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.039992161095142365, 'train/grad_norm_pre_clip': 1.52295982837677, 'timing/data': 0.0003709709271788597, 'timing/model': 0.6236210716888309, 'learning_rate/qwen_vl_interface': 8.34504934677366e-06, 'learning_rate/action_model': 8.34504934677366e-05, 'epoch': 0.4})
+[2026-07-28 14:25:18,768][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.03780405968427658, 'train/grad_norm_pre_clip': 0.7293254733085632, 'timing/data': 0.0003806166350841522, 'timing/model': 0.7971890019252896, 'learning_rate/qwen_vl_interface': 8.342145820182795e-06, 'learning_rate/action_model': 8.342145820182793e-05, 'epoch': 0.4})
+[2026-07-28 14:25:28,129][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.04703196883201599, 'train/grad_norm_pre_clip': 1.0827046632766724, 'timing/data': 0.000392179936170578, 'timing/model': 0.6250379085540771, 'learning_rate/qwen_vl_interface': 8.339240287132073e-06, 'learning_rate/action_model': 8.339240287132073e-05, 'epoch': 0.4})
+[2026-07-28 14:25:37,259][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.03149626776576042, 'train/grad_norm_pre_clip': 0.7566050887107849, 'timing/data': 0.0003916192799806595, 'timing/model': 0.6245898315683007, 'learning_rate/qwen_vl_interface': 8.336332749506872e-06, 'learning_rate/action_model': 8.336332749506872e-05, 'epoch': 0.4})
+[2026-07-28 14:25:46,119][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.07534314692020416, 'train/grad_norm_pre_clip': 1.916695237159729, 'timing/data': 0.00039319973438978195, 'timing/model': 0.6186583498492837, 'learning_rate/qwen_vl_interface': 8.333423209193857e-06, 'learning_rate/action_model': 8.333423209193856e-05, 'epoch': 0.4})
+[2026-07-28 14:25:55,440][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.03865018114447594, 'train/grad_norm_pre_clip': 1.5793204307556152, 'timing/data': 0.0004434147849678993, 'timing/model': 0.8618121892213821, 'learning_rate/qwen_vl_interface': 8.330511668080997e-06, 'learning_rate/action_model': 8.330511668080997e-05, 'epoch': 0.4})
+[2026-07-28 14:26:04,310][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.028617050498723984, 'train/grad_norm_pre_clip': 0.7736155986785889, 'timing/data': 0.0003803754225373268, 'timing/model': 0.6171924481168389, 'learning_rate/qwen_vl_interface': 8.32759812805756e-06, 'learning_rate/action_model': 8.32759812805756e-05, 'epoch': 0.4})
+[2026-07-28 14:26:13,368][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.02539396844804287, 'train/grad_norm_pre_clip': 0.7711194157600403, 'timing/data': 0.0003766249865293503, 'timing/model': 0.6168937562033534, 'learning_rate/qwen_vl_interface': 8.324682591014111e-06, 'learning_rate/action_model': 8.32468259101411e-05, 'epoch': 0.4})
+[2026-07-28 14:26:22,425][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.03825268894433975, 'train/grad_norm_pre_clip': 1.5705622434616089, 'timing/data': 0.00036127306520938873, 'timing/model': 0.6156654311344028, 'learning_rate/qwen_vl_interface': 8.32176505884251e-06, 'learning_rate/action_model': 8.32176505884251e-05, 'epoch': 0.4})
+[2026-07-28 14:26:31,472][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.0331360287964344, 'train/grad_norm_pre_clip': 1.414442539215088, 'timing/data': 0.0003887545317411423, 'timing/model': 0.6162198251113296, 'learning_rate/qwen_vl_interface': 8.318845533435913e-06, 'learning_rate/action_model': 8.318845533435913e-05, 'epoch': 0.4})
+[2026-07-28 14:26:40,987][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.016269970685243607, 'train/grad_norm_pre_clip': 0.6927940249443054, 'timing/data': 0.00036204420030117035, 'timing/model': 0.8402510993182659, 'learning_rate/qwen_vl_interface': 8.315924016688767e-06, 'learning_rate/action_model': 8.315924016688766e-05, 'epoch': 0.4})
+[2026-07-28 14:26:50,053][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.03265802189707756, 'train/grad_norm_pre_clip': 0.9755000472068787, 'timing/data': 0.00037672463804483414, 'timing/model': 0.6177951404824853, 'learning_rate/qwen_vl_interface': 8.313000510496811e-06, 'learning_rate/action_model': 8.313000510496811e-05, 'epoch': 0.4})
+[2026-07-28 14:26:59,101][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.10847996175289154, 'train/grad_norm_pre_clip': 1.82179594039917, 'timing/data': 0.0003734566271305084, 'timing/model': 0.6161158420145512, 'learning_rate/qwen_vl_interface': 8.31007501675708e-06, 'learning_rate/action_model': 8.310075016757079e-05, 'epoch': 0.4})
+[2026-07-28 14:27:08,339][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.02822226658463478, 'train/grad_norm_pre_clip': 1.1570066213607788, 'timing/data': 0.0003500077873468399, 'timing/model': 0.6151460194960237, 'learning_rate/qwen_vl_interface': 8.307147537367894e-06, 'learning_rate/action_model': 8.307147537367893e-05, 'epoch': 0.4})
+[2026-07-28 14:27:17,419][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.01907210610806942, 'train/grad_norm_pre_clip': 0.7666533589363098, 'timing/data': 0.00040844641625881195, 'timing/model': 0.6347295809537172, 'learning_rate/qwen_vl_interface': 8.30421807422886e-06, 'learning_rate/action_model': 8.30421807422886e-05, 'epoch': 0.4})
+[2026-07-28 14:27:26,697][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.014263911172747612, 'train/grad_norm_pre_clip': 0.7847466468811035, 'timing/data': 0.0005892030894756317, 'timing/model': 0.8230843730270863, 'learning_rate/qwen_vl_interface': 8.301286629240876e-06, 'learning_rate/action_model': 8.301286629240876e-05, 'epoch': 0.4})
+[2026-07-28 14:27:35,756][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.04247213155031204, 'train/grad_norm_pre_clip': 1.7801790237426758, 'timing/data': 0.00031599123030900955, 'timing/model': 0.613945983350277, 'learning_rate/qwen_vl_interface': 8.298353204306127e-06, 'learning_rate/action_model': 8.298353204306127e-05, 'epoch': 0.4})
+[2026-07-28 14:27:45,040][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.0647733062505722, 'train/grad_norm_pre_clip': 2.3621931076049805, 'timing/data': 0.0003939596936106682, 'timing/model': 0.6141200866550207, 'learning_rate/qwen_vl_interface': 8.295417801328081e-06, 'learning_rate/action_model': 8.29541780132808e-05, 'epoch': 0.4})
+[2026-07-28 14:27:54,104][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.029674530029296875, 'train/grad_norm_pre_clip': 1.6292728185653687, 'timing/data': 0.00031504593789577484, 'timing/model': 0.6145726526156068, 'learning_rate/qwen_vl_interface': 8.292480422211486e-06, 'learning_rate/action_model': 8.292480422211485e-05, 'epoch': 0.4})
+[2026-07-28 14:28:03,134][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.029874198138713837, 'train/grad_norm_pre_clip': 1.7087702751159668, 'timing/data': 0.00038578175008296967, 'timing/model': 0.6151670850813389, 'learning_rate/qwen_vl_interface': 8.289541068862377e-06, 'learning_rate/action_model': 8.289541068862377e-05, 'epoch': 0.4})
+[2026-07-28 14:28:12,409][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.05191720277070999, 'train/grad_norm_pre_clip': 1.5046154260635376, 'timing/data': 0.00031855516135692596, 'timing/model': 0.6143575496971607, 'learning_rate/qwen_vl_interface': 8.286599743188068e-06, 'learning_rate/action_model': 8.286599743188069e-05, 'epoch': 0.4})
+[2026-07-28 14:28:21,470][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.040597110986709595, 'train/grad_norm_pre_clip': 1.7194756269454956, 'timing/data': 0.00031953398138284683, 'timing/model': 0.614879229106009, 'learning_rate/qwen_vl_interface': 8.283656447097154e-06, 'learning_rate/action_model': 8.283656447097153e-05, 'epoch': 0.4})
+[2026-07-28 14:28:30,739][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.018976673483848572, 'train/grad_norm_pre_clip': 1.099686622619629, 'timing/data': 0.0003256527706980705, 'timing/model': 0.6153787896037102, 'learning_rate/qwen_vl_interface': 8.280711182499507e-06, 'learning_rate/action_model': 8.280711182499506e-05, 'epoch': 0.4})
+[2026-07-28 14:28:39,996][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.05146721005439758, 'train/grad_norm_pre_clip': 1.9581921100616455, 'timing/data': 0.0007456894963979721, 'timing/model': 0.6179751716554165, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.4})
+[2026-07-28 14:28:49,048][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.06885457038879395, 'train/grad_norm_pre_clip': 2.5213682651519775, 'timing/data': 0.000403587706387043, 'timing/model': 0.6148848365992308, 'learning_rate/qwen_vl_interface': 8.274814755429892e-06, 'learning_rate/action_model': 8.274814755429891e-05, 'epoch': 0.4})
+[2026-07-28 14:28:58,101][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.03618698567152023, 'train/grad_norm_pre_clip': 1.0807147026062012, 'timing/data': 0.0004054782912135124, 'timing/model': 0.6161008104681969, 'learning_rate/qwen_vl_interface': 8.271863596784053e-06, 'learning_rate/action_model': 8.271863596784051e-05, 'epoch': 0.4})
+[2026-07-28 14:29:07,469][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.03215985745191574, 'train/grad_norm_pre_clip': 0.9289394021034241, 'timing/data': 0.00033087097108364105, 'timing/model': 0.6182403862476349, 'learning_rate/qwen_vl_interface': 8.268910477283732e-06, 'learning_rate/action_model': 8.26891047728373e-05, 'epoch': 0.4})
+[2026-07-28 14:29:16,753][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.026872962713241577, 'train/grad_norm_pre_clip': 1.0778104066848755, 'timing/data': 0.00032997503876686096, 'timing/model': 0.6160191977396607, 'learning_rate/qwen_vl_interface': 8.265955398845177e-06, 'learning_rate/action_model': 8.265955398845177e-05, 'epoch': 0.4})
+[2026-07-28 14:29:25,584][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.044933393597602844, 'train/grad_norm_pre_clip': 0.78702312707901, 'timing/data': 0.0003419974818825722, 'timing/model': 0.6158815072849393, 'learning_rate/qwen_vl_interface': 8.26299836338591e-06, 'learning_rate/action_model': 8.26299836338591e-05, 'epoch': 0.4})
+[2026-07-28 14:29:34,626][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.030910873785614967, 'train/grad_norm_pre_clip': 0.9798862934112549, 'timing/data': 0.0003779195249080658, 'timing/model': 0.6154316673055291, 'learning_rate/qwen_vl_interface': 8.260039372824715e-06, 'learning_rate/action_model': 8.260039372824715e-05, 'epoch': 0.4})
+[2026-07-28 14:29:43,458][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.05084393173456192, 'train/grad_norm_pre_clip': 1.2430922985076904, 'timing/data': 0.00037869345396757126, 'timing/model': 0.6149739054962993, 'learning_rate/qwen_vl_interface': 8.257078429081652e-06, 'learning_rate/action_model': 8.257078429081651e-05, 'epoch': 0.4})
+[2026-07-28 14:29:52,967][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.06412255764007568, 'train/grad_norm_pre_clip': 1.3148727416992188, 'timing/data': 0.0003184378147125244, 'timing/model': 0.6144653130322695, 'learning_rate/qwen_vl_interface': 8.254115534078041e-06, 'learning_rate/action_model': 8.254115534078042e-05, 'epoch': 0.4})
+[2026-07-28 14:30:01,804][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.08752356469631195, 'train/grad_norm_pre_clip': 1.4444243907928467, 'timing/data': 0.0003378670662641525, 'timing/model': 0.6162582198157907, 'learning_rate/qwen_vl_interface': 8.25115068973648e-06, 'learning_rate/action_model': 8.25115068973648e-05, 'epoch': 0.4})
+[2026-07-28 14:30:10,988][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.036033667623996735, 'train/grad_norm_pre_clip': 0.9578714966773987, 'timing/data': 0.00038596708327531815, 'timing/model': 0.6257123677060008, 'learning_rate/qwen_vl_interface': 8.248183897980815e-06, 'learning_rate/action_model': 8.248183897980815e-05, 'epoch': 0.4})
+[2026-07-28 14:30:20,006][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.12126020342111588, 'train/grad_norm_pre_clip': 2.6589157581329346, 'timing/data': 0.00038819480687379837, 'timing/model': 0.6163268806412816, 'learning_rate/qwen_vl_interface': 8.24521516073617e-06, 'learning_rate/action_model': 8.24521516073617e-05, 'epoch': 0.4})
+[2026-07-28 14:30:29,314][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.042646624147892, 'train/grad_norm_pre_clip': 0.8801462054252625, 'timing/data': 0.0003238581120967865, 'timing/model': 0.8307421440258622, 'learning_rate/qwen_vl_interface': 8.24224447992893e-06, 'learning_rate/action_model': 8.242244479928928e-05, 'epoch': 0.4})
+[2026-07-28 14:30:38,406][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.03532326593995094, 'train/grad_norm_pre_clip': 0.6692920327186584, 'timing/data': 0.00033862702548503876, 'timing/model': 0.6258309921249747, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.41})
+[2026-07-28 14:30:47,473][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.021814167499542236, 'train/grad_norm_pre_clip': 1.2605526447296143, 'timing/data': 0.00041985325515270233, 'timing/model': 0.6169030871242285, 'learning_rate/qwen_vl_interface': 8.236297295338474e-06, 'learning_rate/action_model': 8.236297295338474e-05, 'epoch': 0.41})
+[2026-07-28 14:30:56,541][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.04715413972735405, 'train/grad_norm_pre_clip': 1.2150369882583618, 'timing/data': 0.00040442030876874924, 'timing/model': 0.6152479443699121, 'learning_rate/qwen_vl_interface': 8.233320795414327e-06, 'learning_rate/action_model': 8.233320795414327e-05, 'epoch': 0.41})
+[2026-07-28 14:31:05,639][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.051494136452674866, 'train/grad_norm_pre_clip': 1.2429007291793823, 'timing/data': 0.00033200066536664963, 'timing/model': 0.615825722925365, 'learning_rate/qwen_vl_interface': 8.230342359645705e-06, 'learning_rate/action_model': 8.230342359645705e-05, 'epoch': 0.41})
+[2026-07-28 14:31:14,948][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.05594387650489807, 'train/grad_norm_pre_clip': 0.9629532098770142, 'timing/data': 0.00034138839691877365, 'timing/model': 0.6161687662824988, 'learning_rate/qwen_vl_interface': 8.227361989965285e-06, 'learning_rate/action_model': 8.227361989965284e-05, 'epoch': 0.41})
+[2026-07-28 14:31:24,231][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.08484421670436859, 'train/grad_norm_pre_clip': 1.2400059700012207, 'timing/data': 0.00038141850382089615, 'timing/model': 0.6148017337545753, 'learning_rate/qwen_vl_interface': 8.22437968830699e-06, 'learning_rate/action_model': 8.22437968830699e-05, 'epoch': 0.41})
+[2026-07-28 14:31:33,308][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.04821212589740753, 'train/grad_norm_pre_clip': 0.8959061503410339, 'timing/data': 0.00038551632314920425, 'timing/model': 0.6178399156779051, 'learning_rate/qwen_vl_interface': 8.221395456606012e-06, 'learning_rate/action_model': 8.221395456606011e-05, 'epoch': 0.41})
+[2026-07-28 14:31:42,365][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.05021598935127258, 'train/grad_norm_pre_clip': 1.1762988567352295, 'timing/data': 0.0003244364634156227, 'timing/model': 0.6161281038075686, 'learning_rate/qwen_vl_interface': 8.21840929679878e-06, 'learning_rate/action_model': 8.218409296798778e-05, 'epoch': 0.41})
+[2026-07-28 14:31:51,660][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.05417246371507645, 'train/grad_norm_pre_clip': 0.88996821641922, 'timing/data': 0.00032952800393104553, 'timing/model': 0.6150574628263712, 'learning_rate/qwen_vl_interface': 8.215421210822982e-06, 'learning_rate/action_model': 8.215421210822982e-05, 'epoch': 0.41})
+[2026-07-28 14:32:00,945][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.059202730655670166, 'train/grad_norm_pre_clip': 3.363121509552002, 'timing/data': 0.00035768840461969376, 'timing/model': 0.833020850084722, 'learning_rate/qwen_vl_interface': 8.212431200617557e-06, 'learning_rate/action_model': 8.212431200617556e-05, 'epoch': 0.41})
+[2026-07-28 15:18:37,174][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 15:18:37,174][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21'
+[2026-07-28 15:18:37,175][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 15:18:37,397][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 15:18:37,397][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 15:18:37,484][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 15:18:37,484][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 15:18:37,805][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 15:18:37,820][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 15:18:38,646][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 15:18:38,661][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 15:18:40,460][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 15:18:40,466][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 15:18:41,417][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 15:18:41,418][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 15:18:41,418][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 15:18:41,418][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 15:18:41,418][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 15:18:41,418][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 15:18:41,418][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 15:18:41,419][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 15:18:41,419][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 15:18:41,419][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 15:18:41,420][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 15:18:41,420][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 15:18:41,420][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 15:18:41,421][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 15:18:41,421][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 15:18:41,421][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 15:18:41,422][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 15:18:41,422][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 15:18:41,422][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 15:18:41,423][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 15:18:41,423][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 15:18:41,423][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 15:18:41,424][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 15:18:41,424][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 15:18:41,424][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 15:18:41,424][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 15:18:41,424][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 15:18:41,424][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 15:18:41,425][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 15:18:41,425][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 15:18:41,425][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 15:18:41,425][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 15:18:41,426][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 15:18:41,426][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 15:18:41,426][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 15:18:41,427][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 15:18:41,427][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 15:18:41,427][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 15:18:41,428][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 15:18:41,428][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 15:18:41,428][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 15:18:41,428][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 15:18:41,428][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 15:18:41,428][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 15:18:41,430][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 15:18:41,430][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 15:18:41,430][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 15:18:41,431][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 15:18:41,431][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 15:18:41,431][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 15:18:41,431][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 15:18:41,431][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 15:18:41,431][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 15:18:41,432][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 15:18:41,432][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 15:18:41,432][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 15:18:41,433][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 15:18:41,433][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 15:18:41,433][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 15:18:41,433][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 15:18:41,434][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 15:18:41,434][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 15:18:42,819][starVLA.dataloader][INFO] - Created train VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 15:18:42,826][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 15:18:42,842][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 15:18:42,858][starVLA.dataloader][INFO] - Created eval VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 15:18:42,868][starVLA.training.train_starvla][INFO] - LR Group qwen_vl_interface: lr=1e-05, num_params=713
+[2026-07-28 15:18:42,868][starVLA.training.train_starvla][INFO] - LR Group action_model: lr=0.0001, num_params=16
+[2026-07-28 15:18:44,230][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2000_state, steps: 2000
+[2026-07-28 15:18:44,247][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/config.full.yaml
+[2026-07-28 15:18:44,259][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/config.yaml
+[2026-07-28 15:18:44,259][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2000_state, steps: 2000
+[2026-07-28 15:18:44,264][accelerate.accelerator][INFO] - Since you passed both train and evaluation dataloader, `is_train_batch_min` (here True will decide the `train_batch_size` (8).
+[2026-07-28 15:19:20,221][accelerate.accelerator][INFO] - Loading states from /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2000_state
+[2026-07-28 15:19:20,221][accelerate.accelerator][INFO] - Loading DeepSpeed Model and Optimizer
+[2026-07-28 15:19:51,813][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer loaded from input dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2000_state/pytorch_model
+[2026-07-28 15:19:51,813][accelerate.checkpointing][INFO] - All model weights loaded successfully
+[2026-07-28 15:19:51,813][accelerate.checkpointing][INFO] - All optimizer states loaded successfully
+[2026-07-28 15:19:51,813][accelerate.checkpointing][INFO] - All scheduler states loaded successfully
+[2026-07-28 15:19:51,813][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully
+[2026-07-28 15:19:51,821][accelerate.checkpointing][INFO] - All random states loaded successfully
+[2026-07-28 15:19:51,823][accelerate.accelerator][INFO] - Loading in 0 custom states
+[2026-07-28 15:19:54,283][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-07-28 15:19:54,284][starVLA.training.train_starvla][INFO] -   Total optimization steps = 4000
+[2026-07-28 15:19:54,284][starVLA.training.train_starvla][INFO] -   Per device batch size = 8
+[2026-07-28 15:19:54,284][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-07-28 15:19:54,284][starVLA.training.train_starvla][INFO] -   Total batch size = 256
+[2026-07-28 15:21:14,571][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.03468550369143486, 'train/grad_norm_pre_clip': 0.925728440284729, 'timing/data': 0.000586298294365406, 'timing/model': 0.7805198477581143, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.37})
+[2026-07-28 15:21:21,930][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.04357017204165459, 'train/grad_norm_pre_clip': 0.8156403303146362, 'timing/data': 0.0004213443025946617, 'timing/model': 0.5462719704955816, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.37})
+[2026-07-28 15:21:29,273][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.04717927426099777, 'train/grad_norm_pre_clip': 1.2745213508605957, 'timing/data': 0.0003352481871843338, 'timing/model': 0.5328503958880901, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.37})
+[2026-07-28 15:21:36,684][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.035376399755477905, 'train/grad_norm_pre_clip': 1.1507762670516968, 'timing/data': 0.00046814046800136566, 'timing/model': 0.5306711997836828, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.37})
+[2026-07-28 15:21:43,951][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.028154559433460236, 'train/grad_norm_pre_clip': 0.9715232849121094, 'timing/data': 0.0004140874370932579, 'timing/model': 0.5299150291830301, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.37})
+[2026-07-28 15:21:51,251][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.08441030234098434, 'train/grad_norm_pre_clip': 1.4905760288238525, 'timing/data': 0.0005083782598376274, 'timing/model': 0.5380556592717767, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.37})
+[2026-07-28 15:21:58,601][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.04575571045279503, 'train/grad_norm_pre_clip': 1.6182183027267456, 'timing/data': 0.00031956471502780914, 'timing/model': 0.5310693746432662, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.37})
+[2026-07-28 15:22:05,906][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.06580092757940292, 'train/grad_norm_pre_clip': 2.0963964462280273, 'timing/data': 0.0003304155543446541, 'timing/model': 0.529910852201283, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.37})
+[2026-07-28 15:22:13,229][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.03931510075926781, 'train/grad_norm_pre_clip': 1.4082854986190796, 'timing/data': 0.000469856895506382, 'timing/model': 0.5352877816185355, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.37})
+[2026-07-28 15:22:20,518][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.04396063834428787, 'train/grad_norm_pre_clip': 0.940176784992218, 'timing/data': 0.0003191530704498291, 'timing/model': 0.5298360418528318, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.37})
+[2026-07-28 15:22:27,808][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.08184143900871277, 'train/grad_norm_pre_clip': 3.3887739181518555, 'timing/data': 0.0004412727430462837, 'timing/model': 0.5302613098174334, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.37})
+[2026-07-28 15:22:35,102][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.055854953825473785, 'train/grad_norm_pre_clip': 1.0338964462280273, 'timing/data': 0.0004831729456782341, 'timing/model': 0.5323248654603958, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.37})
+[2026-07-28 15:22:42,461][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.022380702197551727, 'train/grad_norm_pre_clip': 0.6375057697296143, 'timing/data': 0.0003700572997331619, 'timing/model': 0.5336840432137251, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.37})
+[2026-07-28 15:22:49,765][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.08674152195453644, 'train/grad_norm_pre_clip': 1.3505347967147827, 'timing/data': 0.00045573990792036057, 'timing/model': 0.536839347332716, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.37})
+[2026-07-28 15:22:57,109][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.044593971222639084, 'train/grad_norm_pre_clip': 1.1625794172286987, 'timing/data': 0.0003480222076177597, 'timing/model': 0.5334917511790991, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.37})
+[2026-07-28 15:23:04,462][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.05823709815740585, 'train/grad_norm_pre_clip': 1.8592675924301147, 'timing/data': 0.0003335103392601013, 'timing/model': 0.5390631891787052, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.37})
+[2026-07-28 15:23:11,733][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.0668758675456047, 'train/grad_norm_pre_clip': 1.5438235998153687, 'timing/data': 0.00034614652395248413, 'timing/model': 0.5385852064937353, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.37})
+[2026-07-28 15:23:19,131][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.04965759068727493, 'train/grad_norm_pre_clip': 1.479886531829834, 'timing/data': 0.0004412662237882614, 'timing/model': 0.5379236191511154, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.37})
+[2026-07-28 15:23:26,429][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.045403219759464264, 'train/grad_norm_pre_clip': 1.2085034847259521, 'timing/data': 0.0004219161346554756, 'timing/model': 0.5305017037317157, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.37})
+[2026-07-28 15:23:33,816][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.041642867028713226, 'train/grad_norm_pre_clip': 1.4677640199661255, 'timing/data': 0.0003489740192890167, 'timing/model': 0.5392365995794535, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.37})
+[2026-07-28 15:23:41,079][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.060120224952697754, 'train/grad_norm_pre_clip': 1.305343508720398, 'timing/data': 0.0003218110650777817, 'timing/model': 0.5338522111997008, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.37})
+[2026-07-28 15:23:48,338][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.043062660843133926, 'train/grad_norm_pre_clip': 1.7582859992980957, 'timing/data': 0.00034755095839500427, 'timing/model': 0.5330777298659086, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.37})
+[2026-07-28 15:23:55,596][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.034907568246126175, 'train/grad_norm_pre_clip': 1.8461055755615234, 'timing/data': 0.00043292250484228134, 'timing/model': 0.5323438346385956, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.37})
+[2026-07-28 15:24:02,857][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.04442945122718811, 'train/grad_norm_pre_clip': 1.9261796474456787, 'timing/data': 0.00033798906952142715, 'timing/model': 0.532971354201436, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.37})
+[2026-07-28 15:24:10,184][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.04893618822097778, 'train/grad_norm_pre_clip': 1.1785991191864014, 'timing/data': 0.0004608631134033203, 'timing/model': 0.5463729370385408, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.37})
+[2026-07-28 15:24:17,585][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.03747375309467316, 'train/grad_norm_pre_clip': 1.1088720560073853, 'timing/data': 0.00038340315222740173, 'timing/model': 0.5459176879376173, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.37})
+[2026-07-28 15:24:25,009][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.03397957980632782, 'train/grad_norm_pre_clip': 0.7030494213104248, 'timing/data': 0.0003446107730269432, 'timing/model': 0.5443028165027499, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.37})
+[2026-07-28 15:24:32,592][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.04954560101032257, 'train/grad_norm_pre_clip': 2.0172219276428223, 'timing/data': 0.00036307424306869507, 'timing/model': 0.5302649345248938, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.37})
+[2026-07-28 15:24:40,155][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.03607230260968208, 'train/grad_norm_pre_clip': 0.9129666686058044, 'timing/data': 0.0003153746947646141, 'timing/model': 0.538836382329464, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.37})
+[2026-07-28 15:24:47,423][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.03173132985830307, 'train/grad_norm_pre_clip': 0.7596876621246338, 'timing/data': 0.0003362717106938362, 'timing/model': 0.54038564581424, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.37})
+[2026-07-28 15:24:54,663][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.022928327322006226, 'train/grad_norm_pre_clip': 0.5092325210571289, 'timing/data': 0.0007309857755899429, 'timing/model': 0.5307040363550186, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.37})
+[2026-07-28 15:25:01,912][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.03613080456852913, 'train/grad_norm_pre_clip': 0.8013371229171753, 'timing/data': 0.00037031713873147964, 'timing/model': 0.5338630406185985, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.37})
+[2026-07-28 15:25:09,157][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.04886210709810257, 'train/grad_norm_pre_clip': 1.9068812131881714, 'timing/data': 0.0003278655931353569, 'timing/model': 0.5317291868850589, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.37})
+[2026-07-28 15:25:16,420][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.058112017810344696, 'train/grad_norm_pre_clip': 1.5602153539657593, 'timing/data': 0.00046428851783275604, 'timing/model': 0.5327604282647371, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.37})
+[2026-07-28 15:25:23,672][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.019615691155195236, 'train/grad_norm_pre_clip': 0.8957082033157349, 'timing/data': 0.0003167325630784035, 'timing/model': 0.5296951662749052, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.37})
+[2026-07-28 15:25:30,928][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.030016224831342697, 'train/grad_norm_pre_clip': 1.1754319667816162, 'timing/data': 0.000352310948073864, 'timing/model': 0.5305695421993732, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.37})
+[2026-07-28 15:25:38,168][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.019739577546715736, 'train/grad_norm_pre_clip': 0.6363978385925293, 'timing/data': 0.000403357669711113, 'timing/model': 0.5306398896500468, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.37})
+[2026-07-28 15:25:45,447][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.018884122371673584, 'train/grad_norm_pre_clip': 0.5176187753677368, 'timing/data': 0.00041300710290670395, 'timing/model': 0.530427061021328, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.37})
+[2026-07-28 15:25:52,681][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.06788157671689987, 'train/grad_norm_pre_clip': 1.9372413158416748, 'timing/data': 0.00032962020486593246, 'timing/model': 0.5312325777485967, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.37})
+[2026-07-28 15:25:59,941][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.024114709347486496, 'train/grad_norm_pre_clip': 0.8486211895942688, 'timing/data': 0.00033875182271003723, 'timing/model': 0.5316402250900865, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.37})
+[2026-07-28 15:26:07,251][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.021273672580718994, 'train/grad_norm_pre_clip': 0.9014625549316406, 'timing/data': 0.0003634011372923851, 'timing/model': 0.5583146847784519, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.37})
+[2026-07-28 15:26:14,495][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.0585610494017601, 'train/grad_norm_pre_clip': 2.8157474994659424, 'timing/data': 0.0004262179136276245, 'timing/model': 0.5310592614114285, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.38})
+[2026-07-28 15:26:21,788][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.042207442224025726, 'train/grad_norm_pre_clip': 1.0864235162734985, 'timing/data': 0.0003445977345108986, 'timing/model': 0.5404543755576015, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.38})
+[2026-07-28 15:26:29,088][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.02124730870127678, 'train/grad_norm_pre_clip': 1.3122707605361938, 'timing/data': 0.00033687613904476166, 'timing/model': 0.5540181780233979, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.38})
+[2026-07-28 15:26:36,409][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.022812409326434135, 'train/grad_norm_pre_clip': 1.1908031702041626, 'timing/data': 0.00034138932824134827, 'timing/model': 0.5773161137476563, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.38})
+[2026-07-28 15:26:43,689][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.021864980459213257, 'train/grad_norm_pre_clip': 1.553052544593811, 'timing/data': 0.0004067057743668556, 'timing/model': 0.5342673128470778, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:26:50,943][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.07821576297283173, 'train/grad_norm_pre_clip': 1.8669824600219727, 'timing/data': 0.0003777164965867996, 'timing/model': 0.530027843080461, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.38})
+[2026-07-28 15:26:58,263][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.07519981265068054, 'train/grad_norm_pre_clip': 2.606154203414917, 'timing/data': 0.00040718354284763336, 'timing/model': 0.5334280440583825, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.38})
+[2026-07-28 15:27:05,535][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.03848348185420036, 'train/grad_norm_pre_clip': 1.7645783424377441, 'timing/data': 0.0004941904917359352, 'timing/model': 0.5401995303109288, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.38})
+[2026-07-28 15:27:12,797][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.022509291768074036, 'train/grad_norm_pre_clip': 1.5336090326309204, 'timing/data': 0.00045764166861772537, 'timing/model': 0.5314250160008669, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.38})
+[2026-07-28 15:27:20,042][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.0997600331902504, 'train/grad_norm_pre_clip': 3.8346221446990967, 'timing/data': 0.00035298988223075867, 'timing/model': 0.5331057673320174, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.38})
+[2026-07-28 15:27:27,331][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.024548372253775597, 'train/grad_norm_pre_clip': 1.4138678312301636, 'timing/data': 0.0003701653331518173, 'timing/model': 0.534405848942697, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.38})
+[2026-07-28 15:27:34,631][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.033441994339227676, 'train/grad_norm_pre_clip': 1.4047439098358154, 'timing/data': 0.0004069944843649864, 'timing/model': 0.5311871375888586, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:27:41,895][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.05140557885169983, 'train/grad_norm_pre_clip': 2.0882833003997803, 'timing/data': 0.000359133817255497, 'timing/model': 0.53239765483886, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.38})
+[2026-07-28 15:27:49,148][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.021800227463245392, 'train/grad_norm_pre_clip': 0.8901355862617493, 'timing/data': 0.00037290342152118683, 'timing/model': 0.5344845103099942, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:27:56,489][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.025970380753278732, 'train/grad_norm_pre_clip': 0.6774354577064514, 'timing/data': 0.0004641357809305191, 'timing/model': 0.5396522469818592, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.38})
+[2026-07-28 15:28:03,850][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.017131049185991287, 'train/grad_norm_pre_clip': 0.7785165905952454, 'timing/data': 0.00042021647095680237, 'timing/model': 0.5328163262456656, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.38})
+[2026-07-28 15:28:11,313][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.06285019218921661, 'train/grad_norm_pre_clip': 2.04984450340271, 'timing/data': 0.0008834749460220337, 'timing/model': 0.5420073503628373, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.38})
+[2026-07-28 15:28:18,772][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.08354269713163376, 'train/grad_norm_pre_clip': 2.9106338024139404, 'timing/data': 0.0005453964695334435, 'timing/model': 0.5384923117235303, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.38})
+[2026-07-28 15:28:26,206][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.031976595520973206, 'train/grad_norm_pre_clip': 0.8787277340888977, 'timing/data': 0.0004464993253350258, 'timing/model': 0.5353415543213487, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.38})
+[2026-07-28 15:28:33,494][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.036672331392765045, 'train/grad_norm_pre_clip': 0.9131202697753906, 'timing/data': 0.00040011946111917496, 'timing/model': 0.5359375579282641, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.38})
+[2026-07-28 15:28:40,809][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.02511260285973549, 'train/grad_norm_pre_clip': 1.05647873878479, 'timing/data': 0.0003688959404826164, 'timing/model': 0.5370328482240438, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.38})
+[2026-07-28 15:28:48,297][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.032835569232702255, 'train/grad_norm_pre_clip': 1.6614130735397339, 'timing/data': 0.0004153316840529442, 'timing/model': 0.535280498676002, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.38})
+[2026-07-28 15:28:55,682][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.03188108652830124, 'train/grad_norm_pre_clip': 1.0361955165863037, 'timing/data': 0.0004781801253557205, 'timing/model': 0.5389497503638268, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:29:03,021][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.0416664183139801, 'train/grad_norm_pre_clip': 1.147891640663147, 'timing/data': 0.00034085847437381744, 'timing/model': 0.5296417595818639, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:29:10,674][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.04686348885297775, 'train/grad_norm_pre_clip': 1.7210954427719116, 'timing/data': 0.00035486742854118347, 'timing/model': 0.5320211872458458, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.38})
+[2026-07-28 15:29:17,966][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.019851692020893097, 'train/grad_norm_pre_clip': 0.7121147513389587, 'timing/data': 0.000869477167725563, 'timing/model': 0.5359585424885154, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.38})
+[2026-07-28 15:29:25,239][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.03596467152237892, 'train/grad_norm_pre_clip': 1.6955389976501465, 'timing/data': 0.00037935376167297363, 'timing/model': 0.5301764877513051, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:29:32,806][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.04664650559425354, 'train/grad_norm_pre_clip': 1.106522560119629, 'timing/data': 0.0004171421751379967, 'timing/model': 0.8161728195846081, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.38})
+[2026-07-28 15:29:40,128][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.026349730789661407, 'train/grad_norm_pre_clip': 0.7756046652793884, 'timing/data': 0.0003399476408958435, 'timing/model': 0.5361755564808846, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.38})
+[2026-07-28 15:29:47,419][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.06062085181474686, 'train/grad_norm_pre_clip': 2.373678684234619, 'timing/data': 0.0004685232415795326, 'timing/model': 0.5348890973255038, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.38})
+[2026-07-28 15:29:54,865][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.07811425626277924, 'train/grad_norm_pre_clip': 1.6525006294250488, 'timing/data': 0.00048511289060115814, 'timing/model': 0.5682934010401368, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.38})
+[2026-07-28 15:30:02,688][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.05409148335456848, 'train/grad_norm_pre_clip': 1.5313701629638672, 'timing/data': 0.000540110282599926, 'timing/model': 0.6018500793725252, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.38})
+[2026-07-28 15:30:10,264][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.03524423763155937, 'train/grad_norm_pre_clip': 0.828067421913147, 'timing/data': 0.0004933876916766167, 'timing/model': 0.5361164333298802, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.38})
+[2026-07-28 15:30:17,567][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.03733540326356888, 'train/grad_norm_pre_clip': 1.6713463068008423, 'timing/data': 0.00037949811667203903, 'timing/model': 0.5305704977363348, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:30:24,823][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.06281822919845581, 'train/grad_norm_pre_clip': 2.3024256229400635, 'timing/data': 0.0005101757124066353, 'timing/model': 0.5325276236981153, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.38})
+[2026-07-28 15:30:32,086][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.029273545369505882, 'train/grad_norm_pre_clip': 0.778872549533844, 'timing/data': 0.0005222950130701065, 'timing/model': 0.5344435079023242, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.38})
+[2026-07-28 15:30:39,379][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.030099105089902878, 'train/grad_norm_pre_clip': 0.889316976070404, 'timing/data': 0.0004971222952008247, 'timing/model': 0.5332864886149764, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:30:46,648][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.02635851874947548, 'train/grad_norm_pre_clip': 0.8225778937339783, 'timing/data': 0.0004231240600347519, 'timing/model': 0.534866226837039, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:30:53,970][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.055348463356494904, 'train/grad_norm_pre_clip': 1.349693775177002, 'timing/data': 0.0003764582797884941, 'timing/model': 0.5382388196885586, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.38})
+[2026-07-28 15:31:01,292][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.05679360777139664, 'train/grad_norm_pre_clip': 1.1325792074203491, 'timing/data': 0.00041295867413282394, 'timing/model': 0.5396028719842434, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.38})
+[2026-07-28 15:31:08,572][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.04618226736783981, 'train/grad_norm_pre_clip': 1.5945851802825928, 'timing/data': 0.0003540804609656334, 'timing/model': 0.5321815796196461, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.38})
+[2026-07-28 15:31:15,843][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.03224409371614456, 'train/grad_norm_pre_clip': 1.1820343732833862, 'timing/data': 0.0006021792069077492, 'timing/model': 0.5292118433862925, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.38})
+[2026-07-28 15:31:23,075][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.028432197868824005, 'train/grad_norm_pre_clip': 0.9490693807601929, 'timing/data': 0.0004015285521745682, 'timing/model': 0.5295161549001932, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.38})
+[2026-07-28 15:31:30,351][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.031856097280979156, 'train/grad_norm_pre_clip': 1.0530537366867065, 'timing/data': 0.0008782437071204185, 'timing/model': 0.5318124573677778, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.38})
+[2026-07-28 15:31:37,603][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.03341302648186684, 'train/grad_norm_pre_clip': 1.1911848783493042, 'timing/data': 0.000443972647190094, 'timing/model': 0.5549439806491137, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.38})
+[2026-07-28 15:31:44,860][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.04797651618719101, 'train/grad_norm_pre_clip': 1.3045990467071533, 'timing/data': 0.00039172545075416565, 'timing/model': 0.5336830662563443, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.38})
+[2026-07-28 15:31:52,092][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.037388063967227936, 'train/grad_norm_pre_clip': 1.0593688488006592, 'timing/data': 0.0003843754529953003, 'timing/model': 0.5333652095869184, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:31:59,356][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.011365830898284912, 'train/grad_norm_pre_clip': 1.005179524421692, 'timing/data': 0.0003448259085416794, 'timing/model': 0.5356486365199089, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:32:06,605][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.05785380303859711, 'train/grad_norm_pre_clip': 1.3022087812423706, 'timing/data': 0.0003507668152451515, 'timing/model': 0.5287129143252969, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.38})
+[2026-07-28 15:32:13,857][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.044980622828006744, 'train/grad_norm_pre_clip': 1.183374047279358, 'timing/data': 0.00043167080730199814, 'timing/model': 0.5372774861752987, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.38})
+[2026-07-28 15:32:21,088][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.06328441947698593, 'train/grad_norm_pre_clip': 2.36122727394104, 'timing/data': 0.00039008352905511856, 'timing/model': 0.5317696146667004, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:32:28,367][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.04409616068005562, 'train/grad_norm_pre_clip': 1.3866584300994873, 'timing/data': 0.00035974476486444473, 'timing/model': 0.5380271086469293, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.38})
+[2026-07-28 15:32:35,758][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.05050221085548401, 'train/grad_norm_pre_clip': 0.9815462827682495, 'timing/data': 0.00033945124596357346, 'timing/model': 0.5322057511657476, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.38})
+[2026-07-28 15:32:43,117][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.04230416938662529, 'train/grad_norm_pre_clip': 0.851513683795929, 'timing/data': 0.0004178425297141075, 'timing/model': 0.5278818942606449, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.38})
+[2026-07-28 15:32:50,416][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.043995730578899384, 'train/grad_norm_pre_clip': 1.058597445487976, 'timing/data': 0.0003877393901348114, 'timing/model': 0.5302922707051039, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.38})
+[2026-07-28 15:32:57,649][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.04017972946166992, 'train/grad_norm_pre_clip': 0.8753109574317932, 'timing/data': 0.00041006412357091904, 'timing/model': 0.529854990541935, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.39})
+[2026-07-28 15:33:04,974][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.05242381989955902, 'train/grad_norm_pre_clip': 2.408271074295044, 'timing/data': 0.0003975359722971916, 'timing/model': 0.5311024403199553, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.39})
+[2026-07-28 15:33:12,294][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.04554346948862076, 'train/grad_norm_pre_clip': 0.8400155305862427, 'timing/data': 0.0004059094935655594, 'timing/model': 0.5341446409001946, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.39})
+[2026-07-28 15:33:19,622][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.035590775310993195, 'train/grad_norm_pre_clip': 2.1814732551574707, 'timing/data': 0.00046640634536743164, 'timing/model': 0.537503476254642, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.39})
+[2026-07-28 15:33:26,944][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.02519216015934944, 'train/grad_norm_pre_clip': 1.326616644859314, 'timing/data': 0.0003843819722533226, 'timing/model': 0.5343091897666454, 'learning_rate/qwen_vl_interface': 9.999998458888285e-06, 'learning_rate/action_model': 9.999998458888285e-05, 'epoch': 0.39})
+[2026-07-28 15:33:34,297][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.056120965629816055, 'train/grad_norm_pre_clip': 1.185153841972351, 'timing/data': 0.0003172019496560097, 'timing/model': 0.5338618420064449, 'learning_rate/qwen_vl_interface': 9.999993835554141e-06, 'learning_rate/action_model': 9.999993835554141e-05, 'epoch': 0.39})
+[2026-07-28 15:33:41,619][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.07290206849575043, 'train/grad_norm_pre_clip': 1.632485270500183, 'timing/data': 0.00033326074481010437, 'timing/model': 0.5340768862515688, 'learning_rate/qwen_vl_interface': 9.999986130000566e-06, 'learning_rate/action_model': 9.999986130000566e-05, 'epoch': 0.39})
+[2026-07-28 15:33:49,159][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.024156376719474792, 'train/grad_norm_pre_clip': 0.7892479300498962, 'timing/data': 0.0004196399822831154, 'timing/model': 0.5319995796307921, 'learning_rate/qwen_vl_interface': 9.99997534223256e-06, 'learning_rate/action_model': 9.99997534223256e-05, 'epoch': 0.39})
+[2026-07-28 15:33:56,424][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.04050162434577942, 'train/grad_norm_pre_clip': 0.986901044845581, 'timing/data': 0.0004130275920033455, 'timing/model': 0.5320593472570181, 'learning_rate/qwen_vl_interface': 9.999961472257127e-06, 'learning_rate/action_model': 9.999961472257127e-05, 'epoch': 0.39})
+[2026-07-28 15:34:03,695][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.04850831627845764, 'train/grad_norm_pre_clip': 1.3113758563995361, 'timing/data': 0.0005612103268504143, 'timing/model': 0.533462724648416, 'learning_rate/qwen_vl_interface': 9.999944520083262e-06, 'learning_rate/action_model': 9.999944520083261e-05, 'epoch': 0.39})
+[2026-07-28 15:34:11,261][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.0878688395023346, 'train/grad_norm_pre_clip': 3.4464545249938965, 'timing/data': 0.000351480208337307, 'timing/model': 0.5303091015666723, 'learning_rate/qwen_vl_interface': 9.999924485721968e-06, 'learning_rate/action_model': 9.999924485721968e-05, 'epoch': 0.39})
+[2026-07-28 15:34:18,546][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.06864087283611298, 'train/grad_norm_pre_clip': 2.9981582164764404, 'timing/data': 0.0003781961277127266, 'timing/model': 0.5373949510976672, 'learning_rate/qwen_vl_interface': 9.999901369186244e-06, 'learning_rate/action_model': 9.999901369186244e-05, 'epoch': 0.39})
+[2026-07-28 15:34:25,838][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.04322962090373039, 'train/grad_norm_pre_clip': 3.1344685554504395, 'timing/data': 0.0003861980512738228, 'timing/model': 0.5334701472893357, 'learning_rate/qwen_vl_interface': 9.999875170491092e-06, 'learning_rate/action_model': 9.999875170491091e-05, 'epoch': 0.39})
+[2026-07-28 15:34:33,572][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.0448228120803833, 'train/grad_norm_pre_clip': 2.9519810676574707, 'timing/data': 0.00036559533327817917, 'timing/model': 0.5380603400990367, 'learning_rate/qwen_vl_interface': 9.999845889653509e-06, 'learning_rate/action_model': 9.999845889653508e-05, 'epoch': 0.39})
+[2026-07-28 15:34:40,899][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.045556724071502686, 'train/grad_norm_pre_clip': 1.9282082319259644, 'timing/data': 0.000348750501871109, 'timing/model': 0.5367874149233103, 'learning_rate/qwen_vl_interface': 9.999813526692496e-06, 'learning_rate/action_model': 9.999813526692495e-05, 'epoch': 0.39})
+[2026-07-28 15:34:48,361][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.01523914746940136, 'train/grad_norm_pre_clip': 0.4392057955265045, 'timing/data': 0.0007787700742483139, 'timing/model': 0.528502032160759, 'learning_rate/qwen_vl_interface': 9.999778081629055e-06, 'learning_rate/action_model': 9.999778081629054e-05, 'epoch': 0.39})
+[2026-07-28 15:34:55,575][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.02481282316148281, 'train/grad_norm_pre_clip': 0.5556161403656006, 'timing/data': 0.00038342736661434174, 'timing/model': 0.5301755229011178, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.39})
+[2026-07-28 15:35:02,823][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.01359324436634779, 'train/grad_norm_pre_clip': 0.45917248725891113, 'timing/data': 0.0003561573103070259, 'timing/model': 0.5252799913287163, 'learning_rate/qwen_vl_interface': 9.999697945288881e-06, 'learning_rate/action_model': 9.999697945288881e-05, 'epoch': 0.39})
+[2026-07-28 15:35:10,053][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.056349363178014755, 'train/grad_norm_pre_clip': 1.515984296798706, 'timing/data': 0.00041941460222005844, 'timing/model': 0.530616051517427, 'learning_rate/qwen_vl_interface': 9.999653254064153e-06, 'learning_rate/action_model': 9.999653254064152e-05, 'epoch': 0.39})
+[2026-07-28 15:35:17,302][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.030918003991246223, 'train/grad_norm_pre_clip': 0.9591362476348877, 'timing/data': 0.0003582453355193138, 'timing/model': 0.5300891809165478, 'learning_rate/qwen_vl_interface': 9.99960548084099e-06, 'learning_rate/action_model': 9.99960548084099e-05, 'epoch': 0.39})
+[2026-07-28 15:35:24,541][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.03148108348250389, 'train/grad_norm_pre_clip': 0.6955251693725586, 'timing/data': 0.00035261642187833786, 'timing/model': 0.5322107961401343, 'learning_rate/qwen_vl_interface': 9.999554625650398e-06, 'learning_rate/action_model': 9.999554625650399e-05, 'epoch': 0.39})
+[2026-07-28 15:35:31,758][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.03410656005144119, 'train/grad_norm_pre_clip': 0.8694064617156982, 'timing/data': 0.0003262721002101898, 'timing/model': 0.5303754881024361, 'learning_rate/qwen_vl_interface': 9.999500688525376e-06, 'learning_rate/action_model': 9.999500688525375e-05, 'epoch': 0.39})
+[2026-07-28 15:35:38,996][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.04431995749473572, 'train/grad_norm_pre_clip': 1.8532350063323975, 'timing/data': 0.00039958301931619644, 'timing/model': 0.5331122158095241, 'learning_rate/qwen_vl_interface': 9.99944366950092e-06, 'learning_rate/action_model': 9.999443669500919e-05, 'epoch': 0.39})
+[2026-07-28 15:35:46,230][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.06113385781645775, 'train/grad_norm_pre_clip': 2.8794846534729004, 'timing/data': 0.000395381823182106, 'timing/model': 0.5330700343474746, 'learning_rate/qwen_vl_interface': 9.99938356861403e-06, 'learning_rate/action_model': 9.99938356861403e-05, 'epoch': 0.39})
+[2026-07-28 15:35:53,478][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.04122105985879898, 'train/grad_norm_pre_clip': 1.6941031217575073, 'timing/data': 0.0004565967246890068, 'timing/model': 0.5330086229369044, 'learning_rate/qwen_vl_interface': 9.999320385903709e-06, 'learning_rate/action_model': 9.999320385903708e-05, 'epoch': 0.39})
+[2026-07-28 15:36:00,732][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.016877856105566025, 'train/grad_norm_pre_clip': 1.3324028253555298, 'timing/data': 0.00034899916499853134, 'timing/model': 0.5315221613273025, 'learning_rate/qwen_vl_interface': 9.999254121410948e-06, 'learning_rate/action_model': 9.999254121410949e-05, 'epoch': 0.39})
+[2026-07-28 15:36:07,956][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.05675370246171951, 'train/grad_norm_pre_clip': 2.4005889892578125, 'timing/data': 0.00035365112125873566, 'timing/model': 0.5288529209792614, 'learning_rate/qwen_vl_interface': 9.999184775178751e-06, 'learning_rate/action_model': 9.999184775178751e-05, 'epoch': 0.39})
+[2026-07-28 15:36:15,203][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.07436564564704895, 'train/grad_norm_pre_clip': 4.042505741119385, 'timing/data': 0.00033850502222776413, 'timing/model': 0.5313267512246966, 'learning_rate/qwen_vl_interface': 9.999112347252117e-06, 'learning_rate/action_model': 9.999112347252116e-05, 'epoch': 0.39})
+[2026-07-28 15:36:22,506][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.06997665762901306, 'train/grad_norm_pre_clip': 2.734384536743164, 'timing/data': 0.000458727590739727, 'timing/model': 0.5311913676559925, 'learning_rate/qwen_vl_interface': 9.99903683767804e-06, 'learning_rate/action_model': 9.99903683767804e-05, 'epoch': 0.39})
+[2026-07-28 15:36:29,898][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.02910064160823822, 'train/grad_norm_pre_clip': 2.484365224838257, 'timing/data': 0.0003316905349493027, 'timing/model': 0.5290346890687943, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.39})
+[2026-07-28 15:36:37,329][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.015420360490679741, 'train/grad_norm_pre_clip': 1.4436743259429932, 'timing/data': 0.0003958996385335922, 'timing/model': 0.546823762357235, 'learning_rate/qwen_vl_interface': 9.998876573785547e-06, 'learning_rate/action_model': 9.998876573785547e-05, 'epoch': 0.39})
+[2026-07-28 15:36:44,870][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.09609450399875641, 'train/grad_norm_pre_clip': 2.346593141555786, 'timing/data': 0.00048698391765356064, 'timing/model': 0.5471243038773537, 'learning_rate/qwen_vl_interface': 9.998791819571126e-06, 'learning_rate/action_model': 9.998791819571127e-05, 'epoch': 0.39})
+[2026-07-28 15:36:52,177][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.06579124927520752, 'train/grad_norm_pre_clip': 1.7817639112472534, 'timing/data': 0.00033932924270629883, 'timing/model': 0.5304036643356085, 'learning_rate/qwen_vl_interface': 9.998703983917252e-06, 'learning_rate/action_model': 9.998703983917252e-05, 'epoch': 0.39})
+[2026-07-28 15:36:59,468][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.037368543446063995, 'train/grad_norm_pre_clip': 2.2734580039978027, 'timing/data': 0.0003418000414967537, 'timing/model': 0.531717742793262, 'learning_rate/qwen_vl_interface': 9.998613066880919e-06, 'learning_rate/action_model': 9.998613066880918e-05, 'epoch': 0.39})
+[2026-07-28 15:37:06,747][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.022317778319120407, 'train/grad_norm_pre_clip': 0.7450053691864014, 'timing/data': 0.00038756802678108215, 'timing/model': 0.5366506762802601, 'learning_rate/qwen_vl_interface': 9.99851906852112e-06, 'learning_rate/action_model': 9.998519068521119e-05, 'epoch': 0.39})
+[2026-07-28 15:37:14,104][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.10369514673948288, 'train/grad_norm_pre_clip': 2.1854982376098633, 'timing/data': 0.00033660419285297394, 'timing/model': 0.5293383989483118, 'learning_rate/qwen_vl_interface': 9.998421988898851e-06, 'learning_rate/action_model': 9.998421988898851e-05, 'epoch': 0.39})
+[2026-07-28 15:37:21,340][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.04086914658546448, 'train/grad_norm_pre_clip': 1.3297865390777588, 'timing/data': 0.000343923456966877, 'timing/model': 0.5306112468242645, 'learning_rate/qwen_vl_interface': 9.998321828077107e-06, 'learning_rate/action_model': 9.998321828077106e-05, 'epoch': 0.39})
+[2026-07-28 15:37:28,590][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.04237423464655876, 'train/grad_norm_pre_clip': 1.0356028079986572, 'timing/data': 0.00041105225682258606, 'timing/model': 0.5300384555011988, 'learning_rate/qwen_vl_interface': 9.99821858612088e-06, 'learning_rate/action_model': 9.99821858612088e-05, 'epoch': 0.39})
+[2026-07-28 15:37:35,819][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.03746289759874344, 'train/grad_norm_pre_clip': 1.4541136026382446, 'timing/data': 0.00033244676887989044, 'timing/model': 0.5290347244590521, 'learning_rate/qwen_vl_interface': 9.998112263097164e-06, 'learning_rate/action_model': 9.998112263097165e-05, 'epoch': 0.39})
+[2026-07-28 15:37:43,060][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.030401235446333885, 'train/grad_norm_pre_clip': 1.7231754064559937, 'timing/data': 0.0003389669582247734, 'timing/model': 0.5398406265303493, 'learning_rate/qwen_vl_interface': 9.99800285907495e-06, 'learning_rate/action_model': 9.99800285907495e-05, 'epoch': 0.39})
+[2026-07-28 15:37:50,309][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.028306404128670692, 'train/grad_norm_pre_clip': 1.0875266790390015, 'timing/data': 0.00037571974098682404, 'timing/model': 0.5285867070779204, 'learning_rate/qwen_vl_interface': 9.997890374125228e-06, 'learning_rate/action_model': 9.997890374125227e-05, 'epoch': 0.39})
+[2026-07-28 15:37:57,542][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.1066274642944336, 'train/grad_norm_pre_clip': 1.4573712348937988, 'timing/data': 0.00040194112807512283, 'timing/model': 0.5351983522996306, 'learning_rate/qwen_vl_interface': 9.997774808320992e-06, 'learning_rate/action_model': 9.997774808320992e-05, 'epoch': 0.39})
+[2026-07-28 15:38:04,807][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.05392862856388092, 'train/grad_norm_pre_clip': 2.1913111209869385, 'timing/data': 0.0008079148828983307, 'timing/model': 0.5398943787440658, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.39})
+[2026-07-28 15:38:12,046][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.03344664350152016, 'train/grad_norm_pre_clip': 0.7914711236953735, 'timing/data': 0.0003925934433937073, 'timing/model': 0.530647506006062, 'learning_rate/qwen_vl_interface': 9.997534434450922e-06, 'learning_rate/action_model': 9.997534434450921e-05, 'epoch': 0.39})
+[2026-07-28 15:38:19,618][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.05532793328166008, 'train/grad_norm_pre_clip': 1.47849702835083, 'timing/data': 0.000395357608795166, 'timing/model': 0.8372930064797401, 'learning_rate/qwen_vl_interface': 9.997409626541065e-06, 'learning_rate/action_model': 9.997409626541066e-05, 'epoch': 0.39})
+[2026-07-28 15:38:26,869][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.05317814648151398, 'train/grad_norm_pre_clip': 1.2644097805023193, 'timing/data': 0.0003459993749856949, 'timing/model': 0.5301097454503179, 'learning_rate/qwen_vl_interface': 9.997281738088647e-06, 'learning_rate/action_model': 9.997281738088646e-05, 'epoch': 0.39})
+[2026-07-28 15:38:34,088][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.032919399440288544, 'train/grad_norm_pre_clip': 1.1347508430480957, 'timing/data': 0.00037659797817468643, 'timing/model': 0.5292199533432722, 'learning_rate/qwen_vl_interface': 9.997150769176648e-06, 'learning_rate/action_model': 9.997150769176647e-05, 'epoch': 0.39})
+[2026-07-28 15:38:41,331][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.035064879804849625, 'train/grad_norm_pre_clip': 0.7000488042831421, 'timing/data': 0.00030194036662578583, 'timing/model': 0.52898411732167, 'learning_rate/qwen_vl_interface': 9.997016719890052e-06, 'learning_rate/action_model': 9.997016719890052e-05, 'epoch': 0.39})
+[2026-07-28 15:38:48,622][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.03600417077541351, 'train/grad_norm_pre_clip': 1.3123635053634644, 'timing/data': 0.0003964519128203392, 'timing/model': 0.5302159562706947, 'learning_rate/qwen_vl_interface': 9.996879590315847e-06, 'learning_rate/action_model': 9.996879590315846e-05, 'epoch': 0.39})
+[2026-07-28 15:38:55,940][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.07556459307670593, 'train/grad_norm_pre_clip': 1.6736024618148804, 'timing/data': 0.00038979947566986084, 'timing/model': 0.5309728737920523, 'learning_rate/qwen_vl_interface': 9.99673938054301e-06, 'learning_rate/action_model': 9.996739380543009e-05, 'epoch': 0.39})
+[2026-07-28 15:39:03,199][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.06718370318412781, 'train/grad_norm_pre_clip': 1.4328042268753052, 'timing/data': 0.00034079793840646744, 'timing/model': 0.5337670799344778, 'learning_rate/qwen_vl_interface': 9.996596090662525e-06, 'learning_rate/action_model': 9.996596090662525e-05, 'epoch': 0.39})
+[2026-07-28 15:39:10,457][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.019925562664866447, 'train/grad_norm_pre_clip': 1.063801646232605, 'timing/data': 0.00041938573122024536, 'timing/model': 0.5301241371780634, 'learning_rate/qwen_vl_interface': 9.996449720767369e-06, 'learning_rate/action_model': 9.996449720767368e-05, 'epoch': 0.39})
+[2026-07-28 15:39:18,084][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.04145451635122299, 'train/grad_norm_pre_clip': 0.6517598628997803, 'timing/data': 0.00039483606815338135, 'timing/model': 0.8727907799184322, 'learning_rate/qwen_vl_interface': 9.99630027095252e-06, 'learning_rate/action_model': 9.99630027095252e-05, 'epoch': 0.39})
+[2026-07-28 15:39:25,298][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.02958323247730732, 'train/grad_norm_pre_clip': 1.00413978099823, 'timing/data': 0.00036769546568393707, 'timing/model': 0.5298935920000076, 'learning_rate/qwen_vl_interface': 9.996147741314956e-06, 'learning_rate/action_model': 9.996147741314957e-05, 'epoch': 0.39})
+[2026-07-28 15:39:32,548][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.04720788449048996, 'train/grad_norm_pre_clip': 1.3991131782531738, 'timing/data': 0.00031082797795534134, 'timing/model': 0.5315321506932378, 'learning_rate/qwen_vl_interface': 9.99599213195365e-06, 'learning_rate/action_model': 9.99599213195365e-05, 'epoch': 0.4})
+[2026-07-28 15:39:39,835][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.06796377897262573, 'train/grad_norm_pre_clip': 1.5996030569076538, 'timing/data': 0.00040371716022491455, 'timing/model': 0.5432419879361987, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.4})
+[2026-07-28 15:39:47,098][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.06365202367305756, 'train/grad_norm_pre_clip': 1.9130513668060303, 'timing/data': 0.0005642473697662354, 'timing/model': 0.5383195281028748, 'learning_rate/qwen_vl_interface': 9.995671674465709e-06, 'learning_rate/action_model': 9.995671674465709e-05, 'epoch': 0.4})
+[2026-07-28 15:39:54,370][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.04617976397275925, 'train/grad_norm_pre_clip': 2.120243787765503, 'timing/data': 0.00048019830137491226, 'timing/model': 0.5378360617905855, 'learning_rate/qwen_vl_interface': 9.99550682654701e-06, 'learning_rate/action_model': 9.995506826547011e-05, 'epoch': 0.4})
+[2026-07-28 15:40:01,644][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.061196811497211456, 'train/grad_norm_pre_clip': 1.6693321466445923, 'timing/data': 0.00039614737033843994, 'timing/model': 0.530670452862978, 'learning_rate/qwen_vl_interface': 9.995338899320457e-06, 'learning_rate/action_model': 9.995338899320456e-05, 'epoch': 0.4})
+[2026-07-28 15:40:08,930][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.05383338779211044, 'train/grad_norm_pre_clip': 1.0599545240402222, 'timing/data': 0.0004772823303937912, 'timing/model': 0.53857462387532, 'learning_rate/qwen_vl_interface': 9.995167892895011e-06, 'learning_rate/action_model': 9.99516789289501e-05, 'epoch': 0.4})
+[2026-07-28 15:40:16,204][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.020121371373534203, 'train/grad_norm_pre_clip': 0.7777532935142517, 'timing/data': 0.0004264293238520622, 'timing/model': 0.5358260292559862, 'learning_rate/qwen_vl_interface': 9.994993807381635e-06, 'learning_rate/action_model': 9.994993807381635e-05, 'epoch': 0.4})
+[2026-07-28 15:40:23,467][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.061947137117385864, 'train/grad_norm_pre_clip': 2.0461530685424805, 'timing/data': 0.0004868116229772568, 'timing/model': 0.5319307316094637, 'learning_rate/qwen_vl_interface': 9.994816642893294e-06, 'learning_rate/action_model': 9.994816642893293e-05, 'epoch': 0.4})
+[2026-07-28 15:40:30,769][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.028840137645602226, 'train/grad_norm_pre_clip': 1.1186015605926514, 'timing/data': 0.0003742557018995285, 'timing/model': 0.532506993971765, 'learning_rate/qwen_vl_interface': 9.994636399544948e-06, 'learning_rate/action_model': 9.994636399544947e-05, 'epoch': 0.4})
+[2026-07-28 15:40:38,139][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.047319445759058, 'train/grad_norm_pre_clip': 1.5413755178451538, 'timing/data': 0.0004677576944231987, 'timing/model': 0.5341145750135183, 'learning_rate/qwen_vl_interface': 9.994453077453552e-06, 'learning_rate/action_model': 9.994453077453552e-05, 'epoch': 0.4})
+[2026-07-28 15:40:45,389][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.04622029513120651, 'train/grad_norm_pre_clip': 1.4053244590759277, 'timing/data': 0.00037136301398277283, 'timing/model': 0.5323135536164045, 'learning_rate/qwen_vl_interface': 9.994266676738065e-06, 'learning_rate/action_model': 9.994266676738065e-05, 'epoch': 0.4})
+[2026-07-28 15:40:52,645][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.047564174979925156, 'train/grad_norm_pre_clip': 2.4882514476776123, 'timing/data': 0.0003706812858581543, 'timing/model': 0.5365150487050414, 'learning_rate/qwen_vl_interface': 9.99407719751944e-06, 'learning_rate/action_model': 9.994077197519439e-05, 'epoch': 0.4})
+[2026-07-28 15:40:59,912][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.04723910242319107, 'train/grad_norm_pre_clip': 1.7692039012908936, 'timing/data': 0.0003854483366012573, 'timing/model': 0.5321756461635232, 'learning_rate/qwen_vl_interface': 9.993884639920626e-06, 'learning_rate/action_model': 9.993884639920626e-05, 'epoch': 0.4})
+[2026-07-28 15:41:07,179][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.04457967355847359, 'train/grad_norm_pre_clip': 1.3522002696990967, 'timing/data': 0.0004197787493467331, 'timing/model': 0.5314649222418666, 'learning_rate/qwen_vl_interface': 9.993689004066573e-06, 'learning_rate/action_model': 9.993689004066573e-05, 'epoch': 0.4})
+[2026-07-28 15:41:14,446][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.022359982132911682, 'train/grad_norm_pre_clip': 0.572404682636261, 'timing/data': 0.0003379872068762779, 'timing/model': 0.5305391242727637, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.4})
+[2026-07-28 15:41:21,716][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.02552023157477379, 'train/grad_norm_pre_clip': 0.7428421974182129, 'timing/data': 0.0008062999695539474, 'timing/model': 0.5311007462441921, 'learning_rate/qwen_vl_interface': 9.993288498102531e-06, 'learning_rate/action_model': 9.993288498102532e-05, 'epoch': 0.4})
+[2026-07-28 15:41:28,989][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.06792043149471283, 'train/grad_norm_pre_clip': 1.3248547315597534, 'timing/data': 0.0003717821091413498, 'timing/model': 0.536490498110652, 'learning_rate/qwen_vl_interface': 9.993083628252424e-06, 'learning_rate/action_model': 9.993083628252424e-05, 'epoch': 0.4})
+[2026-07-28 15:41:36,233][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.04303919896483421, 'train/grad_norm_pre_clip': 1.7554181814193726, 'timing/data': 0.0003973953425884247, 'timing/model': 0.5317180985584855, 'learning_rate/qwen_vl_interface': 9.992875680666849e-06, 'learning_rate/action_model': 9.992875680666849e-05, 'epoch': 0.4})
+[2026-07-28 15:41:43,527][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.04040007293224335, 'train/grad_norm_pre_clip': 1.1047048568725586, 'timing/data': 0.00047302618622779846, 'timing/model': 0.538140214048326, 'learning_rate/qwen_vl_interface': 9.992664655480736e-06, 'learning_rate/action_model': 9.992664655480735e-05, 'epoch': 0.4})
+[2026-07-28 15:41:50,776][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.06535838544368744, 'train/grad_norm_pre_clip': 1.532152771949768, 'timing/data': 0.0003771977499127388, 'timing/model': 0.5298427045345306, 'learning_rate/qwen_vl_interface': 9.992450552831018e-06, 'learning_rate/action_model': 9.992450552831017e-05, 'epoch': 0.4})
+[2026-07-28 15:41:58,032][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.02792167477309704, 'train/grad_norm_pre_clip': 1.1906028985977173, 'timing/data': 0.00034427735954523087, 'timing/model': 0.5281995320692658, 'learning_rate/qwen_vl_interface': 9.992233372856623e-06, 'learning_rate/action_model': 9.992233372856624e-05, 'epoch': 0.4})
+[2026-07-28 15:42:05,381][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.07715624570846558, 'train/grad_norm_pre_clip': 0.9745553731918335, 'timing/data': 0.0003660125657916069, 'timing/model': 0.5297418814152479, 'learning_rate/qwen_vl_interface': 9.99201311569848e-06, 'learning_rate/action_model': 9.99201311569848e-05, 'epoch': 0.4})
+[2026-07-28 15:42:12,615][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.05172150582075119, 'train/grad_norm_pre_clip': 1.819972276687622, 'timing/data': 0.00037078745663166046, 'timing/model': 0.5324003146961331, 'learning_rate/qwen_vl_interface': 9.99178978149951e-06, 'learning_rate/action_model': 9.99178978149951e-05, 'epoch': 0.4})
+[2026-07-28 15:42:19,849][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.011346658691763878, 'train/grad_norm_pre_clip': 0.7829039692878723, 'timing/data': 0.00042561255395412445, 'timing/model': 0.5349608706310391, 'learning_rate/qwen_vl_interface': 9.99156337040463e-06, 'learning_rate/action_model': 9.991563370404629e-05, 'epoch': 0.4})
+[2026-07-28 15:42:27,072][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.025804657489061356, 'train/grad_norm_pre_clip': 0.8477775454521179, 'timing/data': 0.00034696608781814575, 'timing/model': 0.5295551251620054, 'learning_rate/qwen_vl_interface': 9.991333882560759e-06, 'learning_rate/action_model': 9.991333882560759e-05, 'epoch': 0.4})
+[2026-07-28 15:42:34,365][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.11537854373455048, 'train/grad_norm_pre_clip': 3.1481409072875977, 'timing/data': 0.0003748340532183647, 'timing/model': 0.5407432820647955, 'learning_rate/qwen_vl_interface': 9.991101318116806e-06, 'learning_rate/action_model': 9.991101318116806e-05, 'epoch': 0.4})
+[2026-07-28 15:42:41,671][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.10387084633111954, 'train/grad_norm_pre_clip': 2.5002939701080322, 'timing/data': 0.00044876057654619217, 'timing/model': 0.5454321270808578, 'learning_rate/qwen_vl_interface': 9.990865677223684e-06, 'learning_rate/action_model': 9.990865677223684e-05, 'epoch': 0.4})
+[2026-07-28 15:42:49,053][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.07665880769491196, 'train/grad_norm_pre_clip': 2.2689104080200195, 'timing/data': 0.00034743640571832657, 'timing/model': 0.5296119386330247, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.4})
+[2026-07-28 15:42:56,292][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.04189909249544144, 'train/grad_norm_pre_clip': 1.1416417360305786, 'timing/data': 0.0005216822028160095, 'timing/model': 0.5337640224024653, 'learning_rate/qwen_vl_interface': 9.990385166703534e-06, 'learning_rate/action_model': 9.990385166703533e-05, 'epoch': 0.4})
+[2026-07-28 15:43:03,808][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.02937345765531063, 'train/grad_norm_pre_clip': 0.7575137615203857, 'timing/data': 0.0003306148573756218, 'timing/model': 0.5317255714908242, 'learning_rate/qwen_vl_interface': 9.990140297388306e-06, 'learning_rate/action_model': 9.990140297388307e-05, 'epoch': 0.4})
+[2026-07-28 15:43:11,252][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.03752793371677399, 'train/grad_norm_pre_clip': 0.9178026914596558, 'timing/data': 0.0005537467077374458, 'timing/model': 0.5345003698021173, 'learning_rate/qwen_vl_interface': 9.989892352247505e-06, 'learning_rate/action_model': 9.989892352247504e-05, 'epoch': 0.4})
+[2026-07-28 15:43:18,531][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.0492805615067482, 'train/grad_norm_pre_clip': 0.9211851358413696, 'timing/data': 0.00036531034857034683, 'timing/model': 0.5312402145937085, 'learning_rate/qwen_vl_interface': 9.989641331442014e-06, 'learning_rate/action_model': 9.989641331442013e-05, 'epoch': 0.4})
+[2026-07-28 15:43:25,782][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.0656994953751564, 'train/grad_norm_pre_clip': 2.012566566467285, 'timing/data': 0.0003672931343317032, 'timing/model': 0.5300243878737092, 'learning_rate/qwen_vl_interface': 9.989387235134722e-06, 'learning_rate/action_model': 9.989387235134722e-05, 'epoch': 0.4})
+[2026-07-28 15:43:33,091][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.0518302284181118, 'train/grad_norm_pre_clip': 3.0971879959106445, 'timing/data': 0.00039652641862630844, 'timing/model': 0.5347847733646631, 'learning_rate/qwen_vl_interface': 9.98913006349051e-06, 'learning_rate/action_model': 9.989130063490508e-05, 'epoch': 0.4})
+[2026-07-28 15:43:40,399][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.04725123196840286, 'train/grad_norm_pre_clip': 1.1935917139053345, 'timing/data': 0.0003347238525748253, 'timing/model': 0.5300772869959474, 'learning_rate/qwen_vl_interface': 9.988869816676247e-06, 'learning_rate/action_model': 9.988869816676246e-05, 'epoch': 0.4})
+[2026-07-28 15:43:47,655][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.023817429319024086, 'train/grad_norm_pre_clip': 0.8629550337791443, 'timing/data': 0.0003328351303935051, 'timing/model': 0.5302516547963023, 'learning_rate/qwen_vl_interface': 9.988606494860811e-06, 'learning_rate/action_model': 9.988606494860812e-05, 'epoch': 0.4})
+[2026-07-28 15:43:54,937][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.06613222509622574, 'train/grad_norm_pre_clip': 1.8239420652389526, 'timing/data': 0.0004210192710161209, 'timing/model': 0.5344767812639475, 'learning_rate/qwen_vl_interface': 9.98834009821507e-06, 'learning_rate/action_model': 9.988340098215069e-05, 'epoch': 0.4})
+[2026-07-28 15:44:02,522][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.037736523896455765, 'train/grad_norm_pre_clip': 0.8841400146484375, 'timing/data': 0.0003464939072728157, 'timing/model': 0.536413992755115, 'learning_rate/qwen_vl_interface': 9.98807062691188e-06, 'learning_rate/action_model': 9.988070626911879e-05, 'epoch': 0.4})
+[2026-07-28 15:44:09,761][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.06902289390563965, 'train/grad_norm_pre_clip': 1.715972900390625, 'timing/data': 0.00035443808883428574, 'timing/model': 0.5303618088364601, 'learning_rate/qwen_vl_interface': 9.987798081126099e-06, 'learning_rate/action_model': 9.987798081126099e-05, 'epoch': 0.4})
+[2026-07-28 15:44:17,020][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.06788544356822968, 'train/grad_norm_pre_clip': 2.1041548252105713, 'timing/data': 0.00035179685801267624, 'timing/model': 0.543653067201376, 'learning_rate/qwen_vl_interface': 9.987522461034584e-06, 'learning_rate/action_model': 9.987522461034584e-05, 'epoch': 0.4})
+[2026-07-28 15:44:24,257][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.03197188675403595, 'train/grad_norm_pre_clip': 1.5206520557403564, 'timing/data': 0.0003321683034300804, 'timing/model': 0.5309034390375018, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.4})
+[2026-07-28 15:44:31,495][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.038859836757183075, 'train/grad_norm_pre_clip': 1.2905956506729126, 'timing/data': 0.00037062913179397583, 'timing/model': 0.5288950158283114, 'learning_rate/qwen_vl_interface': 9.98696199865172e-06, 'learning_rate/action_model': 9.98696199865172e-05, 'epoch': 0.4})
+[2026-07-28 15:44:38,823][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.08860288560390472, 'train/grad_norm_pre_clip': 2.797689437866211, 'timing/data': 0.0005601206794381142, 'timing/model': 0.5420573996379972, 'learning_rate/qwen_vl_interface': 9.986677156724054e-06, 'learning_rate/action_model': 9.986677156724054e-05, 'epoch': 0.4})
+[2026-07-28 15:44:46,128][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.03036656230688095, 'train/grad_norm_pre_clip': 1.1751811504364014, 'timing/data': 0.0003305906429886818, 'timing/model': 0.5308513240888715, 'learning_rate/qwen_vl_interface': 9.986389241218003e-06, 'learning_rate/action_model': 9.986389241218003e-05, 'epoch': 0.4})
+[2026-07-28 15:44:53,541][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.07932887971401215, 'train/grad_norm_pre_clip': 1.4709467887878418, 'timing/data': 0.0003397073596715927, 'timing/model': 0.5553998993709683, 'learning_rate/qwen_vl_interface': 9.986098252320397e-06, 'learning_rate/action_model': 9.986098252320397e-05, 'epoch': 0.4})
+[2026-07-28 15:45:01,042][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.06763139367103577, 'train/grad_norm_pre_clip': 2.225151777267456, 'timing/data': 0.0003395276144146919, 'timing/model': 0.5453940220177174, 'learning_rate/qwen_vl_interface': 9.985804190220054e-06, 'learning_rate/action_model': 9.985804190220053e-05, 'epoch': 0.4})
+[2026-07-28 15:45:08,551][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.03553992509841919, 'train/grad_norm_pre_clip': 0.7175177335739136, 'timing/data': 0.0003247307613492012, 'timing/model': 0.5448208628222346, 'learning_rate/qwen_vl_interface': 9.985507055107788e-06, 'learning_rate/action_model': 9.985507055107788e-05, 'epoch': 0.4})
+[2026-07-28 15:45:15,980][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.05490148440003395, 'train/grad_norm_pre_clip': 1.462206244468689, 'timing/data': 0.00033309590071439743, 'timing/model': 0.5315130967646837, 'learning_rate/qwen_vl_interface': 9.985206847176406e-06, 'learning_rate/action_model': 9.985206847176406e-05, 'epoch': 0.4})
+[2026-07-28 15:45:23,264][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.0367375984787941, 'train/grad_norm_pre_clip': 0.7968822121620178, 'timing/data': 0.0004214216023683548, 'timing/model': 0.5329308807849884, 'learning_rate/qwen_vl_interface': 9.984903566620711e-06, 'learning_rate/action_model': 9.984903566620712e-05, 'epoch': 0.4})
+[2026-07-28 15:45:30,549][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.036362502723932266, 'train/grad_norm_pre_clip': 1.5734748840332031, 'timing/data': 0.00036044418811798096, 'timing/model': 0.5312156975269318, 'learning_rate/qwen_vl_interface': 9.984597213637497e-06, 'learning_rate/action_model': 9.984597213637497e-05, 'epoch': 0.4})
+[2026-07-28 15:45:37,808][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.04812820628285408, 'train/grad_norm_pre_clip': 0.8343642354011536, 'timing/data': 0.00036362092941999435, 'timing/model': 0.5315682319924235, 'learning_rate/qwen_vl_interface': 9.984287788425555e-06, 'learning_rate/action_model': 9.984287788425555e-05, 'epoch': 0.4})
+[2026-07-28 15:45:45,068][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.022262969985604286, 'train/grad_norm_pre_clip': 0.8176707029342651, 'timing/data': 0.000338604673743248, 'timing/model': 0.550645999610424, 'learning_rate/qwen_vl_interface': 9.983975291185666e-06, 'learning_rate/action_model': 9.983975291185667e-05, 'epoch': 0.4})
+[2026-07-28 15:45:52,326][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.0473911389708519, 'train/grad_norm_pre_clip': 1.5886262655258179, 'timing/data': 0.0003300420939922333, 'timing/model': 0.528581720776856, 'learning_rate/qwen_vl_interface': 9.983659722120607e-06, 'learning_rate/action_model': 9.983659722120608e-05, 'epoch': 0.4})
+[2026-07-28 15:45:59,608][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.01595466397702694, 'train/grad_norm_pre_clip': 0.6992018818855286, 'timing/data': 0.00033496879041194916, 'timing/model': 0.5294823441654444, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.4})
+[2026-07-28 15:46:06,845][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.03451397269964218, 'train/grad_norm_pre_clip': 0.8441051244735718, 'timing/data': 0.0003178222104907036, 'timing/model': 0.533022090792656, 'learning_rate/qwen_vl_interface': 9.983019369336048e-06, 'learning_rate/action_model': 9.983019369336048e-05, 'epoch': 0.41})
+[2026-07-28 15:46:14,120][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.0716361254453659, 'train/grad_norm_pre_clip': 2.3151817321777344, 'timing/data': 0.00033155176788568497, 'timing/model': 0.5327349435538054, 'learning_rate/qwen_vl_interface': 9.982694586032066e-06, 'learning_rate/action_model': 9.982694586032065e-05, 'epoch': 0.41})
+[2026-07-28 15:46:21,375][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.04579415172338486, 'train/grad_norm_pre_clip': 1.3705612421035767, 'timing/data': 0.0004582498222589493, 'timing/model': 0.5297673512250185, 'learning_rate/qwen_vl_interface': 9.982366731733948e-06, 'learning_rate/action_model': 9.982366731733947e-05, 'epoch': 0.41})
+[2026-07-28 15:46:28,628][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.04489146173000336, 'train/grad_norm_pre_clip': 1.139643907546997, 'timing/data': 0.0003827158361673355, 'timing/model': 0.5297608189284801, 'learning_rate/qwen_vl_interface': 9.982035806654437e-06, 'learning_rate/action_model': 9.982035806654437e-05, 'epoch': 0.41})
+[2026-07-28 15:46:35,895][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.021852366626262665, 'train/grad_norm_pre_clip': 0.9205999374389648, 'timing/data': 0.0003709997981786728, 'timing/model': 0.534233033657074, 'learning_rate/qwen_vl_interface': 9.981701811008267e-06, 'learning_rate/action_model': 9.981701811008267e-05, 'epoch': 0.41})
+[2026-07-28 15:46:43,139][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.023151881992816925, 'train/grad_norm_pre_clip': 1.7967363595962524, 'timing/data': 0.00044190511107444763, 'timing/model': 0.5296668978407979, 'learning_rate/qwen_vl_interface': 9.981364745012163e-06, 'learning_rate/action_model': 9.981364745012163e-05, 'epoch': 0.41})
+[2026-07-28 15:46:50,371][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.08029674738645554, 'train/grad_norm_pre_clip': 1.8267779350280762, 'timing/data': 0.00037456024438142776, 'timing/model': 0.5304149845615029, 'learning_rate/qwen_vl_interface': 9.981024608884842e-06, 'learning_rate/action_model': 9.981024608884843e-05, 'epoch': 0.41})
+[2026-07-28 15:46:57,616][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.05810409039258957, 'train/grad_norm_pre_clip': 2.1532394886016846, 'timing/data': 0.0006716735661029816, 'timing/model': 0.5307759214192629, 'learning_rate/qwen_vl_interface': 9.980681402847017e-06, 'learning_rate/action_model': 9.980681402847016e-05, 'epoch': 0.41})
+[2026-07-28 15:47:04,864][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.06644447892904282, 'train/grad_norm_pre_clip': 2.2240731716156006, 'timing/data': 0.0003510070964694023, 'timing/model': 0.5359053136780858, 'learning_rate/qwen_vl_interface': 9.980335127121391e-06, 'learning_rate/action_model': 9.98033512712139e-05, 'epoch': 0.41})
+[2026-07-28 15:47:12,104][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.041699253022670746, 'train/grad_norm_pre_clip': 0.7844966053962708, 'timing/data': 0.00032644905149936676, 'timing/model': 0.5330937039107084, 'learning_rate/qwen_vl_interface': 9.979985781932658e-06, 'learning_rate/action_model': 9.979985781932657e-05, 'epoch': 0.41})
+[2026-07-28 15:47:19,327][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.03576415032148361, 'train/grad_norm_pre_clip': 1.198784589767456, 'timing/data': 0.00033483467996120453, 'timing/model': 0.5295537291094661, 'learning_rate/qwen_vl_interface': 9.979633367507501e-06, 'learning_rate/action_model': 9.979633367507502e-05, 'epoch': 0.41})
+[2026-07-28 15:47:26,560][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.04944661632180214, 'train/grad_norm_pre_clip': 1.8158997297286987, 'timing/data': 0.00035286974161863327, 'timing/model': 0.528666271828115, 'learning_rate/qwen_vl_interface': 9.979277884074604e-06, 'learning_rate/action_model': 9.979277884074604e-05, 'epoch': 0.41})
+[2026-07-28 15:47:33,793][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.07321474701166153, 'train/grad_norm_pre_clip': 2.5488879680633545, 'timing/data': 0.0004055416211485863, 'timing/model': 0.5285127647221088, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.41})
+[2026-07-28 15:47:41,305][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.04035252705216408, 'train/grad_norm_pre_clip': 1.396397352218628, 'timing/data': 0.0003887070342898369, 'timing/model': 0.5321167334914207, 'learning_rate/qwen_vl_interface': 9.978557711110245e-06, 'learning_rate/action_model': 9.978557711110244e-05, 'epoch': 0.41})
+[2026-07-28 15:47:48,526][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.030560512095689774, 'train/grad_norm_pre_clip': 1.0248287916183472, 'timing/data': 0.0003936607390642166, 'timing/model': 0.5288327401503921, 'learning_rate/qwen_vl_interface': 9.978193022046097e-06, 'learning_rate/action_model': 9.978193022046097e-05, 'epoch': 0.41})
+[2026-07-28 15:47:55,751][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.05686177313327789, 'train/grad_norm_pre_clip': 1.6806706190109253, 'timing/data': 0.0006959224119782448, 'timing/model': 0.5317008038982749, 'learning_rate/qwen_vl_interface': 9.977825264908832e-06, 'learning_rate/action_model': 9.977825264908832e-05, 'epoch': 0.41})
+[2026-07-28 15:48:03,015][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.047337211668491364, 'train/grad_norm_pre_clip': 1.5560832023620605, 'timing/data': 0.00041049346327781677, 'timing/model': 0.5427261609584093, 'learning_rate/qwen_vl_interface': 9.97745443993708e-06, 'learning_rate/action_model': 9.97745443993708e-05, 'epoch': 0.41})
+[2026-07-28 15:48:10,271][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.017045948654413223, 'train/grad_norm_pre_clip': 0.4471178352832794, 'timing/data': 0.00035945605486631393, 'timing/model': 0.5315748499706388, 'learning_rate/qwen_vl_interface': 9.977080547371468e-06, 'learning_rate/action_model': 9.977080547371468e-05, 'epoch': 0.41})
+[2026-07-28 15:48:17,549][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.021905509755015373, 'train/grad_norm_pre_clip': 0.774879515171051, 'timing/data': 0.00035548582673072815, 'timing/model': 0.5315421670675278, 'learning_rate/qwen_vl_interface': 9.97670358745461e-06, 'learning_rate/action_model': 9.97670358745461e-05, 'epoch': 0.41})
+[2026-07-28 15:48:24,813][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.05284368619322777, 'train/grad_norm_pre_clip': 1.486061453819275, 'timing/data': 0.0003619445487856865, 'timing/model': 0.5319324173033237, 'learning_rate/qwen_vl_interface': 9.976323560431112e-06, 'learning_rate/action_model': 9.976323560431111e-05, 'epoch': 0.41})
+[2026-07-28 15:48:32,061][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.036344192922115326, 'train/grad_norm_pre_clip': 1.0843067169189453, 'timing/data': 0.00038591306656599045, 'timing/model': 0.5306919077411294, 'learning_rate/qwen_vl_interface': 9.975940466547567e-06, 'learning_rate/action_model': 9.975940466547567e-05, 'epoch': 0.41})
+[2026-07-28 15:48:39,310][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.03809487819671631, 'train/grad_norm_pre_clip': 0.8645200729370117, 'timing/data': 0.0003535626456141472, 'timing/model': 0.5335373310372233, 'learning_rate/qwen_vl_interface': 9.975554306052563e-06, 'learning_rate/action_model': 9.975554306052562e-05, 'epoch': 0.41})
+[2026-07-28 15:48:46,941][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.04119434952735901, 'train/grad_norm_pre_clip': 1.1430975198745728, 'timing/data': 0.00033402349799871445, 'timing/model': 0.5298499530181289, 'learning_rate/qwen_vl_interface': 9.975165079196674e-06, 'learning_rate/action_model': 9.975165079196673e-05, 'epoch': 0.41})
+[2026-07-28 15:48:54,305][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.030243827030062675, 'train/grad_norm_pre_clip': 0.8890534043312073, 'timing/data': 0.0003541242331266403, 'timing/model': 0.5307067576795816, 'learning_rate/qwen_vl_interface': 9.974772786232466e-06, 'learning_rate/action_model': 9.974772786232466e-05, 'epoch': 0.41})
+[2026-07-28 15:49:01,565][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.048314616084098816, 'train/grad_norm_pre_clip': 1.2832001447677612, 'timing/data': 0.0003857789561152458, 'timing/model': 0.5287693524733186, 'learning_rate/qwen_vl_interface': 9.974377427414493e-06, 'learning_rate/action_model': 9.974377427414493e-05, 'epoch': 0.41})
+[2026-07-28 15:49:08,844][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.019995326176285744, 'train/grad_norm_pre_clip': 0.5972661972045898, 'timing/data': 0.0003306455910205841, 'timing/model': 0.5297339363023639, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.41})
+[2026-07-28 15:49:16,120][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.052066974341869354, 'train/grad_norm_pre_clip': 0.9554343819618225, 'timing/data': 0.0004228847101330757, 'timing/model': 0.5310048609972, 'learning_rate/qwen_vl_interface': 9.973577513245417e-06, 'learning_rate/action_model': 9.973577513245417e-05, 'epoch': 0.41})
+[2026-07-28 15:49:23,351][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.027570530772209167, 'train/grad_norm_pre_clip': 0.7947295308113098, 'timing/data': 0.0003639254719018936, 'timing/model': 0.5307514481246471, 'learning_rate/qwen_vl_interface': 9.973172958413372e-06, 'learning_rate/action_model': 9.97317295841337e-05, 'epoch': 0.41})
+[2026-07-28 15:49:30,636][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.026464302092790604, 'train/grad_norm_pre_clip': 0.9331695437431335, 'timing/data': 0.0004481952637434006, 'timing/model': 0.5320026408880949, 'learning_rate/qwen_vl_interface': 9.97276533876567e-06, 'learning_rate/action_model': 9.97276533876567e-05, 'epoch': 0.41})
+[2026-07-28 15:49:37,915][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.032850705087184906, 'train/grad_norm_pre_clip': 1.1327815055847168, 'timing/data': 0.00039452966302633286, 'timing/model': 0.5312773054465652, 'learning_rate/qwen_vl_interface': 9.972354654566816e-06, 'learning_rate/action_model': 9.972354654566816e-05, 'epoch': 0.41})
+[2026-07-28 15:49:45,166][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.02793865092098713, 'train/grad_norm_pre_clip': 1.3947519063949585, 'timing/data': 0.00039299577474594116, 'timing/model': 0.5332207391038537, 'learning_rate/qwen_vl_interface': 9.971940906083295e-06, 'learning_rate/action_model': 9.971940906083295e-05, 'epoch': 0.41})
+[2026-07-28 15:49:52,399][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.04922318831086159, 'train/grad_norm_pre_clip': 1.5596516132354736, 'timing/data': 0.000372154638171196, 'timing/model': 0.5304972818121314, 'learning_rate/qwen_vl_interface': 9.971524093583586e-06, 'learning_rate/action_model': 9.971524093583586e-05, 'epoch': 0.41})
+[2026-07-28 15:49:59,651][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.05372775346040726, 'train/grad_norm_pre_clip': 2.3836867809295654, 'timing/data': 0.00040981918573379517, 'timing/model': 0.5342901293188334, 'learning_rate/qwen_vl_interface': 9.971104217338154e-06, 'learning_rate/action_model': 9.971104217338153e-05, 'epoch': 0.41})
+[2026-07-28 15:50:06,887][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.049136288464069366, 'train/grad_norm_pre_clip': 1.6815824508666992, 'timing/data': 0.0003744903951883316, 'timing/model': 0.5342784747481346, 'learning_rate/qwen_vl_interface': 9.970681277619452e-06, 'learning_rate/action_model': 9.970681277619452e-05, 'epoch': 0.41})
+[2026-07-28 15:50:14,126][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.031922999769449234, 'train/grad_norm_pre_clip': 1.2797741889953613, 'timing/data': 0.00039768125861883163, 'timing/model': 0.5306333787739277, 'learning_rate/qwen_vl_interface': 9.97025527470192e-06, 'learning_rate/action_model': 9.970255274701919e-05, 'epoch': 0.41})
+[2026-07-28 15:50:21,363][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.046356894075870514, 'train/grad_norm_pre_clip': 1.6759617328643799, 'timing/data': 0.00037561915814876556, 'timing/model': 0.5325682088732719, 'learning_rate/qwen_vl_interface': 9.969826208861987e-06, 'learning_rate/action_model': 9.969826208861987e-05, 'epoch': 0.41})
+[2026-07-28 15:50:28,584][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.03743598610162735, 'train/grad_norm_pre_clip': 1.3330827951431274, 'timing/data': 0.00037769973278045654, 'timing/model': 0.5306339478120208, 'learning_rate/qwen_vl_interface': 9.96939408037807e-06, 'learning_rate/action_model': 9.96939408037807e-05, 'epoch': 0.41})
+[2026-07-28 15:50:35,820][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.021729175001382828, 'train/grad_norm_pre_clip': 1.2112679481506348, 'timing/data': 0.00044032465666532516, 'timing/model': 0.5296338573098183, 'learning_rate/qwen_vl_interface': 9.968958889530573e-06, 'learning_rate/action_model': 9.968958889530573e-05, 'epoch': 0.41})
+[2026-07-28 15:50:43,052][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.04235829412937164, 'train/grad_norm_pre_clip': 1.0920813083648682, 'timing/data': 0.00034354254603385925, 'timing/model': 0.5285363318398595, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.41})
+[2026-07-28 15:50:50,265][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.05257076397538185, 'train/grad_norm_pre_clip': 2.8188908100128174, 'timing/data': 0.0003602057695388794, 'timing/model': 0.5293344026431441, 'learning_rate/qwen_vl_interface': 9.968079321876382e-06, 'learning_rate/action_model': 9.968079321876382e-05, 'epoch': 0.41})
+[2026-07-28 15:50:57,487][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.04747951030731201, 'train/grad_norm_pre_clip': 2.061389446258545, 'timing/data': 0.00037309806793928146, 'timing/model': 0.5295053906738758, 'learning_rate/qwen_vl_interface': 9.967634945640432e-06, 'learning_rate/action_model': 9.967634945640431e-05, 'epoch': 0.41})
+[2026-07-28 15:51:04,725][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.09678968042135239, 'train/grad_norm_pre_clip': 1.6581497192382812, 'timing/data': 0.00038107484579086304, 'timing/model': 0.5297473277896643, 'learning_rate/qwen_vl_interface': 9.967187508182385e-06, 'learning_rate/action_model': 9.967187508182385e-05, 'epoch': 0.41})
+[2026-07-28 15:51:11,961][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.03576964884996414, 'train/grad_norm_pre_clip': 0.9550207257270813, 'timing/data': 0.0006757061928510666, 'timing/model': 0.5324016511440277, 'learning_rate/qwen_vl_interface': 9.966737009792577e-06, 'learning_rate/action_model': 9.966737009792577e-05, 'epoch': 0.41})
+[2026-07-28 15:51:19,180][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.05898217111825943, 'train/grad_norm_pre_clip': 2.2787091732025146, 'timing/data': 0.00036392081528902054, 'timing/model': 0.5308595933020115, 'learning_rate/qwen_vl_interface': 9.966283450763333e-06, 'learning_rate/action_model': 9.966283450763333e-05, 'epoch': 0.41})
+[2026-07-28 15:51:26,424][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.10141865909099579, 'train/grad_norm_pre_clip': 1.7717647552490234, 'timing/data': 0.00048022158443927765, 'timing/model': 0.5318911857903004, 'learning_rate/qwen_vl_interface': 9.965826831388962e-06, 'learning_rate/action_model': 9.965826831388961e-05, 'epoch': 0.41})
+[2026-07-28 15:51:33,649][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2250
+[2026-07-28 15:52:52,296][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2250
+[2026-07-28 16:08:10,400][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2250
+[2026-07-28 16:08:10,402][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.0366421602666378, 'train/grad_norm_pre_clip': 0.662065327167511, 'eval/action_loss/seconds': 78.64266879018396, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.04586682468652725, 'eval/latency_3/loss': 0.04586682468652725, 'eval/flappy/loss': 0.04586682468652725, 'eval/flappy/latency_3/loss': 0.04586682468652725, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 354.3150041695684, 'rl_games_eval/mid_train/mean_length': 2873.35, 'rl_games_eval/mid_train/std_reward': 135.2535348989007, 'rl_games_eval/mid_train/std_length': 1089.2268025989813, 'rl_games_eval/mid_train/macro_mean_reward': 354.3150041695684, 'rl_games_eval/mid_train/macro_mean_length': 2873.35, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 354.3150041695684, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2873.35, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 135.2535348989007, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1089.2268025989813, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 354.3150041695684, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00035171955823898315, 'timing/model': 0.5292435074225068, 'learning_rate/qwen_vl_interface': 9.965367151965757e-06, 'learning_rate/action_model': 9.965367151965756e-05, 'epoch': 0.41})
+[2026-07-28 16:08:19,669][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.040849246084690094, 'train/grad_norm_pre_clip': 1.323276162147522, 'timing/data': 0.00038697104901075363, 'timing/model': 0.6213085521012545, 'learning_rate/qwen_vl_interface': 9.964904412792002e-06, 'learning_rate/action_model': 9.964904412792002e-05, 'epoch': 0.41})
+[2026-07-28 16:08:28,871][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.045545801520347595, 'train/grad_norm_pre_clip': 2.3921594619750977, 'timing/data': 0.0003360658884048462, 'timing/model': 0.6198758222162724, 'learning_rate/qwen_vl_interface': 9.964438614167964e-06, 'learning_rate/action_model': 9.964438614167964e-05, 'epoch': 0.41})
+[2026-07-28 16:08:38,062][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.029848074540495872, 'train/grad_norm_pre_clip': 1.8626818656921387, 'timing/data': 0.00036227423697710037, 'timing/model': 0.6237430386245251, 'learning_rate/qwen_vl_interface': 9.963969756395892e-06, 'learning_rate/action_model': 9.96396975639589e-05, 'epoch': 0.41})
+[2026-07-28 16:08:47,576][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.055391620844602585, 'train/grad_norm_pre_clip': 1.7515465021133423, 'timing/data': 0.00033674389123916626, 'timing/model': 0.6317001804709435, 'learning_rate/qwen_vl_interface': 9.963497839780025e-06, 'learning_rate/action_model': 9.963497839780024e-05, 'epoch': 0.41})
+[2026-07-28 16:08:56,722][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.03614667057991028, 'train/grad_norm_pre_clip': 1.5606961250305176, 'timing/data': 0.0003161383792757988, 'timing/model': 0.6174130840227008, 'learning_rate/qwen_vl_interface': 9.963022864626582e-06, 'learning_rate/action_model': 9.963022864626581e-05, 'epoch': 0.41})
+[2026-07-28 16:09:06,067][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.04843205213546753, 'train/grad_norm_pre_clip': 1.2251235246658325, 'timing/data': 0.00035468582063913345, 'timing/model': 0.8500122744590044, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.41})
+[2026-07-28 16:09:15,338][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.056318238377571106, 'train/grad_norm_pre_clip': 0.9383368492126465, 'timing/data': 0.00037986133247613907, 'timing/model': 0.6209011115133762, 'learning_rate/qwen_vl_interface': 9.962063739941783e-06, 'learning_rate/action_model': 9.962063739941783e-05, 'epoch': 0.41})
+[2026-07-28 16:09:24,489][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.03530174493789673, 'train/grad_norm_pre_clip': 0.9859486222267151, 'timing/data': 0.00046854373067617416, 'timing/model': 0.6208827691152692, 'learning_rate/qwen_vl_interface': 9.961579591032791e-06, 'learning_rate/action_model': 9.96157959103279e-05, 'epoch': 0.41})
+[2026-07-28 16:09:33,874][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.03214311599731445, 'train/grad_norm_pre_clip': 0.7345787882804871, 'timing/data': 0.0003318497911095619, 'timing/model': 0.6210901401937008, 'learning_rate/qwen_vl_interface': 9.961092384830956e-06, 'learning_rate/action_model': 9.961092384830956e-05, 'epoch': 0.41})
+[2026-07-28 16:09:43,300][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.09143383800983429, 'train/grad_norm_pre_clip': 2.5318822860717773, 'timing/data': 0.0004018731415271759, 'timing/model': 0.8476345827803016, 'learning_rate/qwen_vl_interface': 9.96060212165242e-06, 'learning_rate/action_model': 9.96060212165242e-05, 'epoch': 0.42})
+[2026-07-28 16:09:52,438][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.031286727637052536, 'train/grad_norm_pre_clip': 1.1735467910766602, 'timing/data': 0.00035676173865795135, 'timing/model': 0.6207145461812615, 'learning_rate/qwen_vl_interface': 9.96010880181531e-06, 'learning_rate/action_model': 9.960108801815309e-05, 'epoch': 0.42})
+[2026-07-28 16:10:01,658][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.047432392835617065, 'train/grad_norm_pre_clip': 1.3867112398147583, 'timing/data': 0.0003893822431564331, 'timing/model': 0.6196505259722471, 'learning_rate/qwen_vl_interface': 9.959612425639736e-06, 'learning_rate/action_model': 9.959612425639735e-05, 'epoch': 0.42})
+[2026-07-28 16:10:10,803][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.021910490468144417, 'train/grad_norm_pre_clip': 0.8077556490898132, 'timing/data': 0.0004097428172826767, 'timing/model': 0.6201017070561647, 'learning_rate/qwen_vl_interface': 9.95911299344779e-06, 'learning_rate/action_model': 9.959112993447789e-05, 'epoch': 0.42})
+[2026-07-28 16:10:20,186][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.02984153851866722, 'train/grad_norm_pre_clip': 1.1959301233291626, 'timing/data': 0.00036821048706769943, 'timing/model': 0.6197294984012842, 'learning_rate/qwen_vl_interface': 9.958610505563546e-06, 'learning_rate/action_model': 9.958610505563546e-05, 'epoch': 0.42})
+[2026-07-28 16:10:29,424][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.05472061038017273, 'train/grad_norm_pre_clip': 1.214171051979065, 'timing/data': 0.00034225359559059143, 'timing/model': 0.6398170478641987, 'learning_rate/qwen_vl_interface': 9.95810496231307e-06, 'learning_rate/action_model': 9.95810496231307e-05, 'epoch': 0.42})
+[2026-07-28 16:10:38,669][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.00990486703813076, 'train/grad_norm_pre_clip': 0.5506146550178528, 'timing/data': 0.0003677988424897194, 'timing/model': 0.6205004444345832, 'learning_rate/qwen_vl_interface': 9.957596364024396e-06, 'learning_rate/action_model': 9.957596364024396e-05, 'epoch': 0.42})
+[2026-07-28 16:10:48,086][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.03502360358834267, 'train/grad_norm_pre_clip': 1.691384196281433, 'timing/data': 0.00037472881376743317, 'timing/model': 0.8453809171915054, 'learning_rate/qwen_vl_interface': 9.957084711027553e-06, 'learning_rate/action_model': 9.957084711027552e-05, 'epoch': 0.42})
+[2026-07-28 16:10:57,486][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.02566937357187271, 'train/grad_norm_pre_clip': 1.54030179977417, 'timing/data': 0.0003365222364664078, 'timing/model': 0.6240645665675402, 'learning_rate/qwen_vl_interface': 9.956570003654544e-06, 'learning_rate/action_model': 9.956570003654544e-05, 'epoch': 0.42})
+[2026-07-28 16:11:06,629][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.06796786189079285, 'train/grad_norm_pre_clip': 1.6312612295150757, 'timing/data': 0.00031573139131069183, 'timing/model': 0.6194430813193321, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.42})
+[2026-07-28 16:11:15,757][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.05830606445670128, 'train/grad_norm_pre_clip': 2.446080207824707, 'timing/data': 0.0003349464386701584, 'timing/model': 0.6217466136440635, 'learning_rate/qwen_vl_interface': 9.955531427117964e-06, 'learning_rate/action_model': 9.955531427117964e-05, 'epoch': 0.42})
+[2026-07-28 16:11:24,907][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.02430199831724167, 'train/grad_norm_pre_clip': 1.7323298454284668, 'timing/data': 0.00032489653676748276, 'timing/model': 0.6212574895471334, 'learning_rate/qwen_vl_interface': 9.955007558628316e-06, 'learning_rate/action_model': 9.955007558628316e-05, 'epoch': 0.42})
+[2026-07-28 16:11:34,275][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.028900979086756706, 'train/grad_norm_pre_clip': 1.3545750379562378, 'timing/data': 0.00032983534038066864, 'timing/model': 0.6233575399965048, 'learning_rate/qwen_vl_interface': 9.954480637110346e-06, 'learning_rate/action_model': 9.954480637110344e-05, 'epoch': 0.42})
+[2026-07-28 16:11:43,436][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.11261260509490967, 'train/grad_norm_pre_clip': 2.7478103637695312, 'timing/data': 0.0005166046321392059, 'timing/model': 0.6213433211669326, 'learning_rate/qwen_vl_interface': 9.953950662905963e-06, 'learning_rate/action_model': 9.953950662905964e-05, 'epoch': 0.42})
+[2026-07-28 16:11:52,549][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.038487616926431656, 'train/grad_norm_pre_clip': 1.5680885314941406, 'timing/data': 0.0004599904641509056, 'timing/model': 0.6195632526651025, 'learning_rate/qwen_vl_interface': 9.953417636359067e-06, 'learning_rate/action_model': 9.953417636359067e-05, 'epoch': 0.42})
+[2026-07-28 16:12:01,674][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.012319695204496384, 'train/grad_norm_pre_clip': 1.0814412832260132, 'timing/data': 0.0003244718536734581, 'timing/model': 0.6192236309871078, 'learning_rate/qwen_vl_interface': 9.95288155781553e-06, 'learning_rate/action_model': 9.95288155781553e-05, 'epoch': 0.42})
+[2026-07-28 16:12:11,050][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.07783814519643784, 'train/grad_norm_pre_clip': 3.097377061843872, 'timing/data': 0.0005524344742298126, 'timing/model': 0.6212355643510818, 'learning_rate/qwen_vl_interface': 9.952342427623209e-06, 'learning_rate/action_model': 9.952342427623208e-05, 'epoch': 0.42})
+[2026-07-28 16:12:20,401][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.06807069480419159, 'train/grad_norm_pre_clip': 2.9986484050750732, 'timing/data': 0.0003306018188595772, 'timing/model': 0.8478605253621936, 'learning_rate/qwen_vl_interface': 9.95180024613194e-06, 'learning_rate/action_model': 9.951800246131939e-05, 'epoch': 0.42})
+[2026-07-28 16:12:29,557][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.08045069873332977, 'train/grad_norm_pre_clip': 2.1729607582092285, 'timing/data': 0.00031891465187072754, 'timing/model': 0.6194436941295862, 'learning_rate/qwen_vl_interface': 9.951255013693537e-06, 'learning_rate/action_model': 9.951255013693536e-05, 'epoch': 0.42})
+[2026-07-28 16:12:38,693][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.0418081060051918, 'train/grad_norm_pre_clip': 1.2773851156234741, 'timing/data': 0.00033588986843824387, 'timing/model': 0.6215401338413358, 'learning_rate/qwen_vl_interface': 9.950706730661797e-06, 'learning_rate/action_model': 9.950706730661797e-05, 'epoch': 0.42})
+[2026-07-28 16:12:48,119][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.012551160529255867, 'train/grad_norm_pre_clip': 0.7994308471679688, 'timing/data': 0.00042532850056886673, 'timing/model': 0.6188641954213381, 'learning_rate/qwen_vl_interface': 9.950155397392494e-06, 'learning_rate/action_model': 9.950155397392493e-05, 'epoch': 0.42})
+[2026-07-28 16:12:57,366][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.02260763756930828, 'train/grad_norm_pre_clip': 0.8586871027946472, 'timing/data': 0.00032517407089471817, 'timing/model': 0.6189431697130203, 'learning_rate/qwen_vl_interface': 9.949601014243382e-06, 'learning_rate/action_model': 9.949601014243382e-05, 'epoch': 0.42})
+[2026-07-28 16:13:06,824][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.02755456231534481, 'train/grad_norm_pre_clip': 0.8301401138305664, 'timing/data': 0.0003239652141928673, 'timing/model': 0.618879821151495, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.42})
+[2026-07-28 16:13:16,143][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.07149709016084671, 'train/grad_norm_pre_clip': 1.54339599609375, 'timing/data': 0.0004174010828137398, 'timing/model': 0.8040676349774003, 'learning_rate/qwen_vl_interface': 9.948483099746645e-06, 'learning_rate/action_model': 9.948483099746645e-05, 'epoch': 0.42})
+[2026-07-28 16:13:25,262][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.12077875435352325, 'train/grad_norm_pre_clip': 2.728398084640503, 'timing/data': 0.00032595545053482056, 'timing/model': 0.6188228707760572, 'learning_rate/qwen_vl_interface': 9.947919569124423e-06, 'learning_rate/action_model': 9.947919569124422e-05, 'epoch': 0.42})
+[2026-07-28 16:13:34,655][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.03951438143849373, 'train/grad_norm_pre_clip': 1.0937544107437134, 'timing/data': 0.0004115225747227669, 'timing/model': 0.6212027287110686, 'learning_rate/qwen_vl_interface': 9.947352990073196e-06, 'learning_rate/action_model': 9.947352990073195e-05, 'epoch': 0.42})
+[2026-07-28 16:13:43,858][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.04373961314558983, 'train/grad_norm_pre_clip': 1.3370847702026367, 'timing/data': 0.00032537616789340973, 'timing/model': 0.6219680337235332, 'learning_rate/qwen_vl_interface': 9.946783362960612e-06, 'learning_rate/action_model': 9.946783362960612e-05, 'epoch': 0.42})
+[2026-07-28 16:13:53,262][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.025244398042559624, 'train/grad_norm_pre_clip': 0.9148552417755127, 'timing/data': 0.00040223076939582825, 'timing/model': 0.8647260935977101, 'learning_rate/qwen_vl_interface': 9.946210688156297e-06, 'learning_rate/action_model': 9.946210688156297e-05, 'epoch': 0.42})
+[2026-07-28 16:14:02,527][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.04569389671087265, 'train/grad_norm_pre_clip': 1.143133282661438, 'timing/data': 0.00044123176485300064, 'timing/model': 0.625002296641469, 'learning_rate/qwen_vl_interface': 9.945634966031852e-06, 'learning_rate/action_model': 9.945634966031851e-05, 'epoch': 0.42})
+[2026-07-28 16:14:11,937][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.05710926651954651, 'train/grad_norm_pre_clip': 1.3816981315612793, 'timing/data': 0.000328991562128067, 'timing/model': 0.6214954247698188, 'learning_rate/qwen_vl_interface': 9.945056196960858e-06, 'learning_rate/action_model': 9.945056196960858e-05, 'epoch': 0.42})
+[2026-07-28 16:14:21,069][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.02471545897424221, 'train/grad_norm_pre_clip': 0.9224631190299988, 'timing/data': 0.0005389107391238213, 'timing/model': 0.618770182132721, 'learning_rate/qwen_vl_interface': 9.944474381318871e-06, 'learning_rate/action_model': 9.94447438131887e-05, 'epoch': 0.42})
+[2026-07-28 16:14:30,481][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.04061923921108246, 'train/grad_norm_pre_clip': 0.9418254494667053, 'timing/data': 0.00037835538387298584, 'timing/model': 0.6205855896696448, 'learning_rate/qwen_vl_interface': 9.943889519483424e-06, 'learning_rate/action_model': 9.943889519483424e-05, 'epoch': 0.42})
+[2026-07-28 16:14:39,661][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.052358388900756836, 'train/grad_norm_pre_clip': 1.9412122964859009, 'timing/data': 0.0003403620794415474, 'timing/model': 0.62206613086164, 'learning_rate/qwen_vl_interface': 9.94330161183403e-06, 'learning_rate/action_model': 9.94330161183403e-05, 'epoch': 0.42})
+[2026-07-28 16:14:49,053][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.055413179099559784, 'train/grad_norm_pre_clip': 1.5153254270553589, 'timing/data': 0.0003203507512807846, 'timing/model': 0.6204233560711145, 'learning_rate/qwen_vl_interface': 9.942710658752175e-06, 'learning_rate/action_model': 9.942710658752174e-05, 'epoch': 0.42})
+[2026-07-28 16:14:58,272][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.013252843171358109, 'train/grad_norm_pre_clip': 0.7070509195327759, 'timing/data': 0.0003264583647251129, 'timing/model': 0.6260243402794003, 'learning_rate/qwen_vl_interface': 9.942116660621319e-06, 'learning_rate/action_model': 9.94211666062132e-05, 'epoch': 0.42})
+[2026-07-28 16:15:07,672][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.02972521260380745, 'train/grad_norm_pre_clip': 1.0577131509780884, 'timing/data': 0.00037952419370412827, 'timing/model': 0.6183391455560923, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.42})
+[2026-07-28 16:15:16,799][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.03887798264622688, 'train/grad_norm_pre_clip': 1.41355562210083, 'timing/data': 0.00039720628410577774, 'timing/model': 0.6192636555060744, 'learning_rate/qwen_vl_interface': 9.940919530756347e-06, 'learning_rate/action_model': 9.940919530756347e-05, 'epoch': 0.42})
+[2026-07-28 16:15:26,216][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.041012875735759735, 'train/grad_norm_pre_clip': 1.294641375541687, 'timing/data': 0.00032776687294244766, 'timing/model': 0.8413426373153925, 'learning_rate/qwen_vl_interface': 9.940316399799032e-06, 'learning_rate/action_model': 9.940316399799033e-05, 'epoch': 0.42})
+[2026-07-28 16:15:35,623][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.02927933633327484, 'train/grad_norm_pre_clip': 1.0221132040023804, 'timing/data': 0.00047844555228948593, 'timing/model': 0.6265511969104409, 'learning_rate/qwen_vl_interface': 9.939710225346329e-06, 'learning_rate/action_model': 9.939710225346329e-05, 'epoch': 0.42})
+[2026-07-28 16:15:44,606][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.04569631069898605, 'train/grad_norm_pre_clip': 1.0294408798217773, 'timing/data': 0.00036919116973876953, 'timing/model': 0.623228813521564, 'learning_rate/qwen_vl_interface': 9.939101007791575e-06, 'learning_rate/action_model': 9.939101007791575e-05, 'epoch': 0.42})
+[2026-07-28 16:15:53,570][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.028222698718309402, 'train/grad_norm_pre_clip': 1.3660629987716675, 'timing/data': 0.0003324691206216812, 'timing/model': 0.6228475067764521, 'learning_rate/qwen_vl_interface': 9.938488747530087e-06, 'learning_rate/action_model': 9.938488747530086e-05, 'epoch': 0.42})
+[2026-07-28 16:16:02,780][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.0565204918384552, 'train/grad_norm_pre_clip': 2.568741798400879, 'timing/data': 0.00036634411662817, 'timing/model': 0.6238082870841026, 'learning_rate/qwen_vl_interface': 9.937873444959151e-06, 'learning_rate/action_model': 9.937873444959151e-05, 'epoch': 0.42})
+[2026-07-28 16:16:12,009][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.037928249686956406, 'train/grad_norm_pre_clip': 1.6442854404449463, 'timing/data': 0.0003343326970934868, 'timing/model': 0.874948101118207, 'learning_rate/qwen_vl_interface': 9.937255100478036e-06, 'learning_rate/action_model': 9.937255100478036e-05, 'epoch': 0.42})
+[2026-07-28 16:16:21,476][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.07933686673641205, 'train/grad_norm_pre_clip': 2.916017770767212, 'timing/data': 0.00036931969225406647, 'timing/model': 0.6282176850363612, 'learning_rate/qwen_vl_interface': 9.93663371448797e-06, 'learning_rate/action_model': 9.93663371448797e-05, 'epoch': 0.42})
+[2026-07-28 16:16:30,648][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.07246159017086029, 'train/grad_norm_pre_clip': 3.3733320236206055, 'timing/data': 0.0003545936197042465, 'timing/model': 0.622549582272768, 'learning_rate/qwen_vl_interface': 9.936009287392173e-06, 'learning_rate/action_model': 9.936009287392171e-05, 'epoch': 0.42})
+[2026-07-28 16:16:39,833][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.019531818106770515, 'train/grad_norm_pre_clip': 0.6580865979194641, 'timing/data': 0.00032824277877807617, 'timing/model': 0.6234583361074328, 'learning_rate/qwen_vl_interface': 9.935381819595822e-06, 'learning_rate/action_model': 9.935381819595822e-05, 'epoch': 0.42})
+[2026-07-28 16:16:49,008][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.03454261273145676, 'train/grad_norm_pre_clip': 1.6307909488677979, 'timing/data': 0.0004104161635041237, 'timing/model': 0.6242496706545353, 'learning_rate/qwen_vl_interface': 9.934751311506078e-06, 'learning_rate/action_model': 9.934751311506077e-05, 'epoch': 0.42})
+[2026-07-28 16:16:58,516][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.07816490530967712, 'train/grad_norm_pre_clip': 1.6736725568771362, 'timing/data': 0.00045010726898908615, 'timing/model': 0.8753588385879993, 'learning_rate/qwen_vl_interface': 9.93411776353207e-06, 'learning_rate/action_model': 9.934117763532069e-05, 'epoch': 0.42})
+[2026-07-28 16:17:07,638][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.018944822251796722, 'train/grad_norm_pre_clip': 0.4025486409664154, 'timing/data': 0.00040073320269584656, 'timing/model': 0.6208268739283085, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.42})
+[2026-07-28 16:17:16,779][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.023907825350761414, 'train/grad_norm_pre_clip': 1.1499515771865845, 'timing/data': 0.00032910145819187164, 'timing/model': 0.6190123111009598, 'learning_rate/qwen_vl_interface': 9.932841549577642e-06, 'learning_rate/action_model': 9.932841549577641e-05, 'epoch': 0.42})
+[2026-07-28 16:17:25,892][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.05056998133659363, 'train/grad_norm_pre_clip': 2.141705274581909, 'timing/data': 0.00044728536158800125, 'timing/model': 0.6182164875790477, 'learning_rate/qwen_vl_interface': 9.932198884425343e-06, 'learning_rate/action_model': 9.932198884425342e-05, 'epoch': 0.42})
+[2026-07-28 16:17:35,267][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.03526143357157707, 'train/grad_norm_pre_clip': 1.06675386428833, 'timing/data': 0.00039995182305574417, 'timing/model': 0.6183884665369987, 'learning_rate/qwen_vl_interface': 9.931553181045022e-06, 'learning_rate/action_model': 9.931553181045022e-05, 'epoch': 0.42})
+[2026-07-28 16:17:44,419][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.03876683488488197, 'train/grad_norm_pre_clip': 1.2250720262527466, 'timing/data': 0.00032834429293870926, 'timing/model': 0.6173746911808848, 'learning_rate/qwen_vl_interface': 9.930904439855667e-06, 'learning_rate/action_model': 9.930904439855667e-05, 'epoch': 0.42})
+[2026-07-28 16:17:53,776][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.0415005087852478, 'train/grad_norm_pre_clip': 1.6039272546768188, 'timing/data': 0.0003184238448739052, 'timing/model': 0.8537269765511155, 'learning_rate/qwen_vl_interface': 9.930252661278245e-06, 'learning_rate/action_model': 9.930252661278244e-05, 'epoch': 0.42})
+[2026-07-28 16:18:03,160][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.042790476232767105, 'train/grad_norm_pre_clip': 1.1639628410339355, 'timing/data': 0.0003927880898118019, 'timing/model': 0.8464930327609181, 'learning_rate/qwen_vl_interface': 9.92959784573568e-06, 'learning_rate/action_model': 9.92959784573568e-05, 'epoch': 0.43})
+[2026-07-28 16:18:12,303][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.052394259721040726, 'train/grad_norm_pre_clip': 1.4011082649230957, 'timing/data': 0.00042987801134586334, 'timing/model': 0.6233116630464792, 'learning_rate/qwen_vl_interface': 9.92893999365288e-06, 'learning_rate/action_model': 9.92893999365288e-05, 'epoch': 0.43})
+[2026-07-28 16:18:21,217][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.047638438642024994, 'train/grad_norm_pre_clip': 1.6579781770706177, 'timing/data': 0.00036149565130472183, 'timing/model': 0.6237539267167449, 'learning_rate/qwen_vl_interface': 9.928279105456717e-06, 'learning_rate/action_model': 9.928279105456718e-05, 'epoch': 0.43})
+[2026-07-28 16:18:30,397][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.02620416320860386, 'train/grad_norm_pre_clip': 1.2130622863769531, 'timing/data': 0.0003242790699005127, 'timing/model': 0.6183776538819075, 'learning_rate/qwen_vl_interface': 9.927615181576034e-06, 'learning_rate/action_model': 9.927615181576034e-05, 'epoch': 0.43})
+[2026-07-28 16:18:39,769][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.03309909999370575, 'train/grad_norm_pre_clip': 0.8366254568099976, 'timing/data': 0.00038684625178575516, 'timing/model': 0.6230386765673757, 'learning_rate/qwen_vl_interface': 9.926948222441644e-06, 'learning_rate/action_model': 9.926948222441643e-05, 'epoch': 0.43})
+[2026-07-28 16:18:49,219][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.013376476243138313, 'train/grad_norm_pre_clip': 1.2662651538848877, 'timing/data': 0.0006160344928503036, 'timing/model': 0.6288322014734149, 'learning_rate/qwen_vl_interface': 9.926278228486328e-06, 'learning_rate/action_model': 9.926278228486327e-05, 'epoch': 0.43})
+[2026-07-28 16:18:58,376][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.05073367431759834, 'train/grad_norm_pre_clip': 1.2775578498840332, 'timing/data': 0.0003561927005648613, 'timing/model': 0.6211178945377469, 'learning_rate/qwen_vl_interface': 9.92560520014484e-06, 'learning_rate/action_model': 9.92560520014484e-05, 'epoch': 0.43})
+[2026-07-28 16:19:07,519][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.04277646169066429, 'train/grad_norm_pre_clip': 1.0860819816589355, 'timing/data': 0.00034610647708177567, 'timing/model': 0.6217603357508779, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.43})
+[2026-07-28 16:19:16,917][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.054730549454689026, 'train/grad_norm_pre_clip': 1.2475608587265015, 'timing/data': 0.0003293035551905632, 'timing/model': 0.6228958340361714, 'learning_rate/qwen_vl_interface': 9.924250042052197e-06, 'learning_rate/action_model': 9.924250042052197e-05, 'epoch': 0.43})
+[2026-07-28 16:19:26,067][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.040324218571186066, 'train/grad_norm_pre_clip': 1.2096072435379028, 'timing/data': 0.0003279969096183777, 'timing/model': 0.6205056942999363, 'learning_rate/qwen_vl_interface': 9.92356791318039e-06, 'learning_rate/action_model': 9.92356791318039e-05, 'epoch': 0.43})
+[2026-07-28 16:19:35,438][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.04359094053506851, 'train/grad_norm_pre_clip': 1.194636344909668, 'timing/data': 0.00033149681985378265, 'timing/model': 0.8485786821693182, 'learning_rate/qwen_vl_interface': 9.922882751681104e-06, 'learning_rate/action_model': 9.922882751681104e-05, 'epoch': 0.43})
+[2026-07-28 16:19:44,620][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.026332378387451172, 'train/grad_norm_pre_clip': 0.9517666101455688, 'timing/data': 0.00035229790955781937, 'timing/model': 0.6403623828664422, 'learning_rate/qwen_vl_interface': 9.922194557998932e-06, 'learning_rate/action_model': 9.922194557998933e-05, 'epoch': 0.43})
+[2026-07-28 16:19:54,054][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.017828665673732758, 'train/grad_norm_pre_clip': 0.6682933568954468, 'timing/data': 0.0003747306764125824, 'timing/model': 0.6202875068411231, 'learning_rate/qwen_vl_interface': 9.921503332580439e-06, 'learning_rate/action_model': 9.921503332580439e-05, 'epoch': 0.43})
+[2026-07-28 16:20:03,272][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.024929456412792206, 'train/grad_norm_pre_clip': 0.9874531626701355, 'timing/data': 0.00043155811727046967, 'timing/model': 0.6213047038763762, 'learning_rate/qwen_vl_interface': 9.92080907587415e-06, 'learning_rate/action_model': 9.92080907587415e-05, 'epoch': 0.43})
+[2026-07-28 16:20:12,431][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.05174846202135086, 'train/grad_norm_pre_clip': 1.590179681777954, 'timing/data': 0.000527537427842617, 'timing/model': 0.625536628998816, 'learning_rate/qwen_vl_interface': 9.92011178833056e-06, 'learning_rate/action_model': 9.92011178833056e-05, 'epoch': 0.43})
+[2026-07-28 16:20:22,048][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.05815041437745094, 'train/grad_norm_pre_clip': 1.5396350622177124, 'timing/data': 0.00039253663271665573, 'timing/model': 0.8056799434125423, 'learning_rate/qwen_vl_interface': 9.919411470402136e-06, 'learning_rate/action_model': 9.919411470402136e-05, 'epoch': 0.43})
+[2026-07-28 16:20:31,283][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.031078418716788292, 'train/grad_norm_pre_clip': 1.2747502326965332, 'timing/data': 0.0003765597939491272, 'timing/model': 0.6206165216863155, 'learning_rate/qwen_vl_interface': 9.918708122543305e-06, 'learning_rate/action_model': 9.918708122543305e-05, 'epoch': 0.43})
+[2026-07-28 16:20:40,423][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.034097474068403244, 'train/grad_norm_pre_clip': 1.9616036415100098, 'timing/data': 0.0012176595628261566, 'timing/model': 0.6200499068945646, 'learning_rate/qwen_vl_interface': 9.918001745210458e-06, 'learning_rate/action_model': 9.918001745210456e-05, 'epoch': 0.43})
+[2026-07-28 16:20:49,843][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.09555894881486893, 'train/grad_norm_pre_clip': 2.9740986824035645, 'timing/data': 0.00039565470069646835, 'timing/model': 0.6237299423664808, 'learning_rate/qwen_vl_interface': 9.917292338861958e-06, 'learning_rate/action_model': 9.917292338861957e-05, 'epoch': 0.43})
+[2026-07-28 16:20:59,288][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.049699701368808746, 'train/grad_norm_pre_clip': 2.3117101192474365, 'timing/data': 0.000478353351354599, 'timing/model': 0.9249837566167116, 'learning_rate/qwen_vl_interface': 9.916579903958134e-06, 'learning_rate/action_model': 9.916579903958133e-05, 'epoch': 0.43})
+[2026-07-28 16:21:08,541][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.03251117840409279, 'train/grad_norm_pre_clip': 0.915778398513794, 'timing/data': 0.0004805503413081169, 'timing/model': 0.6211986113339663, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.43})
+[2026-07-28 16:21:17,903][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.04967561364173889, 'train/grad_norm_pre_clip': 1.3826817274093628, 'timing/data': 0.00038716942071914673, 'timing/model': 0.6208071727305651, 'learning_rate/qwen_vl_interface': 9.915145950335631e-06, 'learning_rate/action_model': 9.915145950335632e-05, 'epoch': 0.43})
+[2026-07-28 16:21:27,349][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.03320537135004997, 'train/grad_norm_pre_clip': 1.1128828525543213, 'timing/data': 0.0003857426345348358, 'timing/model': 0.6171385422348976, 'learning_rate/qwen_vl_interface': 9.914424432547434e-06, 'learning_rate/action_model': 9.914424432547433e-05, 'epoch': 0.43})
+[2026-07-28 16:21:36,747][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.0511334054172039, 'train/grad_norm_pre_clip': 1.015278697013855, 'timing/data': 0.00041866395622491837, 'timing/model': 0.622531758621335, 'learning_rate/qwen_vl_interface': 9.913699888064862e-06, 'learning_rate/action_model': 9.913699888064861e-05, 'epoch': 0.43})
+[2026-07-28 16:21:45,972][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.06022129952907562, 'train/grad_norm_pre_clip': 1.6384131908416748, 'timing/data': 0.0003804601728916168, 'timing/model': 0.6242645727470517, 'learning_rate/qwen_vl_interface': 9.912972317358066e-06, 'learning_rate/action_model': 9.912972317358066e-05, 'epoch': 0.43})
+[2026-07-28 16:21:55,253][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.04931846261024475, 'train/grad_norm_pre_clip': 1.2954528331756592, 'timing/data': 0.00033957790583372116, 'timing/model': 0.6219146568328142, 'learning_rate/qwen_vl_interface': 9.912241720899157e-06, 'learning_rate/action_model': 9.912241720899157e-05, 'epoch': 0.43})
+[2026-07-28 16:22:04,695][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.013097940012812614, 'train/grad_norm_pre_clip': 0.4256625175476074, 'timing/data': 0.0003613680601119995, 'timing/model': 0.6195204444229603, 'learning_rate/qwen_vl_interface': 9.911508099162215e-06, 'learning_rate/action_model': 9.911508099162214e-05, 'epoch': 0.43})
+[2026-07-28 16:22:14,144][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.033996984362602234, 'train/grad_norm_pre_clip': 1.0035032033920288, 'timing/data': 0.0003611389547586441, 'timing/model': 0.620609569363296, 'learning_rate/qwen_vl_interface': 9.910771452623276e-06, 'learning_rate/action_model': 9.910771452623275e-05, 'epoch': 0.43})
+[2026-07-28 16:22:23,298][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.08236677199602127, 'train/grad_norm_pre_clip': 2.665048122406006, 'timing/data': 0.00038511306047439575, 'timing/model': 0.6217375760897994, 'learning_rate/qwen_vl_interface': 9.910031781760342e-06, 'learning_rate/action_model': 9.910031781760342e-05, 'epoch': 0.43})
+[2026-07-28 16:22:32,498][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.025142956525087357, 'train/grad_norm_pre_clip': 0.757050096988678, 'timing/data': 0.0003699176013469696, 'timing/model': 0.6229003518819809, 'learning_rate/qwen_vl_interface': 9.909289087053378e-06, 'learning_rate/action_model': 9.909289087053378e-05, 'epoch': 0.43})
+[2026-07-28 16:22:42,069][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.03311894088983536, 'train/grad_norm_pre_clip': 1.289070725440979, 'timing/data': 0.0003434605896472931, 'timing/model': 0.8120353147387505, 'learning_rate/qwen_vl_interface': 9.90854336898431e-06, 'learning_rate/action_model': 9.90854336898431e-05, 'epoch': 0.43})
+[2026-07-28 16:22:51,379][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.031819961965084076, 'train/grad_norm_pre_clip': 0.6623920798301697, 'timing/data': 0.0003517046570777893, 'timing/model': 0.6209356812760234, 'learning_rate/qwen_vl_interface': 9.907794628037028e-06, 'learning_rate/action_model': 9.907794628037028e-05, 'epoch': 0.43})
+[2026-07-28 16:23:00,533][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.034385085105895996, 'train/grad_norm_pre_clip': 1.7395706176757812, 'timing/data': 0.0003375830128788948, 'timing/model': 0.6193715939298272, 'learning_rate/qwen_vl_interface': 9.90704286469738e-06, 'learning_rate/action_model': 9.90704286469738e-05, 'epoch': 0.43})
+[2026-07-28 16:23:09,698][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.048205241560935974, 'train/grad_norm_pre_clip': 1.2336242198944092, 'timing/data': 0.00038449373096227646, 'timing/model': 0.6206953674554825, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.43})
+[2026-07-28 16:23:19,127][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.05192987620830536, 'train/grad_norm_pre_clip': 2.3210361003875732, 'timing/data': 0.00034718774259090424, 'timing/model': 0.8605627007782459, 'learning_rate/qwen_vl_interface': 9.905530272794195e-06, 'learning_rate/action_model': 9.905530272794195e-05, 'epoch': 0.43})
+[2026-07-28 16:23:28,314][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.02891433797776699, 'train/grad_norm_pre_clip': 1.010502576828003, 'timing/data': 0.000372384674847126, 'timing/model': 0.6214571157470345, 'learning_rate/qwen_vl_interface': 9.904769445212163e-06, 'learning_rate/action_model': 9.904769445212162e-05, 'epoch': 0.43})
+[2026-07-28 16:23:37,848][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.020730648189783096, 'train/grad_norm_pre_clip': 0.895220935344696, 'timing/data': 0.00039570312947034836, 'timing/model': 0.6243306510150433, 'learning_rate/qwen_vl_interface': 9.904005597200772e-06, 'learning_rate/action_model': 9.904005597200772e-05, 'epoch': 0.43})
+[2026-07-28 16:23:47,050][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.01701139472424984, 'train/grad_norm_pre_clip': 1.1853537559509277, 'timing/data': 0.00040489528328180313, 'timing/model': 0.6211638692766428, 'learning_rate/qwen_vl_interface': 9.903238729255678e-06, 'learning_rate/action_model': 9.903238729255677e-05, 'epoch': 0.43})
+[2026-07-28 16:23:56,257][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.040460653603076935, 'train/grad_norm_pre_clip': 0.8612768650054932, 'timing/data': 0.00033103302121162415, 'timing/model': 0.6376809226348996, 'learning_rate/qwen_vl_interface': 9.902468841874492e-06, 'learning_rate/action_model': 9.902468841874492e-05, 'epoch': 0.43})
+[2026-07-28 16:24:05,670][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.06098207086324692, 'train/grad_norm_pre_clip': 1.399384617805481, 'timing/data': 0.00046675093472003937, 'timing/model': 0.8565705791115761, 'learning_rate/qwen_vl_interface': 9.901695935556784e-06, 'learning_rate/action_model': 9.901695935556783e-05, 'epoch': 0.43})
+[2026-07-28 16:24:15,147][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.04764058440923691, 'train/grad_norm_pre_clip': 1.3516079187393188, 'timing/data': 0.0003685289993882179, 'timing/model': 0.8411074290052056, 'learning_rate/qwen_vl_interface': 9.900920010804088e-06, 'learning_rate/action_model': 9.900920010804088e-05, 'epoch': 0.43})
+[2026-07-28 16:24:24,079][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.05277624726295471, 'train/grad_norm_pre_clip': 2.207463264465332, 'timing/data': 0.0005736518651247025, 'timing/model': 0.621682807803154, 'learning_rate/qwen_vl_interface': 9.900141068119891e-06, 'learning_rate/action_model': 9.900141068119891e-05, 'epoch': 0.43})
+[2026-07-28 16:24:33,280][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.014389649964869022, 'train/grad_norm_pre_clip': 0.5969755053520203, 'timing/data': 0.00035236310213804245, 'timing/model': 0.6265207594260573, 'learning_rate/qwen_vl_interface': 9.89935910800964e-06, 'learning_rate/action_model': 9.89935910800964e-05, 'epoch': 0.43})
+[2026-07-28 16:24:42,523][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.048038557171821594, 'train/grad_norm_pre_clip': 2.2967066764831543, 'timing/data': 0.00035516172647476196, 'timing/model': 0.6230439757928252, 'learning_rate/qwen_vl_interface': 9.898574130980743e-06, 'learning_rate/action_model': 9.898574130980742e-05, 'epoch': 0.43})
+[2026-07-28 16:24:51,914][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.03695524111390114, 'train/grad_norm_pre_clip': 2.233917474746704, 'timing/data': 0.0003939177840948105, 'timing/model': 0.6201481791213155, 'learning_rate/qwen_vl_interface': 9.89778613754256e-06, 'learning_rate/action_model': 9.89778613754256e-05, 'epoch': 0.43})
+[2026-07-28 16:25:01,041][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.022668717429041862, 'train/grad_norm_pre_clip': 1.2231473922729492, 'timing/data': 0.00032722577452659607, 'timing/model': 0.6193994274362922, 'learning_rate/qwen_vl_interface': 9.896995128206412e-06, 'learning_rate/action_model': 9.896995128206411e-05, 'epoch': 0.43})
+[2026-07-28 16:25:10,434][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.0723872035741806, 'train/grad_norm_pre_clip': 2.2928271293640137, 'timing/data': 0.000354863703250885, 'timing/model': 0.6204536240547895, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.43})
+[2026-07-28 16:25:19,967][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.041487060487270355, 'train/grad_norm_pre_clip': 1.960448980331421, 'timing/data': 0.00035383179783821106, 'timing/model': 0.8978117248043418, 'learning_rate/qwen_vl_interface': 9.89540406389529e-06, 'learning_rate/action_model': 9.89540406389529e-05, 'epoch': 0.43})
+[2026-07-28 16:25:29,180][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.05613182857632637, 'train/grad_norm_pre_clip': 1.3008465766906738, 'timing/data': 0.0003247847780585289, 'timing/model': 0.6275726826861501, 'learning_rate/qwen_vl_interface': 9.894604009952742e-06, 'learning_rate/action_model': 9.89460400995274e-05, 'epoch': 0.43})
+[2026-07-28 16:25:38,272][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.013822091743350029, 'train/grad_norm_pre_clip': 0.5413633584976196, 'timing/data': 0.00041611213237047195, 'timing/model': 0.6195777002722025, 'learning_rate/qwen_vl_interface': 9.893800942177073e-06, 'learning_rate/action_model': 9.893800942177073e-05, 'epoch': 0.43})
+[2026-07-28 16:25:47,424][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.05119750648736954, 'train/grad_norm_pre_clip': 2.5300345420837402, 'timing/data': 0.00035018566995859146, 'timing/model': 0.6207767454907298, 'learning_rate/qwen_vl_interface': 9.892994861089393e-06, 'learning_rate/action_model': 9.892994861089393e-05, 'epoch': 0.43})
+[2026-07-28 16:25:56,919][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.05985453352332115, 'train/grad_norm_pre_clip': 2.1278083324432373, 'timing/data': 0.00033992156386375427, 'timing/model': 0.620803595520556, 'learning_rate/qwen_vl_interface': 9.892185767212756e-06, 'learning_rate/action_model': 9.892185767212756e-05, 'epoch': 0.43})
+[2026-07-28 16:26:06,289][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.02348129265010357, 'train/grad_norm_pre_clip': 0.9185712337493896, 'timing/data': 0.0003364812582731247, 'timing/model': 0.6206658072769642, 'learning_rate/qwen_vl_interface': 9.891373661072172e-06, 'learning_rate/action_model': 9.891373661072173e-05, 'epoch': 0.43})
+[2026-07-28 16:26:15,571][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.07830393314361572, 'train/grad_norm_pre_clip': 5.349674224853516, 'timing/data': 0.0004545971751213074, 'timing/model': 0.6217765090987086, 'learning_rate/qwen_vl_interface': 9.89055854319461e-06, 'learning_rate/action_model': 9.890558543194611e-05, 'epoch': 0.43})
+[2026-07-28 16:26:24,721][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.05209089815616608, 'train/grad_norm_pre_clip': 0.9751145243644714, 'timing/data': 0.00037303101271390915, 'timing/model': 0.6206615287810564, 'learning_rate/qwen_vl_interface': 9.889740414108994e-06, 'learning_rate/action_model': 9.889740414108995e-05, 'epoch': 0.43})
+[2026-07-28 16:26:34,141][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.04465119168162346, 'train/grad_norm_pre_clip': 1.2500951290130615, 'timing/data': 0.0003520427271723747, 'timing/model': 0.6215276587754488, 'learning_rate/qwen_vl_interface': 9.888919274346196e-06, 'learning_rate/action_model': 9.888919274346195e-05, 'epoch': 0.44})
+[2026-07-28 16:26:43,298][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.06975482404232025, 'train/grad_norm_pre_clip': 2.1350009441375732, 'timing/data': 0.00032774172723293304, 'timing/model': 0.620024430565536, 'learning_rate/qwen_vl_interface': 9.888095124439043e-06, 'learning_rate/action_model': 9.888095124439043e-05, 'epoch': 0.44})
+[2026-07-28 16:26:52,720][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.04439391940832138, 'train/grad_norm_pre_clip': 2.087684154510498, 'timing/data': 0.000342334620654583, 'timing/model': 0.8409881498664618, 'learning_rate/qwen_vl_interface': 9.88726796492232e-06, 'learning_rate/action_model': 9.88726796492232e-05, 'epoch': 0.44})
+[2026-07-28 16:27:01,879][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.03225592151284218, 'train/grad_norm_pre_clip': 1.950365424156189, 'timing/data': 0.0003420533612370491, 'timing/model': 0.6186507539823651, 'learning_rate/qwen_vl_interface': 9.886437796332759e-06, 'learning_rate/action_model': 9.88643779633276e-05, 'epoch': 0.44})
+[2026-07-28 16:27:11,316][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.049069665372371674, 'train/grad_norm_pre_clip': 1.1328136920928955, 'timing/data': 0.0004539741203188896, 'timing/model': 0.918174528516829, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.44})
+[2026-07-28 16:27:20,457][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.042334821075201035, 'train/grad_norm_pre_clip': 1.1741549968719482, 'timing/data': 0.00034165382385253906, 'timing/model': 0.6206963500007987, 'learning_rate/qwen_vl_interface': 9.884768434091832e-06, 'learning_rate/action_model': 9.884768434091832e-05, 'epoch': 0.44})
+[2026-07-28 16:27:29,610][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.01564827933907509, 'train/grad_norm_pre_clip': 0.48830872774124146, 'timing/data': 0.0003261258825659752, 'timing/model': 0.6262096418067813, 'learning_rate/qwen_vl_interface': 9.883929241523698e-06, 'learning_rate/action_model': 9.883929241523697e-05, 'epoch': 0.44})
+[2026-07-28 16:27:39,066][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.05051552504301071, 'train/grad_norm_pre_clip': 1.0463844537734985, 'timing/data': 0.00034215766936540604, 'timing/model': 0.625275076366961, 'learning_rate/qwen_vl_interface': 9.883087042049185e-06, 'learning_rate/action_model': 9.883087042049186e-05, 'epoch': 0.44})
+[2026-07-28 16:27:48,588][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.054681409150362015, 'train/grad_norm_pre_clip': 1.057440996170044, 'timing/data': 0.0004594894126057625, 'timing/model': 0.623571733944118, 'learning_rate/qwen_vl_interface': 9.882241836214794e-06, 'learning_rate/action_model': 9.882241836214794e-05, 'epoch': 0.44})
+[2026-07-28 16:27:57,845][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.04364785924553871, 'train/grad_norm_pre_clip': 1.2015773057937622, 'timing/data': 0.00033190101385116577, 'timing/model': 0.6224241107702255, 'learning_rate/qwen_vl_interface': 9.881393624568965e-06, 'learning_rate/action_model': 9.881393624568965e-05, 'epoch': 0.44})
+[2026-07-28 16:28:07,359][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.05279325693845749, 'train/grad_norm_pre_clip': 2.8152072429656982, 'timing/data': 0.00045389868319034576, 'timing/model': 0.6828166954219341, 'learning_rate/qwen_vl_interface': 9.880542407662098e-06, 'learning_rate/action_model': 9.880542407662097e-05, 'epoch': 0.44})
+[2026-07-28 16:28:16,910][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.04413723945617676, 'train/grad_norm_pre_clip': 1.938817024230957, 'timing/data': 0.00037553906440734863, 'timing/model': 0.6777632338926196, 'learning_rate/qwen_vl_interface': 9.879688186046535e-06, 'learning_rate/action_model': 9.879688186046534e-05, 'epoch': 0.44})
+[2026-07-28 16:28:26,510][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.05302806943655014, 'train/grad_norm_pre_clip': 1.398870587348938, 'timing/data': 0.0004191221669316292, 'timing/model': 0.6210140231996775, 'learning_rate/qwen_vl_interface': 9.87883096027657e-06, 'learning_rate/action_model': 9.87883096027657e-05, 'epoch': 0.44})
+[2026-07-28 16:28:35,720][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.06678614020347595, 'train/grad_norm_pre_clip': 2.098121166229248, 'timing/data': 0.0005384143441915512, 'timing/model': 0.6204419210553169, 'learning_rate/qwen_vl_interface': 9.87797073090845e-06, 'learning_rate/action_model': 9.87797073090845e-05, 'epoch': 0.44})
+[2026-07-28 16:28:45,213][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.04423532634973526, 'train/grad_norm_pre_clip': 0.9576132893562317, 'timing/data': 0.00040740519762039185, 'timing/model': 0.624655825085938, 'learning_rate/qwen_vl_interface': 9.877107498500368e-06, 'learning_rate/action_model': 9.877107498500367e-05, 'epoch': 0.44})
+[2026-07-28 16:28:54,612][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.04363775998353958, 'train/grad_norm_pre_clip': 1.3038177490234375, 'timing/data': 0.0003505069762468338, 'timing/model': 0.8010326279327273, 'learning_rate/qwen_vl_interface': 9.876241263612467e-06, 'learning_rate/action_model': 9.876241263612465e-05, 'epoch': 0.44})
+[2026-07-28 16:29:03,826][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.05046503245830536, 'train/grad_norm_pre_clip': 1.503172516822815, 'timing/data': 0.0004166187718510628, 'timing/model': 0.6214476432651281, 'learning_rate/qwen_vl_interface': 9.875372026806835e-06, 'learning_rate/action_model': 9.875372026806835e-05, 'epoch': 0.44})
+[2026-07-28 16:29:13,043][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.025837896391749382, 'train/grad_norm_pre_clip': 0.7839919328689575, 'timing/data': 0.000359545461833477, 'timing/model': 0.6202890714630485, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.44})
+[2026-07-28 16:29:22,434][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.058765485882759094, 'train/grad_norm_pre_clip': 1.725678563117981, 'timing/data': 0.00035830680280923843, 'timing/model': 0.6197505677118897, 'learning_rate/qwen_vl_interface': 9.873624549700485e-06, 'learning_rate/action_model': 9.873624549700484e-05, 'epoch': 0.44})
+[2026-07-28 16:29:31,966][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.08988586068153381, 'train/grad_norm_pre_clip': 3.291714906692505, 'timing/data': 0.00044195540249347687, 'timing/model': 0.9267873261123896, 'learning_rate/qwen_vl_interface': 9.872746310533683e-06, 'learning_rate/action_model': 9.872746310533684e-05, 'epoch': 0.44})
+[2026-07-28 16:29:41,100][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.038069963455200195, 'train/grad_norm_pre_clip': 1.5518609285354614, 'timing/data': 0.00037117861211299896, 'timing/model': 0.6200141534209251, 'learning_rate/qwen_vl_interface': 9.871865071716991e-06, 'learning_rate/action_model': 9.87186507171699e-05, 'epoch': 0.44})
+[2026-07-28 16:29:50,267][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.06801506876945496, 'train/grad_norm_pre_clip': 1.8746305704116821, 'timing/data': 0.00036103278398513794, 'timing/model': 0.6221531759947538, 'learning_rate/qwen_vl_interface': 9.870980833822231e-06, 'learning_rate/action_model': 9.87098083382223e-05, 'epoch': 0.44})
+[2026-07-28 16:29:59,731][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.03390824794769287, 'train/grad_norm_pre_clip': 1.004377007484436, 'timing/data': 0.0003872225061058998, 'timing/model': 0.8439945532009006, 'learning_rate/qwen_vl_interface': 9.87009359742318e-06, 'learning_rate/action_model': 9.87009359742318e-05, 'epoch': 0.44})
+[2026-07-28 16:30:09,130][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.05142028629779816, 'train/grad_norm_pre_clip': 1.1433414220809937, 'timing/data': 0.00044313445687294006, 'timing/model': 0.6306860391050577, 'learning_rate/qwen_vl_interface': 9.86920336309555e-06, 'learning_rate/action_model': 9.86920336309555e-05, 'epoch': 0.44})
+[2026-07-28 16:30:18,397][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.044195353984832764, 'train/grad_norm_pre_clip': 0.884253740310669, 'timing/data': 0.00034814607352018356, 'timing/model': 0.6216190457344055, 'learning_rate/qwen_vl_interface': 9.86831013141701e-06, 'learning_rate/action_model': 9.86831013141701e-05, 'epoch': 0.44})
+[2026-07-28 16:30:27,555][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.033768296241760254, 'train/grad_norm_pre_clip': 1.4757746458053589, 'timing/data': 0.000433540903031826, 'timing/model': 0.6200182856991887, 'learning_rate/qwen_vl_interface': 9.867413902967166e-06, 'learning_rate/action_model': 9.867413902967166e-05, 'epoch': 0.44})
+[2026-07-28 16:30:36,744][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.044162385165691376, 'train/grad_norm_pre_clip': 1.2174949645996094, 'timing/data': 0.0003728894516825676, 'timing/model': 0.6254851808771491, 'learning_rate/qwen_vl_interface': 9.86651467832757e-06, 'learning_rate/action_model': 9.86651467832757e-05, 'epoch': 0.44})
+[2026-07-28 16:30:46,215][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.05063248053193092, 'train/grad_norm_pre_clip': 3.0292141437530518, 'timing/data': 0.0003292914479970932, 'timing/model': 0.6236854121088982, 'learning_rate/qwen_vl_interface': 9.865612458081718e-06, 'learning_rate/action_model': 9.865612458081719e-05, 'epoch': 0.44})
+[2026-07-28 16:30:55,668][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.0749654769897461, 'train/grad_norm_pre_clip': 3.8615593910217285, 'timing/data': 0.0004448341205716133, 'timing/model': 0.6354045514017344, 'learning_rate/qwen_vl_interface': 9.864707242815057e-06, 'learning_rate/action_model': 9.864707242815056e-05, 'epoch': 0.44})
+[2026-07-28 16:31:04,881][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.04539242759346962, 'train/grad_norm_pre_clip': 2.6443605422973633, 'timing/data': 0.00039550941437482834, 'timing/model': 0.6272548418492079, 'learning_rate/qwen_vl_interface': 9.863799033114966e-06, 'learning_rate/action_model': 9.863799033114966e-05, 'epoch': 0.44})
+[2026-07-28 16:31:14,377][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.030401133000850677, 'train/grad_norm_pre_clip': 2.1572816371917725, 'timing/data': 0.0003510620445013046, 'timing/model': 0.6234099557623267, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.44})
+[2026-07-28 16:31:23,834][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.05425301566720009, 'train/grad_norm_pre_clip': 1.2827643156051636, 'timing/data': 0.00035741087049245834, 'timing/model': 0.6215232769027352, 'learning_rate/qwen_vl_interface': 9.861973632773748e-06, 'learning_rate/action_model': 9.861973632773748e-05, 'epoch': 0.44})
+[2026-07-28 16:31:33,066][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.04010322317481041, 'train/grad_norm_pre_clip': 1.2312039136886597, 'timing/data': 0.00040942244231700897, 'timing/model': 0.6210668841376901, 'learning_rate/qwen_vl_interface': 9.861056443317105e-06, 'learning_rate/action_model': 9.861056443317105e-05, 'epoch': 0.44})
+[2026-07-28 16:31:42,294][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.0627739429473877, 'train/grad_norm_pre_clip': 1.6974682807922363, 'timing/data': 0.00036529917269945145, 'timing/model': 0.6200567949563265, 'learning_rate/qwen_vl_interface': 9.860136261796e-06, 'learning_rate/action_model': 9.860136261795998e-05, 'epoch': 0.44})
+[2026-07-28 16:31:51,952][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.06073963642120361, 'train/grad_norm_pre_clip': 1.985013723373413, 'timing/data': 0.0005684010684490204, 'timing/model': 1.0615004347637296, 'learning_rate/qwen_vl_interface': 9.859213088807521e-06, 'learning_rate/action_model': 9.85921308880752e-05, 'epoch': 0.44})
+[2026-07-28 16:32:01,104][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.04394806548953056, 'train/grad_norm_pre_clip': 1.7495949268341064, 'timing/data': 0.00045538879930973053, 'timing/model': 0.6225778609514236, 'learning_rate/qwen_vl_interface': 9.858286924950714e-06, 'learning_rate/action_model': 9.858286924950713e-05, 'epoch': 0.44})
+[2026-07-28 16:32:10,062][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.024994216859340668, 'train/grad_norm_pre_clip': 0.8363289833068848, 'timing/data': 0.00031714513897895813, 'timing/model': 0.6189002878963947, 'learning_rate/qwen_vl_interface': 9.85735777082655e-06, 'learning_rate/action_model': 9.857357770826551e-05, 'epoch': 0.44})
+[2026-07-28 16:32:19,194][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.04080728814005852, 'train/grad_norm_pre_clip': 1.8185031414031982, 'timing/data': 0.0004905220121145248, 'timing/model': 0.6207882864400744, 'learning_rate/qwen_vl_interface': 9.856425627037951e-06, 'learning_rate/action_model': 9.856425627037952e-05, 'epoch': 0.44})
+[2026-07-28 16:32:28,614][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.035274606198072433, 'train/grad_norm_pre_clip': 1.2174805402755737, 'timing/data': 0.0003598444163799286, 'timing/model': 0.6234987052157521, 'learning_rate/qwen_vl_interface': 9.855490494189777e-06, 'learning_rate/action_model': 9.855490494189776e-05, 'epoch': 0.44})
+[2026-07-28 16:32:38,019][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.035100847482681274, 'train/grad_norm_pre_clip': 0.9576188921928406, 'timing/data': 0.00034298840910196304, 'timing/model': 0.8680936563760042, 'learning_rate/qwen_vl_interface': 9.854552372888819e-06, 'learning_rate/action_model': 9.854552372888819e-05, 'epoch': 0.44})
+[2026-07-28 16:32:47,154][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.06610353291034698, 'train/grad_norm_pre_clip': 1.6680516004562378, 'timing/data': 0.0003366246819496155, 'timing/model': 0.6201208718121052, 'learning_rate/qwen_vl_interface': 9.853611263743817e-06, 'learning_rate/action_model': 9.853611263743817e-05, 'epoch': 0.44})
+[2026-07-28 16:32:56,591][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.037427470088005066, 'train/grad_norm_pre_clip': 1.2493231296539307, 'timing/data': 0.0003347620368003845, 'timing/model': 0.8601335836574435, 'learning_rate/qwen_vl_interface': 9.85266716736545e-06, 'learning_rate/action_model': 9.852667167365449e-05, 'epoch': 0.44})
+[2026-07-28 16:33:05,993][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.10380218923091888, 'train/grad_norm_pre_clip': 5.80035924911499, 'timing/data': 0.0004909876734018326, 'timing/model': 0.6232793992385268, 'learning_rate/qwen_vl_interface': 9.851720084366324e-06, 'learning_rate/action_model': 9.851720084366324e-05, 'epoch': 0.44})
+[2026-07-28 16:33:14,997][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.08790460228919983, 'train/grad_norm_pre_clip': 3.168816566467285, 'timing/data': 0.0003648446872830391, 'timing/model': 0.6213658228516579, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.44})
+[2026-07-28 16:33:23,999][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.14707422256469727, 'train/grad_norm_pre_clip': 4.818798065185547, 'timing/data': 0.0003330027684569359, 'timing/model': 0.6195470131933689, 'learning_rate/qwen_vl_interface': 9.849816960965959e-06, 'learning_rate/action_model': 9.849816960965959e-05, 'epoch': 0.44})
+[2026-07-28 16:33:32,917][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.07216203957796097, 'train/grad_norm_pre_clip': 4.516938209533691, 'timing/data': 0.0003341306000947952, 'timing/model': 0.6199960177764297, 'learning_rate/qwen_vl_interface': 9.848860921799633e-06, 'learning_rate/action_model': 9.848860921799632e-05, 'epoch': 0.44})
+[2026-07-28 16:33:42,077][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.05950639024376869, 'train/grad_norm_pre_clip': 2.3098397254943848, 'timing/data': 0.00036216992884874344, 'timing/model': 0.6208513546735048, 'learning_rate/qwen_vl_interface': 9.847901898482385e-06, 'learning_rate/action_model': 9.847901898482385e-05, 'epoch': 0.44})
+[2026-07-28 16:33:51,037][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.09220534563064575, 'train/grad_norm_pre_clip': 3.3347136974334717, 'timing/data': 0.00035841204226017, 'timing/model': 0.6208248743787408, 'learning_rate/qwen_vl_interface': 9.846939891636512e-06, 'learning_rate/action_model': 9.846939891636512e-05, 'epoch': 0.44})
+[2026-07-28 16:34:00,172][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.0588114857673645, 'train/grad_norm_pre_clip': 1.4036847352981567, 'timing/data': 0.00037498120218515396, 'timing/model': 0.6228410946205258, 'learning_rate/qwen_vl_interface': 9.845974901886253e-06, 'learning_rate/action_model': 9.845974901886252e-05, 'epoch': 0.44})
+[2026-07-28 16:34:09,319][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.11000365763902664, 'train/grad_norm_pre_clip': 2.538088083267212, 'timing/data': 0.00035390257835388184, 'timing/model': 0.6215661931782961, 'learning_rate/qwen_vl_interface': 9.845006929857777e-06, 'learning_rate/action_model': 9.845006929857776e-05, 'epoch': 0.44})
+[2026-07-28 16:34:18,493][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.041137974709272385, 'train/grad_norm_pre_clip': 0.7825319766998291, 'timing/data': 0.0003409218043088913, 'timing/model': 0.6206308398395777, 'learning_rate/qwen_vl_interface': 9.844035976179193e-06, 'learning_rate/action_model': 9.844035976179192e-05, 'epoch': 0.44})
+[2026-07-28 16:34:27,936][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.03593618422746658, 'train/grad_norm_pre_clip': 0.7411158084869385, 'timing/data': 0.0003543207421898842, 'timing/model': 0.866175421513617, 'learning_rate/qwen_vl_interface': 9.843062041480538e-06, 'learning_rate/action_model': 9.843062041480539e-05, 'epoch': 0.44})
+[2026-07-28 16:34:37,402][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.04850900173187256, 'train/grad_norm_pre_clip': 1.3961766958236694, 'timing/data': 0.0003334786742925644, 'timing/model': 0.6215040264651179, 'learning_rate/qwen_vl_interface': 9.842085126393793e-06, 'learning_rate/action_model': 9.842085126393792e-05, 'epoch': 0.44})
+[2026-07-28 16:34:46,561][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.0405740812420845, 'train/grad_norm_pre_clip': 1.977672815322876, 'timing/data': 0.0003237975761294365, 'timing/model': 0.6201312094926834, 'learning_rate/qwen_vl_interface': 9.841105231552863e-06, 'learning_rate/action_model': 9.841105231552864e-05, 'epoch': 0.44})
+[2026-07-28 16:34:55,714][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.07504837960004807, 'train/grad_norm_pre_clip': 2.9171125888824463, 'timing/data': 0.0003889854997396469, 'timing/model': 0.6275396095588803, 'learning_rate/qwen_vl_interface': 9.840122357593594e-06, 'learning_rate/action_model': 9.840122357593594e-05, 'epoch': 0.45})
+[2026-07-28 16:35:04,847][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.052155718207359314, 'train/grad_norm_pre_clip': 2.5272138118743896, 'timing/data': 0.000350254587829113, 'timing/model': 0.6194590544328094, 'learning_rate/qwen_vl_interface': 9.839136505153762e-06, 'learning_rate/action_model': 9.839136505153762e-05, 'epoch': 0.45})
+[2026-07-28 16:35:14,232][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.03536827862262726, 'train/grad_norm_pre_clip': 1.6117655038833618, 'timing/data': 0.00042384304106235504, 'timing/model': 0.620683940127492, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.45})
+[2026-07-28 16:35:23,699][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.06083597242832184, 'train/grad_norm_pre_clip': 2.3981990814208984, 'timing/data': 0.00034396257251501083, 'timing/model': 0.6243203142657876, 'learning_rate/qwen_vl_interface': 9.837155867393176e-06, 'learning_rate/action_model': 9.837155867393175e-05, 'epoch': 0.45})
+[2026-07-28 16:35:33,012][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.07185912877321243, 'train/grad_norm_pre_clip': 3.340131998062134, 'timing/data': 0.0003343392163515091, 'timing/model': 0.620130573399365, 'learning_rate/qwen_vl_interface': 9.836161083357635e-06, 'learning_rate/action_model': 9.836161083357635e-05, 'epoch': 0.45})
+[2026-07-28 16:35:42,500][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.05907541513442993, 'train/grad_norm_pre_clip': 1.348834753036499, 'timing/data': 0.0003380095586180687, 'timing/model': 0.6188637148588896, 'learning_rate/qwen_vl_interface': 9.835163323411957e-06, 'learning_rate/action_model': 9.835163323411958e-05, 'epoch': 0.45})
+[2026-07-28 16:35:51,735][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.06171417236328125, 'train/grad_norm_pre_clip': 1.2238554954528809, 'timing/data': 0.00035280268639326096, 'timing/model': 0.6219604155048728, 'learning_rate/qwen_vl_interface': 9.834162588203581e-06, 'learning_rate/action_model': 9.83416258820358e-05, 'epoch': 0.45})
+[2026-07-28 16:36:01,227][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.03935721516609192, 'train/grad_norm_pre_clip': 1.1668466329574585, 'timing/data': 0.00033472664654254913, 'timing/model': 0.6206482015550137, 'learning_rate/qwen_vl_interface': 9.833158878381872e-06, 'learning_rate/action_model': 9.83315887838187e-05, 'epoch': 0.45})
+[2026-07-28 16:36:10,443][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.0356413908302784, 'train/grad_norm_pre_clip': 1.9036921262741089, 'timing/data': 0.00035195983946323395, 'timing/model': 0.6193248117342591, 'learning_rate/qwen_vl_interface': 9.832152194598123e-06, 'learning_rate/action_model': 9.832152194598122e-05, 'epoch': 0.45})
+[2026-07-28 16:36:19,832][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.03736996650695801, 'train/grad_norm_pre_clip': 0.8977457880973816, 'timing/data': 0.00033668894320726395, 'timing/model': 0.6211163848638535, 'learning_rate/qwen_vl_interface': 9.831142537505563e-06, 'learning_rate/action_model': 9.831142537505562e-05, 'epoch': 0.45})
+[2026-07-28 16:36:29,210][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.03436555713415146, 'train/grad_norm_pre_clip': 1.4664396047592163, 'timing/data': 0.0004424275830388069, 'timing/model': 0.8549627298489213, 'learning_rate/qwen_vl_interface': 9.830129907759347e-06, 'learning_rate/action_model': 9.830129907759347e-05, 'epoch': 0.45})
+[2026-07-28 16:36:38,387][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.021331578493118286, 'train/grad_norm_pre_clip': 0.9102036356925964, 'timing/data': 0.0003382321447134018, 'timing/model': 0.62019132822752, 'learning_rate/qwen_vl_interface': 9.82911430601656e-06, 'learning_rate/action_model': 9.82911430601656e-05, 'epoch': 0.45})
+[2026-07-28 16:36:47,554][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.04209467023611069, 'train/grad_norm_pre_clip': 1.381377935409546, 'timing/data': 0.0003322446718811989, 'timing/model': 0.6256932886317372, 'learning_rate/qwen_vl_interface': 9.828095732936216e-06, 'learning_rate/action_model': 9.828095732936216e-05, 'epoch': 0.45})
+[2026-07-28 16:36:56,991][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.026661397889256477, 'train/grad_norm_pre_clip': 1.0654549598693848, 'timing/data': 0.0008197585120797157, 'timing/model': 0.6213867217302322, 'learning_rate/qwen_vl_interface': 9.827074189179254e-06, 'learning_rate/action_model': 9.827074189179253e-05, 'epoch': 0.45})
+[2026-07-28 16:37:06,571][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.09921811521053314, 'train/grad_norm_pre_clip': 1.9474947452545166, 'timing/data': 0.0003455318510532379, 'timing/model': 0.8826611060649157, 'learning_rate/qwen_vl_interface': 9.82604967540854e-06, 'learning_rate/action_model': 9.82604967540854e-05, 'epoch': 0.45})
+[2026-07-28 16:37:15,869][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.0794476866722107, 'train/grad_norm_pre_clip': 1.681171178817749, 'timing/data': 0.000338967889547348, 'timing/model': 0.6639410350471735, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.45})
+[2026-07-28 16:37:25,080][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.03326074779033661, 'train/grad_norm_pre_clip': 1.4644688367843628, 'timing/data': 0.00042446795850992203, 'timing/model': 0.6232917672023177, 'learning_rate/qwen_vl_interface': 9.823991740486982e-06, 'learning_rate/action_model': 9.823991740486982e-05, 'epoch': 0.45})
+[2026-07-28 16:37:34,538][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.029539596289396286, 'train/grad_norm_pre_clip': 1.2572197914123535, 'timing/data': 0.0004405267536640167, 'timing/model': 0.8352977074682713, 'learning_rate/qwen_vl_interface': 9.822958320671506e-06, 'learning_rate/action_model': 9.822958320671506e-05, 'epoch': 0.45})
+[2026-07-28 16:37:43,739][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.07123884558677673, 'train/grad_norm_pre_clip': 3.6966960430145264, 'timing/data': 0.00032651424407958984, 'timing/model': 0.6216795975342393, 'learning_rate/qwen_vl_interface': 9.821921933513023e-06, 'learning_rate/action_model': 9.821921933513023e-05, 'epoch': 0.45})
+[2026-07-28 16:37:53,205][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.03539121523499489, 'train/grad_norm_pre_clip': 2.0071122646331787, 'timing/data': 0.0003375690430402756, 'timing/model': 0.8705030167475343, 'learning_rate/qwen_vl_interface': 9.820882579684037e-06, 'learning_rate/action_model': 9.820882579684037e-05, 'epoch': 0.45})
+[2026-07-28 16:38:02,412][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.02535577490925789, 'train/grad_norm_pre_clip': 1.4403526782989502, 'timing/data': 0.0004196902737021446, 'timing/model': 0.62307661306113, 'learning_rate/qwen_vl_interface': 9.819840259858968e-06, 'learning_rate/action_model': 9.819840259858968e-05, 'epoch': 0.45})
+[2026-07-28 16:38:11,681][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.041495431214571, 'train/grad_norm_pre_clip': 1.4506374597549438, 'timing/data': 0.0003312360495328903, 'timing/model': 0.6235961923375726, 'learning_rate/qwen_vl_interface': 9.818794974714167e-06, 'learning_rate/action_model': 9.818794974714168e-05, 'epoch': 0.45})
+[2026-07-28 16:38:20,871][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.03422364220023155, 'train/grad_norm_pre_clip': 1.3594918251037598, 'timing/data': 0.0003638220950961113, 'timing/model': 0.621036090888083, 'learning_rate/qwen_vl_interface': 9.817746724927912e-06, 'learning_rate/action_model': 9.817746724927912e-05, 'epoch': 0.45})
+[2026-07-28 16:38:30,299][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.07987961173057556, 'train/grad_norm_pre_clip': 2.593294858932495, 'timing/data': 0.000333285890519619, 'timing/model': 0.6336498381569982, 'learning_rate/qwen_vl_interface': 9.8166955111804e-06, 'learning_rate/action_model': 9.816695511180399e-05, 'epoch': 0.45})
+[2026-07-28 16:38:39,521][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.045169662684202194, 'train/grad_norm_pre_clip': 1.9212595224380493, 'timing/data': 0.00033133942633867264, 'timing/model': 0.6245820531621575, 'learning_rate/qwen_vl_interface': 9.815641334153746e-06, 'learning_rate/action_model': 9.815641334153746e-05, 'epoch': 0.45})
+[2026-07-28 16:38:48,955][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.024643484503030777, 'train/grad_norm_pre_clip': 1.2394014596939087, 'timing/data': 0.00035569723695516586, 'timing/model': 0.859902722761035, 'learning_rate/qwen_vl_interface': 9.814584194532001e-06, 'learning_rate/action_model': 9.814584194532001e-05, 'epoch': 0.45})
+[2026-07-28 16:38:58,176][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.04191102832555771, 'train/grad_norm_pre_clip': 1.7150628566741943, 'timing/data': 0.0004406468942761421, 'timing/model': 0.6225368930026889, 'learning_rate/qwen_vl_interface': 9.813524093001128e-06, 'learning_rate/action_model': 9.813524093001129e-05, 'epoch': 0.45})
+[2026-07-28 16:39:07,621][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.03580951690673828, 'train/grad_norm_pre_clip': 0.8143829107284546, 'timing/data': 0.000359964556992054, 'timing/model': 0.8567650029435754, 'learning_rate/qwen_vl_interface': 9.81246103024902e-06, 'learning_rate/action_model': 9.81246103024902e-05, 'epoch': 0.45})
+[2026-07-28 16:39:16,869][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.036223169416189194, 'train/grad_norm_pre_clip': 0.8651489019393921, 'timing/data': 0.00033964961767196655, 'timing/model': 0.6203015726059675, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.45})
+[2026-07-28 16:39:26,016][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.05310741066932678, 'train/grad_norm_pre_clip': 1.04817795753479, 'timing/data': 0.00034300796687602997, 'timing/model': 0.620276503264904, 'learning_rate/qwen_vl_interface': 9.810326023842245e-06, 'learning_rate/action_model': 9.810326023842245e-05, 'epoch': 0.45})
+[2026-07-28 16:39:35,409][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.05039156973361969, 'train/grad_norm_pre_clip': 1.7281278371810913, 'timing/data': 0.0003553731366991997, 'timing/model': 0.6210316084325314, 'learning_rate/qwen_vl_interface': 9.809254081572963e-06, 'learning_rate/action_model': 9.809254081572961e-05, 'epoch': 0.45})
+[2026-07-28 16:39:44,793][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.029859919100999832, 'train/grad_norm_pre_clip': 1.07859468460083, 'timing/data': 0.00039921142160892487, 'timing/model': 0.619014423340559, 'learning_rate/qwen_vl_interface': 9.808179180853204e-06, 'learning_rate/action_model': 9.808179180853204e-05, 'epoch': 0.45})
+[2026-07-28 16:39:53,960][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.04361778870224953, 'train/grad_norm_pre_clip': 0.8996740579605103, 'timing/data': 0.00036598090082407, 'timing/model': 0.6189891872927547, 'learning_rate/qwen_vl_interface': 9.807101322380464e-06, 'learning_rate/action_model': 9.807101322380464e-05, 'epoch': 0.45})
+[2026-07-28 16:40:03,113][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.07789228111505508, 'train/grad_norm_pre_clip': 2.9357945919036865, 'timing/data': 0.00033100415021181107, 'timing/model': 0.6197361210361123, 'learning_rate/qwen_vl_interface': 9.806020506854152e-06, 'learning_rate/action_model': 9.806020506854153e-05, 'epoch': 0.45})
+[2026-07-28 16:40:12,748][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.04759616404771805, 'train/grad_norm_pre_clip': 0.9147771596908569, 'timing/data': 0.0003572022542357445, 'timing/model': 0.82852712739259, 'learning_rate/qwen_vl_interface': 9.804936734975596e-06, 'learning_rate/action_model': 9.804936734975596e-05, 'epoch': 0.45})
+[2026-07-28 16:40:21,954][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.0589289627969265, 'train/grad_norm_pre_clip': 1.9945887327194214, 'timing/data': 0.0003861133009195328, 'timing/model': 0.6196576794609427, 'learning_rate/qwen_vl_interface': 9.803850007448046e-06, 'learning_rate/action_model': 9.803850007448045e-05, 'epoch': 0.45})
+[2026-07-28 16:40:31,105][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.046288587152957916, 'train/grad_norm_pre_clip': 1.693444848060608, 'timing/data': 0.00032570958137512207, 'timing/model': 0.6181845804676414, 'learning_rate/qwen_vl_interface': 9.802760324976666e-06, 'learning_rate/action_model': 9.802760324976667e-05, 'epoch': 0.45})
+[2026-07-28 16:40:40,251][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.014715263620018959, 'train/grad_norm_pre_clip': 0.8846513032913208, 'timing/data': 0.0003333697095513344, 'timing/model': 0.6195499710738659, 'learning_rate/qwen_vl_interface': 9.80166768826854e-06, 'learning_rate/action_model': 9.80166768826854e-05, 'epoch': 0.45})
+[2026-07-28 16:40:49,666][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.05948307365179062, 'train/grad_norm_pre_clip': 1.591649055480957, 'timing/data': 0.00035601481795310974, 'timing/model': 0.6214415626600385, 'learning_rate/qwen_vl_interface': 9.800572098032668e-06, 'learning_rate/action_model': 9.800572098032668e-05, 'epoch': 0.45})
+[2026-07-28 16:40:59,082][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.03330899029970169, 'train/grad_norm_pre_clip': 1.5001412630081177, 'timing/data': 0.0003642747178673744, 'timing/model': 0.6219250028952956, 'learning_rate/qwen_vl_interface': 9.799473554979966e-06, 'learning_rate/action_model': 9.799473554979965e-05, 'epoch': 0.45})
+[2026-07-28 16:41:08,287][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.045345038175582886, 'train/grad_norm_pre_clip': 1.2087031602859497, 'timing/data': 0.000466187484562397, 'timing/model': 0.6188093656674027, 'learning_rate/qwen_vl_interface': 9.798372059823268e-06, 'learning_rate/action_model': 9.798372059823268e-05, 'epoch': 0.45})
+[2026-07-28 16:41:17,678][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.01876671612262726, 'train/grad_norm_pre_clip': 0.5902103781700134, 'timing/data': 0.0003403369337320328, 'timing/model': 0.8090130789205432, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.45})
+[2026-07-28 16:41:26,807][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.10535603761672974, 'train/grad_norm_pre_clip': 1.7303807735443115, 'timing/data': 0.000334286130964756, 'timing/model': 0.6219017533585429, 'learning_rate/qwen_vl_interface': 9.79616021605879e-06, 'learning_rate/action_model': 9.796160216058788e-05, 'epoch': 0.45})
+[2026-07-28 16:41:35,963][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.04006138816475868, 'train/grad_norm_pre_clip': 1.9526479244232178, 'timing/data': 0.0003671189770102501, 'timing/model': 0.6212563188746572, 'learning_rate/qwen_vl_interface': 9.795049868886247e-06, 'learning_rate/action_model': 9.795049868886247e-05, 'epoch': 0.45})
+[2026-07-28 16:41:45,348][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.0329122319817543, 'train/grad_norm_pre_clip': 1.2109403610229492, 'timing/data': 0.0003478825092315674, 'timing/model': 0.8464541826397181, 'learning_rate/qwen_vl_interface': 9.793936572480192e-06, 'learning_rate/action_model': 9.793936572480192e-05, 'epoch': 0.45})
+[2026-07-28 16:41:54,773][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.03278309851884842, 'train/grad_norm_pre_clip': 0.9266586303710938, 'timing/data': 0.0003271866589784622, 'timing/model': 0.8628468438982964, 'learning_rate/qwen_vl_interface': 9.792820327563028e-06, 'learning_rate/action_model': 9.792820327563028e-05, 'epoch': 0.45})
+[2026-07-28 16:42:03,897][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.048414479941129684, 'train/grad_norm_pre_clip': 1.067092776298523, 'timing/data': 0.0003935396671295166, 'timing/model': 0.6209458317607641, 'learning_rate/qwen_vl_interface': 9.791701134859076e-06, 'learning_rate/action_model': 9.791701134859075e-05, 'epoch': 0.45})
+[2026-07-28 16:42:13,078][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.035860925912857056, 'train/grad_norm_pre_clip': 1.0783679485321045, 'timing/data': 0.0004141600802540779, 'timing/model': 0.6237087985500693, 'learning_rate/qwen_vl_interface': 9.790578995094565e-06, 'learning_rate/action_model': 9.790578995094563e-05, 'epoch': 0.45})
+[2026-07-28 16:42:22,282][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.025400875136256218, 'train/grad_norm_pre_clip': 1.2987496852874756, 'timing/data': 0.00039242953062057495, 'timing/model': 0.6195159945636988, 'learning_rate/qwen_vl_interface': 9.789453908997639e-06, 'learning_rate/action_model': 9.789453908997638e-05, 'epoch': 0.45})
+[2026-07-28 16:42:31,725][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.039736635982990265, 'train/grad_norm_pre_clip': 1.0297517776489258, 'timing/data': 0.00032517779618501663, 'timing/model': 0.6183209270238876, 'learning_rate/qwen_vl_interface': 9.788325877298355e-06, 'learning_rate/action_model': 9.788325877298355e-05, 'epoch': 0.45})
+[2026-07-28 16:42:41,137][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.031206656247377396, 'train/grad_norm_pre_clip': 1.7103288173675537, 'timing/data': 0.0003298195078969002, 'timing/model': 0.8529806267470121, 'learning_rate/qwen_vl_interface': 9.787194900728682e-06, 'learning_rate/action_model': 9.787194900728683e-05, 'epoch': 0.45})
+[2026-07-28 16:42:50,478][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.039881132543087006, 'train/grad_norm_pre_clip': 1.4622527360916138, 'timing/data': 0.00038564950227737427, 'timing/model': 0.6200254736468196, 'learning_rate/qwen_vl_interface': 9.786060980022499e-06, 'learning_rate/action_model': 9.786060980022498e-05, 'epoch': 0.45})
+[2026-07-28 16:42:59,936][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.027830354869365692, 'train/grad_norm_pre_clip': 1.8167909383773804, 'timing/data': 0.0003297021612524986, 'timing/model': 0.6183895468711853, 'learning_rate/qwen_vl_interface': 9.784924115915592e-06, 'learning_rate/action_model': 9.784924115915592e-05, 'epoch': 0.45})
+[2026-07-28 16:43:09,083][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.024911049753427505, 'train/grad_norm_pre_clip': 1.4841163158416748, 'timing/data': 0.00032529979944229126, 'timing/model': 0.6200100975111127, 'learning_rate/qwen_vl_interface': 9.783784309145663e-06, 'learning_rate/action_model': 9.783784309145661e-05, 'epoch': 0.45})
+[2026-07-28 16:43:18,021][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.024934275075793266, 'train/grad_norm_pre_clip': 1.7809799909591675, 'timing/data': 0.0003508320078253746, 'timing/model': 0.619846947491169, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.45})
+[2026-07-28 16:43:27,226][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.02034943364560604, 'train/grad_norm_pre_clip': 1.0728089809417725, 'timing/data': 0.0003557596355676651, 'timing/model': 0.6228938894346356, 'learning_rate/qwen_vl_interface': 9.78149587057707e-06, 'learning_rate/action_model': 9.78149587057707e-05, 'epoch': 0.46})
+[2026-07-28 16:43:36,744][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.08226098120212555, 'train/grad_norm_pre_clip': 1.5803958177566528, 'timing/data': 0.0004050852730870247, 'timing/model': 0.6189225967973471, 'learning_rate/qwen_vl_interface': 9.780347240263354e-06, 'learning_rate/action_model': 9.780347240263353e-05, 'epoch': 0.46})
+[2026-07-28 16:43:46,177][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.023891620337963104, 'train/grad_norm_pre_clip': 1.5062133073806763, 'timing/data': 0.00032559875398874283, 'timing/model': 0.623182961717248, 'learning_rate/qwen_vl_interface': 9.779195670256498e-06, 'learning_rate/action_model': 9.779195670256497e-05, 'epoch': 0.46})
+[2026-07-28 16:43:55,385][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.034511592239141464, 'train/grad_norm_pre_clip': 1.2325520515441895, 'timing/data': 0.0003366796299815178, 'timing/model': 0.6194727011024952, 'learning_rate/qwen_vl_interface': 9.778041161303746e-06, 'learning_rate/action_model': 9.778041161303745e-05, 'epoch': 0.46})
+[2026-07-28 16:44:04,649][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.03197445347905159, 'train/grad_norm_pre_clip': 0.6480313539505005, 'timing/data': 0.00040752533823251724, 'timing/model': 0.6367054861038923, 'learning_rate/qwen_vl_interface': 9.776883714154242e-06, 'learning_rate/action_model': 9.776883714154241e-05, 'epoch': 0.46})
+[2026-07-28 16:44:14,079][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.020043466240167618, 'train/grad_norm_pre_clip': 1.1156506538391113, 'timing/data': 0.00033170077949762344, 'timing/model': 0.62887226883322, 'learning_rate/qwen_vl_interface': 9.775723329559044e-06, 'learning_rate/action_model': 9.775723329559044e-05, 'epoch': 0.46})
+[2026-07-28 16:44:23,298][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.03239152580499649, 'train/grad_norm_pre_clip': 1.3604532480239868, 'timing/data': 0.0003533121198415756, 'timing/model': 0.6229625325649977, 'learning_rate/qwen_vl_interface': 9.774560008271114e-06, 'learning_rate/action_model': 9.774560008271114e-05, 'epoch': 0.46})
+[2026-07-28 16:44:32,504][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.04930465668439865, 'train/grad_norm_pre_clip': 2.481192111968994, 'timing/data': 0.0003876015543937683, 'timing/model': 0.6225658329203725, 'learning_rate/qwen_vl_interface': 9.773393751045315e-06, 'learning_rate/action_model': 9.773393751045315e-05, 'epoch': 0.46})
+[2026-07-28 16:44:41,948][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.05126471072435379, 'train/grad_norm_pre_clip': 1.2969400882720947, 'timing/data': 0.00037176720798015594, 'timing/model': 0.8693754151463509, 'learning_rate/qwen_vl_interface': 9.772224558638422e-06, 'learning_rate/action_model': 9.772224558638422e-05, 'epoch': 0.46})
+[2026-07-28 16:44:51,330][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.04730755090713501, 'train/grad_norm_pre_clip': 1.041635274887085, 'timing/data': 0.0003437912091612816, 'timing/model': 0.8025551047176123, 'learning_rate/qwen_vl_interface': 9.77105243180911e-06, 'learning_rate/action_model': 9.771052431809109e-05, 'epoch': 0.46})
+[2026-07-28 16:45:00,518][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.060033150017261505, 'train/grad_norm_pre_clip': 1.9383585453033447, 'timing/data': 0.00033815857023000717, 'timing/model': 0.6232755472883582, 'learning_rate/qwen_vl_interface': 9.76987737131796e-06, 'learning_rate/action_model': 9.76987737131796e-05, 'epoch': 0.46})
+[2026-07-28 16:45:09,727][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.042538806796073914, 'train/grad_norm_pre_clip': 1.4317456483840942, 'timing/data': 0.00040260981768369675, 'timing/model': 0.6226861467584968, 'learning_rate/qwen_vl_interface': 9.768699377927454e-06, 'learning_rate/action_model': 9.768699377927453e-05, 'epoch': 0.46})
+[2026-07-28 16:45:18,986][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.043029576539993286, 'train/grad_norm_pre_clip': 1.4046624898910522, 'timing/data': 0.0005703205242753029, 'timing/model': 0.6354390764608979, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.46})
+[2026-07-28 16:45:28,352][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.059961095452308655, 'train/grad_norm_pre_clip': 1.6897451877593994, 'timing/data': 0.0003515072166919708, 'timing/model': 0.6179240876808763, 'learning_rate/qwen_vl_interface': 9.76633459550783e-06, 'learning_rate/action_model': 9.76633459550783e-05, 'epoch': 0.46})
+[2026-07-28 16:45:37,546][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.04832477495074272, 'train/grad_norm_pre_clip': 1.090359091758728, 'timing/data': 0.00048611778765916824, 'timing/model': 0.6205226164311171, 'learning_rate/qwen_vl_interface': 9.765147808013193e-06, 'learning_rate/action_model': 9.765147808013192e-05, 'epoch': 0.46})
+[2026-07-28 16:45:46,974][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.0773165374994278, 'train/grad_norm_pre_clip': 2.52846622467041, 'timing/data': 0.000364009290933609, 'timing/model': 0.6208188785240054, 'learning_rate/qwen_vl_interface': 9.763958090688165e-06, 'learning_rate/action_model': 9.763958090688164e-05, 'epoch': 0.46})
+[2026-07-28 16:45:56,184][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.07069000601768494, 'train/grad_norm_pre_clip': 0.9811394810676575, 'timing/data': 0.00033743400126695633, 'timing/model': 0.6192715782672167, 'learning_rate/qwen_vl_interface': 9.762765444304737e-06, 'learning_rate/action_model': 9.762765444304737e-05, 'epoch': 0.46})
+[2026-07-28 16:46:05,739][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.03441161662340164, 'train/grad_norm_pre_clip': 0.746789813041687, 'timing/data': 0.00032746419310569763, 'timing/model': 0.6201140647754073, 'learning_rate/qwen_vl_interface': 9.761569869636808e-06, 'learning_rate/action_model': 9.761569869636808e-05, 'epoch': 0.46})
+[2026-07-28 16:46:15,386][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.059925395995378494, 'train/grad_norm_pre_clip': 1.621444821357727, 'timing/data': 0.00032208021730184555, 'timing/model': 0.6218444565311074, 'learning_rate/qwen_vl_interface': 9.760371367460172e-06, 'learning_rate/action_model': 9.760371367460171e-05, 'epoch': 0.46})
+[2026-07-28 16:46:24,579][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.022518359124660492, 'train/grad_norm_pre_clip': 0.671491801738739, 'timing/data': 0.0003539016470313072, 'timing/model': 0.6255066646263003, 'learning_rate/qwen_vl_interface': 9.759169938552524e-06, 'learning_rate/action_model': 9.759169938552524e-05, 'epoch': 0.46})
+[2026-07-28 16:46:34,029][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.0392465777695179, 'train/grad_norm_pre_clip': 0.6219671368598938, 'timing/data': 0.0003345133736729622, 'timing/model': 0.8542667226865888, 'learning_rate/qwen_vl_interface': 9.757965583693458e-06, 'learning_rate/action_model': 9.757965583693457e-05, 'epoch': 0.46})
+[2026-07-28 16:46:43,338][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.03958627209067345, 'train/grad_norm_pre_clip': 1.133223533630371, 'timing/data': 0.00032319873571395874, 'timing/model': 0.7580147413536906, 'learning_rate/qwen_vl_interface': 9.756758303664467e-06, 'learning_rate/action_model': 9.756758303664468e-05, 'epoch': 0.46})
+[2026-07-28 16:46:52,520][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2500
+[2026-07-28 16:47:02,787][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2500
+[2026-07-28 17:01:42,212][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500
+[2026-07-28 17:01:42,214][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.03334356099367142, 'train/grad_norm_pre_clip': 0.8026732802391052, 'eval/action_loss/seconds': 10.262307683937252, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.03505444526672363, 'eval/latency_3/loss': 0.03505444526672363, 'eval/flappy/loss': 0.03505444526672363, 'eval/flappy/latency_3/loss': 0.03505444526672363, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 318.88500375412406, 'rl_games_eval/mid_train/mean_length': 2587.3, 'rl_games_eval/mid_train/std_reward': 154.78360002135435, 'rl_games_eval/mid_train/std_length': 1247.67135496492, 'rl_games_eval/mid_train/macro_mean_reward': 318.88500375412406, 'rl_games_eval/mid_train/macro_mean_length': 2587.3, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 318.88500375412406, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2587.3, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 154.78360002135435, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1247.67135496492, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 318.88500375412406, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00035103689879179, 'timing/model': 0.6210416639223695, 'learning_rate/qwen_vl_interface': 9.755548099248942e-06, 'learning_rate/action_model': 9.75554809924894e-05, 'epoch': 0.46})
+[2026-07-28 17:01:42,217][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2500_state
+[2026-07-28 17:01:42,217][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 17:02:20,248][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2500_state/pytorch_model
+[2026-07-28 17:02:20,248][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2500_state/sampler.bin
+[2026-07-28 17:02:20,249][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2500_state/sampler_1.bin
+[2026-07-28 17:02:20,256][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_2500_state/random_states_0.pkl
+[2026-07-28 17:02:28,345][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500
+[2026-07-28 17:02:28,345][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 17:02:28,367][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 17:02:37,672][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.027223993092775345, 'train/grad_norm_pre_clip': 0.6514342427253723, 'timing/data': 0.000356367789208889, 'timing/model': 0.6221224227920175, 'learning_rate/qwen_vl_interface': 9.75433497123217e-06, 'learning_rate/action_model': 9.754334971232171e-05, 'epoch': 0.46})
+[2026-07-28 17:02:47,322][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.07566305249929428, 'train/grad_norm_pre_clip': 2.4898762702941895, 'timing/data': 0.0003427332267165184, 'timing/model': 0.8654396887868643, 'learning_rate/qwen_vl_interface': 9.753118920401341e-06, 'learning_rate/action_model': 9.75311892040134e-05, 'epoch': 0.46})
+[2026-07-28 17:02:56,471][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.019278770312666893, 'train/grad_norm_pre_clip': 0.5616477131843567, 'timing/data': 0.0003285938873887062, 'timing/model': 0.617574505507946, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.46})
+[2026-07-28 17:03:05,615][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.017428074032068253, 'train/grad_norm_pre_clip': 0.6161119341850281, 'timing/data': 0.0003483602777123451, 'timing/model': 0.6211262512952089, 'learning_rate/qwen_vl_interface': 9.750678053455725e-06, 'learning_rate/action_model': 9.750678053455725e-05, 'epoch': 0.46})
+[2026-07-28 17:03:14,766][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.01127829123288393, 'train/grad_norm_pre_clip': 0.4862288534641266, 'timing/data': 0.00035028252750635147, 'timing/model': 0.6193836778402328, 'learning_rate/qwen_vl_interface': 9.749453238924793e-06, 'learning_rate/action_model': 9.749453238924792e-05, 'epoch': 0.46})
+[2026-07-28 17:03:24,143][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.039491377770900726, 'train/grad_norm_pre_clip': 0.9949695467948914, 'timing/data': 0.0004084864631295204, 'timing/model': 0.8373897662386298, 'learning_rate/qwen_vl_interface': 9.7482255047475e-06, 'learning_rate/action_model': 9.7482255047475e-05, 'epoch': 0.46})
+[2026-07-28 17:03:33,542][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.031066596508026123, 'train/grad_norm_pre_clip': 1.9168449640274048, 'timing/data': 0.00033129844814538956, 'timing/model': 0.6200473960489035, 'learning_rate/qwen_vl_interface': 9.746994851720518e-06, 'learning_rate/action_model': 9.746994851720517e-05, 'epoch': 0.46})
+[2026-07-28 17:03:42,446][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.02806440368294716, 'train/grad_norm_pre_clip': 1.530192255973816, 'timing/data': 0.0003289114683866501, 'timing/model': 0.618765689432621, 'learning_rate/qwen_vl_interface': 9.745761280642397e-06, 'learning_rate/action_model': 9.745761280642397e-05, 'epoch': 0.46})
+[2026-07-28 17:03:51,342][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.02022596076130867, 'train/grad_norm_pre_clip': 1.0091552734375, 'timing/data': 0.0003363434225320816, 'timing/model': 0.6212849356234074, 'learning_rate/qwen_vl_interface': 9.744524792313591e-06, 'learning_rate/action_model': 9.74452479231359e-05, 'epoch': 0.46})
+[2026-07-28 17:04:00,248][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.03647444397211075, 'train/grad_norm_pre_clip': 1.728291630744934, 'timing/data': 0.0003456445410847664, 'timing/model': 0.6187842981889844, 'learning_rate/qwen_vl_interface': 9.743285387536445e-06, 'learning_rate/action_model': 9.743285387536444e-05, 'epoch': 0.46})
+[2026-07-28 17:04:09,319][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.03674313426017761, 'train/grad_norm_pre_clip': 1.450946569442749, 'timing/data': 0.0003574443981051445, 'timing/model': 0.6188003551214933, 'learning_rate/qwen_vl_interface': 9.742043067115193e-06, 'learning_rate/action_model': 9.742043067115192e-05, 'epoch': 0.46})
+[2026-07-28 17:04:18,820][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.012219926342368126, 'train/grad_norm_pre_clip': 0.8736963868141174, 'timing/data': 0.0003693085163831711, 'timing/model': 0.6247242800891399, 'learning_rate/qwen_vl_interface': 9.740797831855965e-06, 'learning_rate/action_model': 9.740797831855965e-05, 'epoch': 0.46})
+[2026-07-28 17:04:28,078][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.06984551250934601, 'train/grad_norm_pre_clip': 1.4772793054580688, 'timing/data': 0.00034514348953962326, 'timing/model': 0.6200794465839863, 'learning_rate/qwen_vl_interface': 9.73954968256678e-06, 'learning_rate/action_model': 9.73954968256678e-05, 'epoch': 0.46})
+[2026-07-28 17:04:37,328][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.027044285088777542, 'train/grad_norm_pre_clip': 0.8147941827774048, 'timing/data': 0.0003314260393381119, 'timing/model': 0.6208559181541204, 'learning_rate/qwen_vl_interface': 9.738298620057547e-06, 'learning_rate/action_model': 9.738298620057546e-05, 'epoch': 0.46})
+[2026-07-28 17:04:46,667][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.07684378325939178, 'train/grad_norm_pre_clip': 1.2643699645996094, 'timing/data': 0.0003291657194495201, 'timing/model': 0.6185254082083702, 'learning_rate/qwen_vl_interface': 9.73704464514007e-06, 'learning_rate/action_model': 9.737044645140068e-05, 'epoch': 0.46})
+[2026-07-28 17:04:55,820][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.03911839798092842, 'train/grad_norm_pre_clip': 1.5613468885421753, 'timing/data': 0.00035494565963745117, 'timing/model': 0.6198698543012142, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.46})
+[2026-07-28 17:05:04,954][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.05387838929891586, 'train/grad_norm_pre_clip': 1.1990556716918945, 'timing/data': 0.0007896693423390388, 'timing/model': 0.6186040341854095, 'learning_rate/qwen_vl_interface': 9.734527961337028e-06, 'learning_rate/action_model': 9.734527961337028e-05, 'epoch': 0.46})
+[2026-07-28 17:05:14,067][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.052015628665685654, 'train/grad_norm_pre_clip': 1.6914427280426025, 'timing/data': 0.0003257310017943382, 'timing/model': 0.6197907282039523, 'learning_rate/qwen_vl_interface': 9.733265254084514e-06, 'learning_rate/action_model': 9.733265254084513e-05, 'epoch': 0.46})
+[2026-07-28 17:05:23,395][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.059782497584819794, 'train/grad_norm_pre_clip': 1.6126165390014648, 'timing/data': 0.00034405291080474854, 'timing/model': 0.8296798318624496, 'learning_rate/qwen_vl_interface': 9.73199963768985e-06, 'learning_rate/action_model': 9.731999637689849e-05, 'epoch': 0.46})
+[2026-07-28 17:05:32,542][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.062424156814813614, 'train/grad_norm_pre_clip': 1.4745064973831177, 'timing/data': 0.00040896888822317123, 'timing/model': 0.6208884669467807, 'learning_rate/qwen_vl_interface': 9.73073111297428e-06, 'learning_rate/action_model': 9.73073111297428e-05, 'epoch': 0.46})
+[2026-07-28 17:05:41,919][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.06469947844743729, 'train/grad_norm_pre_clip': 1.5587769746780396, 'timing/data': 0.00033814553171396255, 'timing/model': 0.6173099745064974, 'learning_rate/qwen_vl_interface': 9.72945968076094e-06, 'learning_rate/action_model': 9.729459680760939e-05, 'epoch': 0.46})
+[2026-07-28 17:05:51,052][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.026128727942705154, 'train/grad_norm_pre_clip': 0.9807180166244507, 'timing/data': 0.00041534751653671265, 'timing/model': 0.6212878497317433, 'learning_rate/qwen_vl_interface': 9.728185341874844e-06, 'learning_rate/action_model': 9.728185341874844e-05, 'epoch': 0.46})
+[2026-07-28 17:06:00,219][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.035119496285915375, 'train/grad_norm_pre_clip': 1.0145812034606934, 'timing/data': 0.000333515927195549, 'timing/model': 0.6200557705014944, 'learning_rate/qwen_vl_interface': 9.7269080971429e-06, 'learning_rate/action_model': 9.726908097142899e-05, 'epoch': 0.46})
+[2026-07-28 17:06:09,632][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.0949983075261116, 'train/grad_norm_pre_clip': 2.2287251949310303, 'timing/data': 0.0003489796072244644, 'timing/model': 0.6267335461452603, 'learning_rate/qwen_vl_interface': 9.725627947393896e-06, 'learning_rate/action_model': 9.725627947393896e-05, 'epoch': 0.46})
+[2026-07-28 17:06:19,059][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.03965672105550766, 'train/grad_norm_pre_clip': 1.741659164428711, 'timing/data': 0.0003984309732913971, 'timing/model': 0.6284250793978572, 'learning_rate/qwen_vl_interface': 9.72434489345851e-06, 'learning_rate/action_model': 9.72434489345851e-05, 'epoch': 0.46})
+[2026-07-28 17:06:28,209][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.04785992205142975, 'train/grad_norm_pre_clip': 1.1134952306747437, 'timing/data': 0.0003710128366947174, 'timing/model': 0.6294506285339594, 'learning_rate/qwen_vl_interface': 9.723058936169298e-06, 'learning_rate/action_model': 9.723058936169297e-05, 'epoch': 0.46})
+[2026-07-28 17:06:37,328][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.055783696472644806, 'train/grad_norm_pre_clip': 1.6745985746383667, 'timing/data': 0.00039103254675865173, 'timing/model': 0.6267940551042557, 'learning_rate/qwen_vl_interface': 9.721770076360707e-06, 'learning_rate/action_model': 9.721770076360706e-05, 'epoch': 0.46})
+[2026-07-28 17:06:46,484][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.02492426335811615, 'train/grad_norm_pre_clip': 0.4950510859489441, 'timing/data': 0.00033497344702482224, 'timing/model': 0.6275107897818089, 'learning_rate/qwen_vl_interface': 9.720478314869061e-06, 'learning_rate/action_model': 9.72047831486906e-05, 'epoch': 0.46})
+[2026-07-28 17:06:56,100][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.0500972680747509, 'train/grad_norm_pre_clip': 0.8410481214523315, 'timing/data': 0.00033257994800806046, 'timing/model': 0.8550009466707706, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.46})
+[2026-07-28 17:07:05,476][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.06969378143548965, 'train/grad_norm_pre_clip': 0.8716976046562195, 'timing/data': 0.00037707947194576263, 'timing/model': 0.6191680347546935, 'learning_rate/qwen_vl_interface': 9.717886090191334e-06, 'learning_rate/action_model': 9.717886090191333e-05, 'epoch': 0.46})
+[2026-07-28 17:07:14,365][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.028912780806422234, 'train/grad_norm_pre_clip': 0.8060177564620972, 'timing/data': 0.00033761002123355865, 'timing/model': 0.6192162381485105, 'learning_rate/qwen_vl_interface': 9.716585628687317e-06, 'learning_rate/action_model': 9.716585628687316e-05, 'epoch': 0.46})
+[2026-07-28 17:07:23,482][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.04091706871986389, 'train/grad_norm_pre_clip': 1.172042965888977, 'timing/data': 0.00034457258880138397, 'timing/model': 0.6210417188704014, 'learning_rate/qwen_vl_interface': 9.715282268864379e-06, 'learning_rate/action_model': 9.715282268864378e-05, 'epoch': 0.47})
+[2026-07-28 17:07:32,606][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.03493344038724899, 'train/grad_norm_pre_clip': 0.6789636611938477, 'timing/data': 0.00035900436341762543, 'timing/model': 0.6183890048414469, 'learning_rate/qwen_vl_interface': 9.713976011568256e-06, 'learning_rate/action_model': 9.713976011568256e-05, 'epoch': 0.47})
+[2026-07-28 17:07:42,243][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.02741190418601036, 'train/grad_norm_pre_clip': 1.5471185445785522, 'timing/data': 0.0004369020462036133, 'timing/model': 0.6177451182156801, 'learning_rate/qwen_vl_interface': 9.712666857646565e-06, 'learning_rate/action_model': 9.712666857646564e-05, 'epoch': 0.47})
+[2026-07-28 17:07:51,421][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.041840776801109314, 'train/grad_norm_pre_clip': 0.9401312470436096, 'timing/data': 0.0003389902412891388, 'timing/model': 0.6209803260862827, 'learning_rate/qwen_vl_interface': 9.711354807948798e-06, 'learning_rate/action_model': 9.711354807948798e-05, 'epoch': 0.47})
+[2026-07-28 17:08:00,613][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.0225166417658329, 'train/grad_norm_pre_clip': 0.5635263323783875, 'timing/data': 0.0003204345703125, 'timing/model': 0.6246240586042404, 'learning_rate/qwen_vl_interface': 9.710039863326334e-06, 'learning_rate/action_model': 9.710039863326334e-05, 'epoch': 0.47})
+[2026-07-28 17:08:10,132][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.02907663583755493, 'train/grad_norm_pre_clip': 1.3250248432159424, 'timing/data': 0.00032332539558410645, 'timing/model': 0.6202263804152608, 'learning_rate/qwen_vl_interface': 9.708722024632425e-06, 'learning_rate/action_model': 9.708722024632424e-05, 'epoch': 0.47})
+[2026-07-28 17:08:19,269][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.03922690451145172, 'train/grad_norm_pre_clip': 1.0988668203353882, 'timing/data': 0.0003565531224012375, 'timing/model': 0.619973674416542, 'learning_rate/qwen_vl_interface': 9.7074012927222e-06, 'learning_rate/action_model': 9.7074012927222e-05, 'epoch': 0.47})
+[2026-07-28 17:08:28,664][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.015992609784007072, 'train/grad_norm_pre_clip': 0.7688761353492737, 'timing/data': 0.00032792892307043076, 'timing/model': 0.6200161268934608, 'learning_rate/qwen_vl_interface': 9.706077668452671e-06, 'learning_rate/action_model': 9.706077668452671e-05, 'epoch': 0.47})
+[2026-07-28 17:08:37,814][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.05397286266088486, 'train/grad_norm_pre_clip': 1.6399462223052979, 'timing/data': 0.00032995082437992096, 'timing/model': 0.6194619908928871, 'learning_rate/qwen_vl_interface': 9.704751152682722e-06, 'learning_rate/action_model': 9.704751152682721e-05, 'epoch': 0.47})
+[2026-07-28 17:08:47,318][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.03569404035806656, 'train/grad_norm_pre_clip': 2.125682830810547, 'timing/data': 0.00033983122557401657, 'timing/model': 0.6191934607923031, 'learning_rate/qwen_vl_interface': 9.703421746273115e-06, 'learning_rate/action_model': 9.703421746273114e-05, 'epoch': 0.47})
+[2026-07-28 17:08:56,465][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.03225114196538925, 'train/grad_norm_pre_clip': 1.0398666858673096, 'timing/data': 0.00038192514330148697, 'timing/model': 0.6181223224848509, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.47})
+[2026-07-28 17:09:06,029][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.04083862155675888, 'train/grad_norm_pre_clip': 1.594861388206482, 'timing/data': 0.0003224452957510948, 'timing/model': 0.6168557032942772, 'learning_rate/qwen_vl_interface': 9.700754264987347e-06, 'learning_rate/action_model': 9.700754264987347e-05, 'epoch': 0.47})
+[2026-07-28 17:09:15,249][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.040241342037916183, 'train/grad_norm_pre_clip': 2.0052742958068848, 'timing/data': 0.0005319276824593544, 'timing/model': 0.6363212512806058, 'learning_rate/qwen_vl_interface': 9.69941619184209e-06, 'learning_rate/action_model': 9.699416191842089e-05, 'epoch': 0.47})
+[2026-07-28 17:09:24,632][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.056949108839035034, 'train/grad_norm_pre_clip': 1.4273396730422974, 'timing/data': 0.00042491964995861053, 'timing/model': 0.6200962550938129, 'learning_rate/qwen_vl_interface': 9.69807523151897e-06, 'learning_rate/action_model': 9.698075231518969e-05, 'epoch': 0.47})
+[2026-07-28 17:09:33,786][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.03085365518927574, 'train/grad_norm_pre_clip': 0.9057296514511108, 'timing/data': 0.00032844115048646927, 'timing/model': 0.6187968803569674, 'learning_rate/qwen_vl_interface': 9.696731384888125e-06, 'learning_rate/action_model': 9.696731384888125e-05, 'epoch': 0.47})
+[2026-07-28 17:09:43,372][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.021034296602010727, 'train/grad_norm_pre_clip': 0.6844279170036316, 'timing/data': 0.000332830473780632, 'timing/model': 0.6737982388585806, 'learning_rate/qwen_vl_interface': 9.695384652821562e-06, 'learning_rate/action_model': 9.695384652821561e-05, 'epoch': 0.47})
+[2026-07-28 17:09:52,509][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.06646744906902313, 'train/grad_norm_pre_clip': 1.6051510572433472, 'timing/data': 0.00034181680530309677, 'timing/model': 0.620427917689085, 'learning_rate/qwen_vl_interface': 9.694035036193162e-06, 'learning_rate/action_model': 9.694035036193161e-05, 'epoch': 0.47})
+[2026-07-28 17:10:01,902][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.035471465438604355, 'train/grad_norm_pre_clip': 1.3802093267440796, 'timing/data': 0.0003419620916247368, 'timing/model': 0.8723269822075963, 'learning_rate/qwen_vl_interface': 9.692682535878674e-06, 'learning_rate/action_model': 9.692682535878673e-05, 'epoch': 0.47})
+[2026-07-28 17:10:11,079][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.017773432657122612, 'train/grad_norm_pre_clip': 0.49727433919906616, 'timing/data': 0.00034557562321424484, 'timing/model': 0.6545274341478944, 'learning_rate/qwen_vl_interface': 9.691327152755721e-06, 'learning_rate/action_model': 9.691327152755722e-05, 'epoch': 0.47})
+[2026-07-28 17:10:20,478][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.05702400952577591, 'train/grad_norm_pre_clip': 1.5597970485687256, 'timing/data': 0.00041534285992383957, 'timing/model': 0.621206421405077, 'learning_rate/qwen_vl_interface': 9.6899688877038e-06, 'learning_rate/action_model': 9.689968887703798e-05, 'epoch': 0.47})
+[2026-07-28 17:10:29,689][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.0212557315826416, 'train/grad_norm_pre_clip': 0.7323856353759766, 'timing/data': 0.0003658933565020561, 'timing/model': 0.6192364571616054, 'learning_rate/qwen_vl_interface': 9.688607741604269e-06, 'learning_rate/action_model': 9.688607741604269e-05, 'epoch': 0.47})
+[2026-07-28 17:10:38,844][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.0638524740934372, 'train/grad_norm_pre_clip': 2.0315394401550293, 'timing/data': 0.0003255251795053482, 'timing/model': 0.6210849843919277, 'learning_rate/qwen_vl_interface': 9.687243715340365e-06, 'learning_rate/action_model': 9.687243715340364e-05, 'epoch': 0.47})
+[2026-07-28 17:10:47,979][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.03727493807673454, 'train/grad_norm_pre_clip': 0.9957866072654724, 'timing/data': 0.00036847591400146484, 'timing/model': 0.6215273281559348, 'learning_rate/qwen_vl_interface': 9.685876809797188e-06, 'learning_rate/action_model': 9.685876809797188e-05, 'epoch': 0.47})
+[2026-07-28 17:10:57,785][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.03834719583392143, 'train/grad_norm_pre_clip': 0.8582788109779358, 'timing/data': 0.0004262169823050499, 'timing/model': 0.6260030874982476, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.47})
+[2026-07-28 17:11:07,053][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.05015561729669571, 'train/grad_norm_pre_clip': 1.3968108892440796, 'timing/data': 0.0003833770751953125, 'timing/model': 0.6217204695567489, 'learning_rate/qwen_vl_interface': 9.683134364422766e-06, 'learning_rate/action_model': 9.683134364422765e-05, 'epoch': 0.47})
+[2026-07-28 17:11:16,229][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.04179185628890991, 'train/grad_norm_pre_clip': 1.1117454767227173, 'timing/data': 0.0003775758668780327, 'timing/model': 0.6181668173521757, 'learning_rate/qwen_vl_interface': 9.681758826371064e-06, 'learning_rate/action_model': 9.681758826371063e-05, 'epoch': 0.47})
+[2026-07-28 17:11:25,611][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.02629554644227028, 'train/grad_norm_pre_clip': 0.8098400235176086, 'timing/data': 0.0004138033837080002, 'timing/model': 0.8165736794471741, 'learning_rate/qwen_vl_interface': 9.680380412599174e-06, 'learning_rate/action_model': 9.680380412599173e-05, 'epoch': 0.47})
+[2026-07-28 17:11:35,047][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.028249669820070267, 'train/grad_norm_pre_clip': 0.841768741607666, 'timing/data': 0.0003763372078537941, 'timing/model': 0.8416092377156019, 'learning_rate/qwen_vl_interface': 9.678999124001533e-06, 'learning_rate/action_model': 9.678999124001533e-05, 'epoch': 0.47})
+[2026-07-28 17:11:44,359][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.07278483361005783, 'train/grad_norm_pre_clip': 1.1247807741165161, 'timing/data': 0.0005704779177904129, 'timing/model': 0.6601242627948523, 'learning_rate/qwen_vl_interface': 9.67761496147445e-06, 'learning_rate/action_model': 9.677614961474449e-05, 'epoch': 0.47})
+[2026-07-28 17:11:53,581][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.07053728401660919, 'train/grad_norm_pre_clip': 2.2815325260162354, 'timing/data': 0.0003648344427347183, 'timing/model': 0.6205059699714184, 'learning_rate/qwen_vl_interface': 9.676227925916082e-06, 'learning_rate/action_model': 9.676227925916082e-05, 'epoch': 0.47})
+[2026-07-28 17:12:03,062][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.031144484877586365, 'train/grad_norm_pre_clip': 0.9675036668777466, 'timing/data': 0.00037560146301984787, 'timing/model': 0.9244201043620706, 'learning_rate/qwen_vl_interface': 9.674838018226472e-06, 'learning_rate/action_model': 9.674838018226471e-05, 'epoch': 0.47})
+[2026-07-28 17:12:12,470][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.07437779009342194, 'train/grad_norm_pre_clip': 2.6998002529144287, 'timing/data': 0.00039975158870220184, 'timing/model': 0.6228524809703231, 'learning_rate/qwen_vl_interface': 9.67344523930751e-06, 'learning_rate/action_model': 9.673445239307511e-05, 'epoch': 0.47})
+[2026-07-28 17:12:21,672][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.04749689996242523, 'train/grad_norm_pre_clip': 1.8979041576385498, 'timing/data': 0.0003740563988685608, 'timing/model': 0.6214779270812869, 'learning_rate/qwen_vl_interface': 9.672049590062959e-06, 'learning_rate/action_model': 9.672049590062958e-05, 'epoch': 0.47})
+[2026-07-28 17:12:30,688][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.0600053109228611, 'train/grad_norm_pre_clip': 1.2854350805282593, 'timing/data': 0.0003527803346514702, 'timing/model': 0.6200031228363514, 'learning_rate/qwen_vl_interface': 9.670651071398436e-06, 'learning_rate/action_model': 9.670651071398436e-05, 'epoch': 0.47})
+[2026-07-28 17:12:39,638][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.04635778069496155, 'train/grad_norm_pre_clip': 1.0692976713180542, 'timing/data': 0.00045811478048563004, 'timing/model': 0.6250849990174174, 'learning_rate/qwen_vl_interface': 9.669249684221427e-06, 'learning_rate/action_model': 9.669249684221428e-05, 'epoch': 0.47})
+[2026-07-28 17:12:48,883][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.009018665179610252, 'train/grad_norm_pre_clip': 0.40668147802352905, 'timing/data': 0.0004270533099770546, 'timing/model': 0.8653169805184007, 'learning_rate/qwen_vl_interface': 9.66784542944128e-06, 'learning_rate/action_model': 9.66784542944128e-05, 'epoch': 0.47})
+[2026-07-28 17:12:57,853][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.030531490221619606, 'train/grad_norm_pre_clip': 0.8673892021179199, 'timing/data': 0.0003984915092587471, 'timing/model': 0.6244547152891755, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.47})
+[2026-07-28 17:13:07,002][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.0379088893532753, 'train/grad_norm_pre_clip': 0.5951145887374878, 'timing/data': 0.00039036665111780167, 'timing/model': 0.6204771921038628, 'learning_rate/qwen_vl_interface': 9.66502832071824e-06, 'learning_rate/action_model': 9.66502832071824e-05, 'epoch': 0.47})
+[2026-07-28 17:13:16,213][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.06517201662063599, 'train/grad_norm_pre_clip': 2.091076612472534, 'timing/data': 0.00042368192225694656, 'timing/model': 0.6211212100461125, 'learning_rate/qwen_vl_interface': 9.663615468603344e-06, 'learning_rate/action_model': 9.663615468603343e-05, 'epoch': 0.47})
+[2026-07-28 17:13:25,406][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.04532281681895256, 'train/grad_norm_pre_clip': 1.0435547828674316, 'timing/data': 0.00038416869938373566, 'timing/model': 0.6204664204269648, 'learning_rate/qwen_vl_interface': 9.662199752541287e-06, 'learning_rate/action_model': 9.662199752541287e-05, 'epoch': 0.47})
+[2026-07-28 17:13:34,568][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.06316830217838287, 'train/grad_norm_pre_clip': 2.4018869400024414, 'timing/data': 0.00044103246182203293, 'timing/model': 0.6215605530887842, 'learning_rate/qwen_vl_interface': 9.660781173450711e-06, 'learning_rate/action_model': 9.660781173450712e-05, 'epoch': 0.47})
+[2026-07-28 17:13:43,969][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.05857832357287407, 'train/grad_norm_pre_clip': 1.7124981880187988, 'timing/data': 0.00046582240611314774, 'timing/model': 0.6195273566991091, 'learning_rate/qwen_vl_interface': 9.659359732252121e-06, 'learning_rate/action_model': 9.659359732252121e-05, 'epoch': 0.47})
+[2026-07-28 17:13:53,461][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.030094698071479797, 'train/grad_norm_pre_clip': 0.8538582921028137, 'timing/data': 0.0003882497549057007, 'timing/model': 0.6216547396034002, 'learning_rate/qwen_vl_interface': 9.657935429867872e-06, 'learning_rate/action_model': 9.657935429867873e-05, 'epoch': 0.47})
+[2026-07-28 17:14:02,649][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.016889818012714386, 'train/grad_norm_pre_clip': 1.066085696220398, 'timing/data': 0.0003314819186925888, 'timing/model': 0.6189295714721084, 'learning_rate/qwen_vl_interface': 9.65650826722218e-06, 'learning_rate/action_model': 9.656508267222179e-05, 'epoch': 0.47})
+[2026-07-28 17:14:11,849][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.03232730180025101, 'train/grad_norm_pre_clip': 0.8093465566635132, 'timing/data': 0.0004524942487478256, 'timing/model': 0.6214635213837028, 'learning_rate/qwen_vl_interface': 9.655078245241114e-06, 'learning_rate/action_model': 9.655078245241112e-05, 'epoch': 0.47})
+[2026-07-28 17:14:21,282][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.05432892590761185, 'train/grad_norm_pre_clip': 1.7625514268875122, 'timing/data': 0.000361517071723938, 'timing/model': 0.8699876870959997, 'learning_rate/qwen_vl_interface': 9.653645364852598e-06, 'learning_rate/action_model': 9.653645364852597e-05, 'epoch': 0.47})
+[2026-07-28 17:14:30,486][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.05683262646198273, 'train/grad_norm_pre_clip': 1.4257844686508179, 'timing/data': 0.0003320099785923958, 'timing/model': 0.621357255615294, 'learning_rate/qwen_vl_interface': 9.652209626986414e-06, 'learning_rate/action_model': 9.652209626986413e-05, 'epoch': 0.47})
+[2026-07-28 17:14:39,693][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.012628693133592606, 'train/grad_norm_pre_clip': 0.7681018114089966, 'timing/data': 0.00032148323953151703, 'timing/model': 0.6203741738572717, 'learning_rate/qwen_vl_interface': 9.6507710325742e-06, 'learning_rate/action_model': 9.6507710325742e-05, 'epoch': 0.47})
+[2026-07-28 17:14:49,256][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.016796620562672615, 'train/grad_norm_pre_clip': 0.7215579152107239, 'timing/data': 0.00046784523874521255, 'timing/model': 0.8827232960611582, 'learning_rate/qwen_vl_interface': 9.649329582549439e-06, 'learning_rate/action_model': 9.649329582549439e-05, 'epoch': 0.47})
+[2026-07-28 17:14:58,783][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.0393160879611969, 'train/grad_norm_pre_clip': 1.4579862356185913, 'timing/data': 0.0003731800243258476, 'timing/model': 0.8364271083846688, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.47})
+[2026-07-28 17:15:07,976][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.038892995566129684, 'train/grad_norm_pre_clip': 1.5052552223205566, 'timing/data': 0.0003327690064907074, 'timing/model': 0.6275170082226396, 'learning_rate/qwen_vl_interface': 9.646438119405503e-06, 'learning_rate/action_model': 9.646438119405503e-05, 'epoch': 0.47})
+[2026-07-28 17:15:17,100][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.03707655146718025, 'train/grad_norm_pre_clip': 1.2249130010604858, 'timing/data': 0.0003284979611635208, 'timing/model': 0.6192301549017429, 'learning_rate/qwen_vl_interface': 9.644988108162566e-06, 'learning_rate/action_model': 9.644988108162567e-05, 'epoch': 0.47})
+[2026-07-28 17:15:26,366][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.07652471959590912, 'train/grad_norm_pre_clip': 2.0566163063049316, 'timing/data': 0.00046256184577941895, 'timing/model': 0.6211744649335742, 'learning_rate/qwen_vl_interface': 9.643535245059562e-06, 'learning_rate/action_model': 9.643535245059561e-05, 'epoch': 0.47})
+[2026-07-28 17:15:35,885][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.040865425020456314, 'train/grad_norm_pre_clip': 1.3982375860214233, 'timing/data': 0.0003477390855550766, 'timing/model': 0.627178942784667, 'learning_rate/qwen_vl_interface': 9.642079531039235e-06, 'learning_rate/action_model': 9.642079531039235e-05, 'epoch': 0.47})
+[2026-07-28 17:15:45,051][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.01888158544898033, 'train/grad_norm_pre_clip': 0.7615702152252197, 'timing/data': 0.00035072024911642075, 'timing/model': 0.6211415957659483, 'learning_rate/qwen_vl_interface': 9.640620967046186e-06, 'learning_rate/action_model': 9.640620967046186e-05, 'epoch': 0.47})
+[2026-07-28 17:15:54,433][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.006072934251278639, 'train/grad_norm_pre_clip': 0.24543778598308563, 'timing/data': 0.0003805868327617645, 'timing/model': 0.6185998655855656, 'learning_rate/qwen_vl_interface': 9.63915955402686e-06, 'learning_rate/action_model': 9.63915955402686e-05, 'epoch': 0.48})
+[2026-07-28 17:16:03,865][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.029606567695736885, 'train/grad_norm_pre_clip': 0.9059900045394897, 'timing/data': 0.0003685643896460533, 'timing/model': 0.8460719818249345, 'learning_rate/qwen_vl_interface': 9.637695292929552e-06, 'learning_rate/action_model': 9.637695292929552e-05, 'epoch': 0.48})
+[2026-07-28 17:16:13,044][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.025308165699243546, 'train/grad_norm_pre_clip': 1.6107306480407715, 'timing/data': 0.0003448501229286194, 'timing/model': 0.6191187994554639, 'learning_rate/qwen_vl_interface': 9.636228184704405e-06, 'learning_rate/action_model': 9.636228184704405e-05, 'epoch': 0.48})
+[2026-07-28 17:16:22,169][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.04546472057700157, 'train/grad_norm_pre_clip': 1.6859840154647827, 'timing/data': 0.00033875368535518646, 'timing/model': 0.6198591748252511, 'learning_rate/qwen_vl_interface': 9.634758230303408e-06, 'learning_rate/action_model': 9.634758230303408e-05, 'epoch': 0.48})
+[2026-07-28 17:16:31,587][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.041968509554862976, 'train/grad_norm_pre_clip': 1.0055255889892578, 'timing/data': 0.0003736671060323715, 'timing/model': 0.872276864014566, 'learning_rate/qwen_vl_interface': 9.633285430680401e-06, 'learning_rate/action_model': 9.633285430680401e-05, 'epoch': 0.48})
+[2026-07-28 17:16:40,800][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.05692833662033081, 'train/grad_norm_pre_clip': 1.1117918491363525, 'timing/data': 0.00036225002259016037, 'timing/model': 0.6227566199377179, 'learning_rate/qwen_vl_interface': 9.631809786791066e-06, 'learning_rate/action_model': 9.631809786791066e-05, 'epoch': 0.48})
+[2026-07-28 17:16:50,212][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.0273782666772604, 'train/grad_norm_pre_clip': 0.991828203201294, 'timing/data': 0.000325559638440609, 'timing/model': 0.6367080276831985, 'learning_rate/qwen_vl_interface': 9.630331299592932e-06, 'learning_rate/action_model': 9.630331299592933e-05, 'epoch': 0.48})
+[2026-07-28 17:16:59,381][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.04103948920965195, 'train/grad_norm_pre_clip': 1.062129020690918, 'timing/data': 0.000427858904004097, 'timing/model': 0.6198170091956854, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.48})
+[2026-07-28 17:17:08,550][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.057097431272268295, 'train/grad_norm_pre_clip': 1.3599627017974854, 'timing/data': 0.00036988500505685806, 'timing/model': 0.6198437875136733, 'learning_rate/qwen_vl_interface': 9.627365799109617e-06, 'learning_rate/action_model': 9.627365799109617e-05, 'epoch': 0.48})
+[2026-07-28 17:17:18,041][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.06238272786140442, 'train/grad_norm_pre_clip': 1.6457350254058838, 'timing/data': 0.00038404110819101334, 'timing/model': 0.6236198451370001, 'learning_rate/qwen_vl_interface': 9.625878787748712e-06, 'learning_rate/action_model': 9.625878787748713e-05, 'epoch': 0.48})
+[2026-07-28 17:17:27,192][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.05314085632562637, 'train/grad_norm_pre_clip': 1.2531977891921997, 'timing/data': 0.00034912489354610443, 'timing/model': 0.6237433841452003, 'learning_rate/qwen_vl_interface': 9.624388936927572e-06, 'learning_rate/action_model': 9.624388936927571e-05, 'epoch': 0.48})
+[2026-07-28 17:17:36,604][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.03148365765810013, 'train/grad_norm_pre_clip': 1.0107322931289673, 'timing/data': 0.0004340261220932007, 'timing/model': 0.8464161390438676, 'learning_rate/qwen_vl_interface': 9.622896247612943e-06, 'learning_rate/action_model': 9.622896247612942e-05, 'epoch': 0.48})
+[2026-07-28 17:17:45,761][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.06867657601833344, 'train/grad_norm_pre_clip': 1.0259636640548706, 'timing/data': 0.0004210788756608963, 'timing/model': 0.6193008227273822, 'learning_rate/qwen_vl_interface': 9.621400720773414e-06, 'learning_rate/action_model': 9.621400720773414e-05, 'epoch': 0.48})
+[2026-07-28 17:17:54,934][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.027843158692121506, 'train/grad_norm_pre_clip': 1.5061633586883545, 'timing/data': 0.00041034817695617676, 'timing/model': 0.6200173608958721, 'learning_rate/qwen_vl_interface': 9.619902357379416e-06, 'learning_rate/action_model': 9.619902357379415e-05, 'epoch': 0.48})
+[2026-07-28 17:18:04,382][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.05287064239382744, 'train/grad_norm_pre_clip': 1.3386037349700928, 'timing/data': 0.0004190346226096153, 'timing/model': 0.6226701308041811, 'learning_rate/qwen_vl_interface': 9.618401158403222e-06, 'learning_rate/action_model': 9.618401158403222e-05, 'epoch': 0.48})
+[2026-07-28 17:18:13,554][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.04360159486532211, 'train/grad_norm_pre_clip': 1.1229567527770996, 'timing/data': 0.0004088692367076874, 'timing/model': 0.6201403643935919, 'learning_rate/qwen_vl_interface': 9.616897124818944e-06, 'learning_rate/action_model': 9.616897124818945e-05, 'epoch': 0.48})
+[2026-07-28 17:18:22,949][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.0843416079878807, 'train/grad_norm_pre_clip': 3.120107650756836, 'timing/data': 0.00041887443512678146, 'timing/model': 0.6216466575860977, 'learning_rate/qwen_vl_interface': 9.615390257602532e-06, 'learning_rate/action_model': 9.615390257602531e-05, 'epoch': 0.48})
+[2026-07-28 17:18:32,144][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.01988605223596096, 'train/grad_norm_pre_clip': 1.2375030517578125, 'timing/data': 0.0004246039316058159, 'timing/model': 0.6213242970407009, 'learning_rate/qwen_vl_interface': 9.613880557731776e-06, 'learning_rate/action_model': 9.613880557731776e-05, 'epoch': 0.48})
+[2026-07-28 17:18:41,327][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.04306739568710327, 'train/grad_norm_pre_clip': 0.8670442700386047, 'timing/data': 0.0004156269133090973, 'timing/model': 0.6208207113668323, 'learning_rate/qwen_vl_interface': 9.612368026186304e-06, 'learning_rate/action_model': 9.612368026186304e-05, 'epoch': 0.48})
+[2026-07-28 17:18:51,015][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.04906965792179108, 'train/grad_norm_pre_clip': 1.508607029914856, 'timing/data': 0.0003904290497303009, 'timing/model': 0.6197647983208299, 'learning_rate/qwen_vl_interface': 9.610852663947582e-06, 'learning_rate/action_model': 9.610852663947583e-05, 'epoch': 0.48})
+[2026-07-28 17:19:00,220][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.0514657199382782, 'train/grad_norm_pre_clip': 1.7901288270950317, 'timing/data': 0.00047558173537254333, 'timing/model': 0.6205156864598393, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.48})
+[2026-07-28 17:19:09,240][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.03804170340299606, 'train/grad_norm_pre_clip': 0.9001701474189758, 'timing/data': 0.000426621176302433, 'timing/model': 0.622656506486237, 'learning_rate/qwen_vl_interface': 9.607813451325431e-06, 'learning_rate/action_model': 9.60781345132543e-05, 'epoch': 0.48})
+[2026-07-28 17:19:18,264][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.053795669227838516, 'train/grad_norm_pre_clip': 1.0087436437606812, 'timing/data': 0.00041552726179361343, 'timing/model': 0.624210013076663, 'learning_rate/qwen_vl_interface': 9.606289602914113e-06, 'learning_rate/action_model': 9.606289602914112e-05, 'epoch': 0.48})
+[2026-07-28 17:19:28,027][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.06509105861186981, 'train/grad_norm_pre_clip': 1.856440544128418, 'timing/data': 0.0004097744822502136, 'timing/model': 0.6202355762943625, 'learning_rate/qwen_vl_interface': 9.604762927753768e-06, 'learning_rate/action_model': 9.604762927753768e-05, 'epoch': 0.48})
+[2026-07-28 17:19:37,253][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.029851675033569336, 'train/grad_norm_pre_clip': 0.6572431325912476, 'timing/data': 0.00037337373942136765, 'timing/model': 0.6198860229924321, 'learning_rate/qwen_vl_interface': 9.603233426835037e-06, 'learning_rate/action_model': 9.603233426835036e-05, 'epoch': 0.48})
+[2026-07-28 17:19:46,454][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.032379232347011566, 'train/grad_norm_pre_clip': 1.0845614671707153, 'timing/data': 0.0003652134910225868, 'timing/model': 0.6219884343445301, 'learning_rate/qwen_vl_interface': 9.601701101150399e-06, 'learning_rate/action_model': 9.601701101150398e-05, 'epoch': 0.48})
+[2026-07-28 17:19:55,696][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.038275256752967834, 'train/grad_norm_pre_clip': 0.7653951048851013, 'timing/data': 0.0004749428480863571, 'timing/model': 0.6221339302137494, 'learning_rate/qwen_vl_interface': 9.600165951694162e-06, 'learning_rate/action_model': 9.60016595169416e-05, 'epoch': 0.48})
+[2026-07-28 17:20:05,129][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.0706147849559784, 'train/grad_norm_pre_clip': 2.021550178527832, 'timing/data': 0.00039084814488887787, 'timing/model': 0.6208954341709614, 'learning_rate/qwen_vl_interface': 9.598627979462466e-06, 'learning_rate/action_model': 9.598627979462467e-05, 'epoch': 0.48})
+[2026-07-28 17:20:14,644][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.011397911235690117, 'train/grad_norm_pre_clip': 0.42607179284095764, 'timing/data': 0.000363900326192379, 'timing/model': 0.6226577796041965, 'learning_rate/qwen_vl_interface': 9.59708718545329e-06, 'learning_rate/action_model': 9.597087185453288e-05, 'epoch': 0.48})
+[2026-07-28 17:20:23,853][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.03728604316711426, 'train/grad_norm_pre_clip': 0.7107821106910706, 'timing/data': 0.00045280996710062027, 'timing/model': 0.621613085269928, 'learning_rate/qwen_vl_interface': 9.595543570666433e-06, 'learning_rate/action_model': 9.595543570666432e-05, 'epoch': 0.48})
+[2026-07-28 17:20:33,086][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.054517559707164764, 'train/grad_norm_pre_clip': 1.5548851490020752, 'timing/data': 0.00034562405198812485, 'timing/model': 0.6224086889997125, 'learning_rate/qwen_vl_interface': 9.593997136103533e-06, 'learning_rate/action_model': 9.593997136103533e-05, 'epoch': 0.48})
+[2026-07-28 17:20:42,505][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.05249437317252159, 'train/grad_norm_pre_clip': 1.4913171529769897, 'timing/data': 0.0003340383991599083, 'timing/model': 0.8511708304286003, 'learning_rate/qwen_vl_interface': 9.592447882768053e-06, 'learning_rate/action_model': 9.592447882768052e-05, 'epoch': 0.48})
+[2026-07-28 17:20:52,017][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.017810029909014702, 'train/grad_norm_pre_clip': 0.6759623885154724, 'timing/data': 0.00046949461102485657, 'timing/model': 0.617973580956459, 'learning_rate/qwen_vl_interface': 9.590895811665284e-06, 'learning_rate/action_model': 9.590895811665285e-05, 'epoch': 0.48})
+[2026-07-28 17:21:00,990][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.029049037024378777, 'train/grad_norm_pre_clip': 1.1561429500579834, 'timing/data': 0.00037871673703193665, 'timing/model': 0.6212858567014337, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.48})
+[2026-07-28 17:21:10,150][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.0394710935652256, 'train/grad_norm_pre_clip': 1.5207065343856812, 'timing/data': 0.0003673657774925232, 'timing/model': 0.6240498721599579, 'learning_rate/qwen_vl_interface': 9.587783220188212e-06, 'learning_rate/action_model': 9.587783220188212e-05, 'epoch': 0.48})
+[2026-07-28 17:21:19,316][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.036460697650909424, 'train/grad_norm_pre_clip': 0.8016583323478699, 'timing/data': 0.0003611752763390541, 'timing/model': 0.6204784065485001, 'learning_rate/qwen_vl_interface': 9.586222701833631e-06, 'learning_rate/action_model': 9.58622270183363e-05, 'epoch': 0.48})
+[2026-07-28 17:21:28,928][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.05383481830358505, 'train/grad_norm_pre_clip': 2.737788200378418, 'timing/data': 0.0003560008481144905, 'timing/model': 0.9077582098543644, 'learning_rate/qwen_vl_interface': 9.584659369751216e-06, 'learning_rate/action_model': 9.584659369751215e-05, 'epoch': 0.48})
+[2026-07-28 17:21:38,365][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.019991125911474228, 'train/grad_norm_pre_clip': 0.5884313583374023, 'timing/data': 0.0004081185907125473, 'timing/model': 0.6277634790167212, 'learning_rate/qwen_vl_interface': 9.583093224955397e-06, 'learning_rate/action_model': 9.583093224955397e-05, 'epoch': 0.48})
+[2026-07-28 17:21:47,560][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.047915972769260406, 'train/grad_norm_pre_clip': 2.0346086025238037, 'timing/data': 0.0005909977480769157, 'timing/model': 0.6204667650163174, 'learning_rate/qwen_vl_interface': 9.58152426846243e-06, 'learning_rate/action_model': 9.581524268462429e-05, 'epoch': 0.48})
+[2026-07-28 17:21:56,777][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.042558491230010986, 'train/grad_norm_pre_clip': 0.9890199303627014, 'timing/data': 0.000507112592458725, 'timing/model': 0.6255785441026092, 'learning_rate/qwen_vl_interface': 9.579952501290388e-06, 'learning_rate/action_model': 9.579952501290387e-05, 'epoch': 0.48})
+[2026-07-28 17:22:06,048][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.025338560342788696, 'train/grad_norm_pre_clip': 0.6034095287322998, 'timing/data': 0.0003736605867743492, 'timing/model': 0.6198196271434426, 'learning_rate/qwen_vl_interface': 9.57837792445918e-06, 'learning_rate/action_model': 9.578377924459181e-05, 'epoch': 0.48})
+[2026-07-28 17:22:15,941][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.0346541628241539, 'train/grad_norm_pre_clip': 0.8617803454399109, 'timing/data': 0.00037697330117225647, 'timing/model': 0.9680376332253218, 'learning_rate/qwen_vl_interface': 9.57680053899053e-06, 'learning_rate/action_model': 9.576800538990529e-05, 'epoch': 0.48})
+[2026-07-28 17:22:25,300][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.03753472864627838, 'train/grad_norm_pre_clip': 1.1097818613052368, 'timing/data': 0.0003601405769586563, 'timing/model': 0.6516786459833384, 'learning_rate/qwen_vl_interface': 9.575220345907984e-06, 'learning_rate/action_model': 9.575220345907984e-05, 'epoch': 0.48})
+[2026-07-28 17:22:34,537][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.04729309678077698, 'train/grad_norm_pre_clip': 1.8820682764053345, 'timing/data': 0.00035919900983572006, 'timing/model': 0.6233192570507526, 'learning_rate/qwen_vl_interface': 9.573637346236915e-06, 'learning_rate/action_model': 9.573637346236915e-05, 'epoch': 0.48})
+[2026-07-28 17:22:43,498][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.03775458037853241, 'train/grad_norm_pre_clip': 1.3223755359649658, 'timing/data': 0.00039378367364406586, 'timing/model': 0.6203534826636314, 'learning_rate/qwen_vl_interface': 9.572051541004513e-06, 'learning_rate/action_model': 9.572051541004513e-05, 'epoch': 0.48})
+[2026-07-28 17:22:53,053][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.011479522101581097, 'train/grad_norm_pre_clip': 0.5435776710510254, 'timing/data': 0.00036079157143831253, 'timing/model': 0.6731433812528849, 'learning_rate/qwen_vl_interface': 9.57046293123979e-06, 'learning_rate/action_model': 9.570462931239789e-05, 'epoch': 0.48})
+[2026-07-28 17:23:02,353][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.01282944530248642, 'train/grad_norm_pre_clip': 1.0114268064498901, 'timing/data': 0.0003924556076526642, 'timing/model': 0.6262995088472962, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.48})
+[2026-07-28 17:23:11,615][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.02186857908964157, 'train/grad_norm_pre_clip': 0.7838981747627258, 'timing/data': 0.0003660190850496292, 'timing/model': 0.6219891728833318, 'learning_rate/qwen_vl_interface': 9.567277302238524e-06, 'learning_rate/action_model': 9.567277302238524e-05, 'epoch': 0.48})
+[2026-07-28 17:23:20,812][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.038934264332056046, 'train/grad_norm_pre_clip': 1.3889843225479126, 'timing/data': 0.0003333454951643944, 'timing/model': 0.6281535429880023, 'learning_rate/qwen_vl_interface': 9.565680285069104e-06, 'learning_rate/action_model': 9.565680285069103e-05, 'epoch': 0.48})
+[2026-07-28 17:23:30,314][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.026529211550951004, 'train/grad_norm_pre_clip': 1.2063301801681519, 'timing/data': 0.0004704669117927551, 'timing/model': 0.9037331519648433, 'learning_rate/qwen_vl_interface': 9.564080467501601e-06, 'learning_rate/action_model': 9.5640804675016e-05, 'epoch': 0.48})
+[2026-07-28 17:23:39,505][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.026871709153056145, 'train/grad_norm_pre_clip': 0.9531985521316528, 'timing/data': 0.00034994352608919144, 'timing/model': 0.6198089392855763, 'learning_rate/qwen_vl_interface': 9.562477850574118e-06, 'learning_rate/action_model': 9.562477850574119e-05, 'epoch': 0.48})
+[2026-07-28 17:23:48,753][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.057441793382167816, 'train/grad_norm_pre_clip': 2.06360125541687, 'timing/data': 0.00035242363810539246, 'timing/model': 0.6177581017836928, 'learning_rate/qwen_vl_interface': 9.56087243532658e-06, 'learning_rate/action_model': 9.56087243532658e-05, 'epoch': 0.48})
+[2026-07-28 17:23:58,042][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.024183372035622597, 'train/grad_norm_pre_clip': 1.2136458158493042, 'timing/data': 0.00036099646240472794, 'timing/model': 0.6180834854021668, 'learning_rate/qwen_vl_interface': 9.55926422280072e-06, 'learning_rate/action_model': 9.55926422280072e-05, 'epoch': 0.48})
+[2026-07-28 17:24:07,405][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.06884845346212387, 'train/grad_norm_pre_clip': 2.275374174118042, 'timing/data': 0.0004123356193304062, 'timing/model': 0.8062539249658585, 'learning_rate/qwen_vl_interface': 9.55765321404009e-06, 'learning_rate/action_model': 9.557653214040091e-05, 'epoch': 0.48})
+[2026-07-28 17:24:16,560][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.012359930202364922, 'train/grad_norm_pre_clip': 0.8066580295562744, 'timing/data': 0.00032880622893571854, 'timing/model': 0.6203105049207807, 'learning_rate/qwen_vl_interface': 9.556039410090061e-06, 'learning_rate/action_model': 9.556039410090061e-05, 'epoch': 0.49})
+[2026-07-28 17:24:25,960][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.07312402874231339, 'train/grad_norm_pre_clip': 1.9647371768951416, 'timing/data': 0.00034235045313835144, 'timing/model': 0.6207858063280582, 'learning_rate/qwen_vl_interface': 9.554422811997804e-06, 'learning_rate/action_model': 9.554422811997805e-05, 'epoch': 0.49})
+[2026-07-28 17:24:35,105][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.059147823601961136, 'train/grad_norm_pre_clip': 1.4502568244934082, 'timing/data': 0.000463634729385376, 'timing/model': 0.6199633656069636, 'learning_rate/qwen_vl_interface': 9.55280342081232e-06, 'learning_rate/action_model': 9.55280342081232e-05, 'epoch': 0.49})
+[2026-07-28 17:24:44,544][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.021848611533641815, 'train/grad_norm_pre_clip': 0.6665990352630615, 'timing/data': 0.00035119615495204926, 'timing/model': 0.8689403142780066, 'learning_rate/qwen_vl_interface': 9.55118123758441e-06, 'learning_rate/action_model': 9.551181237584409e-05, 'epoch': 0.49})
+[2026-07-28 17:24:53,745][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.027589991688728333, 'train/grad_norm_pre_clip': 0.7517790198326111, 'timing/data': 0.00042362138628959656, 'timing/model': 0.6204091953113675, 'learning_rate/qwen_vl_interface': 9.549556263366691e-06, 'learning_rate/action_model': 9.549556263366692e-05, 'epoch': 0.49})
+[2026-07-28 17:25:02,911][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.05931227281689644, 'train/grad_norm_pre_clip': 2.2105002403259277, 'timing/data': 0.00038159172981977463, 'timing/model': 0.6186850778758526, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.49})
+[2026-07-28 17:25:12,364][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.019353363662958145, 'train/grad_norm_pre_clip': 1.1178263425827026, 'timing/data': 0.00037543196231126785, 'timing/model': 0.6197647461667657, 'learning_rate/qwen_vl_interface': 9.546297946181354e-06, 'learning_rate/action_model': 9.546297946181354e-05, 'epoch': 0.49})
+[2026-07-28 17:25:21,849][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.04589848965406418, 'train/grad_norm_pre_clip': 1.732958197593689, 'timing/data': 0.00037667620927095413, 'timing/model': 0.6258744709193707, 'learning_rate/qwen_vl_interface': 9.544664605328022e-06, 'learning_rate/action_model': 9.544664605328022e-05, 'epoch': 0.49})
+[2026-07-28 17:25:31,096][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.043951213359832764, 'train/grad_norm_pre_clip': 1.5010130405426025, 'timing/data': 0.00035487115383148193, 'timing/model': 0.6184369651600718, 'learning_rate/qwen_vl_interface': 9.543028477713455e-06, 'learning_rate/action_model': 9.543028477713453e-05, 'epoch': 0.49})
+[2026-07-28 17:25:40,354][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.04113931953907013, 'train/grad_norm_pre_clip': 2.1397459506988525, 'timing/data': 0.00033646170049905777, 'timing/model': 0.6203877935186028, 'learning_rate/qwen_vl_interface': 9.541389564399316e-06, 'learning_rate/action_model': 9.541389564399315e-05, 'epoch': 0.49})
+[2026-07-28 17:25:49,765][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.01939246617257595, 'train/grad_norm_pre_clip': 0.9046984314918518, 'timing/data': 0.0003323480486869812, 'timing/model': 0.6287673376500607, 'learning_rate/qwen_vl_interface': 9.53974786644908e-06, 'learning_rate/action_model': 9.53974786644908e-05, 'epoch': 0.49})
+[2026-07-28 17:25:59,173][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.01853976957499981, 'train/grad_norm_pre_clip': 0.9564579725265503, 'timing/data': 0.0004911534488201141, 'timing/model': 0.6260829288512468, 'learning_rate/qwen_vl_interface': 9.538103384928029e-06, 'learning_rate/action_model': 9.538103384928027e-05, 'epoch': 0.49})
+[2026-07-28 17:26:08,380][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.05840914696455002, 'train/grad_norm_pre_clip': 1.7571368217468262, 'timing/data': 0.0003711758181452751, 'timing/model': 0.6292667370289564, 'learning_rate/qwen_vl_interface': 9.536456120903244e-06, 'learning_rate/action_model': 9.536456120903244e-05, 'epoch': 0.49})
+[2026-07-28 17:26:17,800][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.060864731669425964, 'train/grad_norm_pre_clip': 3.2404956817626953, 'timing/data': 0.00042038969695568085, 'timing/model': 0.8437441270798445, 'learning_rate/qwen_vl_interface': 9.534806075443619e-06, 'learning_rate/action_model': 9.534806075443619e-05, 'epoch': 0.49})
+[2026-07-28 17:26:27,220][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.01582379825413227, 'train/grad_norm_pre_clip': 0.7594085931777954, 'timing/data': 0.00048490893095731735, 'timing/model': 0.8054251987487078, 'learning_rate/qwen_vl_interface': 9.533153249619853e-06, 'learning_rate/action_model': 9.533153249619853e-05, 'epoch': 0.49})
+[2026-07-28 17:26:36,377][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.049980200827121735, 'train/grad_norm_pre_clip': 2.443028450012207, 'timing/data': 0.0004011914134025574, 'timing/model': 0.6180223077535629, 'learning_rate/qwen_vl_interface': 9.531497644504444e-06, 'learning_rate/action_model': 9.531497644504445e-05, 'epoch': 0.49})
+[2026-07-28 17:26:45,526][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.018532592803239822, 'train/grad_norm_pre_clip': 0.9076560735702515, 'timing/data': 0.00041675660759210587, 'timing/model': 0.6199640529230237, 'learning_rate/qwen_vl_interface': 9.529839261171698e-06, 'learning_rate/action_model': 9.529839261171696e-05, 'epoch': 0.49})
+[2026-07-28 17:26:54,687][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.026501808315515518, 'train/grad_norm_pre_clip': 0.7796621918678284, 'timing/data': 0.00033278390765190125, 'timing/model': 0.6180003620684147, 'learning_rate/qwen_vl_interface': 9.52817810069772e-06, 'learning_rate/action_model': 9.52817810069772e-05, 'epoch': 0.49})
+[2026-07-28 17:27:04,313][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.062062885612249374, 'train/grad_norm_pre_clip': 1.684309720993042, 'timing/data': 0.00035346392542123795, 'timing/model': 0.8691748874261975, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.49})
+[2026-07-28 17:27:13,599][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.020104365423321724, 'train/grad_norm_pre_clip': 1.18423330783844, 'timing/data': 0.00046752579510211945, 'timing/model': 0.6257330225780606, 'learning_rate/qwen_vl_interface': 9.524847452639509e-06, 'learning_rate/action_model': 9.524847452639509e-05, 'epoch': 0.49})
+[2026-07-28 17:27:22,809][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.0443241223692894, 'train/grad_norm_pre_clip': 1.1188595294952393, 'timing/data': 0.00036552175879478455, 'timing/model': 0.6224994324147701, 'learning_rate/qwen_vl_interface': 9.523177967216496e-06, 'learning_rate/action_model': 9.523177967216496e-05, 'epoch': 0.49})
+[2026-07-28 17:27:32,045][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.04404916614294052, 'train/grad_norm_pre_clip': 1.565969467163086, 'timing/data': 0.0003499360755085945, 'timing/model': 0.6229620641097426, 'learning_rate/qwen_vl_interface': 9.521505708974694e-06, 'learning_rate/action_model': 9.521505708974694e-05, 'epoch': 0.49})
+[2026-07-28 17:27:41,202][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.030354337766766548, 'train/grad_norm_pre_clip': 0.8066161274909973, 'timing/data': 0.00039743632078170776, 'timing/model': 0.6202689809724689, 'learning_rate/qwen_vl_interface': 9.519830678999214e-06, 'learning_rate/action_model': 9.519830678999213e-05, 'epoch': 0.49})
+[2026-07-28 17:27:50,845][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.026901237666606903, 'train/grad_norm_pre_clip': 1.0189990997314453, 'timing/data': 0.0003704754635691643, 'timing/model': 0.8696014238521457, 'learning_rate/qwen_vl_interface': 9.518152878376961e-06, 'learning_rate/action_model': 9.51815287837696e-05, 'epoch': 0.49})
+[2026-07-28 17:28:00,094][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.029755156487226486, 'train/grad_norm_pre_clip': 0.8727912306785583, 'timing/data': 0.0004378771409392357, 'timing/model': 0.6200621984899044, 'learning_rate/qwen_vl_interface': 9.516472308196646e-06, 'learning_rate/action_model': 9.516472308196645e-05, 'epoch': 0.49})
+[2026-07-28 17:28:09,263][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.052864961326122284, 'train/grad_norm_pre_clip': 1.8462843894958496, 'timing/data': 0.00045035500079393387, 'timing/model': 0.6193247837945819, 'learning_rate/qwen_vl_interface': 9.51478896954877e-06, 'learning_rate/action_model': 9.51478896954877e-05, 'epoch': 0.49})
+[2026-07-28 17:28:18,481][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.03689500689506531, 'train/grad_norm_pre_clip': 1.44051194190979, 'timing/data': 0.00044091977179050446, 'timing/model': 0.6191241145133972, 'learning_rate/qwen_vl_interface': 9.51310286352563e-06, 'learning_rate/action_model': 9.51310286352563e-05, 'epoch': 0.49})
+[2026-07-28 17:28:27,703][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.04403954744338989, 'train/grad_norm_pre_clip': 0.9056105613708496, 'timing/data': 0.0004476262256503105, 'timing/model': 0.6188744660466909, 'learning_rate/qwen_vl_interface': 9.51141399122133e-06, 'learning_rate/action_model': 9.511413991221331e-05, 'epoch': 0.49})
+[2026-07-28 17:28:37,168][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.03605687618255615, 'train/grad_norm_pre_clip': 1.0488100051879883, 'timing/data': 0.00039317086338996887, 'timing/model': 0.6213855799287558, 'learning_rate/qwen_vl_interface': 9.509722353731757e-06, 'learning_rate/action_model': 9.509722353731756e-05, 'epoch': 0.49})
+[2026-07-28 17:28:46,396][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.017102397978305817, 'train/grad_norm_pre_clip': 0.734698474407196, 'timing/data': 0.0005038194358348846, 'timing/model': 0.6214937726035714, 'learning_rate/qwen_vl_interface': 9.508027952154595e-06, 'learning_rate/action_model': 9.508027952154593e-05, 'epoch': 0.49})
+[2026-07-28 17:28:55,843][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.026767291128635406, 'train/grad_norm_pre_clip': 0.8437949419021606, 'timing/data': 0.00036255456507205963, 'timing/model': 0.6186700742691755, 'learning_rate/qwen_vl_interface': 9.506330787589321e-06, 'learning_rate/action_model': 9.506330787589322e-05, 'epoch': 0.49})
+[2026-07-28 17:29:05,242][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.062482692301273346, 'train/grad_norm_pre_clip': 1.392547369003296, 'timing/data': 0.0003676917403936386, 'timing/model': 0.6203821441158652, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.49})
+[2026-07-28 17:29:14,520][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.018451493233442307, 'train/grad_norm_pre_clip': 1.2135640382766724, 'timing/data': 0.00036925170570611954, 'timing/model': 0.6211763080209494, 'learning_rate/qwen_vl_interface': 9.50292817390133e-06, 'learning_rate/action_model': 9.502928173901329e-05, 'epoch': 0.49})
+[2026-07-28 17:29:23,673][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.029790829867124557, 'train/grad_norm_pre_clip': 0.6462365388870239, 'timing/data': 0.0003439914435148239, 'timing/model': 0.6199908554553986, 'learning_rate/qwen_vl_interface': 9.501222726986525e-06, 'learning_rate/action_model': 9.501222726986524e-05, 'epoch': 0.49})
+[2026-07-28 17:29:32,890][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.029786942526698112, 'train/grad_norm_pre_clip': 1.0993340015411377, 'timing/data': 0.0003358796238899231, 'timing/model': 0.6615610271692276, 'learning_rate/qwen_vl_interface': 9.49951452149945e-06, 'learning_rate/action_model': 9.499514521499451e-05, 'epoch': 0.49})
+[2026-07-28 17:29:42,362][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.02905382215976715, 'train/grad_norm_pre_clip': 0.9702856540679932, 'timing/data': 0.00036301277577877045, 'timing/model': 0.6177955493330956, 'learning_rate/qwen_vl_interface': 9.497803558548539e-06, 'learning_rate/action_model': 9.497803558548539e-05, 'epoch': 0.49})
+[2026-07-28 17:29:51,488][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.05528390407562256, 'train/grad_norm_pre_clip': 1.3375952243804932, 'timing/data': 0.00035901181399822235, 'timing/model': 0.6182274725288153, 'learning_rate/qwen_vl_interface': 9.496089839244015e-06, 'learning_rate/action_model': 9.496089839244015e-05, 'epoch': 0.49})
+[2026-07-28 17:30:00,627][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.059383876621723175, 'train/grad_norm_pre_clip': 2.7567298412323, 'timing/data': 0.00043180957436561584, 'timing/model': 0.6218842016533017, 'learning_rate/qwen_vl_interface': 9.494373364697895e-06, 'learning_rate/action_model': 9.494373364697894e-05, 'epoch': 0.49})
+[2026-07-28 17:30:09,759][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.07181711494922638, 'train/grad_norm_pre_clip': 3.084707498550415, 'timing/data': 0.0006446586921811104, 'timing/model': 0.6226191874593496, 'learning_rate/qwen_vl_interface': 9.492654136023978e-06, 'learning_rate/action_model': 9.492654136023979e-05, 'epoch': 0.49})
+[2026-07-28 17:30:19,404][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.016084391623735428, 'train/grad_norm_pre_clip': 0.5710883736610413, 'timing/data': 0.00037608295679092407, 'timing/model': 0.6522887628525496, 'learning_rate/qwen_vl_interface': 9.490932154337855e-06, 'learning_rate/action_model': 9.490932154337855e-05, 'epoch': 0.49})
+[2026-07-28 17:30:28,851][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.02728550136089325, 'train/grad_norm_pre_clip': 0.7939621210098267, 'timing/data': 0.000483168289065361, 'timing/model': 0.6212706714868546, 'learning_rate/qwen_vl_interface': 9.4892074207569e-06, 'learning_rate/action_model': 9.489207420756899e-05, 'epoch': 0.49})
+[2026-07-28 17:30:38,024][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.03623931109905243, 'train/grad_norm_pre_clip': 1.0114595890045166, 'timing/data': 0.00035839155316352844, 'timing/model': 0.6187506159767509, 'learning_rate/qwen_vl_interface': 9.487479936400274e-06, 'learning_rate/action_model': 9.487479936400272e-05, 'epoch': 0.49})
+[2026-07-28 17:30:46,944][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.027577733621001244, 'train/grad_norm_pre_clip': 1.2780746221542358, 'timing/data': 0.0004477696493268013, 'timing/model': 0.6204907055944204, 'learning_rate/qwen_vl_interface': 9.485749702388923e-06, 'learning_rate/action_model': 9.485749702388923e-05, 'epoch': 0.49})
+[2026-07-28 17:30:56,381][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.02939605340361595, 'train/grad_norm_pre_clip': 2.0470292568206787, 'timing/data': 0.0003711320459842682, 'timing/model': 0.8469886165112257, 'learning_rate/qwen_vl_interface': 9.484016719845576e-06, 'learning_rate/action_model': 9.484016719845576e-05, 'epoch': 0.49})
+[2026-07-28 17:31:05,504][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.02362537384033203, 'train/grad_norm_pre_clip': 0.9219607710838318, 'timing/data': 0.0004883212968707085, 'timing/model': 0.6184319779276848, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.49})
+[2026-07-28 17:31:14,773][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.029279347509145737, 'train/grad_norm_pre_clip': 0.6381241679191589, 'timing/data': 0.0003605801612138748, 'timing/model': 0.6335135428234935, 'learning_rate/qwen_vl_interface': 9.480542513662734e-06, 'learning_rate/action_model': 9.480542513662734e-05, 'epoch': 0.49})
+[2026-07-28 17:31:23,905][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.01799488067626953, 'train/grad_norm_pre_clip': 0.7811532020568848, 'timing/data': 0.000349326990544796, 'timing/model': 0.6170207299292088, 'learning_rate/qwen_vl_interface': 9.478801292277617e-06, 'learning_rate/action_model': 9.478801292277616e-05, 'epoch': 0.49})
+[2026-07-28 17:31:33,343][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.06869243830442429, 'train/grad_norm_pre_clip': 1.440587043762207, 'timing/data': 0.0003714337944984436, 'timing/model': 0.6204236494377255, 'learning_rate/qwen_vl_interface': 9.477057326869249e-06, 'learning_rate/action_model': 9.477057326869249e-05, 'epoch': 0.49})
+[2026-07-28 17:31:42,479][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.03605141118168831, 'train/grad_norm_pre_clip': 1.1330686807632446, 'timing/data': 0.0003727646544575691, 'timing/model': 0.6250210460275412, 'learning_rate/qwen_vl_interface': 9.475310618569274e-06, 'learning_rate/action_model': 9.475310618569274e-05, 'epoch': 0.49})
+[2026-07-28 17:31:51,847][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.07377550005912781, 'train/grad_norm_pre_clip': 1.7146507501602173, 'timing/data': 0.00034297723323106766, 'timing/model': 0.6225939486175776, 'learning_rate/qwen_vl_interface': 9.473561168511115e-06, 'learning_rate/action_model': 9.473561168511115e-05, 'epoch': 0.49})
+[2026-07-28 17:32:01,072][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.051371753215789795, 'train/grad_norm_pre_clip': 1.2195236682891846, 'timing/data': 0.0003688354045152664, 'timing/model': 0.6217223340645432, 'learning_rate/qwen_vl_interface': 9.471808977829967e-06, 'learning_rate/action_model': 9.471808977829967e-05, 'epoch': 0.49})
+[2026-07-28 17:32:10,500][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.05923699960112572, 'train/grad_norm_pre_clip': 1.1793298721313477, 'timing/data': 0.00034936703741550446, 'timing/model': 0.6202127141878009, 'learning_rate/qwen_vl_interface': 9.470054047662808e-06, 'learning_rate/action_model': 9.470054047662808e-05, 'epoch': 0.49})
+[2026-07-28 17:32:19,621][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.09138703346252441, 'train/grad_norm_pre_clip': 1.8390531539916992, 'timing/data': 0.00035237986594438553, 'timing/model': 0.6186245810240507, 'learning_rate/qwen_vl_interface': 9.468296379148394e-06, 'learning_rate/action_model': 9.468296379148393e-05, 'epoch': 0.49})
+[2026-07-28 17:32:29,006][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.05132686346769333, 'train/grad_norm_pre_clip': 1.2260432243347168, 'timing/data': 0.00035523343831300735, 'timing/model': 0.849585154093802, 'learning_rate/qwen_vl_interface': 9.466535973427259e-06, 'learning_rate/action_model': 9.466535973427258e-05, 'epoch': 0.49})
+[2026-07-28 17:32:38,439][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.0516740083694458, 'train/grad_norm_pre_clip': 2.2346606254577637, 'timing/data': 0.0003482382744550705, 'timing/model': 0.8388548558577895, 'learning_rate/qwen_vl_interface': 9.464772831641706e-06, 'learning_rate/action_model': 9.464772831641707e-05, 'epoch': 0.5})
+[2026-07-28 17:32:47,571][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.052161555737257004, 'train/grad_norm_pre_clip': 1.8963944911956787, 'timing/data': 0.000360855832695961, 'timing/model': 0.6193886697292328, 'learning_rate/qwen_vl_interface': 9.463006954935823e-06, 'learning_rate/action_model': 9.463006954935822e-05, 'epoch': 0.5})
+[2026-07-28 17:32:56,702][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.029949713498353958, 'train/grad_norm_pre_clip': 1.1801143884658813, 'timing/data': 0.00032803695648908615, 'timing/model': 0.6198081625625491, 'learning_rate/qwen_vl_interface': 9.461238344455468e-06, 'learning_rate/action_model': 9.461238344455468e-05, 'epoch': 0.5})
+[2026-07-28 17:33:06,059][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.06097372993826866, 'train/grad_norm_pre_clip': 2.3525779247283936, 'timing/data': 0.0004011690616607666, 'timing/model': 0.6186606390401721, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.5})
+[2026-07-28 17:33:15,492][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.041282638907432556, 'train/grad_norm_pre_clip': 0.6770138144493103, 'timing/data': 0.0003784317523241043, 'timing/model': 0.9099285779520869, 'learning_rate/qwen_vl_interface': 9.457692926763638e-06, 'learning_rate/action_model': 9.457692926763638e-05, 'epoch': 0.5})
+[2026-07-28 17:33:24,616][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.027312975376844406, 'train/grad_norm_pre_clip': 0.9020497798919678, 'timing/data': 0.0003482634201645851, 'timing/model': 0.619217955507338, 'learning_rate/qwen_vl_interface': 9.455916121852748e-06, 'learning_rate/action_model': 9.455916121852747e-05, 'epoch': 0.5})
+[2026-07-28 17:33:34,016][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.05335528403520584, 'train/grad_norm_pre_clip': 0.7006136178970337, 'timing/data': 0.0003274502232670784, 'timing/model': 0.8523511188104749, 'learning_rate/qwen_vl_interface': 9.454136587768549e-06, 'learning_rate/action_model': 9.454136587768548e-05, 'epoch': 0.5})
+[2026-07-28 17:33:43,159][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.03310827910900116, 'train/grad_norm_pre_clip': 0.5842478275299072, 'timing/data': 0.0003304649144411087, 'timing/model': 0.6216065008193254, 'learning_rate/qwen_vl_interface': 9.452354325665763e-06, 'learning_rate/action_model': 9.452354325665763e-05, 'epoch': 0.5})
+[2026-07-28 17:33:52,523][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.018769871443510056, 'train/grad_norm_pre_clip': 1.011816382408142, 'timing/data': 0.00044325273483991623, 'timing/model': 0.6215015100315213, 'learning_rate/qwen_vl_interface': 9.450569336700878e-06, 'learning_rate/action_model': 9.450569336700878e-05, 'epoch': 0.5})
+[2026-07-28 17:34:01,706][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.03959355503320694, 'train/grad_norm_pre_clip': 0.999953031539917, 'timing/data': 0.0004102131351828575, 'timing/model': 0.6173423612490296, 'learning_rate/qwen_vl_interface': 9.448781622032155e-06, 'learning_rate/action_model': 9.448781622032155e-05, 'epoch': 0.5})
+[2026-07-28 17:34:10,873][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.029179517179727554, 'train/grad_norm_pre_clip': 0.919293224811554, 'timing/data': 0.00037401914596557617, 'timing/model': 0.6195633504539728, 'learning_rate/qwen_vl_interface': 9.446991182819624e-06, 'learning_rate/action_model': 9.446991182819623e-05, 'epoch': 0.5})
+[2026-07-28 17:34:20,256][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.030274413526058197, 'train/grad_norm_pre_clip': 1.808396577835083, 'timing/data': 0.00035202596336603165, 'timing/model': 0.6171335354447365, 'learning_rate/qwen_vl_interface': 9.44519802022508e-06, 'learning_rate/action_model': 9.44519802022508e-05, 'epoch': 0.5})
+[2026-07-28 17:34:29,632][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.017687899991869926, 'train/grad_norm_pre_clip': 0.5474141836166382, 'timing/data': 0.00037448201328516006, 'timing/model': 0.6178755220025778, 'learning_rate/qwen_vl_interface': 9.443402135412088e-06, 'learning_rate/action_model': 9.443402135412088e-05, 'epoch': 0.5})
+[2026-07-28 17:34:38,550][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.045686639845371246, 'train/grad_norm_pre_clip': 1.553598165512085, 'timing/data': 0.00036361441016197205, 'timing/model': 0.6183858327567577, 'learning_rate/qwen_vl_interface': 9.441603529545977e-06, 'learning_rate/action_model': 9.441603529545976e-05, 'epoch': 0.5})
+[2026-07-28 17:34:47,513][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.04489678516983986, 'train/grad_norm_pre_clip': 1.245391845703125, 'timing/data': 0.00034213345497846603, 'timing/model': 0.6182856587693095, 'learning_rate/qwen_vl_interface': 9.439802203793843e-06, 'learning_rate/action_model': 9.439802203793842e-05, 'epoch': 0.5})
+[2026-07-28 17:34:56,848][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.026406915858387947, 'train/grad_norm_pre_clip': 0.9221710562705994, 'timing/data': 0.00035339221358299255, 'timing/model': 0.7945212330669165, 'learning_rate/qwen_vl_interface': 9.437998159324547e-06, 'learning_rate/action_model': 9.437998159324547e-05, 'epoch': 0.5})
+[2026-07-28 17:35:06,207][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.06361966580152512, 'train/grad_norm_pre_clip': 2.229741334915161, 'timing/data': 0.0003322754055261612, 'timing/model': 0.8496164297685027, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.5})
+[2026-07-28 17:35:15,346][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.02190038375556469, 'train/grad_norm_pre_clip': 1.0103201866149902, 'timing/data': 0.00035298336297273636, 'timing/model': 0.6197633482515812, 'learning_rate/qwen_vl_interface': 9.434381918918731e-06, 'learning_rate/action_model': 9.434381918918731e-05, 'epoch': 0.5})
+[2026-07-28 17:35:24,552][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.033152997493743896, 'train/grad_norm_pre_clip': 1.211548089981079, 'timing/data': 0.0003724880516529083, 'timing/model': 0.6207503275945783, 'learning_rate/qwen_vl_interface': 9.432569725328753e-06, 'learning_rate/action_model': 9.432569725328753e-05, 'epoch': 0.5})
+[2026-07-28 17:35:34,001][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.01737341284751892, 'train/grad_norm_pre_clip': 0.780230700969696, 'timing/data': 0.00036292895674705505, 'timing/model': 0.8907879143953323, 'learning_rate/qwen_vl_interface': 9.430754817714686e-06, 'learning_rate/action_model': 9.430754817714687e-05, 'epoch': 0.5})
+[2026-07-28 17:35:43,171][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.07956942915916443, 'train/grad_norm_pre_clip': 2.2452666759490967, 'timing/data': 0.00034185126423835754, 'timing/model': 0.6246242988854647, 'learning_rate/qwen_vl_interface': 9.42893719725421e-06, 'learning_rate/action_model': 9.42893719725421e-05, 'epoch': 0.5})
+[2026-07-28 17:35:52,547][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.13907325267791748, 'train/grad_norm_pre_clip': 3.9091711044311523, 'timing/data': 0.00034416839480400085, 'timing/model': 0.6228688433766365, 'learning_rate/qwen_vl_interface': 9.427116865126757e-06, 'learning_rate/action_model': 9.427116865126756e-05, 'epoch': 0.5})
+[2026-07-28 17:36:01,731][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.031389594078063965, 'train/grad_norm_pre_clip': 1.3472201824188232, 'timing/data': 0.00033770594745874405, 'timing/model': 0.6177778020501137, 'learning_rate/qwen_vl_interface': 9.42529382251352e-06, 'learning_rate/action_model': 9.42529382251352e-05, 'epoch': 0.5})
+[2026-07-28 17:36:10,859][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.04889337718486786, 'train/grad_norm_pre_clip': 3.261641502380371, 'timing/data': 0.0003332551568746567, 'timing/model': 0.6196740511804819, 'learning_rate/qwen_vl_interface': 9.423468070597453e-06, 'learning_rate/action_model': 9.423468070597452e-05, 'epoch': 0.5})
+[2026-07-28 17:36:20,438][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.024412214756011963, 'train/grad_norm_pre_clip': 0.9095411896705627, 'timing/data': 0.00032282527536153793, 'timing/model': 0.8375364495441318, 'learning_rate/qwen_vl_interface': 9.421639610563264e-06, 'learning_rate/action_model': 9.421639610563264e-05, 'epoch': 0.5})
+[2026-07-28 17:36:29,573][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.05326475948095322, 'train/grad_norm_pre_clip': 2.447021245956421, 'timing/data': 0.00033699721097946167, 'timing/model': 0.6183832390233874, 'learning_rate/qwen_vl_interface': 9.419808443597422e-06, 'learning_rate/action_model': 9.419808443597423e-05, 'epoch': 0.5})
+[2026-07-28 17:36:38,739][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.07770539820194244, 'train/grad_norm_pre_clip': 2.9367666244506836, 'timing/data': 0.0003663301467895508, 'timing/model': 0.6224498758092523, 'learning_rate/qwen_vl_interface': 9.417974570888152e-06, 'learning_rate/action_model': 9.417974570888153e-05, 'epoch': 0.5})
+[2026-07-28 17:36:47,996][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.054513297975063324, 'train/grad_norm_pre_clip': 1.6981995105743408, 'timing/data': 0.0003573233261704445, 'timing/model': 0.6221407419070601, 'learning_rate/qwen_vl_interface': 9.416137993625437e-06, 'learning_rate/action_model': 9.416137993625437e-05, 'epoch': 0.5})
+[2026-07-28 17:36:57,446][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.03129785135388374, 'train/grad_norm_pre_clip': 0.6489474177360535, 'timing/data': 0.00034117046743631363, 'timing/model': 0.6182777788490057, 'learning_rate/qwen_vl_interface': 9.414298713001007e-06, 'learning_rate/action_model': 9.414298713001006e-05, 'epoch': 0.5})
+[2026-07-28 17:37:06,597][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.022900698706507683, 'train/grad_norm_pre_clip': 0.46397024393081665, 'timing/data': 0.00034349411725997925, 'timing/model': 0.6222386443987489, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.5})
+[2026-07-28 17:37:15,737][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.03220502287149429, 'train/grad_norm_pre_clip': 0.8644846677780151, 'timing/data': 0.0003382610157132149, 'timing/model': 0.6194267058745027, 'learning_rate/qwen_vl_interface': 9.410612046442719e-06, 'learning_rate/action_model': 9.410612046442718e-05, 'epoch': 0.5})
+[2026-07-28 17:37:25,114][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.04567565768957138, 'train/grad_norm_pre_clip': 0.9179646372795105, 'timing/data': 0.0003387061879038811, 'timing/model': 0.6199620366096497, 'learning_rate/qwen_vl_interface': 9.408764662901097e-06, 'learning_rate/action_model': 9.408764662901098e-05, 'epoch': 0.5})
+[2026-07-28 17:37:34,254][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.028259342536330223, 'train/grad_norm_pre_clip': 0.8269843459129333, 'timing/data': 0.0004080822691321373, 'timing/model': 0.6189132304862142, 'learning_rate/qwen_vl_interface': 9.406914580782238e-06, 'learning_rate/action_model': 9.406914580782238e-05, 'epoch': 0.5})
+[2026-07-28 17:37:43,377][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.04125694930553436, 'train/grad_norm_pre_clip': 1.5187212228775024, 'timing/data': 0.0004363544285297394, 'timing/model': 0.6206660410389304, 'learning_rate/qwen_vl_interface': 9.405061801286638e-06, 'learning_rate/action_model': 9.405061801286638e-05, 'epoch': 0.5})
+[2026-07-28 17:37:52,539][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.07351091504096985, 'train/grad_norm_pre_clip': 1.769410490989685, 'timing/data': 0.00040114764124155045, 'timing/model': 0.6195728555321693, 'learning_rate/qwen_vl_interface': 9.403206325616546e-06, 'learning_rate/action_model': 9.403206325616545e-05, 'epoch': 0.5})
+[2026-07-28 17:38:01,730][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.013868056237697601, 'train/grad_norm_pre_clip': 0.7562159895896912, 'timing/data': 0.0003436524420976639, 'timing/model': 0.6203164653852582, 'learning_rate/qwen_vl_interface': 9.40134815497596e-06, 'learning_rate/action_model': 9.40134815497596e-05, 'epoch': 0.5})
+[2026-07-28 17:38:11,112][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.042592309415340424, 'train/grad_norm_pre_clip': 1.4248390197753906, 'timing/data': 0.0003723409026861191, 'timing/model': 0.8481219243258238, 'learning_rate/qwen_vl_interface': 9.399487290570629e-06, 'learning_rate/action_model': 9.399487290570629e-05, 'epoch': 0.5})
+[2026-07-28 17:38:20,480][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.07765045017004013, 'train/grad_norm_pre_clip': 2.827995538711548, 'timing/data': 0.0003831246867775917, 'timing/model': 0.8460376188158989, 'learning_rate/qwen_vl_interface': 9.397623733608042e-06, 'learning_rate/action_model': 9.397623733608042e-05, 'epoch': 0.5})
+[2026-07-28 17:38:29,637][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.04365839064121246, 'train/grad_norm_pre_clip': 2.373722553253174, 'timing/data': 0.0004120785742998123, 'timing/model': 0.6240352047607303, 'learning_rate/qwen_vl_interface': 9.395757485297447e-06, 'learning_rate/action_model': 9.395757485297446e-05, 'epoch': 0.5})
+[2026-07-28 17:38:38,825][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.04975956305861473, 'train/grad_norm_pre_clip': 1.737956166267395, 'timing/data': 0.00043100398033857346, 'timing/model': 0.6184351556003094, 'learning_rate/qwen_vl_interface': 9.393888546849828e-06, 'learning_rate/action_model': 9.393888546849828e-05, 'epoch': 0.5})
+[2026-07-28 17:38:48,016][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.04533601552248001, 'train/grad_norm_pre_clip': 1.0134743452072144, 'timing/data': 0.00035885442048311234, 'timing/model': 0.6183741092681885, 'learning_rate/qwen_vl_interface': 9.392016919477922e-06, 'learning_rate/action_model': 9.392016919477921e-05, 'epoch': 0.5})
+[2026-07-28 17:38:57,153][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.020539715886116028, 'train/grad_norm_pre_clip': 0.7923555970191956, 'timing/data': 0.000346597284078598, 'timing/model': 0.6190530927851796, 'learning_rate/qwen_vl_interface': 9.390142604396206e-06, 'learning_rate/action_model': 9.390142604396205e-05, 'epoch': 0.5})
+[2026-07-28 17:39:06,294][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.03869748115539551, 'train/grad_norm_pre_clip': 1.7433383464813232, 'timing/data': 0.0003653913736343384, 'timing/model': 0.6202786564826965, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.5})
+[2026-07-28 17:39:15,419][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.019051894545555115, 'train/grad_norm_pre_clip': 1.0130606889724731, 'timing/data': 0.0003386596217751503, 'timing/model': 0.6176054831594229, 'learning_rate/qwen_vl_interface': 9.386385915969975e-06, 'learning_rate/action_model': 9.386385915969976e-05, 'epoch': 0.5})
+[2026-07-28 17:39:24,338][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.058821775019168854, 'train/grad_norm_pre_clip': 1.4247674942016602, 'timing/data': 0.00034356117248535156, 'timing/model': 0.6181088294833899, 'learning_rate/qwen_vl_interface': 9.384503545063138e-06, 'learning_rate/action_model': 9.384503545063137e-05, 'epoch': 0.5})
+[2026-07-28 17:39:33,465][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.019243445247411728, 'train/grad_norm_pre_clip': 0.7609474062919617, 'timing/data': 0.00035903044044971466, 'timing/model': 0.6193481646478176, 'learning_rate/qwen_vl_interface': 9.382618491321837e-06, 'learning_rate/action_model': 9.382618491321837e-05, 'epoch': 0.5})
+[2026-07-28 17:39:43,067][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.037977151572704315, 'train/grad_norm_pre_clip': 0.8461247682571411, 'timing/data': 0.00040322355926036835, 'timing/model': 0.8515767008066177, 'learning_rate/qwen_vl_interface': 9.380730755969265e-06, 'learning_rate/action_model': 9.380730755969264e-05, 'epoch': 0.5})
+[2026-07-28 17:39:52,305][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.031313080340623856, 'train/grad_norm_pre_clip': 1.3750741481781006, 'timing/data': 0.00039967242628335953, 'timing/model': 0.6214354299008846, 'learning_rate/qwen_vl_interface': 9.378840340230353e-06, 'learning_rate/action_model': 9.378840340230352e-05, 'epoch': 0.5})
+[2026-07-28 17:40:01,694][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.0515202060341835, 'train/grad_norm_pre_clip': 1.3190803527832031, 'timing/data': 0.00032943859696388245, 'timing/model': 0.6195923499763012, 'learning_rate/qwen_vl_interface': 9.376947245331769e-06, 'learning_rate/action_model': 9.376947245331767e-05, 'epoch': 0.5})
+[2026-07-28 17:40:11,221][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.04660695791244507, 'train/grad_norm_pre_clip': 1.9175463914871216, 'timing/data': 0.00034546665847301483, 'timing/model': 0.8468554122373462, 'learning_rate/qwen_vl_interface': 9.375051472501922e-06, 'learning_rate/action_model': 9.375051472501923e-05, 'epoch': 0.5})
+[2026-07-28 17:40:20,672][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.032729167491197586, 'train/grad_norm_pre_clip': 1.1242530345916748, 'timing/data': 0.0003272872418165207, 'timing/model': 0.6198677746579051, 'learning_rate/qwen_vl_interface': 9.373153022970961e-06, 'learning_rate/action_model': 9.373153022970961e-05, 'epoch': 0.5})
+[2026-07-28 17:40:29,875][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.04945717006921768, 'train/grad_norm_pre_clip': 1.4108692407608032, 'timing/data': 0.00034153927117586136, 'timing/model': 0.6330002201721072, 'learning_rate/qwen_vl_interface': 9.371251897970768e-06, 'learning_rate/action_model': 9.371251897970768e-05, 'epoch': 0.5})
+[2026-07-28 17:40:39,047][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.028529709205031395, 'train/grad_norm_pre_clip': 0.9912776947021484, 'timing/data': 0.00034050364047288895, 'timing/model': 0.6171530140563846, 'learning_rate/qwen_vl_interface': 9.369348098734961e-06, 'learning_rate/action_model': 9.369348098734961e-05, 'epoch': 0.5})
+[2026-07-28 17:40:48,499][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.009321090765297413, 'train/grad_norm_pre_clip': 0.5333521962165833, 'timing/data': 0.00038711633533239365, 'timing/model': 0.8509434554725885, 'learning_rate/qwen_vl_interface': 9.367441626498897e-06, 'learning_rate/action_model': 9.367441626498896e-05, 'epoch': 0.5})
+[2026-07-28 17:40:57,664][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.03596067428588867, 'train/grad_norm_pre_clip': 1.499686360359192, 'timing/data': 0.0004214160144329071, 'timing/model': 0.6279402663931251, 'learning_rate/qwen_vl_interface': 9.365532482499663e-06, 'learning_rate/action_model': 9.365532482499662e-05, 'epoch': 0.5})
+[2026-07-28 17:41:07,080][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2750
+[2026-07-28 17:41:17,320][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2750
+[2026-07-28 17:53:05,488][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2750
+[2026-07-28 17:53:05,489][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.03419191762804985, 'train/grad_norm_pre_clip': 1.1293278932571411, 'eval/action_loss/seconds': 10.23517141956836, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.035046566277742386, 'eval/latency_3/loss': 0.035046566277742386, 'eval/flappy/loss': 0.035046566277742386, 'eval/flappy/latency_3/loss': 0.035046566277742386, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 256.680003028363, 'rl_games_eval/mid_train/mean_length': 2086.95, 'rl_games_eval/mid_train/std_reward': 167.62305705644414, 'rl_games_eval/mid_train/std_length': 1350.2858762128856, 'rl_games_eval/mid_train/macro_mean_reward': 256.680003028363, 'rl_games_eval/mid_train/macro_mean_length': 2086.95, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 256.680003028363, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2086.95, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 167.62305705644414, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1350.2858762128856, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 256.680003028363, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003370698541402817, 'timing/model': 0.6209904924035072, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.51})
+[2026-07-28 17:53:14,885][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.05924768000841141, 'train/grad_norm_pre_clip': 1.2197905778884888, 'timing/data': 0.0003766817972064018, 'timing/model': 0.6189627796411514, 'learning_rate/qwen_vl_interface': 9.361706184168714e-06, 'learning_rate/action_model': 9.361706184168713e-05, 'epoch': 0.51})
+[2026-07-28 17:53:24,012][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.06784985214471817, 'train/grad_norm_pre_clip': 2.118405818939209, 'timing/data': 0.0003659473732113838, 'timing/model': 0.6179878506809473, 'learning_rate/qwen_vl_interface': 9.359789032319843e-06, 'learning_rate/action_model': 9.359789032319842e-05, 'epoch': 0.51})
+[2026-07-28 17:53:33,380][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.02094777300953865, 'train/grad_norm_pre_clip': 1.5507222414016724, 'timing/data': 0.00033828336745500565, 'timing/model': 0.6209776373580098, 'learning_rate/qwen_vl_interface': 9.357869213673484e-06, 'learning_rate/action_model': 9.357869213673485e-05, 'epoch': 0.51})
+[2026-07-28 17:53:42,751][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.09352578222751617, 'train/grad_norm_pre_clip': 1.6802717447280884, 'timing/data': 0.0003442559391260147, 'timing/model': 0.62523705791682, 'learning_rate/qwen_vl_interface': 9.355946729475393e-06, 'learning_rate/action_model': 9.355946729475393e-05, 'epoch': 0.51})
+[2026-07-28 17:53:51,878][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.021029211580753326, 'train/grad_norm_pre_clip': 0.885733962059021, 'timing/data': 0.0003631496801972389, 'timing/model': 0.6182832503691316, 'learning_rate/qwen_vl_interface': 9.354021580973049e-06, 'learning_rate/action_model': 9.354021580973048e-05, 'epoch': 0.51})
+[2026-07-28 17:54:01,000][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.02287762612104416, 'train/grad_norm_pre_clip': 0.8339571356773376, 'timing/data': 0.0003583719953894615, 'timing/model': 0.6198090389370918, 'learning_rate/qwen_vl_interface': 9.352093769415655e-06, 'learning_rate/action_model': 9.352093769415655e-05, 'epoch': 0.51})
+[2026-07-28 17:54:10,487][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.027122478932142258, 'train/grad_norm_pre_clip': 0.9859631657600403, 'timing/data': 0.0003349706530570984, 'timing/model': 1.0002038832753897, 'learning_rate/qwen_vl_interface': 9.35016329605415e-06, 'learning_rate/action_model': 9.350163296054151e-05, 'epoch': 0.51})
+[2026-07-28 17:54:19,598][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.014866914600133896, 'train/grad_norm_pre_clip': 0.40582141280174255, 'timing/data': 0.00035511981695890427, 'timing/model': 0.6179866027086973, 'learning_rate/qwen_vl_interface': 9.348230162141198e-06, 'learning_rate/action_model': 9.348230162141198e-05, 'epoch': 0.51})
+[2026-07-28 17:54:28,769][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.023338880389928818, 'train/grad_norm_pre_clip': 1.1309700012207031, 'timing/data': 0.00032335519790649414, 'timing/model': 0.617423658259213, 'learning_rate/qwen_vl_interface': 9.346294368931188e-06, 'learning_rate/action_model': 9.346294368931188e-05, 'epoch': 0.51})
+[2026-07-28 17:54:37,960][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.04385489225387573, 'train/grad_norm_pre_clip': 1.8409333229064941, 'timing/data': 0.0004997318610548973, 'timing/model': 0.6188664939254522, 'learning_rate/qwen_vl_interface': 9.344355917680235e-06, 'learning_rate/action_model': 9.344355917680235e-05, 'epoch': 0.51})
+[2026-07-28 17:54:47,602][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.08490632474422455, 'train/grad_norm_pre_clip': 3.263455390930176, 'timing/data': 0.00032008904963731766, 'timing/model': 0.8534672129899263, 'learning_rate/qwen_vl_interface': 9.342414809646179e-06, 'learning_rate/action_model': 9.342414809646179e-05, 'epoch': 0.51})
+[2026-07-28 17:54:56,729][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.06508533656597137, 'train/grad_norm_pre_clip': 1.852298617362976, 'timing/data': 0.00031837355345487595, 'timing/model': 0.6210230821743608, 'learning_rate/qwen_vl_interface': 9.340471046088581e-06, 'learning_rate/action_model': 9.34047104608858e-05, 'epoch': 0.51})
+[2026-07-28 17:55:05,877][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.0567622110247612, 'train/grad_norm_pre_clip': 0.8265429735183716, 'timing/data': 0.0003595016896724701, 'timing/model': 0.6190725630149245, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.51})
+[2026-07-28 17:55:15,273][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.06259895861148834, 'train/grad_norm_pre_clip': 1.2709765434265137, 'timing/data': 0.0003385143354535103, 'timing/model': 0.8501892164349556, 'learning_rate/qwen_vl_interface': 9.336575557449641e-06, 'learning_rate/action_model': 9.336575557449641e-05, 'epoch': 0.51})
+[2026-07-28 17:55:24,439][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.01984269917011261, 'train/grad_norm_pre_clip': 1.182454228401184, 'timing/data': 0.0003372449427843094, 'timing/model': 0.6184513941407204, 'learning_rate/qwen_vl_interface': 9.334623834896039e-06, 'learning_rate/action_model': 9.334623834896038e-05, 'epoch': 0.51})
+[2026-07-28 17:55:33,818][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.0794503390789032, 'train/grad_norm_pre_clip': 3.2467262744903564, 'timing/data': 0.0003518667072057724, 'timing/model': 0.8669307762756944, 'learning_rate/qwen_vl_interface': 9.332669461874376e-06, 'learning_rate/action_model': 9.332669461874374e-05, 'epoch': 0.51})
+[2026-07-28 17:55:42,969][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.06523406505584717, 'train/grad_norm_pre_clip': 1.052198052406311, 'timing/data': 0.0003263000398874283, 'timing/model': 0.6241019368171692, 'learning_rate/qwen_vl_interface': 9.330712439652824e-06, 'learning_rate/action_model': 9.330712439652823e-05, 'epoch': 0.51})
+[2026-07-28 17:55:52,126][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.05681600421667099, 'train/grad_norm_pre_clip': 1.3037453889846802, 'timing/data': 0.0003374209627509117, 'timing/model': 0.6503293234854937, 'learning_rate/qwen_vl_interface': 9.328752769501272e-06, 'learning_rate/action_model': 9.328752769501272e-05, 'epoch': 0.51})
+[2026-07-28 17:56:01,695][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.04236122965812683, 'train/grad_norm_pre_clip': 1.6044609546661377, 'timing/data': 0.0003416081890463829, 'timing/model': 0.620308855548501, 'learning_rate/qwen_vl_interface': 9.326790452691332e-06, 'learning_rate/action_model': 9.326790452691332e-05, 'epoch': 0.51})
+[2026-07-28 17:56:10,904][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.06805220246315002, 'train/grad_norm_pre_clip': 1.7177804708480835, 'timing/data': 0.0005127405747771263, 'timing/model': 0.6246091164648533, 'learning_rate/qwen_vl_interface': 9.324825490496327e-06, 'learning_rate/action_model': 9.324825490496327e-05, 'epoch': 0.51})
+[2026-07-28 17:56:20,222][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.029495656490325928, 'train/grad_norm_pre_clip': 0.8851865530014038, 'timing/data': 0.0004678824916481972, 'timing/model': 0.6239865832030773, 'learning_rate/qwen_vl_interface': 9.322857884191303e-06, 'learning_rate/action_model': 9.322857884191302e-05, 'epoch': 0.51})
+[2026-07-28 17:56:29,489][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.03332826495170593, 'train/grad_norm_pre_clip': 1.9407767057418823, 'timing/data': 0.0003708507865667343, 'timing/model': 0.6206015031784773, 'learning_rate/qwen_vl_interface': 9.320887635053013e-06, 'learning_rate/action_model': 9.320887635053014e-05, 'epoch': 0.51})
+[2026-07-28 17:56:38,962][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.07561454176902771, 'train/grad_norm_pre_clip': 1.4398950338363647, 'timing/data': 0.0003645028918981552, 'timing/model': 0.621308927424252, 'learning_rate/qwen_vl_interface': 9.318914744359936e-06, 'learning_rate/action_model': 9.318914744359937e-05, 'epoch': 0.51})
+[2026-07-28 17:56:48,453][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.045250654220581055, 'train/grad_norm_pre_clip': 1.352763056755066, 'timing/data': 0.0004532141610980034, 'timing/model': 0.8533814270049334, 'learning_rate/qwen_vl_interface': 9.316939213392257e-06, 'learning_rate/action_model': 9.316939213392257e-05, 'epoch': 0.51})
+[2026-07-28 17:56:57,656][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.04658571630716324, 'train/grad_norm_pre_clip': 1.868404507637024, 'timing/data': 0.00043218955397605896, 'timing/model': 0.6233102129772305, 'learning_rate/qwen_vl_interface': 9.314961043431877e-06, 'learning_rate/action_model': 9.314961043431876e-05, 'epoch': 0.51})
+[2026-07-28 17:57:06,911][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.024959295988082886, 'train/grad_norm_pre_clip': 1.0029150247573853, 'timing/data': 0.0003879796713590622, 'timing/model': 0.6260057697072625, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.51})
+[2026-07-28 17:57:16,165][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.027860939502716064, 'train/grad_norm_pre_clip': 1.1865936517715454, 'timing/data': 0.00036753155291080475, 'timing/model': 0.6241608057171106, 'learning_rate/qwen_vl_interface': 9.310996791669175e-06, 'learning_rate/action_model': 9.310996791669175e-05, 'epoch': 0.51})
+[2026-07-28 17:57:25,713][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.03048614412546158, 'train/grad_norm_pre_clip': 0.788794994354248, 'timing/data': 0.0003539137542247772, 'timing/model': 0.6251250524073839, 'learning_rate/qwen_vl_interface': 9.309010712439213e-06, 'learning_rate/action_model': 9.309010712439213e-05, 'epoch': 0.51})
+[2026-07-28 17:57:35,125][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.038059696555137634, 'train/grad_norm_pre_clip': 1.2705076932907104, 'timing/data': 0.0004065725952386856, 'timing/model': 0.620717185549438, 'learning_rate/qwen_vl_interface': 9.307021999361267e-06, 'learning_rate/action_model': 9.307021999361267e-05, 'epoch': 0.51})
+[2026-07-28 17:57:44,416][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.035735029727220535, 'train/grad_norm_pre_clip': 1.0429984331130981, 'timing/data': 0.00039603468030691147, 'timing/model': 0.6235661134123802, 'learning_rate/qwen_vl_interface': 9.305030653725794e-06, 'learning_rate/action_model': 9.305030653725794e-05, 'epoch': 0.51})
+[2026-07-28 17:57:53,729][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.019774217158555984, 'train/grad_norm_pre_clip': 0.7714767456054688, 'timing/data': 0.00041351187974214554, 'timing/model': 0.621283651329577, 'learning_rate/qwen_vl_interface': 9.303036676824954e-06, 'learning_rate/action_model': 9.303036676824953e-05, 'epoch': 0.51})
+[2026-07-28 17:58:03,299][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.030803458765149117, 'train/grad_norm_pre_clip': 1.2855935096740723, 'timing/data': 0.000530565157532692, 'timing/model': 0.6215517250820994, 'learning_rate/qwen_vl_interface': 9.301040069952616e-06, 'learning_rate/action_model': 9.301040069952616e-05, 'epoch': 0.51})
+[2026-07-28 17:58:12,531][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.011468102224171162, 'train/grad_norm_pre_clip': 0.7701811194419861, 'timing/data': 0.0003636833280324936, 'timing/model': 0.6202855603769422, 'learning_rate/qwen_vl_interface': 9.299040834404362e-06, 'learning_rate/action_model': 9.299040834404361e-05, 'epoch': 0.51})
+[2026-07-28 17:58:21,477][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.05191202461719513, 'train/grad_norm_pre_clip': 1.5835371017456055, 'timing/data': 0.0004936177283525467, 'timing/model': 0.6211018431931734, 'learning_rate/qwen_vl_interface': 9.297038971477468e-06, 'learning_rate/action_model': 9.297038971477466e-05, 'epoch': 0.51})
+[2026-07-28 17:58:30,687][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.05237032473087311, 'train/grad_norm_pre_clip': 3.2099087238311768, 'timing/data': 0.00035174284130334854, 'timing/model': 0.6201879121363163, 'learning_rate/qwen_vl_interface': 9.295034482470922e-06, 'learning_rate/action_model': 9.295034482470923e-05, 'epoch': 0.51})
+[2026-07-28 17:58:40,078][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.011749768629670143, 'train/grad_norm_pre_clip': 0.8703935742378235, 'timing/data': 0.00046442076563835144, 'timing/model': 0.6189167685806751, 'learning_rate/qwen_vl_interface': 9.293027368685418e-06, 'learning_rate/action_model': 9.293027368685419e-05, 'epoch': 0.51})
+[2026-07-28 17:58:49,525][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.045568086206912994, 'train/grad_norm_pre_clip': 1.9772255420684814, 'timing/data': 0.00035200174897909164, 'timing/model': 0.6219626814126968, 'learning_rate/qwen_vl_interface': 9.29101763142335e-06, 'learning_rate/action_model': 9.29101763142335e-05, 'epoch': 0.51})
+[2026-07-28 17:58:58,657][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.018093448132276535, 'train/grad_norm_pre_clip': 0.7874137163162231, 'timing/data': 0.000413370318710804, 'timing/model': 0.6190769076347351, 'learning_rate/qwen_vl_interface': 9.28900527198881e-06, 'learning_rate/action_model': 9.289005271988811e-05, 'epoch': 0.51})
+[2026-07-28 17:59:07,827][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.04434431344270706, 'train/grad_norm_pre_clip': 1.6249253749847412, 'timing/data': 0.00043685920536518097, 'timing/model': 0.6173837222158909, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.51})
+[2026-07-28 17:59:17,271][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.07146306335926056, 'train/grad_norm_pre_clip': 2.358247756958008, 'timing/data': 0.00042448099702596664, 'timing/model': 0.620830294676125, 'learning_rate/qwen_vl_interface': 9.284972691827226e-06, 'learning_rate/action_model': 9.284972691827224e-05, 'epoch': 0.51})
+[2026-07-28 17:59:26,672][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.05820935219526291, 'train/grad_norm_pre_clip': 1.4782326221466064, 'timing/data': 0.00034579914063215256, 'timing/model': 0.625618364661932, 'learning_rate/qwen_vl_interface': 9.282952473716873e-06, 'learning_rate/action_model': 9.282952473716874e-05, 'epoch': 0.51})
+[2026-07-28 17:59:35,882][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.07345817983150482, 'train/grad_norm_pre_clip': 2.84379506111145, 'timing/data': 0.0003650933504104614, 'timing/model': 0.617700177244842, 'learning_rate/qwen_vl_interface': 9.280929638667447e-06, 'learning_rate/action_model': 9.280929638667447e-05, 'epoch': 0.51})
+[2026-07-28 17:59:45,034][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.044588059186935425, 'train/grad_norm_pre_clip': 2.4728846549987793, 'timing/data': 0.00044489093124866486, 'timing/model': 0.6199752045795321, 'learning_rate/qwen_vl_interface': 9.278904187991542e-06, 'learning_rate/action_model': 9.278904187991541e-05, 'epoch': 0.51})
+[2026-07-28 17:59:54,464][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.03136441484093666, 'train/grad_norm_pre_clip': 1.4119888544082642, 'timing/data': 0.0003420161083340645, 'timing/model': 0.8664578935131431, 'learning_rate/qwen_vl_interface': 9.27687612300345e-06, 'learning_rate/action_model': 9.27687612300345e-05, 'epoch': 0.51})
+[2026-07-28 18:00:03,906][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.07046251744031906, 'train/grad_norm_pre_clip': 2.5435233116149902, 'timing/data': 0.0003467109054327011, 'timing/model': 0.623071969486773, 'learning_rate/qwen_vl_interface': 9.27484544501916e-06, 'learning_rate/action_model': 9.274845445019159e-05, 'epoch': 0.51})
+[2026-07-28 18:00:12,938][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.04549148678779602, 'train/grad_norm_pre_clip': 2.1024749279022217, 'timing/data': 0.0003749970346689224, 'timing/model': 0.6203492879867554, 'learning_rate/qwen_vl_interface': 9.27281215535636e-06, 'learning_rate/action_model': 9.272812155356358e-05, 'epoch': 0.51})
+[2026-07-28 18:00:22,107][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.05192562937736511, 'train/grad_norm_pre_clip': 1.5963140726089478, 'timing/data': 0.0004350561648607254, 'timing/model': 0.6208723913878202, 'learning_rate/qwen_vl_interface': 9.270776255334425e-06, 'learning_rate/action_model': 9.270776255334425e-05, 'epoch': 0.51})
+[2026-07-28 18:00:31,302][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.05226651951670647, 'train/grad_norm_pre_clip': 2.1347992420196533, 'timing/data': 0.0003317166119813919, 'timing/model': 0.6190075036138296, 'learning_rate/qwen_vl_interface': 9.268737746274432e-06, 'learning_rate/action_model': 9.268737746274432e-05, 'epoch': 0.51})
+[2026-07-28 18:00:40,700][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.0355403795838356, 'train/grad_norm_pre_clip': 1.0787121057510376, 'timing/data': 0.00045990943908691406, 'timing/model': 0.6264295112341642, 'learning_rate/qwen_vl_interface': 9.266696629499146e-06, 'learning_rate/action_model': 9.266696629499145e-05, 'epoch': 0.51})
+[2026-07-28 18:00:50,055][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.04404490813612938, 'train/grad_norm_pre_clip': 1.0210049152374268, 'timing/data': 0.0004450250416994095, 'timing/model': 0.6208227798342705, 'learning_rate/qwen_vl_interface': 9.264652906333026e-06, 'learning_rate/action_model': 9.264652906333026e-05, 'epoch': 0.51})
+[2026-07-28 18:00:59,263][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.019726810976862907, 'train/grad_norm_pre_clip': 0.568091869354248, 'timing/data': 0.00037904921919107437, 'timing/model': 0.6186114773154259, 'learning_rate/qwen_vl_interface': 9.26260657810222e-06, 'learning_rate/action_model': 9.262606578102221e-05, 'epoch': 0.51})
+[2026-07-28 18:01:08,490][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.01814020425081253, 'train/grad_norm_pre_clip': 0.7571526765823364, 'timing/data': 0.0003636591136455536, 'timing/model': 0.6194119146093726, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.51})
+[2026-07-28 18:01:17,931][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.039666127413511276, 'train/grad_norm_pre_clip': 1.591879963874817, 'timing/data': 0.0004504406824707985, 'timing/model': 0.6213751370087266, 'learning_rate/qwen_vl_interface': 9.258506111759608e-06, 'learning_rate/action_model': 9.258506111759609e-05, 'epoch': 0.51})
+[2026-07-28 18:01:27,290][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.025223035365343094, 'train/grad_norm_pre_clip': 1.363328218460083, 'timing/data': 0.00035770609974861145, 'timing/model': 0.8450714247301221, 'learning_rate/qwen_vl_interface': 9.256451976308548e-06, 'learning_rate/action_model': 9.25645197630855e-05, 'epoch': 0.52})
+[2026-07-28 18:01:36,444][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.04917275905609131, 'train/grad_norm_pre_clip': 1.5883445739746094, 'timing/data': 0.00038003362715244293, 'timing/model': 0.6204870045185089, 'learning_rate/qwen_vl_interface': 9.2543952411143e-06, 'learning_rate/action_model': 9.254395241114299e-05, 'epoch': 0.52})
+[2026-07-28 18:01:45,766][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.07056163251399994, 'train/grad_norm_pre_clip': 1.6887704133987427, 'timing/data': 0.0003806445747613907, 'timing/model': 0.7947138845920563, 'learning_rate/qwen_vl_interface': 9.252335907511452e-06, 'learning_rate/action_model': 9.252335907511453e-05, 'epoch': 0.52})
+[2026-07-28 18:01:55,017][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.032712921500205994, 'train/grad_norm_pre_clip': 1.204054594039917, 'timing/data': 0.0004084911197423935, 'timing/model': 0.6175948260352015, 'learning_rate/qwen_vl_interface': 9.250273976836289e-06, 'learning_rate/action_model': 9.250273976836289e-05, 'epoch': 0.52})
+[2026-07-28 18:02:04,187][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.020208118483424187, 'train/grad_norm_pre_clip': 1.001280426979065, 'timing/data': 0.0003379862755537033, 'timing/model': 0.6189944315701723, 'learning_rate/qwen_vl_interface': 9.248209450426772e-06, 'learning_rate/action_model': 9.248209450426772e-05, 'epoch': 0.52})
+[2026-07-28 18:02:13,522][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.024628229439258575, 'train/grad_norm_pre_clip': 0.6785784363746643, 'timing/data': 0.00036896858364343643, 'timing/model': 0.6352713964879513, 'learning_rate/qwen_vl_interface': 9.24614232962255e-06, 'learning_rate/action_model': 9.24614232962255e-05, 'epoch': 0.52})
+[2026-07-28 18:02:23,247][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.042859092354774475, 'train/grad_norm_pre_clip': 1.2312568426132202, 'timing/data': 0.00046100374311208725, 'timing/model': 0.8906297851353884, 'learning_rate/qwen_vl_interface': 9.244072615764957e-06, 'learning_rate/action_model': 9.244072615764956e-05, 'epoch': 0.52})
+[2026-07-28 18:02:32,255][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.04105161502957344, 'train/grad_norm_pre_clip': 0.8662264943122864, 'timing/data': 0.00040816236287355423, 'timing/model': 0.6247516665607691, 'learning_rate/qwen_vl_interface': 9.242000310197005e-06, 'learning_rate/action_model': 9.242000310197004e-05, 'epoch': 0.52})
+[2026-07-28 18:02:41,471][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.019708767533302307, 'train/grad_norm_pre_clip': 1.1806597709655762, 'timing/data': 0.0005019353702664375, 'timing/model': 0.6213339203968644, 'learning_rate/qwen_vl_interface': 9.239925414263392e-06, 'learning_rate/action_model': 9.239925414263392e-05, 'epoch': 0.52})
+[2026-07-28 18:02:50,730][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.04767787829041481, 'train/grad_norm_pre_clip': 2.500889778137207, 'timing/data': 0.00038532912731170654, 'timing/model': 0.6195849571377039, 'learning_rate/qwen_vl_interface': 9.237847929310497e-06, 'learning_rate/action_model': 9.237847929310496e-05, 'epoch': 0.52})
+[2026-07-28 18:03:00,237][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.031177036464214325, 'train/grad_norm_pre_clip': 1.3848028182983398, 'timing/data': 0.0006588306277990341, 'timing/model': 0.6251240400597453, 'learning_rate/qwen_vl_interface': 9.235767856686375e-06, 'learning_rate/action_model': 9.235767856686375e-05, 'epoch': 0.52})
+[2026-07-28 18:03:09,690][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.016227848827838898, 'train/grad_norm_pre_clip': 0.7559672594070435, 'timing/data': 0.0004426725208759308, 'timing/model': 0.8579877763986588, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.52})
+[2026-07-28 18:03:19,120][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.021983128041028976, 'train/grad_norm_pre_clip': 0.9766966104507446, 'timing/data': 0.0005232514813542366, 'timing/model': 0.6225560829043388, 'learning_rate/qwen_vl_interface': 9.23159995382508e-06, 'learning_rate/action_model': 9.231599953825079e-05, 'epoch': 0.52})
+[2026-07-28 18:03:28,392][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.01939057558774948, 'train/grad_norm_pre_clip': 0.7751720547676086, 'timing/data': 0.00036433152854442596, 'timing/model': 0.6343406261876225, 'learning_rate/qwen_vl_interface': 9.22951212629241e-06, 'learning_rate/action_model': 9.229512126292411e-05, 'epoch': 0.52})
+[2026-07-28 18:03:37,804][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.06673237681388855, 'train/grad_norm_pre_clip': 1.545059084892273, 'timing/data': 0.0003585768863558769, 'timing/model': 0.6206636307761073, 'learning_rate/qwen_vl_interface': 9.227421716497531e-06, 'learning_rate/action_model': 9.227421716497531e-05, 'epoch': 0.52})
+[2026-07-28 18:03:47,219][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.03897121921181679, 'train/grad_norm_pre_clip': 1.9080955982208252, 'timing/data': 0.0003820117563009262, 'timing/model': 0.875478932633996, 'learning_rate/qwen_vl_interface': 9.225328725796879e-06, 'learning_rate/action_model': 9.22532872579688e-05, 'epoch': 0.52})
+[2026-07-28 18:03:56,461][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.015117095783352852, 'train/grad_norm_pre_clip': 0.8437961935997009, 'timing/data': 0.0003680940717458725, 'timing/model': 0.6213021790608764, 'learning_rate/qwen_vl_interface': 9.223233155548579e-06, 'learning_rate/action_model': 9.223233155548578e-05, 'epoch': 0.52})
+[2026-07-28 18:04:05,871][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.009830934926867485, 'train/grad_norm_pre_clip': 0.751776397228241, 'timing/data': 0.00041658151894807816, 'timing/model': 0.8572375616058707, 'learning_rate/qwen_vl_interface': 9.22113500711242e-06, 'learning_rate/action_model': 9.221135007112419e-05, 'epoch': 0.52})
+[2026-07-28 18:04:15,045][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.1127147227525711, 'train/grad_norm_pre_clip': 3.533825397491455, 'timing/data': 0.00036904122680425644, 'timing/model': 0.6211496153846383, 'learning_rate/qwen_vl_interface': 9.219034281849867e-06, 'learning_rate/action_model': 9.219034281849867e-05, 'epoch': 0.52})
+[2026-07-28 18:04:24,441][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.02489522285759449, 'train/grad_norm_pre_clip': 0.9304307103157043, 'timing/data': 0.00038137752562761307, 'timing/model': 0.6243085786700249, 'learning_rate/qwen_vl_interface': 9.216930981124062e-06, 'learning_rate/action_model': 9.216930981124062e-05, 'epoch': 0.52})
+[2026-07-28 18:04:33,704][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.05436070263385773, 'train/grad_norm_pre_clip': 1.6189852952957153, 'timing/data': 0.0003486545756459236, 'timing/model': 0.6213607341051102, 'learning_rate/qwen_vl_interface': 9.214825106299808e-06, 'learning_rate/action_model': 9.214825106299808e-05, 'epoch': 0.52})
+[2026-07-28 18:04:43,182][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.06857671588659286, 'train/grad_norm_pre_clip': 2.2715671062469482, 'timing/data': 0.00048707611858844757, 'timing/model': 0.9213000331073999, 'learning_rate/qwen_vl_interface': 9.212716658743592e-06, 'learning_rate/action_model': 9.212716658743592e-05, 'epoch': 0.52})
+[2026-07-28 18:04:52,575][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.017522545531392097, 'train/grad_norm_pre_clip': 1.464289665222168, 'timing/data': 0.00037224404513835907, 'timing/model': 0.8584620719775558, 'learning_rate/qwen_vl_interface': 9.210605639823558e-06, 'learning_rate/action_model': 9.210605639823557e-05, 'epoch': 0.52})
+[2026-07-28 18:05:01,784][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.05990709364414215, 'train/grad_norm_pre_clip': 2.0759329795837402, 'timing/data': 0.0003276020288467407, 'timing/model': 0.6198088340461254, 'learning_rate/qwen_vl_interface': 9.208492050909523e-06, 'learning_rate/action_model': 9.208492050909523e-05, 'epoch': 0.52})
+[2026-07-28 18:05:10,927][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.02253996767103672, 'train/grad_norm_pre_clip': 1.1607935428619385, 'timing/data': 0.00041513796895742416, 'timing/model': 0.619565924629569, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.52})
+[2026-07-28 18:05:20,123][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.038123857229948044, 'train/grad_norm_pre_clip': 1.7726608514785767, 'timing/data': 0.0003404337912797928, 'timing/model': 0.6221562176942825, 'learning_rate/qwen_vl_interface': 9.204257168587061e-06, 'learning_rate/action_model': 9.204257168587061e-05, 'epoch': 0.52})
+[2026-07-28 18:05:29,547][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.07167265564203262, 'train/grad_norm_pre_clip': 1.8924390077590942, 'timing/data': 0.00039341021329164505, 'timing/model': 0.8589571230113506, 'learning_rate/qwen_vl_interface': 9.2021358779266e-06, 'learning_rate/action_model': 9.202135877926601e-05, 'epoch': 0.52})
+[2026-07-28 18:05:38,925][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.08468080312013626, 'train/grad_norm_pre_clip': 2.27229380607605, 'timing/data': 0.00042862631380558014, 'timing/model': 0.6202445654198527, 'learning_rate/qwen_vl_interface': 9.200012022768077e-06, 'learning_rate/action_model': 9.200012022768077e-05, 'epoch': 0.52})
+[2026-07-28 18:05:48,118][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.06920676678419113, 'train/grad_norm_pre_clip': 2.0155787467956543, 'timing/data': 0.0003561852499842644, 'timing/model': 0.6188739808276296, 'learning_rate/qwen_vl_interface': 9.19788560448964e-06, 'learning_rate/action_model': 9.197885604489639e-05, 'epoch': 0.52})
+[2026-07-28 18:05:57,075][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.03775664046406746, 'train/grad_norm_pre_clip': 1.2387580871582031, 'timing/data': 0.00034083612263202667, 'timing/model': 0.618751092813909, 'learning_rate/qwen_vl_interface': 9.195756624471094e-06, 'learning_rate/action_model': 9.195756624471094e-05, 'epoch': 0.52})
+[2026-07-28 18:06:05,981][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.054783180356025696, 'train/grad_norm_pre_clip': 1.9808422327041626, 'timing/data': 0.0003574015572667122, 'timing/model': 0.6221747426316142, 'learning_rate/qwen_vl_interface': 9.193625084093912e-06, 'learning_rate/action_model': 9.193625084093911e-05, 'epoch': 0.52})
+[2026-07-28 18:06:15,126][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.06451059877872467, 'train/grad_norm_pre_clip': 1.7822707891464233, 'timing/data': 0.00034306012094020844, 'timing/model': 0.617877640761435, 'learning_rate/qwen_vl_interface': 9.191490984741227e-06, 'learning_rate/action_model': 9.191490984741227e-05, 'epoch': 0.52})
+[2026-07-28 18:06:24,300][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.018410755321383476, 'train/grad_norm_pre_clip': 1.1921238899230957, 'timing/data': 0.00034252554178237915, 'timing/model': 0.8589436151087284, 'learning_rate/qwen_vl_interface': 9.189354327797835e-06, 'learning_rate/action_model': 9.189354327797836e-05, 'epoch': 0.52})
+[2026-07-28 18:06:33,675][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.025942429900169373, 'train/grad_norm_pre_clip': 1.0135233402252197, 'timing/data': 0.00046750251203775406, 'timing/model': 0.6234537838026881, 'learning_rate/qwen_vl_interface': 9.18721511465019e-06, 'learning_rate/action_model': 9.187215114650191e-05, 'epoch': 0.52})
+[2026-07-28 18:06:42,868][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.03461633622646332, 'train/grad_norm_pre_clip': 1.0152559280395508, 'timing/data': 0.00036611221730709076, 'timing/model': 0.6211823150515556, 'learning_rate/qwen_vl_interface': 9.185073346686402e-06, 'learning_rate/action_model': 9.185073346686401e-05, 'epoch': 0.52})
+[2026-07-28 18:06:52,030][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.05676092579960823, 'train/grad_norm_pre_clip': 1.779205083847046, 'timing/data': 0.00033927522599697113, 'timing/model': 0.620280365459621, 'learning_rate/qwen_vl_interface': 9.182929025296242e-06, 'learning_rate/action_model': 9.182929025296242e-05, 'epoch': 0.52})
+[2026-07-28 18:07:01,195][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.0479765422642231, 'train/grad_norm_pre_clip': 1.1647460460662842, 'timing/data': 0.00035154446959495544, 'timing/model': 0.6201764401048422, 'learning_rate/qwen_vl_interface': 9.180782151871138e-06, 'learning_rate/action_model': 9.180782151871137e-05, 'epoch': 0.52})
+[2026-07-28 18:07:10,584][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.05138245224952698, 'train/grad_norm_pre_clip': 1.110215187072754, 'timing/data': 0.0008010324090719223, 'timing/model': 0.6182889631018043, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.52})
+[2026-07-28 18:07:20,144][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.05478779971599579, 'train/grad_norm_pre_clip': 2.3190388679504395, 'timing/data': 0.00045227445662021637, 'timing/model': 0.6331738652661443, 'learning_rate/qwen_vl_interface': 9.176480754490078e-06, 'learning_rate/action_model': 9.176480754490078e-05, 'epoch': 0.52})
+[2026-07-28 18:07:29,624][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.022512882947921753, 'train/grad_norm_pre_clip': 1.7363611459732056, 'timing/data': 0.00035214517265558243, 'timing/model': 0.8643108168616891, 'learning_rate/qwen_vl_interface': 9.174326233325253e-06, 'learning_rate/action_model': 9.174326233325253e-05, 'epoch': 0.52})
+[2026-07-28 18:07:38,883][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.03145698830485344, 'train/grad_norm_pre_clip': 0.914849579334259, 'timing/data': 0.0003596106544137001, 'timing/model': 0.622376374900341, 'learning_rate/qwen_vl_interface': 9.17216916570774e-06, 'learning_rate/action_model': 9.172169165707741e-05, 'epoch': 0.52})
+[2026-07-28 18:07:48,113][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.0700964704155922, 'train/grad_norm_pre_clip': 1.0153859853744507, 'timing/data': 0.00038417428731918335, 'timing/model': 0.6195675283670425, 'learning_rate/qwen_vl_interface': 9.170009553037239e-06, 'learning_rate/action_model': 9.170009553037238e-05, 'epoch': 0.52})
+[2026-07-28 18:07:57,339][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.048856399953365326, 'train/grad_norm_pre_clip': 1.085017442703247, 'timing/data': 0.0003445185720920563, 'timing/model': 0.6234747488051653, 'learning_rate/qwen_vl_interface': 9.167847396715098e-06, 'learning_rate/action_model': 9.167847396715098e-05, 'epoch': 0.52})
+[2026-07-28 18:08:06,510][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.026948727667331696, 'train/grad_norm_pre_clip': 0.6915175914764404, 'timing/data': 0.0003587203100323677, 'timing/model': 0.618693282827735, 'learning_rate/qwen_vl_interface': 9.165682698144317e-06, 'learning_rate/action_model': 9.165682698144317e-05, 'epoch': 0.52})
+[2026-07-28 18:08:16,145][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.037054721266031265, 'train/grad_norm_pre_clip': 1.9257817268371582, 'timing/data': 0.0003297831863164902, 'timing/model': 0.6231262180954218, 'learning_rate/qwen_vl_interface': 9.163515458729545e-06, 'learning_rate/action_model': 9.163515458729545e-05, 'epoch': 0.52})
+[2026-07-28 18:08:25,300][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.050280649214982986, 'train/grad_norm_pre_clip': 2.1427981853485107, 'timing/data': 0.0003596646711230278, 'timing/model': 0.6201672963798046, 'learning_rate/qwen_vl_interface': 9.16134567987708e-06, 'learning_rate/action_model': 9.16134567987708e-05, 'epoch': 0.52})
+[2026-07-28 18:08:34,552][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.051349032670259476, 'train/grad_norm_pre_clip': 1.8521058559417725, 'timing/data': 0.00042822491377592087, 'timing/model': 0.6188833750784397, 'learning_rate/qwen_vl_interface': 9.159173362994868e-06, 'learning_rate/action_model': 9.159173362994869e-05, 'epoch': 0.52})
+[2026-07-28 18:08:43,810][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.024203961715102196, 'train/grad_norm_pre_clip': 0.7685260772705078, 'timing/data': 0.0004119258373975754, 'timing/model': 0.619767495431006, 'learning_rate/qwen_vl_interface': 9.156998509492503e-06, 'learning_rate/action_model': 9.156998509492503e-05, 'epoch': 0.52})
+[2026-07-28 18:08:53,436][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.03414798155426979, 'train/grad_norm_pre_clip': 1.280422329902649, 'timing/data': 0.0003405418246984482, 'timing/model': 0.6200023647397757, 'learning_rate/qwen_vl_interface': 9.154821120781221e-06, 'learning_rate/action_model': 9.154821120781222e-05, 'epoch': 0.52})
+[2026-07-28 18:09:02,483][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.034956738352775574, 'train/grad_norm_pre_clip': 1.100046157836914, 'timing/data': 0.0003356020897626877, 'timing/model': 0.6185202738270164, 'learning_rate/qwen_vl_interface': 9.152641198273911e-06, 'learning_rate/action_model': 9.15264119827391e-05, 'epoch': 0.52})
+[2026-07-28 18:09:11,741][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.03839505836367607, 'train/grad_norm_pre_clip': 1.540398359298706, 'timing/data': 0.0003533046692609787, 'timing/model': 0.6202288940548897, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.52})
+[2026-07-28 18:09:21,154][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.03688724339008331, 'train/grad_norm_pre_clip': 0.9177883863449097, 'timing/data': 0.000443480908870697, 'timing/model': 0.8595298519358039, 'learning_rate/qwen_vl_interface': 9.148273757530948e-06, 'learning_rate/action_model': 9.148273757530948e-05, 'epoch': 0.52})
+[2026-07-28 18:09:30,592][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.04248768091201782, 'train/grad_norm_pre_clip': 1.1671167612075806, 'timing/data': 0.00035142339766025543, 'timing/model': 0.8446127399802208, 'learning_rate/qwen_vl_interface': 9.146086242129283e-06, 'learning_rate/action_model': 9.146086242129282e-05, 'epoch': 0.52})
+[2026-07-28 18:09:39,755][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.015787510201334953, 'train/grad_norm_pre_clip': 0.699501633644104, 'timing/data': 0.00033180415630340576, 'timing/model': 0.6256558680906892, 'learning_rate/qwen_vl_interface': 9.143896198599553e-06, 'learning_rate/action_model': 9.143896198599553e-05, 'epoch': 0.52})
+[2026-07-28 18:09:48,932][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.018523817881941795, 'train/grad_norm_pre_clip': 0.7214664220809937, 'timing/data': 0.0003421623259782791, 'timing/model': 0.61871422175318, 'learning_rate/qwen_vl_interface': 9.141703628362858e-06, 'learning_rate/action_model': 9.141703628362858e-05, 'epoch': 0.52})
+[2026-07-28 18:09:58,167][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.01844060607254505, 'train/grad_norm_pre_clip': 1.5536459684371948, 'timing/data': 0.000345439650118351, 'timing/model': 0.6178186722099781, 'learning_rate/qwen_vl_interface': 9.139508532841928e-06, 'learning_rate/action_model': 9.139508532841927e-05, 'epoch': 0.53})
+[2026-07-28 18:10:07,538][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.05761561170220375, 'train/grad_norm_pre_clip': 1.4412542581558228, 'timing/data': 0.00032201968133449554, 'timing/model': 0.8459333786740899, 'learning_rate/qwen_vl_interface': 9.137310913461139e-06, 'learning_rate/action_model': 9.137310913461139e-05, 'epoch': 0.53})
+[2026-07-28 18:10:16,663][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.017643090337514877, 'train/grad_norm_pre_clip': 1.2995113134384155, 'timing/data': 0.00035335682332515717, 'timing/model': 0.6191191012039781, 'learning_rate/qwen_vl_interface': 9.135110771646502e-06, 'learning_rate/action_model': 9.135110771646502e-05, 'epoch': 0.53})
+[2026-07-28 18:10:26,121][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.0188855342566967, 'train/grad_norm_pre_clip': 1.4071040153503418, 'timing/data': 0.00032994989305734634, 'timing/model': 0.6243644338101149, 'learning_rate/qwen_vl_interface': 9.132908108825667e-06, 'learning_rate/action_model': 9.132908108825667e-05, 'epoch': 0.53})
+[2026-07-28 18:10:35,294][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.023374885320663452, 'train/grad_norm_pre_clip': 0.9262256026268005, 'timing/data': 0.0003825342282652855, 'timing/model': 0.6213257415220141, 'learning_rate/qwen_vl_interface': 9.130702926427915e-06, 'learning_rate/action_model': 9.130702926427914e-05, 'epoch': 0.53})
+[2026-07-28 18:10:44,547][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.032420143485069275, 'train/grad_norm_pre_clip': 1.7270370721817017, 'timing/data': 0.000415608286857605, 'timing/model': 0.6278470335528255, 'learning_rate/qwen_vl_interface': 9.128495225884167e-06, 'learning_rate/action_model': 9.128495225884166e-05, 'epoch': 0.53})
+[2026-07-28 18:10:53,471][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.07318087667226791, 'train/grad_norm_pre_clip': 1.8781813383102417, 'timing/data': 0.00033458415418863297, 'timing/model': 0.6198373017832637, 'learning_rate/qwen_vl_interface': 9.126285008626975e-06, 'learning_rate/action_model': 9.126285008626974e-05, 'epoch': 0.53})
+[2026-07-28 18:11:02,678][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.08415594696998596, 'train/grad_norm_pre_clip': 2.1046230792999268, 'timing/data': 0.0003334004431962967, 'timing/model': 0.8733371263369918, 'learning_rate/qwen_vl_interface': 9.124072276090524e-06, 'learning_rate/action_model': 9.124072276090524e-05, 'epoch': 0.53})
+[2026-07-28 18:11:12,045][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.05621981620788574, 'train/grad_norm_pre_clip': 2.1187267303466797, 'timing/data': 0.00033343955874443054, 'timing/model': 0.6194916991516948, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.53})
+[2026-07-28 18:11:21,267][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.01853245683014393, 'train/grad_norm_pre_clip': 1.1012139320373535, 'timing/data': 0.00036076828837394714, 'timing/model': 0.6199953379109502, 'learning_rate/qwen_vl_interface': 9.119639270924754e-06, 'learning_rate/action_model': 9.119639270924754e-05, 'epoch': 0.53})
+[2026-07-28 18:11:30,504][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.057067349553108215, 'train/grad_norm_pre_clip': 1.7707124948501587, 'timing/data': 0.00037563592195510864, 'timing/model': 0.6198147423565388, 'learning_rate/qwen_vl_interface': 9.11741900117196e-06, 'learning_rate/action_model': 9.11741900117196e-05, 'epoch': 0.53})
+[2026-07-28 18:11:39,875][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.07414871454238892, 'train/grad_norm_pre_clip': 2.118091106414795, 'timing/data': 0.00040102843195199966, 'timing/model': 0.6196369724348187, 'learning_rate/qwen_vl_interface': 9.115196221892969e-06, 'learning_rate/action_model': 9.115196221892969e-05, 'epoch': 0.53})
+[2026-07-28 18:11:49,107][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.02173030376434326, 'train/grad_norm_pre_clip': 0.7239698171615601, 'timing/data': 0.00033588241785764694, 'timing/model': 0.6223670886829495, 'learning_rate/qwen_vl_interface': 9.112970934530111e-06, 'learning_rate/action_model': 9.112970934530111e-05, 'epoch': 0.53})
+[2026-07-28 18:11:58,581][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.05137474834918976, 'train/grad_norm_pre_clip': 1.2278316020965576, 'timing/data': 0.0003389855846762657, 'timing/model': 0.6263315612450242, 'learning_rate/qwen_vl_interface': 9.110743140527353e-06, 'learning_rate/action_model': 9.110743140527352e-05, 'epoch': 0.53})
+[2026-07-28 18:12:08,131][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.040706880390644073, 'train/grad_norm_pre_clip': 1.25528085231781, 'timing/data': 0.00033292919397354126, 'timing/model': 0.8586893575266004, 'learning_rate/qwen_vl_interface': 9.108512841330288e-06, 'learning_rate/action_model': 9.108512841330287e-05, 'epoch': 0.53})
+[2026-07-28 18:12:17,405][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.03365349769592285, 'train/grad_norm_pre_clip': 1.647558331489563, 'timing/data': 0.0003418410196900368, 'timing/model': 0.6220914609730244, 'learning_rate/qwen_vl_interface': 9.106280038386129e-06, 'learning_rate/action_model': 9.10628003838613e-05, 'epoch': 0.53})
+[2026-07-28 18:12:26,581][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.037569329142570496, 'train/grad_norm_pre_clip': 1.8580536842346191, 'timing/data': 0.0004462422803044319, 'timing/model': 0.6217588391155005, 'learning_rate/qwen_vl_interface': 9.104044733143723e-06, 'learning_rate/action_model': 9.104044733143723e-05, 'epoch': 0.53})
+[2026-07-28 18:12:36,050][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.015584100969135761, 'train/grad_norm_pre_clip': 0.7885138392448425, 'timing/data': 0.0003620954230427742, 'timing/model': 0.6251025293022394, 'learning_rate/qwen_vl_interface': 9.10180692705353e-06, 'learning_rate/action_model': 9.10180692705353e-05, 'epoch': 0.53})
+[2026-07-28 18:12:45,230][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.04306360334157944, 'train/grad_norm_pre_clip': 1.8406742811203003, 'timing/data': 0.00034676678478717804, 'timing/model': 0.6206457000225782, 'learning_rate/qwen_vl_interface': 9.099566621567643e-06, 'learning_rate/action_model': 9.099566621567642e-05, 'epoch': 0.53})
+[2026-07-28 18:12:54,648][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.064466193318367, 'train/grad_norm_pre_clip': 1.9760771989822388, 'timing/data': 0.0003284262493252754, 'timing/model': 0.6189401736482978, 'learning_rate/qwen_vl_interface': 9.097323818139767e-06, 'learning_rate/action_model': 9.097323818139767e-05, 'epoch': 0.53})
+[2026-07-28 18:13:03,822][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.03395187109708786, 'train/grad_norm_pre_clip': 0.9639132022857666, 'timing/data': 0.0003288397565484047, 'timing/model': 0.6196001200005412, 'learning_rate/qwen_vl_interface': 9.095078518225237e-06, 'learning_rate/action_model': 9.095078518225236e-05, 'epoch': 0.53})
+[2026-07-28 18:13:13,002][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.053181927651166916, 'train/grad_norm_pre_clip': 1.6893582344055176, 'timing/data': 0.0004074973985552788, 'timing/model': 0.6211515925824642, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.53})
+[2026-07-28 18:13:22,593][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.014573611319065094, 'train/grad_norm_pre_clip': 0.4409056603908539, 'timing/data': 0.0004799049347639084, 'timing/model': 0.6213585827499628, 'learning_rate/qwen_vl_interface': 9.09058043476563e-06, 'learning_rate/action_model': 9.09058043476563e-05, 'epoch': 0.53})
+[2026-07-28 18:13:31,796][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.033355485647916794, 'train/grad_norm_pre_clip': 1.0604654550552368, 'timing/data': 0.00037098582834005356, 'timing/model': 0.6181800896301866, 'learning_rate/qwen_vl_interface': 9.088327654139311e-06, 'learning_rate/action_model': 9.08832765413931e-05, 'epoch': 0.53})
+[2026-07-28 18:13:41,237][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.0209941528737545, 'train/grad_norm_pre_clip': 0.8932163715362549, 'timing/data': 0.0004321206361055374, 'timing/model': 0.8601424461230636, 'learning_rate/qwen_vl_interface': 9.08607238286385e-06, 'learning_rate/action_model': 9.08607238286385e-05, 'epoch': 0.53})
+[2026-07-28 18:13:50,498][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.034228235483169556, 'train/grad_norm_pre_clip': 0.7262452840805054, 'timing/data': 0.00036439113318920135, 'timing/model': 0.6219495935365558, 'learning_rate/qwen_vl_interface': 9.083814622402667e-06, 'learning_rate/action_model': 9.083814622402666e-05, 'epoch': 0.53})
+[2026-07-28 18:13:59,881][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.04121537134051323, 'train/grad_norm_pre_clip': 0.8783717751502991, 'timing/data': 0.0005047060549259186, 'timing/model': 0.6229810584336519, 'learning_rate/qwen_vl_interface': 9.0815543742208e-06, 'learning_rate/action_model': 9.081554374220799e-05, 'epoch': 0.53})
+[2026-07-28 18:14:09,112][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.06136952340602875, 'train/grad_norm_pre_clip': 1.3716388940811157, 'timing/data': 0.00035412050783634186, 'timing/model': 0.6383804762735963, 'learning_rate/qwen_vl_interface': 9.079291639784897e-06, 'learning_rate/action_model': 9.079291639784896e-05, 'epoch': 0.53})
+[2026-07-28 18:14:18,244][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.028380446135997772, 'train/grad_norm_pre_clip': 3.0560669898986816, 'timing/data': 0.0003588944673538208, 'timing/model': 0.6195357264950871, 'learning_rate/qwen_vl_interface': 9.077026420563223e-06, 'learning_rate/action_model': 9.077026420563223e-05, 'epoch': 0.53})
+[2026-07-28 18:14:27,691][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.062036629766225815, 'train/grad_norm_pre_clip': 1.5119813680648804, 'timing/data': 0.0003612423315644264, 'timing/model': 0.6216994561254978, 'learning_rate/qwen_vl_interface': 9.074758718025652e-06, 'learning_rate/action_model': 9.074758718025652e-05, 'epoch': 0.53})
+[2026-07-28 18:14:37,162][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.03396874666213989, 'train/grad_norm_pre_clip': 1.2987037897109985, 'timing/data': 0.00032550282776355743, 'timing/model': 0.677264210768044, 'learning_rate/qwen_vl_interface': 9.072488533643678e-06, 'learning_rate/action_model': 9.072488533643677e-05, 'epoch': 0.53})
+[2026-07-28 18:14:46,392][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.026658274233341217, 'train/grad_norm_pre_clip': 0.7061249613761902, 'timing/data': 0.00044187530875205994, 'timing/model': 0.6182037079706788, 'learning_rate/qwen_vl_interface': 9.070215868890391e-06, 'learning_rate/action_model': 9.070215868890391e-05, 'epoch': 0.53})
+[2026-07-28 18:14:55,972][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.039101697504520416, 'train/grad_norm_pre_clip': 0.8822153210639954, 'timing/data': 0.0003488576039671898, 'timing/model': 0.6215499816462398, 'learning_rate/qwen_vl_interface': 9.067940725240503e-06, 'learning_rate/action_model': 9.067940725240503e-05, 'epoch': 0.53})
+[2026-07-28 18:15:05,133][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.06569365411996841, 'train/grad_norm_pre_clip': 1.646677851676941, 'timing/data': 0.0004385029897093773, 'timing/model': 0.620017172768712, 'learning_rate/qwen_vl_interface': 9.06566310417033e-06, 'learning_rate/action_model': 9.06566310417033e-05, 'epoch': 0.53})
+[2026-07-28 18:15:14,593][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.046436507254838943, 'train/grad_norm_pre_clip': 0.987740159034729, 'timing/data': 0.00038721971213817596, 'timing/model': 0.6245463816449046, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.53})
+[2026-07-28 18:15:23,756][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.0521060973405838, 'train/grad_norm_pre_clip': 0.9681792259216309, 'timing/data': 0.0004044370725750923, 'timing/model': 0.6206388743594289, 'learning_rate/qwen_vl_interface': 9.061100435682429e-06, 'learning_rate/action_model': 9.061100435682429e-05, 'epoch': 0.53})
+[2026-07-28 18:15:33,270][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.033455561846494675, 'train/grad_norm_pre_clip': 0.7039787173271179, 'timing/data': 0.000521291047334671, 'timing/model': 0.6193886231631041, 'learning_rate/qwen_vl_interface': 9.058815391225363e-06, 'learning_rate/action_model': 9.058815391225362e-05, 'epoch': 0.53})
+[2026-07-28 18:15:42,390][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.032118819653987885, 'train/grad_norm_pre_clip': 1.4322222471237183, 'timing/data': 0.00034264009445905685, 'timing/model': 0.6184778986498713, 'learning_rate/qwen_vl_interface': 9.056527875269344e-06, 'learning_rate/action_model': 9.056527875269344e-05, 'epoch': 0.53})
+[2026-07-28 18:15:51,779][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.03050592914223671, 'train/grad_norm_pre_clip': 0.7266159653663635, 'timing/data': 0.00041684694588184357, 'timing/model': 0.6190851693972945, 'learning_rate/qwen_vl_interface': 9.054237889298714e-06, 'learning_rate/action_model': 9.054237889298714e-05, 'epoch': 0.53})
+[2026-07-28 18:16:00,939][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.0272379070520401, 'train/grad_norm_pre_clip': 0.8572532534599304, 'timing/data': 0.00042996928095817566, 'timing/model': 0.6199121763929725, 'learning_rate/qwen_vl_interface': 9.05194543479942e-06, 'learning_rate/action_model': 9.05194543479942e-05, 'epoch': 0.53})
+[2026-07-28 18:16:10,339][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.01535780169069767, 'train/grad_norm_pre_clip': 0.4550102949142456, 'timing/data': 0.00033401139080524445, 'timing/model': 0.6268448634073138, 'learning_rate/qwen_vl_interface': 9.049650513259007e-06, 'learning_rate/action_model': 9.049650513259008e-05, 'epoch': 0.53})
+[2026-07-28 18:16:19,645][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.04373245686292648, 'train/grad_norm_pre_clip': 1.9960798025131226, 'timing/data': 0.00038811471313238144, 'timing/model': 0.7951317355036736, 'learning_rate/qwen_vl_interface': 9.047353126166632e-06, 'learning_rate/action_model': 9.047353126166632e-05, 'epoch': 0.53})
+[2026-07-28 18:16:28,791][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.01928652636706829, 'train/grad_norm_pre_clip': 0.9038854241371155, 'timing/data': 0.0004262039437890053, 'timing/model': 0.619898903183639, 'learning_rate/qwen_vl_interface': 9.045053275013038e-06, 'learning_rate/action_model': 9.045053275013038e-05, 'epoch': 0.53})
+[2026-07-28 18:16:37,996][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.05513177812099457, 'train/grad_norm_pre_clip': 1.1000540256500244, 'timing/data': 0.0005475766956806183, 'timing/model': 0.6213572444394231, 'learning_rate/qwen_vl_interface': 9.042750961290578e-06, 'learning_rate/action_model': 9.042750961290578e-05, 'epoch': 0.53})
+[2026-07-28 18:16:47,586][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.022883202880620956, 'train/grad_norm_pre_clip': 0.8511697053909302, 'timing/data': 0.00035248324275016785, 'timing/model': 0.8558140974491835, 'learning_rate/qwen_vl_interface': 9.040446186493195e-06, 'learning_rate/action_model': 9.040446186493194e-05, 'epoch': 0.53})
+[2026-07-28 18:16:57,147][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.041388388723134995, 'train/grad_norm_pre_clip': 1.604233741760254, 'timing/data': 0.0003417329862713814, 'timing/model': 0.8630784992128611, 'learning_rate/qwen_vl_interface': 9.038138952116435e-06, 'learning_rate/action_model': 9.038138952116435e-05, 'epoch': 0.53})
+[2026-07-28 18:17:06,087][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.016808148473501205, 'train/grad_norm_pre_clip': 0.8849265575408936, 'timing/data': 0.00040240585803985596, 'timing/model': 0.6198247661814094, 'learning_rate/qwen_vl_interface': 9.035829259657437e-06, 'learning_rate/action_model': 9.035829259657436e-05, 'epoch': 0.53})
+[2026-07-28 18:17:14,991][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.017549023032188416, 'train/grad_norm_pre_clip': 0.4885566532611847, 'timing/data': 0.000352734699845314, 'timing/model': 0.6210044659674168, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.53})
+[2026-07-28 18:17:24,134][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.0160495787858963, 'train/grad_norm_pre_clip': 0.8084225058555603, 'timing/data': 0.00043734628707170486, 'timing/model': 0.6166068613529205, 'learning_rate/qwen_vl_interface': 9.031202506489254e-06, 'learning_rate/action_model': 9.031202506489253e-05, 'epoch': 0.53})
+[2026-07-28 18:17:33,562][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.04689141735434532, 'train/grad_norm_pre_clip': 1.5298619270324707, 'timing/data': 0.0004946347326040268, 'timing/model': 0.6247101128101349, 'learning_rate/qwen_vl_interface': 9.02888544878232e-06, 'learning_rate/action_model': 9.028885448782319e-05, 'epoch': 0.53})
+[2026-07-28 18:17:42,991][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.03352602571249008, 'train/grad_norm_pre_clip': 1.524207353591919, 'timing/data': 0.0003810124471783638, 'timing/model': 0.6218105321750045, 'learning_rate/qwen_vl_interface': 9.026565938997644e-06, 'learning_rate/action_model': 9.026565938997645e-05, 'epoch': 0.53})
+[2026-07-28 18:17:52,171][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.035868193954229355, 'train/grad_norm_pre_clip': 1.168443202972412, 'timing/data': 0.00036185141652822495, 'timing/model': 0.617948635481298, 'learning_rate/qwen_vl_interface': 9.024243978640334e-06, 'learning_rate/action_model': 9.024243978640333e-05, 'epoch': 0.53})
+[2026-07-28 18:18:01,377][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.04824354499578476, 'train/grad_norm_pre_clip': 1.7233827114105225, 'timing/data': 0.0003497442230582237, 'timing/model': 0.6203574053943157, 'learning_rate/qwen_vl_interface': 9.02191956921708e-06, 'learning_rate/action_model': 9.02191956921708e-05, 'epoch': 0.53})
+[2026-07-28 18:18:10,787][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.02558159828186035, 'train/grad_norm_pre_clip': 1.0882418155670166, 'timing/data': 0.00035202689468860626, 'timing/model': 0.6204483015462756, 'learning_rate/qwen_vl_interface': 9.019592712236172e-06, 'learning_rate/action_model': 9.019592712236171e-05, 'epoch': 0.53})
+[2026-07-28 18:18:20,220][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.04587694630026817, 'train/grad_norm_pre_clip': 1.2306838035583496, 'timing/data': 0.0003559477627277374, 'timing/model': 0.8548499550670385, 'learning_rate/qwen_vl_interface': 9.017263409207478e-06, 'learning_rate/action_model': 9.017263409207478e-05, 'epoch': 0.54})
+[2026-07-28 18:18:29,378][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.04783087968826294, 'train/grad_norm_pre_clip': 2.9862494468688965, 'timing/data': 0.00031338073313236237, 'timing/model': 0.6220403742045164, 'learning_rate/qwen_vl_interface': 9.014931661642455e-06, 'learning_rate/action_model': 9.014931661642455e-05, 'epoch': 0.54})
+[2026-07-28 18:18:38,714][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.014491022564470768, 'train/grad_norm_pre_clip': 0.7694932222366333, 'timing/data': 0.00035580992698669434, 'timing/model': 0.7896572006866336, 'learning_rate/qwen_vl_interface': 9.012597471054154e-06, 'learning_rate/action_model': 9.012597471054154e-05, 'epoch': 0.54})
+[2026-07-28 18:18:47,922][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.020571164786815643, 'train/grad_norm_pre_clip': 0.8328706622123718, 'timing/data': 0.00034244172275066376, 'timing/model': 0.6205093050375581, 'learning_rate/qwen_vl_interface': 9.010260838957206e-06, 'learning_rate/action_model': 9.010260838957205e-05, 'epoch': 0.54})
+[2026-07-28 18:18:57,326][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.03777358680963516, 'train/grad_norm_pre_clip': 1.0079009532928467, 'timing/data': 0.00034400448203086853, 'timing/model': 0.6199483014643192, 'learning_rate/qwen_vl_interface': 9.007921766867823e-06, 'learning_rate/action_model': 9.007921766867822e-05, 'epoch': 0.54})
+[2026-07-28 18:19:06,461][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.031205814331769943, 'train/grad_norm_pre_clip': 1.0804641246795654, 'timing/data': 0.000328991562128067, 'timing/model': 0.6186910774558783, 'learning_rate/qwen_vl_interface': 9.005580256303804e-06, 'learning_rate/action_model': 9.005580256303803e-05, 'epoch': 0.54})
+[2026-07-28 18:19:15,859][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.02249792031943798, 'train/grad_norm_pre_clip': 1.1211092472076416, 'timing/data': 0.0003398824483156204, 'timing/model': 0.8572083795443177, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.54})
+[2026-07-28 18:19:25,277][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.029964953660964966, 'train/grad_norm_pre_clip': 1.4753713607788086, 'timing/data': 0.00033543631434440613, 'timing/model': 0.852972848340869, 'learning_rate/qwen_vl_interface': 9.000889925830964e-06, 'learning_rate/action_model': 9.000889925830965e-05, 'epoch': 0.54})
+[2026-07-28 18:19:34,428][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.031738102436065674, 'train/grad_norm_pre_clip': 2.610956907272339, 'timing/data': 0.00032662972807884216, 'timing/model': 0.6198002388700843, 'learning_rate/qwen_vl_interface': 8.998541108965648e-06, 'learning_rate/action_model': 8.998541108965648e-05, 'epoch': 0.54})
+[2026-07-28 18:19:43,559][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.010593906044960022, 'train/grad_norm_pre_clip': 0.5803703665733337, 'timing/data': 0.0005522994324564934, 'timing/model': 0.6186546608805656, 'learning_rate/qwen_vl_interface': 8.996189859712704e-06, 'learning_rate/action_model': 8.996189859712703e-05, 'epoch': 0.54})
+[2026-07-28 18:19:52,934][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.02572239562869072, 'train/grad_norm_pre_clip': 1.118266224861145, 'timing/data': 0.00033045001327991486, 'timing/model': 0.6209132084622979, 'learning_rate/qwen_vl_interface': 8.993836179597833e-06, 'learning_rate/action_model': 8.993836179597833e-05, 'epoch': 0.54})
+[2026-07-28 18:20:02,084][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.04249257594347, 'train/grad_norm_pre_clip': 1.4880328178405762, 'timing/data': 0.00034533627331256866, 'timing/model': 0.6194461034610868, 'learning_rate/qwen_vl_interface': 8.99148007014831e-06, 'learning_rate/action_model': 8.99148007014831e-05, 'epoch': 0.54})
+[2026-07-28 18:20:11,485][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.015916669741272926, 'train/grad_norm_pre_clip': 0.7470299601554871, 'timing/data': 0.0003319615498185158, 'timing/model': 0.6197585677728057, 'learning_rate/qwen_vl_interface': 8.989121532892989e-06, 'learning_rate/action_model': 8.989121532892988e-05, 'epoch': 0.54})
+[2026-07-28 18:20:20,617][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.017933117225766182, 'train/grad_norm_pre_clip': 1.0078091621398926, 'timing/data': 0.0003424547612667084, 'timing/model': 0.6186792375519872, 'learning_rate/qwen_vl_interface': 8.986760569362302e-06, 'learning_rate/action_model': 8.986760569362301e-05, 'epoch': 0.54})
+[2026-07-28 18:20:29,999][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.035580482333898544, 'train/grad_norm_pre_clip': 0.9983949065208435, 'timing/data': 0.00036896392703056335, 'timing/model': 0.620340215973556, 'learning_rate/qwen_vl_interface': 8.984397181088251e-06, 'learning_rate/action_model': 8.984397181088251e-05, 'epoch': 0.54})
+[2026-07-28 18:20:39,310][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.042880743741989136, 'train/grad_norm_pre_clip': 1.2224633693695068, 'timing/data': 0.0005804663524031639, 'timing/model': 0.6575293485075235, 'learning_rate/qwen_vl_interface': 8.982031369604412e-06, 'learning_rate/action_model': 8.982031369604411e-05, 'epoch': 0.54})
+[2026-07-28 18:20:48,560][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.014264209195971489, 'train/grad_norm_pre_clip': 0.8574228286743164, 'timing/data': 0.00041049346327781677, 'timing/model': 0.6209513107314706, 'learning_rate/qwen_vl_interface': 8.979663136445936e-06, 'learning_rate/action_model': 8.979663136445935e-05, 'epoch': 0.54})
+[2026-07-28 18:20:58,247][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.04151367023587227, 'train/grad_norm_pre_clip': 1.6562540531158447, 'timing/data': 0.00037082191556692123, 'timing/model': 1.0425566481426358, 'learning_rate/qwen_vl_interface': 8.977292483149542e-06, 'learning_rate/action_model': 8.977292483149542e-05, 'epoch': 0.54})
+[2026-07-28 18:21:07,208][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.03109370544552803, 'train/grad_norm_pre_clip': 1.7553590536117554, 'timing/data': 0.00041821226477622986, 'timing/model': 0.6215554149821401, 'learning_rate/qwen_vl_interface': 8.974919411253523e-06, 'learning_rate/action_model': 8.974919411253523e-05, 'epoch': 0.54})
+[2026-07-28 18:21:16,423][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.06985076516866684, 'train/grad_norm_pre_clip': 0.7689751386642456, 'timing/data': 0.0004111761227250099, 'timing/model': 0.6195736518129706, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.54})
+[2026-07-28 18:21:25,579][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.024615352973341942, 'train/grad_norm_pre_clip': 1.0629031658172607, 'timing/data': 0.00033949408680200577, 'timing/model': 0.6257075369358063, 'learning_rate/qwen_vl_interface': 8.970166017823621e-06, 'learning_rate/action_model': 8.97016601782362e-05, 'epoch': 0.54})
+[2026-07-28 18:21:35,213][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.028621304780244827, 'train/grad_norm_pre_clip': 1.2587803602218628, 'timing/data': 0.0003415308892726898, 'timing/model': 0.8642582548782229, 'learning_rate/qwen_vl_interface': 8.96778569937416e-06, 'learning_rate/action_model': 8.96778569937416e-05, 'epoch': 0.54})
+[2026-07-28 18:21:44,625][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.03492843359708786, 'train/grad_norm_pre_clip': 2.0305659770965576, 'timing/data': 0.00033225957304239273, 'timing/model': 0.6232390515506268, 'learning_rate/qwen_vl_interface': 8.965402968493922e-06, 'learning_rate/action_model': 8.965402968493922e-05, 'epoch': 0.54})
+[2026-07-28 18:21:53,870][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.05095364898443222, 'train/grad_norm_pre_clip': 1.5925620794296265, 'timing/data': 0.00033944007009267807, 'timing/model': 0.6213564770296216, 'learning_rate/qwen_vl_interface': 8.96301782672904e-06, 'learning_rate/action_model': 8.963017826729038e-05, 'epoch': 0.54})
+[2026-07-28 18:22:02,841][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.026360755786299706, 'train/grad_norm_pre_clip': 0.8446090221405029, 'timing/data': 0.0003721369430422783, 'timing/model': 0.6223845575004816, 'learning_rate/qwen_vl_interface': 8.960630275627197e-06, 'learning_rate/action_model': 8.960630275627197e-05, 'epoch': 0.54})
+[2026-07-28 18:22:12,306][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.031952112913131714, 'train/grad_norm_pre_clip': 1.2586742639541626, 'timing/data': 0.0003513302654027939, 'timing/model': 0.6266385726630688, 'learning_rate/qwen_vl_interface': 8.958240316737655e-06, 'learning_rate/action_model': 8.958240316737656e-05, 'epoch': 0.54})
+[2026-07-28 18:22:21,485][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.09108316898345947, 'train/grad_norm_pre_clip': 3.6938729286193848, 'timing/data': 0.0003495989367365837, 'timing/model': 0.6208327459171414, 'learning_rate/qwen_vl_interface': 8.955847951611233e-06, 'learning_rate/action_model': 8.955847951611232e-05, 'epoch': 0.54})
+[2026-07-28 18:22:30,711][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.053788840770721436, 'train/grad_norm_pre_clip': 1.379928469657898, 'timing/data': 0.0003575049340724945, 'timing/model': 0.6213928125798702, 'learning_rate/qwen_vl_interface': 8.953453181800307e-06, 'learning_rate/action_model': 8.953453181800306e-05, 'epoch': 0.54})
+[2026-07-28 18:22:39,708][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.041472479701042175, 'train/grad_norm_pre_clip': 1.313007116317749, 'timing/data': 0.0003478126600384712, 'timing/model': 0.6223741751164198, 'learning_rate/qwen_vl_interface': 8.951056008858819e-06, 'learning_rate/action_model': 8.95105600885882e-05, 'epoch': 0.54})
+[2026-07-28 18:22:49,153][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.016097625717520714, 'train/grad_norm_pre_clip': 1.0800129175186157, 'timing/data': 0.00038082245737314224, 'timing/model': 0.6234363820403814, 'learning_rate/qwen_vl_interface': 8.94865643434227e-06, 'learning_rate/action_model': 8.948656434342269e-05, 'epoch': 0.54})
+[2026-07-28 18:22:58,276][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.03480524569749832, 'train/grad_norm_pre_clip': 0.893538773059845, 'timing/data': 0.0003403695300221443, 'timing/model': 0.6204529842361808, 'learning_rate/qwen_vl_interface': 8.946254459807717e-06, 'learning_rate/action_model': 8.946254459807717e-05, 'epoch': 0.54})
+[2026-07-28 18:23:07,488][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.05065162852406502, 'train/grad_norm_pre_clip': 1.2098301649093628, 'timing/data': 0.00033574458211660385, 'timing/model': 0.6201685108244419, 'learning_rate/qwen_vl_interface': 8.943850086813773e-06, 'learning_rate/action_model': 8.943850086813773e-05, 'epoch': 0.54})
+[2026-07-28 18:23:16,611][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.09473002701997757, 'train/grad_norm_pre_clip': 2.106614589691162, 'timing/data': 0.0003466494381427765, 'timing/model': 0.8001803085207939, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.54})
+[2026-07-28 18:23:25,802][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.0429597832262516, 'train/grad_norm_pre_clip': 1.5307646989822388, 'timing/data': 0.00041873473674058914, 'timing/model': 0.619978261180222, 'learning_rate/qwen_vl_interface': 8.93903415168996e-06, 'learning_rate/action_model': 8.93903415168996e-05, 'epoch': 0.54})
+[2026-07-28 18:23:34,757][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.022053562104701996, 'train/grad_norm_pre_clip': 0.537125289440155, 'timing/data': 0.00033408496528863907, 'timing/model': 0.6178032997995615, 'learning_rate/qwen_vl_interface': 8.9366225926851e-06, 'learning_rate/action_model': 8.936622592685098e-05, 'epoch': 0.54})
+[2026-07-28 18:23:43,707][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.06912876665592194, 'train/grad_norm_pre_clip': 1.412779688835144, 'timing/data': 0.0003551645204424858, 'timing/model': 0.6220110263675451, 'learning_rate/qwen_vl_interface': 8.934208641470863e-06, 'learning_rate/action_model': 8.934208641470863e-05, 'epoch': 0.54})
+[2026-07-28 18:23:53,358][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.05812890827655792, 'train/grad_norm_pre_clip': 1.3258458375930786, 'timing/data': 0.0008735274896025658, 'timing/model': 0.8653415208682418, 'learning_rate/qwen_vl_interface': 8.931792299613637e-06, 'learning_rate/action_model': 8.931792299613636e-05, 'epoch': 0.54})
+[2026-07-28 18:24:02,767][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.020803939551115036, 'train/grad_norm_pre_clip': 1.2129572629928589, 'timing/data': 0.0004726443439722061, 'timing/model': 0.8521583089604974, 'learning_rate/qwen_vl_interface': 8.92937356868136e-06, 'learning_rate/action_model': 8.92937356868136e-05, 'epoch': 0.54})
+[2026-07-28 18:24:11,989][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.03474821150302887, 'train/grad_norm_pre_clip': 0.9443567395210266, 'timing/data': 0.0003955475986003876, 'timing/model': 0.6197316804900765, 'learning_rate/qwen_vl_interface': 8.926952450243522e-06, 'learning_rate/action_model': 8.926952450243522e-05, 'epoch': 0.54})
+[2026-07-28 18:24:21,200][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.02655467949807644, 'train/grad_norm_pre_clip': 0.6968210339546204, 'timing/data': 0.0003354167565703392, 'timing/model': 0.6186296213418245, 'learning_rate/qwen_vl_interface': 8.924528945871157e-06, 'learning_rate/action_model': 8.924528945871156e-05, 'epoch': 0.54})
+[2026-07-28 18:24:30,358][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.04051068052649498, 'train/grad_norm_pre_clip': 1.5550113916397095, 'timing/data': 0.00035366974771022797, 'timing/model': 0.6200416442006826, 'learning_rate/qwen_vl_interface': 8.922103057136857e-06, 'learning_rate/action_model': 8.922103057136856e-05, 'epoch': 0.54})
+[2026-07-28 18:24:39,750][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.02747683972120285, 'train/grad_norm_pre_clip': 1.0831859111785889, 'timing/data': 0.0003737211227416992, 'timing/model': 0.8741599358618259, 'learning_rate/qwen_vl_interface': 8.919674785614747e-06, 'learning_rate/action_model': 8.919674785614747e-05, 'epoch': 0.54})
+[2026-07-28 18:24:48,925][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.05807603895664215, 'train/grad_norm_pre_clip': 1.4808728694915771, 'timing/data': 0.0003640521317720413, 'timing/model': 0.6208505462855101, 'learning_rate/qwen_vl_interface': 8.91724413288051e-06, 'learning_rate/action_model': 8.91724413288051e-05, 'epoch': 0.54})
+[2026-07-28 18:24:58,345][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.03851576894521713, 'train/grad_norm_pre_clip': 1.6129426956176758, 'timing/data': 0.00043845921754837036, 'timing/model': 0.6200838293880224, 'learning_rate/qwen_vl_interface': 8.91481110051137e-06, 'learning_rate/action_model': 8.91481110051137e-05, 'epoch': 0.54})
+[2026-07-28 18:25:07,517][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.01889313943684101, 'train/grad_norm_pre_clip': 0.8072128891944885, 'timing/data': 0.00036535877734422684, 'timing/model': 0.6198080955073237, 'learning_rate/qwen_vl_interface': 8.912375690086097e-06, 'learning_rate/action_model': 8.912375690086096e-05, 'epoch': 0.54})
+[2026-07-28 18:25:16,715][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.04149189218878746, 'train/grad_norm_pre_clip': 1.0856773853302002, 'timing/data': 0.0003527281805872917, 'timing/model': 0.6210066163912416, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.54})
+[2026-07-28 18:25:25,661][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.029544387012720108, 'train/grad_norm_pre_clip': 1.0499498844146729, 'timing/data': 0.00044855009764432907, 'timing/model': 0.6191316740587354, 'learning_rate/qwen_vl_interface': 8.907497741389932e-06, 'learning_rate/action_model': 8.907497741389932e-05, 'epoch': 0.54})
+[2026-07-28 18:25:34,838][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.030850078910589218, 'train/grad_norm_pre_clip': 1.3173928260803223, 'timing/data': 0.000360928475856781, 'timing/model': 0.8498816397041082, 'learning_rate/qwen_vl_interface': 8.905055206284289e-06, 'learning_rate/action_model': 8.905055206284288e-05, 'epoch': 0.54})
+[2026-07-28 18:25:44,319][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.014727792702615261, 'train/grad_norm_pre_clip': 0.6184087991714478, 'timing/data': 0.0003331294283270836, 'timing/model': 0.6208404153585434, 'learning_rate/qwen_vl_interface': 8.902610299453005e-06, 'learning_rate/action_model': 8.902610299453004e-05, 'epoch': 0.54})
+[2026-07-28 18:25:53,474][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.03924446925520897, 'train/grad_norm_pre_clip': 1.2874850034713745, 'timing/data': 0.0003609834238886833, 'timing/model': 0.620929628610611, 'learning_rate/qwen_vl_interface': 8.900163022482554e-06, 'learning_rate/action_model': 8.900163022482553e-05, 'epoch': 0.54})
+[2026-07-28 18:26:02,656][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.03705275431275368, 'train/grad_norm_pre_clip': 1.1333417892456055, 'timing/data': 0.0004451647400856018, 'timing/model': 0.6175784934312105, 'learning_rate/qwen_vl_interface': 8.897713376960946e-06, 'learning_rate/action_model': 8.897713376960946e-05, 'epoch': 0.54})
+[2026-07-28 18:26:12,073][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.0482003279030323, 'train/grad_norm_pre_clip': 1.118820071220398, 'timing/data': 0.000344201922416687, 'timing/model': 0.6183665301650763, 'learning_rate/qwen_vl_interface': 8.895261364477731e-06, 'learning_rate/action_model': 8.895261364477731e-05, 'epoch': 0.54})
+[2026-07-28 18:26:21,246][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.04291616007685661, 'train/grad_norm_pre_clip': 1.2449790239334106, 'timing/data': 0.0004689600318670273, 'timing/model': 0.621298992075026, 'learning_rate/qwen_vl_interface': 8.892806986623993e-06, 'learning_rate/action_model': 8.892806986623992e-05, 'epoch': 0.54})
+[2026-07-28 18:26:30,743][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.05284909904003143, 'train/grad_norm_pre_clip': 1.8614088296890259, 'timing/data': 0.0003575338050723076, 'timing/model': 0.6234006015583873, 'learning_rate/qwen_vl_interface': 8.890350244992349e-06, 'learning_rate/action_model': 8.890350244992348e-05, 'epoch': 0.54})
+[2026-07-28 18:26:40,127][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.023644182831048965, 'train/grad_norm_pre_clip': 0.9701257348060608, 'timing/data': 0.0004387851804494858, 'timing/model': 0.851997135207057, 'learning_rate/qwen_vl_interface': 8.887891141176954e-06, 'learning_rate/action_model': 8.887891141176954e-05, 'epoch': 0.54})
+[2026-07-28 18:26:49,411][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.048901427537202835, 'train/grad_norm_pre_clip': 2.1528615951538086, 'timing/data': 0.0003638584166765213, 'timing/model': 0.6234999457374215, 'learning_rate/qwen_vl_interface': 8.885429676773495e-06, 'learning_rate/action_model': 8.885429676773495e-05, 'epoch': 0.55})
+[2026-07-28 18:26:58,631][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.027438733726739883, 'train/grad_norm_pre_clip': 1.0833441019058228, 'timing/data': 0.00041877664625644684, 'timing/model': 0.6221700673922896, 'learning_rate/qwen_vl_interface': 8.882965853379188e-06, 'learning_rate/action_model': 8.882965853379187e-05, 'epoch': 0.55})
+[2026-07-28 18:27:08,074][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.06302846223115921, 'train/grad_norm_pre_clip': 1.3226505517959595, 'timing/data': 0.0003305794671177864, 'timing/model': 0.623419270850718, 'learning_rate/qwen_vl_interface': 8.88049967259278e-06, 'learning_rate/action_model': 8.880499672592778e-05, 'epoch': 0.55})
+[2026-07-28 18:27:17,263][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.017789535224437714, 'train/grad_norm_pre_clip': 0.8922605514526367, 'timing/data': 0.00039051566272974014, 'timing/model': 0.62129230145365, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.55})
+[2026-07-28 18:27:26,768][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.03134981542825699, 'train/grad_norm_pre_clip': 1.1276806592941284, 'timing/data': 0.0003922227770090103, 'timing/model': 0.6294085495173931, 'learning_rate/qwen_vl_interface': 8.875560245246307e-06, 'learning_rate/action_model': 8.875560245246307e-05, 'epoch': 0.55})
+[2026-07-28 18:27:35,980][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.04305867478251457, 'train/grad_norm_pre_clip': 1.4176808595657349, 'timing/data': 0.0003355126827955246, 'timing/model': 0.6213074438273907, 'learning_rate/qwen_vl_interface': 8.873087001891381e-06, 'learning_rate/action_model': 8.87308700189138e-05, 'epoch': 0.55})
+[2026-07-28 18:27:45,163][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.030404552817344666, 'train/grad_norm_pre_clip': 0.9022057056427002, 'timing/data': 0.00034598447382450104, 'timing/model': 0.6207305397838354, 'learning_rate/qwen_vl_interface': 8.870611407554635e-06, 'learning_rate/action_model': 8.870611407554636e-05, 'epoch': 0.55})
+[2026-07-28 18:27:54,734][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.025431692600250244, 'train/grad_norm_pre_clip': 0.7698675990104675, 'timing/data': 0.0005613705143332481, 'timing/model': 0.6206160187721252, 'learning_rate/qwen_vl_interface': 8.868133463842456e-06, 'learning_rate/action_model': 8.868133463842457e-05, 'epoch': 0.55})
+[2026-07-28 18:28:03,955][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.03107764944434166, 'train/grad_norm_pre_clip': 1.1911007165908813, 'timing/data': 0.000558461993932724, 'timing/model': 0.6193029228597879, 'learning_rate/qwen_vl_interface': 8.865653172362755e-06, 'learning_rate/action_model': 8.865653172362754e-05, 'epoch': 0.55})
+[2026-07-28 18:28:13,382][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.032520413398742676, 'train/grad_norm_pre_clip': 1.3909984827041626, 'timing/data': 0.00032638758420944214, 'timing/model': 0.8614557450637221, 'learning_rate/qwen_vl_interface': 8.863170534724962e-06, 'learning_rate/action_model': 8.863170534724961e-05, 'epoch': 0.55})
+[2026-07-28 18:28:22,529][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.041830919682979584, 'train/grad_norm_pre_clip': 1.126373529434204, 'timing/data': 0.0003986191004514694, 'timing/model': 0.6194651061668992, 'learning_rate/qwen_vl_interface': 8.860685552540038e-06, 'learning_rate/action_model': 8.860685552540037e-05, 'epoch': 0.55})
+[2026-07-28 18:28:31,972][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.021048758178949356, 'train/grad_norm_pre_clip': 0.9971883296966553, 'timing/data': 0.00045904889702796936, 'timing/model': 0.6365345632657409, 'learning_rate/qwen_vl_interface': 8.858198227420461e-06, 'learning_rate/action_model': 8.85819822742046e-05, 'epoch': 0.55})
+[2026-07-28 18:28:41,207][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.03784005716443062, 'train/grad_norm_pre_clip': 1.0203994512557983, 'timing/data': 0.00035448186099529266, 'timing/model': 0.6217091269791126, 'learning_rate/qwen_vl_interface': 8.855708560980225e-06, 'learning_rate/action_model': 8.855708560980225e-05, 'epoch': 0.55})
+[2026-07-28 18:28:50,366][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.01959667168557644, 'train/grad_norm_pre_clip': 0.6824490427970886, 'timing/data': 0.00034572649747133255, 'timing/model': 0.6204167986288667, 'learning_rate/qwen_vl_interface': 8.853216554834852e-06, 'learning_rate/action_model': 8.85321655483485e-05, 'epoch': 0.55})
+[2026-07-28 18:28:59,779][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.034201085567474365, 'train/grad_norm_pre_clip': 1.3644461631774902, 'timing/data': 0.00042839162051677704, 'timing/model': 0.6188184255734086, 'learning_rate/qwen_vl_interface': 8.850722210601376e-06, 'learning_rate/action_model': 8.850722210601377e-05, 'epoch': 0.55})
+[2026-07-28 18:29:08,999][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.036178942769765854, 'train/grad_norm_pre_clip': 1.055618405342102, 'timing/data': 0.0003586290404200554, 'timing/model': 0.6203356198966503, 'learning_rate/qwen_vl_interface': 8.848225529898351e-06, 'learning_rate/action_model': 8.84822552989835e-05, 'epoch': 0.55})
+[2026-07-28 18:29:18,436][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.02125270664691925, 'train/grad_norm_pre_clip': 0.8110888600349426, 'timing/data': 0.00037407036870718, 'timing/model': 0.6228077672421932, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.55})
+[2026-07-28 18:29:27,877][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.02391083538532257, 'train/grad_norm_pre_clip': 1.1818403005599976, 'timing/data': 0.00034056883305311203, 'timing/model': 0.6246231803670526, 'learning_rate/qwen_vl_interface': 8.84322516556544e-06, 'learning_rate/action_model': 8.84322516556544e-05, 'epoch': 0.55})
+[2026-07-28 18:29:37,062][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.05838143825531006, 'train/grad_norm_pre_clip': 1.2950178384780884, 'timing/data': 0.00046508852392435074, 'timing/model': 0.6181061761453748, 'learning_rate/qwen_vl_interface': 8.840721485180237e-06, 'learning_rate/action_model': 8.840721485180236e-05, 'epoch': 0.55})
+[2026-07-28 18:29:46,342][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.034549131989479065, 'train/grad_norm_pre_clip': 2.0172715187072754, 'timing/data': 0.00041570793837308884, 'timing/model': 0.618943895213306, 'learning_rate/qwen_vl_interface': 8.838215474814848e-06, 'learning_rate/action_model': 8.838215474814847e-05, 'epoch': 0.55})
+[2026-07-28 18:29:55,759][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.031133392825722694, 'train/grad_norm_pre_clip': 0.9437042474746704, 'timing/data': 0.00035498570650815964, 'timing/model': 0.6206349702551961, 'learning_rate/qwen_vl_interface': 8.835707136095392e-06, 'learning_rate/action_model': 8.835707136095391e-05, 'epoch': 0.55})
+[2026-07-28 18:30:05,139][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.026149261742830276, 'train/grad_norm_pre_clip': 1.0053095817565918, 'timing/data': 0.00035193655639886856, 'timing/model': 0.6212424971163273, 'learning_rate/qwen_vl_interface': 8.833196470649505e-06, 'learning_rate/action_model': 8.833196470649505e-05, 'epoch': 0.55})
+[2026-07-28 18:30:14,175][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.08395862579345703, 'train/grad_norm_pre_clip': 2.3031861782073975, 'timing/data': 0.0003706933930516243, 'timing/model': 0.6180436313152313, 'learning_rate/qwen_vl_interface': 8.83068348010633e-06, 'learning_rate/action_model': 8.83068348010633e-05, 'epoch': 0.55})
+[2026-07-28 18:30:23,362][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.05073101073503494, 'train/grad_norm_pre_clip': 1.4594815969467163, 'timing/data': 0.00035231467336416245, 'timing/model': 0.6178216235712171, 'learning_rate/qwen_vl_interface': 8.82816816609652e-06, 'learning_rate/action_model': 8.828168166096521e-05, 'epoch': 0.55})
+[2026-07-28 18:30:32,297][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.049697041511535645, 'train/grad_norm_pre_clip': 1.4262596368789673, 'timing/data': 0.0003379322588443756, 'timing/model': 0.6218229234218597, 'learning_rate/qwen_vl_interface': 8.825650530252238e-06, 'learning_rate/action_model': 8.825650530252238e-05, 'epoch': 0.55})
+[2026-07-28 18:30:41,920][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.05163412168622017, 'train/grad_norm_pre_clip': 1.487830400466919, 'timing/data': 0.0004191771149635315, 'timing/model': 0.6172505225986242, 'learning_rate/qwen_vl_interface': 8.823130574207144e-06, 'learning_rate/action_model': 8.823130574207143e-05, 'epoch': 0.55})
+[2026-07-28 18:30:50,861][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.055370159447193146, 'train/grad_norm_pre_clip': 1.8646303415298462, 'timing/data': 0.00034300703555345535, 'timing/model': 0.6197088826447725, 'learning_rate/qwen_vl_interface': 8.82060829959641e-06, 'learning_rate/action_model': 8.820608299596411e-05, 'epoch': 0.55})
+[2026-07-28 18:31:00,013][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.021755510941147804, 'train/grad_norm_pre_clip': 0.6961013078689575, 'timing/data': 0.0004062606021761894, 'timing/model': 0.6213146178051829, 'learning_rate/qwen_vl_interface': 8.81808370805672e-06, 'learning_rate/action_model': 8.818083708056721e-05, 'epoch': 0.55})
+[2026-07-28 18:31:09,473][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.047595810145139694, 'train/grad_norm_pre_clip': 1.8070849180221558, 'timing/data': 0.00039611104875802994, 'timing/model': 0.8747398629784584, 'learning_rate/qwen_vl_interface': 8.815556801226251e-06, 'learning_rate/action_model': 8.815556801226251e-05, 'epoch': 0.55})
+[2026-07-28 18:31:18,891][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.020518088713288307, 'train/grad_norm_pre_clip': 1.5735152959823608, 'timing/data': 0.000372864305973053, 'timing/model': 0.8424699893221259, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.55})
+[2026-07-28 18:31:28,139][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.05216250941157341, 'train/grad_norm_pre_clip': 1.9235299825668335, 'timing/data': 0.000428101047873497, 'timing/model': 0.6203304845839739, 'learning_rate/qwen_vl_interface': 8.810496048253203e-06, 'learning_rate/action_model': 8.810496048253202e-05, 'epoch': 0.55})
+[2026-07-28 18:31:37,373][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.01812039501965046, 'train/grad_norm_pre_clip': 0.47971588373184204, 'timing/data': 0.00045966170728206635, 'timing/model': 0.6233951663598418, 'learning_rate/qwen_vl_interface': 8.80796220539449e-06, 'learning_rate/action_model': 8.80796220539449e-05, 'epoch': 0.55})
+[2026-07-28 18:31:46,556][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3000
+[2026-07-28 18:31:56,777][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3000
+[2026-07-28 18:44:22,567][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000
+[2026-07-28 18:44:22,568][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.03766188025474548, 'train/grad_norm_pre_clip': 1.3144978284835815, 'eval/action_loss/seconds': 10.217314990237355, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.03712918609380722, 'eval/latency_3/loss': 0.03712918609380722, 'eval/flappy/loss': 0.03712918609380722, 'eval/flappy/latency_3/loss': 0.03712918609380722, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 283.24500333927574, 'rl_games_eval/mid_train/mean_length': 2301.4, 'rl_games_eval/mid_train/std_reward': 161.9693615741844, 'rl_games_eval/mid_train/std_length': 1305.6022135397902, 'rl_games_eval/mid_train/macro_mean_reward': 283.24500333927574, 'rl_games_eval/mid_train/macro_mean_length': 2301.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 283.24500333927574, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2301.4, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 161.9693615741844, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1305.6022135397902, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 283.24500333927574, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003377161920070648, 'timing/model': 0.6239254483953118, 'learning_rate/qwen_vl_interface': 8.805426053812731e-06, 'learning_rate/action_model': 8.805426053812731e-05, 'epoch': 0.55})
+[2026-07-28 18:44:22,571][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3000_state
+[2026-07-28 18:44:22,571][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 18:45:00,320][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3000_state/pytorch_model
+[2026-07-28 18:45:00,321][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3000_state/sampler.bin
+[2026-07-28 18:45:00,321][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3000_state/sampler_1.bin
+[2026-07-28 18:45:00,323][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3000_state/random_states_0.pkl
+[2026-07-28 18:45:08,522][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000
+[2026-07-28 18:45:08,522][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 18:45:08,544][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 18:45:17,976][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.05076492950320244, 'train/grad_norm_pre_clip': 1.3923163414001465, 'timing/data': 0.00045276153832674026, 'timing/model': 0.6181969558820128, 'learning_rate/qwen_vl_interface': 8.802887595153607e-06, 'learning_rate/action_model': 8.802887595153606e-05, 'epoch': 0.55})
+[2026-07-28 18:45:27,367][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.0376812182366848, 'train/grad_norm_pre_clip': 1.1219979524612427, 'timing/data': 0.00035473424941301346, 'timing/model': 0.6222213776782155, 'learning_rate/qwen_vl_interface': 8.800346831064294e-06, 'learning_rate/action_model': 8.800346831064294e-05, 'epoch': 0.55})
+[2026-07-28 18:45:36,509][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.04815838485956192, 'train/grad_norm_pre_clip': 1.1185708045959473, 'timing/data': 0.0006319563835859299, 'timing/model': 0.6204603286460042, 'learning_rate/qwen_vl_interface': 8.797803763193471e-06, 'learning_rate/action_model': 8.797803763193471e-05, 'epoch': 0.55})
+[2026-07-28 18:45:45,984][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.024821724742650986, 'train/grad_norm_pre_clip': 0.9418717622756958, 'timing/data': 0.0004220949485898018, 'timing/model': 0.6241429438814521, 'learning_rate/qwen_vl_interface': 8.795258393191302e-06, 'learning_rate/action_model': 8.795258393191301e-05, 'epoch': 0.55})
+[2026-07-28 18:45:55,459][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.02733663097023964, 'train/grad_norm_pre_clip': 1.1216925382614136, 'timing/data': 0.0003623310476541519, 'timing/model': 0.6234103497117758, 'learning_rate/qwen_vl_interface': 8.792710722709451e-06, 'learning_rate/action_model': 8.792710722709451e-05, 'epoch': 0.55})
+[2026-07-28 18:46:04,672][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.06433392316102982, 'train/grad_norm_pre_clip': 1.838845133781433, 'timing/data': 0.00046566687524318695, 'timing/model': 0.6187326302751899, 'learning_rate/qwen_vl_interface': 8.790160753401076e-06, 'learning_rate/action_model': 8.790160753401077e-05, 'epoch': 0.55})
+[2026-07-28 18:46:14,092][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.04734271764755249, 'train/grad_norm_pre_clip': 2.09649658203125, 'timing/data': 0.000364154577255249, 'timing/model': 0.6225871909409761, 'learning_rate/qwen_vl_interface': 8.787608486920824e-06, 'learning_rate/action_model': 8.787608486920824e-05, 'epoch': 0.55})
+[2026-07-28 18:46:23,627][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.054602645337581635, 'train/grad_norm_pre_clip': 1.423776626586914, 'timing/data': 0.00041840970516204834, 'timing/model': 0.6291638007387519, 'learning_rate/qwen_vl_interface': 8.78505392492483e-06, 'learning_rate/action_model': 8.78505392492483e-05, 'epoch': 0.55})
+[2026-07-28 18:46:33,106][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.020825929939746857, 'train/grad_norm_pre_clip': 0.6668676733970642, 'timing/data': 0.00037722010165452957, 'timing/model': 0.8554002465680242, 'learning_rate/qwen_vl_interface': 8.782497069070725e-06, 'learning_rate/action_model': 8.782497069070725e-05, 'epoch': 0.55})
+[2026-07-28 18:46:42,324][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.042035363614559174, 'train/grad_norm_pre_clip': 1.5956966876983643, 'timing/data': 0.0003327280282974243, 'timing/model': 0.6222369903698564, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.55})
+[2026-07-28 18:46:51,794][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.06931286305189133, 'train/grad_norm_pre_clip': 1.593408465385437, 'timing/data': 0.00043365824967622757, 'timing/model': 0.8687651744112372, 'learning_rate/qwen_vl_interface': 8.777376482426129e-06, 'learning_rate/action_model': 8.777376482426128e-05, 'epoch': 0.55})
+[2026-07-28 18:47:01,080][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.05108204111456871, 'train/grad_norm_pre_clip': 1.0665761232376099, 'timing/data': 0.0003393096849322319, 'timing/model': 0.6223462456837296, 'learning_rate/qwen_vl_interface': 8.774812754958332e-06, 'learning_rate/action_model': 8.774812754958331e-05, 'epoch': 0.55})
+[2026-07-28 18:47:10,251][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.03289619833230972, 'train/grad_norm_pre_clip': 1.003088116645813, 'timing/data': 0.00032227952033281326, 'timing/model': 0.6215475276112556, 'learning_rate/qwen_vl_interface': 8.772246740277805e-06, 'learning_rate/action_model': 8.772246740277805e-05, 'epoch': 0.55})
+[2026-07-28 18:47:19,428][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.04954686760902405, 'train/grad_norm_pre_clip': 1.2351689338684082, 'timing/data': 0.000370771624147892, 'timing/model': 0.6241882275789976, 'learning_rate/qwen_vl_interface': 8.769678440049608e-06, 'learning_rate/action_model': 8.769678440049608e-05, 'epoch': 0.55})
+[2026-07-28 18:47:28,909][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.07658272981643677, 'train/grad_norm_pre_clip': 1.2864534854888916, 'timing/data': 0.00044041965156793594, 'timing/model': 0.6221677167341113, 'learning_rate/qwen_vl_interface': 8.767107855940287e-06, 'learning_rate/action_model': 8.767107855940286e-05, 'epoch': 0.55})
+[2026-07-28 18:47:38,442][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.04221227765083313, 'train/grad_norm_pre_clip': 1.235154628753662, 'timing/data': 0.00035896431654691696, 'timing/model': 0.6226784959435463, 'learning_rate/qwen_vl_interface': 8.764534989617861e-06, 'learning_rate/action_model': 8.76453498961786e-05, 'epoch': 0.55})
+[2026-07-28 18:47:47,726][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.02664502151310444, 'train/grad_norm_pre_clip': 0.6405002474784851, 'timing/data': 0.0003550508990883827, 'timing/model': 0.6227792724967003, 'learning_rate/qwen_vl_interface': 8.761959842751837e-06, 'learning_rate/action_model': 8.761959842751837e-05, 'epoch': 0.55})
+[2026-07-28 18:47:57,213][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.05446784943342209, 'train/grad_norm_pre_clip': 1.8623913526535034, 'timing/data': 0.00033594854176044464, 'timing/model': 0.6218779915943742, 'learning_rate/qwen_vl_interface': 8.7593824170132e-06, 'learning_rate/action_model': 8.7593824170132e-05, 'epoch': 0.55})
+[2026-07-28 18:48:06,452][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.027785900980234146, 'train/grad_norm_pre_clip': 1.5073802471160889, 'timing/data': 0.00046094972640275955, 'timing/model': 0.6225278219208121, 'learning_rate/qwen_vl_interface': 8.756802714074412e-06, 'learning_rate/action_model': 8.756802714074412e-05, 'epoch': 0.55})
+[2026-07-28 18:48:15,936][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.042585745453834534, 'train/grad_norm_pre_clip': 1.1101782321929932, 'timing/data': 0.000354020856320858, 'timing/model': 0.6240202840417624, 'learning_rate/qwen_vl_interface': 8.75422073560942e-06, 'learning_rate/action_model': 8.754220735609419e-05, 'epoch': 0.55})
+[2026-07-28 18:48:25,316][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.013246087357401848, 'train/grad_norm_pre_clip': 0.42588910460472107, 'timing/data': 0.0003486489877104759, 'timing/model': 0.7992630191147327, 'learning_rate/qwen_vl_interface': 8.751636483293635e-06, 'learning_rate/action_model': 8.751636483293634e-05, 'epoch': 0.55})
+[2026-07-28 18:48:34,569][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.06009653955698013, 'train/grad_norm_pre_clip': 1.5941439867019653, 'timing/data': 0.00035848747938871384, 'timing/model': 0.6191685628145933, 'learning_rate/qwen_vl_interface': 8.749049958803952e-06, 'learning_rate/action_model': 8.749049958803951e-05, 'epoch': 0.56})
+[2026-07-28 18:48:43,914][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.03418605774641037, 'train/grad_norm_pre_clip': 2.091102123260498, 'timing/data': 0.00040738191455602646, 'timing/model': 0.7955051064491272, 'learning_rate/qwen_vl_interface': 8.74646116381874e-06, 'learning_rate/action_model': 8.74646116381874e-05, 'epoch': 0.56})
+[2026-07-28 18:48:53,086][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.019042769446969032, 'train/grad_norm_pre_clip': 0.44362369179725647, 'timing/data': 0.000470542348921299, 'timing/model': 0.6205613482743502, 'learning_rate/qwen_vl_interface': 8.74387010001784e-06, 'learning_rate/action_model': 8.743870100017838e-05, 'epoch': 0.56})
+[2026-07-28 18:49:02,606][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.05999414622783661, 'train/grad_norm_pre_clip': 1.5699703693389893, 'timing/data': 0.0003509204834699631, 'timing/model': 0.6228125151246786, 'learning_rate/qwen_vl_interface': 8.741276769082562e-06, 'learning_rate/action_model': 8.741276769082562e-05, 'epoch': 0.56})
+[2026-07-28 18:49:11,769][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.03790887072682381, 'train/grad_norm_pre_clip': 0.9978107810020447, 'timing/data': 0.000369182787835598, 'timing/model': 0.6215099953114986, 'learning_rate/qwen_vl_interface': 8.738681172695694e-06, 'learning_rate/action_model': 8.738681172695694e-05, 'epoch': 0.56})
+[2026-07-28 18:49:21,193][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.033767253160476685, 'train/grad_norm_pre_clip': 1.3928883075714111, 'timing/data': 0.00033711548894643784, 'timing/model': 0.8628134336322546, 'learning_rate/qwen_vl_interface': 8.73608331254149e-06, 'learning_rate/action_model': 8.736083312541489e-05, 'epoch': 0.56})
+[2026-07-28 18:49:30,587][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.04339536651968956, 'train/grad_norm_pre_clip': 1.0632503032684326, 'timing/data': 0.00034065917134284973, 'timing/model': 0.8581074643880129, 'learning_rate/qwen_vl_interface': 8.733483190305671e-06, 'learning_rate/action_model': 8.733483190305671e-05, 'epoch': 0.56})
+[2026-07-28 18:49:39,790][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.03715870529413223, 'train/grad_norm_pre_clip': 0.6124219298362732, 'timing/data': 0.0003964090719819069, 'timing/model': 0.6233608555048704, 'learning_rate/qwen_vl_interface': 8.730880807675431e-06, 'learning_rate/action_model': 8.73088080767543e-05, 'epoch': 0.56})
+[2026-07-28 18:49:48,973][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.022828448563814163, 'train/grad_norm_pre_clip': 0.4795480966567993, 'timing/data': 0.0005344850942492485, 'timing/model': 0.6203004885464907, 'learning_rate/qwen_vl_interface': 8.728276166339426e-06, 'learning_rate/action_model': 8.728276166339426e-05, 'epoch': 0.56})
+[2026-07-28 18:49:58,105][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.025766711682081223, 'train/grad_norm_pre_clip': 1.2140417098999023, 'timing/data': 0.0003331117331981659, 'timing/model': 0.6184895541518927, 'learning_rate/qwen_vl_interface': 8.725669267987779e-06, 'learning_rate/action_model': 8.725669267987778e-05, 'epoch': 0.56})
+[2026-07-28 18:50:07,521][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.04796339198946953, 'train/grad_norm_pre_clip': 2.1053757667541504, 'timing/data': 0.0003547072410583496, 'timing/model': 0.8635207079350948, 'learning_rate/qwen_vl_interface': 8.723060114312079e-06, 'learning_rate/action_model': 8.723060114312079e-05, 'epoch': 0.56})
+[2026-07-28 18:50:16,987][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.03340023010969162, 'train/grad_norm_pre_clip': 1.0320656299591064, 'timing/data': 0.0003549531102180481, 'timing/model': 0.6214973544701934, 'learning_rate/qwen_vl_interface': 8.720448707005377e-06, 'learning_rate/action_model': 8.720448707005376e-05, 'epoch': 0.56})
+[2026-07-28 18:50:26,262][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.04029272869229317, 'train/grad_norm_pre_clip': 1.0841821432113647, 'timing/data': 0.0003915047273039818, 'timing/model': 0.6204922236502171, 'learning_rate/qwen_vl_interface': 8.717835047762188e-06, 'learning_rate/action_model': 8.717835047762187e-05, 'epoch': 0.56})
+[2026-07-28 18:50:35,177][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.04616229236125946, 'train/grad_norm_pre_clip': 1.017989993095398, 'timing/data': 0.0004351278766989708, 'timing/model': 0.6211397685110569, 'learning_rate/qwen_vl_interface': 8.715219138278486e-06, 'learning_rate/action_model': 8.715219138278486e-05, 'epoch': 0.56})
+[2026-07-28 18:50:44,165][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.0329999178647995, 'train/grad_norm_pre_clip': 1.5942832231521606, 'timing/data': 0.0003480454906821251, 'timing/model': 0.6191364387050271, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.56})
+[2026-07-28 18:50:53,335][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.028915761038661003, 'train/grad_norm_pre_clip': 1.1149718761444092, 'timing/data': 0.00034874584525823593, 'timing/model': 0.6202109633013606, 'learning_rate/qwen_vl_interface': 8.709980575380741e-06, 'learning_rate/action_model': 8.709980575380741e-05, 'epoch': 0.56})
+[2026-07-28 18:51:02,537][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.02655896358191967, 'train/grad_norm_pre_clip': 1.1451621055603027, 'timing/data': 0.00037878379225730896, 'timing/model': 0.8541565295308828, 'learning_rate/qwen_vl_interface': 8.707357925365949e-06, 'learning_rate/action_model': 8.707357925365948e-05, 'epoch': 0.56})
+[2026-07-28 18:51:11,994][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.03720001131296158, 'train/grad_norm_pre_clip': 1.1488606929779053, 'timing/data': 0.0003425534814596176, 'timing/model': 0.6229594144970179, 'learning_rate/qwen_vl_interface': 8.704733031909134e-06, 'learning_rate/action_model': 8.704733031909133e-05, 'epoch': 0.56})
+[2026-07-28 18:51:21,229][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.07554815709590912, 'train/grad_norm_pre_clip': 1.9945635795593262, 'timing/data': 0.00045641884207725525, 'timing/model': 0.6225822363048792, 'learning_rate/qwen_vl_interface': 8.70210589671356e-06, 'learning_rate/action_model': 8.702105896713561e-05, 'epoch': 0.56})
+[2026-07-28 18:51:30,388][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.032901979982852936, 'train/grad_norm_pre_clip': 0.8486046195030212, 'timing/data': 0.0003539426252245903, 'timing/model': 0.6188947688788176, 'learning_rate/qwen_vl_interface': 8.699476521483952e-06, 'learning_rate/action_model': 8.699476521483952e-05, 'epoch': 0.56})
+[2026-07-28 18:51:39,566][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.05571455508470535, 'train/grad_norm_pre_clip': 1.4428138732910156, 'timing/data': 0.0003378354012966156, 'timing/model': 0.6219377899542451, 'learning_rate/qwen_vl_interface': 8.696844907926478e-06, 'learning_rate/action_model': 8.696844907926477e-05, 'epoch': 0.56})
+[2026-07-28 18:51:48,957][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.03188231959939003, 'train/grad_norm_pre_clip': 0.7811952233314514, 'timing/data': 0.00032549165189266205, 'timing/model': 0.6220663879066706, 'learning_rate/qwen_vl_interface': 8.694211057748766e-06, 'learning_rate/action_model': 8.694211057748765e-05, 'epoch': 0.56})
+[2026-07-28 18:51:58,164][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.017217809334397316, 'train/grad_norm_pre_clip': 0.4838328957557678, 'timing/data': 0.0003427630290389061, 'timing/model': 0.6190224187448621, 'learning_rate/qwen_vl_interface': 8.69157497265989e-06, 'learning_rate/action_model': 8.69157497265989e-05, 'epoch': 0.56})
+[2026-07-28 18:52:07,705][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.10547037422657013, 'train/grad_norm_pre_clip': 6.449232578277588, 'timing/data': 0.00044842343777418137, 'timing/model': 0.8898223815485835, 'learning_rate/qwen_vl_interface': 8.68893665437038e-06, 'learning_rate/action_model': 8.688936654370378e-05, 'epoch': 0.56})
+[2026-07-28 18:52:16,878][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.02724476531147957, 'train/grad_norm_pre_clip': 1.130826711654663, 'timing/data': 0.0003718715161085129, 'timing/model': 0.6259560203179717, 'learning_rate/qwen_vl_interface': 8.68629610459221e-06, 'learning_rate/action_model': 8.686296104592209e-05, 'epoch': 0.56})
+[2026-07-28 18:52:26,079][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.02183668315410614, 'train/grad_norm_pre_clip': 0.5171266198158264, 'timing/data': 0.000344129279255867, 'timing/model': 0.6205809386447072, 'learning_rate/qwen_vl_interface': 8.683653325038805e-06, 'learning_rate/action_model': 8.683653325038804e-05, 'epoch': 0.56})
+[2026-07-28 18:52:35,245][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.04317671060562134, 'train/grad_norm_pre_clip': 0.9982659816741943, 'timing/data': 0.0003737742081284523, 'timing/model': 0.6209119921550155, 'learning_rate/qwen_vl_interface': 8.681008317425037e-06, 'learning_rate/action_model': 8.681008317425037e-05, 'epoch': 0.56})
+[2026-07-28 18:52:44,487][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.026917777955532074, 'train/grad_norm_pre_clip': 1.3694510459899902, 'timing/data': 0.00036466214805841446, 'timing/model': 0.6211849879473448, 'learning_rate/qwen_vl_interface': 8.678361083467221e-06, 'learning_rate/action_model': 8.67836108346722e-05, 'epoch': 0.56})
+[2026-07-28 18:52:54,155][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.03360145539045334, 'train/grad_norm_pre_clip': 0.8998404145240784, 'timing/data': 0.00036103930324316025, 'timing/model': 0.6209034295752645, 'learning_rate/qwen_vl_interface': 8.67571162488312e-06, 'learning_rate/action_model': 8.67571162488312e-05, 'epoch': 0.56})
+[2026-07-28 18:53:03,288][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.051849886775016785, 'train/grad_norm_pre_clip': 1.2506287097930908, 'timing/data': 0.00033307261765003204, 'timing/model': 0.6207693330943584, 'learning_rate/qwen_vl_interface': 8.673059943391937e-06, 'learning_rate/action_model': 8.673059943391936e-05, 'epoch': 0.56})
+[2026-07-28 18:53:12,436][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.04868452250957489, 'train/grad_norm_pre_clip': 1.5006537437438965, 'timing/data': 0.0003510620445013046, 'timing/model': 0.6192072266712785, 'learning_rate/qwen_vl_interface': 8.670406040714321e-06, 'learning_rate/action_model': 8.670406040714321e-05, 'epoch': 0.56})
+[2026-07-28 18:53:21,585][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.03216782584786415, 'train/grad_norm_pre_clip': 1.4677170515060425, 'timing/data': 0.00037167780101299286, 'timing/model': 0.6195477424189448, 'learning_rate/qwen_vl_interface': 8.667749918572361e-06, 'learning_rate/action_model': 8.667749918572361e-05, 'epoch': 0.56})
+[2026-07-28 18:53:31,203][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.02983967959880829, 'train/grad_norm_pre_clip': 0.6910191178321838, 'timing/data': 0.00043709948658943176, 'timing/model': 0.622164717875421, 'learning_rate/qwen_vl_interface': 8.665091578689584e-06, 'learning_rate/action_model': 8.665091578689584e-05, 'epoch': 0.56})
+[2026-07-28 18:53:40,164][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.038755372166633606, 'train/grad_norm_pre_clip': 1.6007813215255737, 'timing/data': 0.00041819922626018524, 'timing/model': 0.6183620374649763, 'learning_rate/qwen_vl_interface': 8.662431022790958e-06, 'learning_rate/action_model': 8.662431022790958e-05, 'epoch': 0.56})
+[2026-07-28 18:53:49,308][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.041847020387649536, 'train/grad_norm_pre_clip': 1.4759482145309448, 'timing/data': 0.00033671874552965164, 'timing/model': 0.61927873454988, 'learning_rate/qwen_vl_interface': 8.65976825260289e-06, 'learning_rate/action_model': 8.659768252602888e-05, 'epoch': 0.56})
+[2026-07-28 18:53:58,660][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.05816596373915672, 'train/grad_norm_pre_clip': 1.851089596748352, 'timing/data': 0.0004888027906417847, 'timing/model': 0.8458598339930177, 'learning_rate/qwen_vl_interface': 8.657103269853222e-06, 'learning_rate/action_model': 8.657103269853222e-05, 'epoch': 0.56})
+[2026-07-28 18:54:08,252][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.03618254512548447, 'train/grad_norm_pre_clip': 0.8101382851600647, 'timing/data': 0.0004035942256450653, 'timing/model': 0.8443628745153546, 'learning_rate/qwen_vl_interface': 8.654436076271232e-06, 'learning_rate/action_model': 8.654436076271232e-05, 'epoch': 0.56})
+[2026-07-28 18:54:17,423][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.016628362238407135, 'train/grad_norm_pre_clip': 0.3660445511341095, 'timing/data': 0.0005694972351193428, 'timing/model': 0.6204473758116364, 'learning_rate/qwen_vl_interface': 8.651766673587635e-06, 'learning_rate/action_model': 8.651766673587635e-05, 'epoch': 0.56})
+[2026-07-28 18:54:26,571][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.0535229817032814, 'train/grad_norm_pre_clip': 1.4079712629318237, 'timing/data': 0.0003407895565032959, 'timing/model': 0.6193549670279026, 'learning_rate/qwen_vl_interface': 8.649095063534576e-06, 'learning_rate/action_model': 8.649095063534576e-05, 'epoch': 0.56})
+[2026-07-28 18:54:35,714][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.04241885989904404, 'train/grad_norm_pre_clip': 0.8700300455093384, 'timing/data': 0.0003340030089020729, 'timing/model': 0.618882542476058, 'learning_rate/qwen_vl_interface': 8.646421247845633e-06, 'learning_rate/action_model': 8.646421247845633e-05, 'epoch': 0.56})
+[2026-07-28 18:54:45,109][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.028011122718453407, 'train/grad_norm_pre_clip': 1.2435401678085327, 'timing/data': 0.00042031239718198776, 'timing/model': 0.856303590349853, 'learning_rate/qwen_vl_interface': 8.643745228255816e-06, 'learning_rate/action_model': 8.643745228255817e-05, 'epoch': 0.56})
+[2026-07-28 18:54:54,258][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.03325497731566429, 'train/grad_norm_pre_clip': 0.7491611838340759, 'timing/data': 0.00039234571158885956, 'timing/model': 0.6221484281122684, 'learning_rate/qwen_vl_interface': 8.641067006501568e-06, 'learning_rate/action_model': 8.641067006501567e-05, 'epoch': 0.56})
+[2026-07-28 18:55:03,663][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.03140971064567566, 'train/grad_norm_pre_clip': 0.772297203540802, 'timing/data': 0.0004281485453248024, 'timing/model': 0.8556120274588466, 'learning_rate/qwen_vl_interface': 8.638386584320753e-06, 'learning_rate/action_model': 8.638386584320752e-05, 'epoch': 0.56})
+[2026-07-28 18:55:12,819][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.03966730833053589, 'train/grad_norm_pre_clip': 1.9107195138931274, 'timing/data': 0.000344645231962204, 'timing/model': 0.6178842037916183, 'learning_rate/qwen_vl_interface': 8.635703963452673e-06, 'learning_rate/action_model': 8.635703963452673e-05, 'epoch': 0.56})
+[2026-07-28 18:55:22,068][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.040240079164505005, 'train/grad_norm_pre_clip': 0.9807478189468384, 'timing/data': 0.00033806636929512024, 'timing/model': 0.6172162955626845, 'learning_rate/qwen_vl_interface': 8.633019145638048e-06, 'learning_rate/action_model': 8.633019145638048e-05, 'epoch': 0.56})
+[2026-07-28 18:55:31,228][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.04157913848757744, 'train/grad_norm_pre_clip': 1.0612156391143799, 'timing/data': 0.00036270078271627426, 'timing/model': 0.6212156666442752, 'learning_rate/qwen_vl_interface': 8.630332132619029e-06, 'learning_rate/action_model': 8.630332132619028e-05, 'epoch': 0.56})
+[2026-07-28 18:55:40,632][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.049342237412929535, 'train/grad_norm_pre_clip': 2.306096315383911, 'timing/data': 0.0004161195829510689, 'timing/model': 0.8503153752535582, 'learning_rate/qwen_vl_interface': 8.627642926139187e-06, 'learning_rate/action_model': 8.627642926139187e-05, 'epoch': 0.56})
+[2026-07-28 18:55:49,761][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.033218882977962494, 'train/grad_norm_pre_clip': 0.9259905815124512, 'timing/data': 0.00036175549030303955, 'timing/model': 0.6214128416031599, 'learning_rate/qwen_vl_interface': 8.624951527943522e-06, 'learning_rate/action_model': 8.624951527943522e-05, 'epoch': 0.56})
+[2026-07-28 18:55:59,183][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.031052976846694946, 'train/grad_norm_pre_clip': 0.7981271743774414, 'timing/data': 0.00042607448995113373, 'timing/model': 0.6340164737775922, 'learning_rate/qwen_vl_interface': 8.622257939778453e-06, 'learning_rate/action_model': 8.622257939778453e-05, 'epoch': 0.56})
+[2026-07-28 18:56:08,391][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.01297454908490181, 'train/grad_norm_pre_clip': 0.46953514218330383, 'timing/data': 0.00035959668457508087, 'timing/model': 0.6210471196100116, 'learning_rate/qwen_vl_interface': 8.619562163391818e-06, 'learning_rate/action_model': 8.619562163391817e-05, 'epoch': 0.56})
+[2026-07-28 18:56:17,546][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.03461647406220436, 'train/grad_norm_pre_clip': 1.0632116794586182, 'timing/data': 0.0003276541829109192, 'timing/model': 0.6208640346303582, 'learning_rate/qwen_vl_interface': 8.616864200532877e-06, 'learning_rate/action_model': 8.616864200532877e-05, 'epoch': 0.56})
+[2026-07-28 18:56:26,915][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.040866218507289886, 'train/grad_norm_pre_clip': 0.9923615455627441, 'timing/data': 0.0003381110727787018, 'timing/model': 0.621848814189434, 'learning_rate/qwen_vl_interface': 8.61416405295231e-06, 'learning_rate/action_model': 8.61416405295231e-05, 'epoch': 0.56})
+[2026-07-28 18:56:36,033][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.0569106824696064, 'train/grad_norm_pre_clip': 1.1160708665847778, 'timing/data': 0.0003659632056951523, 'timing/model': 0.6214069128036499, 'learning_rate/qwen_vl_interface': 8.611461722402214e-06, 'learning_rate/action_model': 8.611461722402215e-05, 'epoch': 0.56})
+[2026-07-28 18:56:45,476][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.05347675830125809, 'train/grad_norm_pre_clip': 1.5716501474380493, 'timing/data': 0.0003476683050394058, 'timing/model': 0.6204340551048517, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.56})
+[2026-07-28 18:56:54,619][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.0820574089884758, 'train/grad_norm_pre_clip': 2.5333752632141113, 'timing/data': 0.000378328375518322, 'timing/model': 0.6207348434254527, 'learning_rate/qwen_vl_interface': 8.606050519408901e-06, 'learning_rate/action_model': 8.606050519408901e-05, 'epoch': 0.56})
+[2026-07-28 18:57:04,043][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.05011022090911865, 'train/grad_norm_pre_clip': 2.155618190765381, 'timing/data': 0.00034294556826353073, 'timing/model': 0.6212740503251553, 'learning_rate/qwen_vl_interface': 8.603341650476953e-06, 'learning_rate/action_model': 8.603341650476952e-05, 'epoch': 0.57})
+[2026-07-28 18:57:13,188][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.03143911436200142, 'train/grad_norm_pre_clip': 0.829081118106842, 'timing/data': 0.0003328351303935051, 'timing/model': 0.6181806372478604, 'learning_rate/qwen_vl_interface': 8.600630605598016e-06, 'learning_rate/action_model': 8.600630605598015e-05, 'epoch': 0.57})
+[2026-07-28 18:57:22,385][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.06757216155529022, 'train/grad_norm_pre_clip': 1.743546485900879, 'timing/data': 0.000345463864505291, 'timing/model': 0.623023328371346, 'learning_rate/qwen_vl_interface': 8.597917386531255e-06, 'learning_rate/action_model': 8.597917386531254e-05, 'epoch': 0.57})
+[2026-07-28 18:57:31,969][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.046052418649196625, 'train/grad_norm_pre_clip': 1.4579848051071167, 'timing/data': 0.00043490808457136154, 'timing/model': 0.8781490987166762, 'learning_rate/qwen_vl_interface': 8.595201995037252e-06, 'learning_rate/action_model': 8.595201995037251e-05, 'epoch': 0.57})
+[2026-07-28 18:57:41,182][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.03767823427915573, 'train/grad_norm_pre_clip': 1.6883671283721924, 'timing/data': 0.0004362771287560463, 'timing/model': 0.6224758317694068, 'learning_rate/qwen_vl_interface': 8.59248443287799e-06, 'learning_rate/action_model': 8.592484432877989e-05, 'epoch': 0.57})
+[2026-07-28 18:57:50,314][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.025059139356017113, 'train/grad_norm_pre_clip': 0.8134491443634033, 'timing/data': 0.0003527756780385971, 'timing/model': 0.6188705004751682, 'learning_rate/qwen_vl_interface': 8.589764701816868e-06, 'learning_rate/action_model': 8.589764701816868e-05, 'epoch': 0.57})
+[2026-07-28 18:57:59,242][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.050313565880060196, 'train/grad_norm_pre_clip': 1.3923636674880981, 'timing/data': 0.0003967992961406708, 'timing/model': 0.6206050114706159, 'learning_rate/qwen_vl_interface': 8.587042803618693e-06, 'learning_rate/action_model': 8.587042803618693e-05, 'epoch': 0.57})
+[2026-07-28 18:58:08,562][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.02653641253709793, 'train/grad_norm_pre_clip': 0.8414167165756226, 'timing/data': 0.0008372608572244644, 'timing/model': 0.7930506067350507, 'learning_rate/qwen_vl_interface': 8.584318740049672e-06, 'learning_rate/action_model': 8.584318740049672e-05, 'epoch': 0.57})
+[2026-07-28 18:58:17,973][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.047086767852306366, 'train/grad_norm_pre_clip': 1.166757583618164, 'timing/data': 0.0003428347408771515, 'timing/model': 0.6185523867607117, 'learning_rate/qwen_vl_interface': 8.581592512877419e-06, 'learning_rate/action_model': 8.581592512877418e-05, 'epoch': 0.57})
+[2026-07-28 18:58:26,901][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.02576466090977192, 'train/grad_norm_pre_clip': 0.6821935176849365, 'timing/data': 0.0004010852426290512, 'timing/model': 0.6186157874763012, 'learning_rate/qwen_vl_interface': 8.578864123870955e-06, 'learning_rate/action_model': 8.578864123870955e-05, 'epoch': 0.57})
+[2026-07-28 18:58:35,833][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.023684851825237274, 'train/grad_norm_pre_clip': 1.7532737255096436, 'timing/data': 0.0003467472270131111, 'timing/model': 0.620005751028657, 'learning_rate/qwen_vl_interface': 8.576133574800703e-06, 'learning_rate/action_model': 8.576133574800703e-05, 'epoch': 0.57})
+[2026-07-28 18:58:45,228][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.09950371086597443, 'train/grad_norm_pre_clip': 1.2063101530075073, 'timing/data': 0.000327124260365963, 'timing/model': 0.846012151800096, 'learning_rate/qwen_vl_interface': 8.573400867438486e-06, 'learning_rate/action_model': 8.573400867438486e-05, 'epoch': 0.57})
+[2026-07-28 18:58:54,347][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.05747562646865845, 'train/grad_norm_pre_clip': 1.18840754032135, 'timing/data': 0.00037391297519207, 'timing/model': 0.6189834047108889, 'learning_rate/qwen_vl_interface': 8.570666003557528e-06, 'learning_rate/action_model': 8.570666003557528e-05, 'epoch': 0.57})
+[2026-07-28 18:59:03,500][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.026327088475227356, 'train/grad_norm_pre_clip': 0.8961987495422363, 'timing/data': 0.0004218835383653641, 'timing/model': 0.6207413449883461, 'learning_rate/qwen_vl_interface': 8.567928984932452e-06, 'learning_rate/action_model': 8.567928984932452e-05, 'epoch': 0.57})
+[2026-07-28 18:59:12,659][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.029096495360136032, 'train/grad_norm_pre_clip': 0.798058271408081, 'timing/data': 0.00039482396095991135, 'timing/model': 0.6183404512703419, 'learning_rate/qwen_vl_interface': 8.565189813339278e-06, 'learning_rate/action_model': 8.565189813339278e-05, 'epoch': 0.57})
+[2026-07-28 18:59:22,039][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.05184444040060043, 'train/grad_norm_pre_clip': 0.8818650245666504, 'timing/data': 0.00035386253148317337, 'timing/model': 0.8519156761467457, 'learning_rate/qwen_vl_interface': 8.562448490555429e-06, 'learning_rate/action_model': 8.562448490555429e-05, 'epoch': 0.57})
+[2026-07-28 18:59:31,213][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.048520855605602264, 'train/grad_norm_pre_clip': 1.3455921411514282, 'timing/data': 0.0003569023683667183, 'timing/model': 0.6183823272585869, 'learning_rate/qwen_vl_interface': 8.559705018359716e-06, 'learning_rate/action_model': 8.559705018359716e-05, 'epoch': 0.57})
+[2026-07-28 18:59:40,637][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.031002506613731384, 'train/grad_norm_pre_clip': 0.7728013396263123, 'timing/data': 0.00032801181077957153, 'timing/model': 0.6223857514560223, 'learning_rate/qwen_vl_interface': 8.55695939853235e-06, 'learning_rate/action_model': 8.55695939853235e-05, 'epoch': 0.57})
+[2026-07-28 18:59:49,813][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.03194492682814598, 'train/grad_norm_pre_clip': 1.3637003898620605, 'timing/data': 0.0003632912412285805, 'timing/model': 0.6225811522454023, 'learning_rate/qwen_vl_interface': 8.554211632854932e-06, 'learning_rate/action_model': 8.554211632854932e-05, 'epoch': 0.57})
+[2026-07-28 18:59:58,951][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.017857341095805168, 'train/grad_norm_pre_clip': 0.42015987634658813, 'timing/data': 0.0003475947305560112, 'timing/model': 0.6187001960352063, 'learning_rate/qwen_vl_interface': 8.551461723110461e-06, 'learning_rate/action_model': 8.551461723110461e-05, 'epoch': 0.57})
+[2026-07-28 19:00:08,289][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.022396914660930634, 'train/grad_norm_pre_clip': 0.7648042440414429, 'timing/data': 0.0003504473716020584, 'timing/model': 0.6178900860249996, 'learning_rate/qwen_vl_interface': 8.548709671083321e-06, 'learning_rate/action_model': 8.54870967108332e-05, 'epoch': 0.57})
+[2026-07-28 19:00:17,815][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.07417505234479904, 'train/grad_norm_pre_clip': 1.5971906185150146, 'timing/data': 0.000344717875123024, 'timing/model': 0.6194293154403567, 'learning_rate/qwen_vl_interface': 8.545955478559288e-06, 'learning_rate/action_model': 8.545955478559288e-05, 'epoch': 0.57})
+[2026-07-28 19:00:27,134][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.03643542528152466, 'train/grad_norm_pre_clip': 1.6064174175262451, 'timing/data': 0.00036803819239139557, 'timing/model': 0.6209885757416487, 'learning_rate/qwen_vl_interface': 8.54319914732553e-06, 'learning_rate/action_model': 8.543199147325529e-05, 'epoch': 0.57})
+[2026-07-28 19:00:36,544][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.044879309833049774, 'train/grad_norm_pre_clip': 0.8762770891189575, 'timing/data': 0.00041343364864587784, 'timing/model': 0.8730461010709405, 'learning_rate/qwen_vl_interface': 8.540440679170599e-06, 'learning_rate/action_model': 8.540440679170598e-05, 'epoch': 0.57})
+[2026-07-28 19:00:46,001][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.03521699830889702, 'train/grad_norm_pre_clip': 0.9807860851287842, 'timing/data': 0.00040597841143608093, 'timing/model': 0.6244622822850943, 'learning_rate/qwen_vl_interface': 8.537680075884433e-06, 'learning_rate/action_model': 8.537680075884433e-05, 'epoch': 0.57})
+[2026-07-28 19:00:55,249][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.03431900590658188, 'train/grad_norm_pre_clip': 1.1166507005691528, 'timing/data': 0.0003487849608063698, 'timing/model': 0.6198166320100427, 'learning_rate/qwen_vl_interface': 8.53491733925836e-06, 'learning_rate/action_model': 8.53491733925836e-05, 'epoch': 0.57})
+[2026-07-28 19:01:04,397][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.04334517940878868, 'train/grad_norm_pre_clip': 2.175419569015503, 'timing/data': 0.0003281654790043831, 'timing/model': 0.622194548137486, 'learning_rate/qwen_vl_interface': 8.532152471085092e-06, 'learning_rate/action_model': 8.53215247108509e-05, 'epoch': 0.57})
+[2026-07-28 19:01:13,731][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.04801570251584053, 'train/grad_norm_pre_clip': 1.5232515335083008, 'timing/data': 0.000374404713511467, 'timing/model': 0.797832346521318, 'learning_rate/qwen_vl_interface': 8.529385473158717e-06, 'learning_rate/action_model': 8.529385473158717e-05, 'epoch': 0.57})
+[2026-07-28 19:01:22,868][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.0410570353269577, 'train/grad_norm_pre_clip': 4.583446025848389, 'timing/data': 0.0003401348367333412, 'timing/model': 0.6199961844831705, 'learning_rate/qwen_vl_interface': 8.526616347274712e-06, 'learning_rate/action_model': 8.526616347274712e-05, 'epoch': 0.57})
+[2026-07-28 19:01:32,022][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.03893045708537102, 'train/grad_norm_pre_clip': 1.139875054359436, 'timing/data': 0.0003719199448823929, 'timing/model': 0.618935345672071, 'learning_rate/qwen_vl_interface': 8.523845095229935e-06, 'learning_rate/action_model': 8.523845095229935e-05, 'epoch': 0.57})
+[2026-07-28 19:01:41,399][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.046764396131038666, 'train/grad_norm_pre_clip': 1.1037218570709229, 'timing/data': 0.000354117713868618, 'timing/model': 0.8613946801051497, 'learning_rate/qwen_vl_interface': 8.521071718822616e-06, 'learning_rate/action_model': 8.521071718822616e-05, 'epoch': 0.57})
+[2026-07-28 19:01:50,762][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.009063247591257095, 'train/grad_norm_pre_clip': 0.6939932107925415, 'timing/data': 0.0003790855407714844, 'timing/model': 0.6245139250531793, 'learning_rate/qwen_vl_interface': 8.518296219852375e-06, 'learning_rate/action_model': 8.518296219852375e-05, 'epoch': 0.57})
+[2026-07-28 19:01:59,846][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.027711397036910057, 'train/grad_norm_pre_clip': 0.8391202688217163, 'timing/data': 0.0003533940762281418, 'timing/model': 0.6183608444407582, 'learning_rate/qwen_vl_interface': 8.515518600120201e-06, 'learning_rate/action_model': 8.5155186001202e-05, 'epoch': 0.57})
+[2026-07-28 19:02:08,775][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.028723910450935364, 'train/grad_norm_pre_clip': 1.1309354305267334, 'timing/data': 0.0003356076776981354, 'timing/model': 0.6191436238586903, 'learning_rate/qwen_vl_interface': 8.512738861428458e-06, 'learning_rate/action_model': 8.512738861428458e-05, 'epoch': 0.57})
+[2026-07-28 19:02:18,149][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.020569143816828728, 'train/grad_norm_pre_clip': 0.65218186378479, 'timing/data': 0.0005942340940237045, 'timing/model': 0.838765605352819, 'learning_rate/qwen_vl_interface': 8.509957005580892e-06, 'learning_rate/action_model': 8.509957005580892e-05, 'epoch': 0.57})
+[2026-07-28 19:02:27,536][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.04046135023236275, 'train/grad_norm_pre_clip': 0.9894651770591736, 'timing/data': 0.00032916106283664703, 'timing/model': 0.6421580836176872, 'learning_rate/qwen_vl_interface': 8.507173034382617e-06, 'learning_rate/action_model': 8.507173034382617e-05, 'epoch': 0.57})
+[2026-07-28 19:02:36,667][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.018747486174106598, 'train/grad_norm_pre_clip': 1.431212306022644, 'timing/data': 0.0003402195870876312, 'timing/model': 0.6191846001893282, 'learning_rate/qwen_vl_interface': 8.504386949640124e-06, 'learning_rate/action_model': 8.504386949640124e-05, 'epoch': 0.57})
+[2026-07-28 19:02:45,780][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.035717763006687164, 'train/grad_norm_pre_clip': 1.1032453775405884, 'timing/data': 0.00033923424780368805, 'timing/model': 0.6211925921961665, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.57})
+[2026-07-28 19:02:54,869][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.04708099365234375, 'train/grad_norm_pre_clip': 0.9144235849380493, 'timing/data': 0.00033304933458566666, 'timing/model': 0.6183000169694424, 'learning_rate/qwen_vl_interface': 8.498808446755291e-06, 'learning_rate/action_model': 8.498808446755291e-05, 'epoch': 0.57})
+[2026-07-28 19:03:04,259][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.02759513258934021, 'train/grad_norm_pre_clip': 0.8259603381156921, 'timing/data': 0.00034499913454055786, 'timing/model': 0.6207710094749928, 'learning_rate/qwen_vl_interface': 8.49601603223278e-06, 'learning_rate/action_model': 8.49601603223278e-05, 'epoch': 0.57})
+[2026-07-28 19:03:13,365][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.035696834325790405, 'train/grad_norm_pre_clip': 0.9491897225379944, 'timing/data': 0.0003880569711327553, 'timing/model': 0.617039923556149, 'learning_rate/qwen_vl_interface': 8.493221511405708e-06, 'learning_rate/action_model': 8.493221511405707e-05, 'epoch': 0.57})
+[2026-07-28 19:03:22,475][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.016779636964201927, 'train/grad_norm_pre_clip': 0.7106224298477173, 'timing/data': 0.0003351457417011261, 'timing/model': 0.6190543975681067, 'learning_rate/qwen_vl_interface': 8.490424886087407e-06, 'learning_rate/action_model': 8.490424886087407e-05, 'epoch': 0.57})
+[2026-07-28 19:03:31,595][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.024310803040862083, 'train/grad_norm_pre_clip': 0.5772150158882141, 'timing/data': 0.00045421160757541656, 'timing/model': 0.6174442889168859, 'learning_rate/qwen_vl_interface': 8.487626158092578e-06, 'learning_rate/action_model': 8.487626158092579e-05, 'epoch': 0.57})
+[2026-07-28 19:03:40,953][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.08009845018386841, 'train/grad_norm_pre_clip': 1.5680561065673828, 'timing/data': 0.00035315100103616714, 'timing/model': 0.6194952139630914, 'learning_rate/qwen_vl_interface': 8.484825329237287e-06, 'learning_rate/action_model': 8.484825329237287e-05, 'epoch': 0.57})
+[2026-07-28 19:03:50,350][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.033180754631757736, 'train/grad_norm_pre_clip': 1.0253214836120605, 'timing/data': 0.00032818689942359924, 'timing/model': 0.8352765217423439, 'learning_rate/qwen_vl_interface': 8.48202240133896e-06, 'learning_rate/action_model': 8.482022401338959e-05, 'epoch': 0.57})
+[2026-07-28 19:03:59,519][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.05990646034479141, 'train/grad_norm_pre_clip': 2.845963954925537, 'timing/data': 0.0004225457087159157, 'timing/model': 0.6179971965029836, 'learning_rate/qwen_vl_interface': 8.479217376216385e-06, 'learning_rate/action_model': 8.479217376216385e-05, 'epoch': 0.57})
+[2026-07-28 19:04:08,623][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.029227115213871002, 'train/grad_norm_pre_clip': 0.9802504181861877, 'timing/data': 0.00035026948899030685, 'timing/model': 0.6183918006718159, 'learning_rate/qwen_vl_interface': 8.476410255689714e-06, 'learning_rate/action_model': 8.476410255689713e-05, 'epoch': 0.57})
+[2026-07-28 19:04:18,029][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.05734454095363617, 'train/grad_norm_pre_clip': 1.1348814964294434, 'timing/data': 0.00030668918043375015, 'timing/model': 0.6211490901187062, 'learning_rate/qwen_vl_interface': 8.473601041580457e-06, 'learning_rate/action_model': 8.473601041580457e-05, 'epoch': 0.57})
+[2026-07-28 19:04:27,151][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.04457101970911026, 'train/grad_norm_pre_clip': 1.5051178932189941, 'timing/data': 0.00032549258321523666, 'timing/model': 0.6180408801883459, 'learning_rate/qwen_vl_interface': 8.470789735711483e-06, 'learning_rate/action_model': 8.470789735711483e-05, 'epoch': 0.57})
+[2026-07-28 19:04:36,503][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.052180495113134384, 'train/grad_norm_pre_clip': 0.9646311402320862, 'timing/data': 0.0003533940762281418, 'timing/model': 0.8504900243133307, 'learning_rate/qwen_vl_interface': 8.467976339907015e-06, 'learning_rate/action_model': 8.467976339907016e-05, 'epoch': 0.57})
+[2026-07-28 19:04:45,813][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.051892273128032684, 'train/grad_norm_pre_clip': 0.8319413661956787, 'timing/data': 0.00033753272145986557, 'timing/model': 0.7911511594429612, 'learning_rate/qwen_vl_interface': 8.465160855992639e-06, 'learning_rate/action_model': 8.465160855992638e-05, 'epoch': 0.57})
+[2026-07-28 19:04:54,959][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.05399450287222862, 'train/grad_norm_pre_clip': 1.0086181163787842, 'timing/data': 0.00032741203904151917, 'timing/model': 0.6184675004333258, 'learning_rate/qwen_vl_interface': 8.462343285795289e-06, 'learning_rate/action_model': 8.462343285795288e-05, 'epoch': 0.57})
+[2026-07-28 19:05:04,081][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.023831330239772797, 'train/grad_norm_pre_clip': 0.6511184573173523, 'timing/data': 0.0004134140908718109, 'timing/model': 0.616716236807406, 'learning_rate/qwen_vl_interface': 8.459523631143259e-06, 'learning_rate/action_model': 8.459523631143258e-05, 'epoch': 0.57})
+[2026-07-28 19:05:13,212][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.024336302652955055, 'train/grad_norm_pre_clip': 0.9890561103820801, 'timing/data': 0.0003618188202381134, 'timing/model': 0.618952551856637, 'learning_rate/qwen_vl_interface': 8.456701893866188e-06, 'learning_rate/action_model': 8.456701893866187e-05, 'epoch': 0.57})
+[2026-07-28 19:05:22,851][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.024120774120092392, 'train/grad_norm_pre_clip': 1.0503734350204468, 'timing/data': 0.0003945678472518921, 'timing/model': 0.8563434593379498, 'learning_rate/qwen_vl_interface': 8.453878075795072e-06, 'learning_rate/action_model': 8.453878075795072e-05, 'epoch': 0.58})
+[2026-07-28 19:05:31,758][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.02318330481648445, 'train/grad_norm_pre_clip': 1.057453989982605, 'timing/data': 0.00033623725175857544, 'timing/model': 0.6187481507658958, 'learning_rate/qwen_vl_interface': 8.451052178762258e-06, 'learning_rate/action_model': 8.451052178762258e-05, 'epoch': 0.58})
+[2026-07-28 19:05:40,882][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.03330119326710701, 'train/grad_norm_pre_clip': 1.247074842453003, 'timing/data': 0.0003323359414935112, 'timing/model': 0.6195300566032529, 'learning_rate/qwen_vl_interface': 8.448224204601439e-06, 'learning_rate/action_model': 8.448224204601438e-05, 'epoch': 0.58})
+[2026-07-28 19:05:49,992][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.0486002042889595, 'train/grad_norm_pre_clip': 1.1051294803619385, 'timing/data': 0.0004522223025560379, 'timing/model': 0.6192054152488708, 'learning_rate/qwen_vl_interface': 8.445394155147655e-06, 'learning_rate/action_model': 8.445394155147653e-05, 'epoch': 0.58})
+[2026-07-28 19:05:59,450][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.02973560430109501, 'train/grad_norm_pre_clip': 1.1353402137756348, 'timing/data': 0.00036299414932727814, 'timing/model': 0.620746610686183, 'learning_rate/qwen_vl_interface': 8.442562032237295e-06, 'learning_rate/action_model': 8.442562032237296e-05, 'epoch': 0.58})
+[2026-07-28 19:06:08,808][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.04558718949556351, 'train/grad_norm_pre_clip': 0.6316165924072266, 'timing/data': 0.00040351971983909607, 'timing/model': 0.6244679223746061, 'learning_rate/qwen_vl_interface': 8.439727837708094e-06, 'learning_rate/action_model': 8.439727837708095e-05, 'epoch': 0.58})
+[2026-07-28 19:06:17,989][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.04349276050925255, 'train/grad_norm_pre_clip': 0.9126506447792053, 'timing/data': 0.00037034135311841965, 'timing/model': 0.6190974451601505, 'learning_rate/qwen_vl_interface': 8.436891573399132e-06, 'learning_rate/action_model': 8.436891573399131e-05, 'epoch': 0.58})
+[2026-07-28 19:06:27,153][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.03130380064249039, 'train/grad_norm_pre_clip': 1.3252379894256592, 'timing/data': 0.0004712836816906929, 'timing/model': 0.6186606856063008, 'learning_rate/qwen_vl_interface': 8.434053241150826e-06, 'learning_rate/action_model': 8.434053241150825e-05, 'epoch': 0.58})
+[2026-07-28 19:06:36,530][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.030328867956995964, 'train/grad_norm_pre_clip': 0.9981837272644043, 'timing/data': 0.0004048030823469162, 'timing/model': 0.6179007040336728, 'learning_rate/qwen_vl_interface': 8.43121284280494e-06, 'learning_rate/action_model': 8.431212842804939e-05, 'epoch': 0.58})
+[2026-07-28 19:06:45,664][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.07024041563272476, 'train/grad_norm_pre_clip': 2.569450616836548, 'timing/data': 0.0003749588504433632, 'timing/model': 0.6197749311104417, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.58})
+[2026-07-28 19:06:55,039][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.03859000653028488, 'train/grad_norm_pre_clip': 1.330307960510254, 'timing/data': 0.00035165902227163315, 'timing/model': 0.8400676157325506, 'learning_rate/qwen_vl_interface': 8.425525855194184e-06, 'learning_rate/action_model': 8.425525855194183e-05, 'epoch': 0.58})
+[2026-07-28 19:07:04,392][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.04711374640464783, 'train/grad_norm_pre_clip': 1.751242756843567, 'timing/data': 0.00034153927117586136, 'timing/model': 0.7913505639880896, 'learning_rate/qwen_vl_interface': 8.422679269619538e-06, 'learning_rate/action_model': 8.422679269619538e-05, 'epoch': 0.58})
+[2026-07-28 19:07:13,541][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.04560420662164688, 'train/grad_norm_pre_clip': 1.1992958784103394, 'timing/data': 0.00036358926445245743, 'timing/model': 0.6198634365573525, 'learning_rate/qwen_vl_interface': 8.419830625327758e-06, 'learning_rate/action_model': 8.419830625327758e-05, 'epoch': 0.58})
+[2026-07-28 19:07:22,691][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.04002852365374565, 'train/grad_norm_pre_clip': 0.6447264552116394, 'timing/data': 0.0003643454983830452, 'timing/model': 0.6183100640773773, 'learning_rate/qwen_vl_interface': 8.416979924167303e-06, 'learning_rate/action_model': 8.416979924167303e-05, 'epoch': 0.58})
+[2026-07-28 19:07:31,847][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.04947274923324585, 'train/grad_norm_pre_clip': 0.8632583022117615, 'timing/data': 0.0003788936883211136, 'timing/model': 0.6278194328770041, 'learning_rate/qwen_vl_interface': 8.414127167987957e-06, 'learning_rate/action_model': 8.414127167987957e-05, 'epoch': 0.58})
+[2026-07-28 19:07:41,499][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.05608968436717987, 'train/grad_norm_pre_clip': 1.9816007614135742, 'timing/data': 0.00039515644311904907, 'timing/model': 0.8835922004655004, 'learning_rate/qwen_vl_interface': 8.411272358640843e-06, 'learning_rate/action_model': 8.411272358640844e-05, 'epoch': 0.58})
+[2026-07-28 19:07:50,442][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.019081447273492813, 'train/grad_norm_pre_clip': 0.8409662246704102, 'timing/data': 0.00044758711010217667, 'timing/model': 0.6174199739471078, 'learning_rate/qwen_vl_interface': 8.408415497978418e-06, 'learning_rate/action_model': 8.408415497978418e-05, 'epoch': 0.58})
+[2026-07-28 19:07:59,477][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.03761604428291321, 'train/grad_norm_pre_clip': 0.9436408877372742, 'timing/data': 0.00035468675196170807, 'timing/model': 0.6195085700601339, 'learning_rate/qwen_vl_interface': 8.405556587854468e-06, 'learning_rate/action_model': 8.405556587854467e-05, 'epoch': 0.58})
+[2026-07-28 19:08:08,649][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.011730212718248367, 'train/grad_norm_pre_clip': 0.46843892335891724, 'timing/data': 0.00035268906503915787, 'timing/model': 0.6200336394831538, 'learning_rate/qwen_vl_interface': 8.402695630124106e-06, 'learning_rate/action_model': 8.402695630124106e-05, 'epoch': 0.58})
+[2026-07-28 19:08:18,175][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.04197008162736893, 'train/grad_norm_pre_clip': 1.2216541767120361, 'timing/data': 0.0003476077690720558, 'timing/model': 0.9143955474719405, 'learning_rate/qwen_vl_interface': 8.399832626643778e-06, 'learning_rate/action_model': 8.399832626643778e-05, 'epoch': 0.58})
+[2026-07-28 19:08:27,684][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.05249388515949249, 'train/grad_norm_pre_clip': 1.0004695653915405, 'timing/data': 0.00041618943214416504, 'timing/model': 0.8537743473425508, 'learning_rate/qwen_vl_interface': 8.396967579271256e-06, 'learning_rate/action_model': 8.396967579271257e-05, 'epoch': 0.58})
+[2026-07-28 19:08:36,871][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.04266315698623657, 'train/grad_norm_pre_clip': 0.730681836605072, 'timing/data': 0.00035960134118795395, 'timing/model': 0.6176346661522985, 'learning_rate/qwen_vl_interface': 8.394100489865638e-06, 'learning_rate/action_model': 8.394100489865638e-05, 'epoch': 0.58})
+[2026-07-28 19:08:45,992][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.05526294931769371, 'train/grad_norm_pre_clip': 2.090707302093506, 'timing/data': 0.00035544857382774353, 'timing/model': 0.6197736226022243, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.58})
+[2026-07-28 19:08:55,351][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.09785494953393936, 'train/grad_norm_pre_clip': 3.6441822052001953, 'timing/data': 0.00035246647894382477, 'timing/model': 0.6172725772485137, 'learning_rate/qwen_vl_interface': 8.388360192398131e-06, 'learning_rate/action_model': 8.38836019239813e-05, 'epoch': 0.58})
+[2026-07-28 19:09:04,553][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.06283597648143768, 'train/grad_norm_pre_clip': 1.7157683372497559, 'timing/data': 0.00035582203418016434, 'timing/model': 0.6190246203914285, 'learning_rate/qwen_vl_interface': 8.385486988061058e-06, 'learning_rate/action_model': 8.385486988061058e-05, 'epoch': 0.58})
+[2026-07-28 19:09:13,684][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.01990625448524952, 'train/grad_norm_pre_clip': 0.47319576144218445, 'timing/data': 0.0003304695710539818, 'timing/model': 0.6204174868762493, 'learning_rate/qwen_vl_interface': 8.38261174914052e-06, 'learning_rate/action_model': 8.38261174914052e-05, 'epoch': 0.58})
+[2026-07-28 19:09:23,061][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.042680688202381134, 'train/grad_norm_pre_clip': 1.5078563690185547, 'timing/data': 0.0003362353891134262, 'timing/model': 0.8512915214523673, 'learning_rate/qwen_vl_interface': 8.379734477502229e-06, 'learning_rate/action_model': 8.379734477502228e-05, 'epoch': 0.58})
+[2026-07-28 19:09:32,448][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.023538071662187576, 'train/grad_norm_pre_clip': 0.9621345400810242, 'timing/data': 0.00041925255209207535, 'timing/model': 0.6247788425534964, 'learning_rate/qwen_vl_interface': 8.376855175013214e-06, 'learning_rate/action_model': 8.376855175013213e-05, 'epoch': 0.58})
+[2026-07-28 19:09:41,636][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.04523564130067825, 'train/grad_norm_pre_clip': 0.9209403991699219, 'timing/data': 0.0004641106352210045, 'timing/model': 0.6204047808423638, 'learning_rate/qwen_vl_interface': 8.373973843541822e-06, 'learning_rate/action_model': 8.373973843541822e-05, 'epoch': 0.58})
+[2026-07-28 19:09:50,816][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.041014663875103, 'train/grad_norm_pre_clip': 1.0775903463363647, 'timing/data': 0.00035417452454566956, 'timing/model': 0.6183924414217472, 'learning_rate/qwen_vl_interface': 8.371090484957722e-06, 'learning_rate/action_model': 8.371090484957722e-05, 'epoch': 0.58})
+[2026-07-28 19:09:59,975][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.03460168093442917, 'train/grad_norm_pre_clip': 0.9466537237167358, 'timing/data': 0.00036008376628160477, 'timing/model': 0.6190652148798108, 'learning_rate/qwen_vl_interface': 8.368205101131891e-06, 'learning_rate/action_model': 8.368205101131891e-05, 'epoch': 0.58})
+[2026-07-28 19:10:09,358][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.03632526099681854, 'train/grad_norm_pre_clip': 0.7172222137451172, 'timing/data': 0.00042052101343870163, 'timing/model': 0.6251811003312469, 'learning_rate/qwen_vl_interface': 8.365317693936624e-06, 'learning_rate/action_model': 8.365317693936623e-05, 'epoch': 0.58})
+[2026-07-28 19:10:18,573][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.025339679792523384, 'train/grad_norm_pre_clip': 0.6376157402992249, 'timing/data': 0.00036615878343582153, 'timing/model': 0.6179301925003529, 'learning_rate/qwen_vl_interface': 8.362428265245526e-06, 'learning_rate/action_model': 8.362428265245526e-05, 'epoch': 0.58})
+[2026-07-28 19:10:27,910][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.06256172060966492, 'train/grad_norm_pre_clip': 2.219508171081543, 'timing/data': 0.0003395192325115204, 'timing/model': 0.8429443556815386, 'learning_rate/qwen_vl_interface': 8.359536816933518e-06, 'learning_rate/action_model': 8.359536816933518e-05, 'epoch': 0.58})
+[2026-07-28 19:10:37,052][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.062302328646183014, 'train/grad_norm_pre_clip': 1.8008954524993896, 'timing/data': 0.00047437846660614014, 'timing/model': 0.6239246753975749, 'learning_rate/qwen_vl_interface': 8.356643350876832e-06, 'learning_rate/action_model': 8.356643350876832e-05, 'epoch': 0.58})
+[2026-07-28 19:10:46,198][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.05711790919303894, 'train/grad_norm_pre_clip': 1.1762356758117676, 'timing/data': 0.0003821626305580139, 'timing/model': 0.6212795674800873, 'learning_rate/qwen_vl_interface': 8.353747868953002e-06, 'learning_rate/action_model': 8.353747868953003e-05, 'epoch': 0.58})
+[2026-07-28 19:10:55,629][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.026547159999608994, 'train/grad_norm_pre_clip': 0.9721435308456421, 'timing/data': 0.0005981456488370895, 'timing/model': 0.8501893319189548, 'learning_rate/qwen_vl_interface': 8.350850373040876e-06, 'learning_rate/action_model': 8.350850373040876e-05, 'epoch': 0.58})
+[2026-07-28 19:11:04,873][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.047703325748443604, 'train/grad_norm_pre_clip': 1.1854511499404907, 'timing/data': 0.0003875391557812691, 'timing/model': 0.6229612389579415, 'learning_rate/qwen_vl_interface': 8.34795086502061e-06, 'learning_rate/action_model': 8.347950865020609e-05, 'epoch': 0.58})
+[2026-07-28 19:11:14,322][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.036469876766204834, 'train/grad_norm_pre_clip': 1.7067102193832397, 'timing/data': 0.00041425321251153946, 'timing/model': 0.6218371456488967, 'learning_rate/qwen_vl_interface': 8.34504934677366e-06, 'learning_rate/action_model': 8.34504934677366e-05, 'epoch': 0.58})
+[2026-07-28 19:11:23,527][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.029128126800060272, 'train/grad_norm_pre_clip': 0.6541845798492432, 'timing/data': 0.00038930121809244156, 'timing/model': 0.6193251665681601, 'learning_rate/qwen_vl_interface': 8.342145820182795e-06, 'learning_rate/action_model': 8.342145820182793e-05, 'epoch': 0.58})
+[2026-07-28 19:11:32,761][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.032551735639572144, 'train/grad_norm_pre_clip': 0.5700979828834534, 'timing/data': 0.0003817537799477577, 'timing/model': 0.6249355580657721, 'learning_rate/qwen_vl_interface': 8.339240287132073e-06, 'learning_rate/action_model': 8.339240287132073e-05, 'epoch': 0.58})
+[2026-07-28 19:11:42,155][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.02203325182199478, 'train/grad_norm_pre_clip': 0.7518137693405151, 'timing/data': 0.0004684310406446457, 'timing/model': 0.6206308295950294, 'learning_rate/qwen_vl_interface': 8.336332749506872e-06, 'learning_rate/action_model': 8.336332749506872e-05, 'epoch': 0.58})
+[2026-07-28 19:11:51,626][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.05717901885509491, 'train/grad_norm_pre_clip': 1.1070259809494019, 'timing/data': 0.000450928695499897, 'timing/model': 0.6205996135249734, 'learning_rate/qwen_vl_interface': 8.333423209193857e-06, 'learning_rate/action_model': 8.333423209193856e-05, 'epoch': 0.58})
+[2026-07-28 19:12:00,822][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.025898007676005363, 'train/grad_norm_pre_clip': 1.1073814630508423, 'timing/data': 0.0003740685060620308, 'timing/model': 0.6190670635551214, 'learning_rate/qwen_vl_interface': 8.330511668080997e-06, 'learning_rate/action_model': 8.330511668080997e-05, 'epoch': 0.58})
+[2026-07-28 19:12:10,310][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.026987075805664062, 'train/grad_norm_pre_clip': 0.6689611673355103, 'timing/data': 0.0003365492448210716, 'timing/model': 0.619855291210115, 'learning_rate/qwen_vl_interface': 8.32759812805756e-06, 'learning_rate/action_model': 8.32759812805756e-05, 'epoch': 0.58})
+[2026-07-28 19:12:19,785][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.026303360238671303, 'train/grad_norm_pre_clip': 0.72269207239151, 'timing/data': 0.00047524087131023407, 'timing/model': 0.896786792203784, 'learning_rate/qwen_vl_interface': 8.324682591014111e-06, 'learning_rate/action_model': 8.32468259101411e-05, 'epoch': 0.58})
+[2026-07-28 19:12:28,960][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.022418513894081116, 'train/grad_norm_pre_clip': 0.6094434857368469, 'timing/data': 0.0003897203132510185, 'timing/model': 0.6283646812662482, 'learning_rate/qwen_vl_interface': 8.32176505884251e-06, 'learning_rate/action_model': 8.32176505884251e-05, 'epoch': 0.58})
+[2026-07-28 19:12:38,147][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.015357929281890392, 'train/grad_norm_pre_clip': 0.7714360356330872, 'timing/data': 0.0006033126264810562, 'timing/model': 0.6205198764801025, 'learning_rate/qwen_vl_interface': 8.318845533435913e-06, 'learning_rate/action_model': 8.318845533435913e-05, 'epoch': 0.58})
+[2026-07-28 19:12:47,654][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.011682254262268543, 'train/grad_norm_pre_clip': 0.34634843468666077, 'timing/data': 0.0004000682383775711, 'timing/model': 0.6192507687956095, 'learning_rate/qwen_vl_interface': 8.315924016688767e-06, 'learning_rate/action_model': 8.315924016688766e-05, 'epoch': 0.58})
+[2026-07-28 19:12:56,803][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.032864317297935486, 'train/grad_norm_pre_clip': 0.8153150677680969, 'timing/data': 0.00035755522549152374, 'timing/model': 0.6192064099013805, 'learning_rate/qwen_vl_interface': 8.313000510496811e-06, 'learning_rate/action_model': 8.313000510496811e-05, 'epoch': 0.58})
+[2026-07-28 19:13:06,231][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.10127641260623932, 'train/grad_norm_pre_clip': 1.1865721940994263, 'timing/data': 0.000454598106443882, 'timing/model': 0.9000245248898864, 'learning_rate/qwen_vl_interface': 8.31007501675708e-06, 'learning_rate/action_model': 8.310075016757079e-05, 'epoch': 0.58})
+[2026-07-28 19:13:15,360][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.01718483678996563, 'train/grad_norm_pre_clip': 0.965999960899353, 'timing/data': 0.0003779260441660881, 'timing/model': 0.6197589710354805, 'learning_rate/qwen_vl_interface': 8.307147537367894e-06, 'learning_rate/action_model': 8.307147537367893e-05, 'epoch': 0.58})
+[2026-07-28 19:13:24,744][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.008169487118721008, 'train/grad_norm_pre_clip': 0.5310561060905457, 'timing/data': 0.0003579491749405861, 'timing/model': 0.6215864643454552, 'learning_rate/qwen_vl_interface': 8.30421807422886e-06, 'learning_rate/action_model': 8.30421807422886e-05, 'epoch': 0.58})
+[2026-07-28 19:13:33,683][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.016059236600995064, 'train/grad_norm_pre_clip': 0.7722341418266296, 'timing/data': 0.00035983603447675705, 'timing/model': 0.6200645342469215, 'learning_rate/qwen_vl_interface': 8.301286629240876e-06, 'learning_rate/action_model': 8.301286629240876e-05, 'epoch': 0.58})
+[2026-07-28 19:13:42,841][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.039555639028549194, 'train/grad_norm_pre_clip': 1.7216402292251587, 'timing/data': 0.0003531845286488533, 'timing/model': 0.6204032376408577, 'learning_rate/qwen_vl_interface': 8.298353204306127e-06, 'learning_rate/action_model': 8.298353204306127e-05, 'epoch': 0.59})
+[2026-07-28 19:13:51,987][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.05195430666208267, 'train/grad_norm_pre_clip': 1.980302095413208, 'timing/data': 0.00044489186257123947, 'timing/model': 0.6200551809743047, 'learning_rate/qwen_vl_interface': 8.295417801328081e-06, 'learning_rate/action_model': 8.29541780132808e-05, 'epoch': 0.59})
+[2026-07-28 19:17:23,258][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 19:17:23,261][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21'
+[2026-07-28 19:17:23,261][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 19:17:23,478][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 19:17:23,478][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 19:17:23,482][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 19:17:23,482][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 19:17:24,040][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 19:17:24,055][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 19:17:24,214][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 19:17:24,229][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 19:17:26,572][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 19:17:26,572][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 19:17:27,563][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 19:17:27,564][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 19:17:27,564][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 19:17:27,564][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 19:17:27,564][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 19:17:27,564][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 19:17:27,564][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 19:17:27,565][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 19:17:27,565][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 19:17:27,565][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 19:17:27,566][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 19:17:27,566][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 19:17:27,566][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 19:17:27,567][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 19:17:27,567][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 19:17:27,567][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 19:17:27,568][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 19:17:27,568][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 19:17:27,568][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 19:17:27,569][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 19:17:27,569][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 19:17:27,569][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 19:17:27,570][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 19:17:27,570][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 19:17:27,570][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 19:17:27,570][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 19:17:27,570][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 19:17:27,570][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 19:17:27,571][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 19:17:27,571][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 19:17:27,571][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 19:17:27,638][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 19:17:27,638][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 19:17:27,638][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 19:17:27,638][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 19:17:27,639][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 19:17:27,639][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 19:17:27,639][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 19:17:27,640][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 19:17:27,640][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 19:17:27,640][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 19:17:27,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 19:17:27,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 19:17:27,641][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 19:17:27,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 19:17:27,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 19:17:27,642][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 19:17:27,643][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 19:17:27,643][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 19:17:27,643][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 19:17:27,644][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 19:17:27,644][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 19:17:27,644][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 19:17:27,645][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 19:17:27,645][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 19:17:27,645][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 19:17:27,645][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 19:17:27,645][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 19:17:27,645][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 19:17:27,646][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 19:17:27,646][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 19:17:27,646][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 19:17:28,928][starVLA.dataloader][INFO] - Created train VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 19:17:28,930][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 19:17:28,936][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 19:17:28,953][starVLA.dataloader][INFO] - Created eval VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 19:17:28,960][starVLA.training.train_starvla][INFO] - LR Group qwen_vl_interface: lr=1e-05, num_params=713
+[2026-07-28 19:17:28,960][starVLA.training.train_starvla][INFO] - LR Group action_model: lr=0.0001, num_params=16
+[2026-07-28 19:17:30,285][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3000_state, steps: 3000
+[2026-07-28 19:17:30,297][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/config.full.yaml
+[2026-07-28 19:17:30,310][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/config.yaml
+[2026-07-28 19:17:30,310][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3000_state, steps: 3000
+[2026-07-28 19:17:30,315][accelerate.accelerator][INFO] - Since you passed both train and evaluation dataloader, `is_train_batch_min` (here True will decide the `train_batch_size` (8).
+[2026-07-28 19:18:06,455][accelerate.accelerator][INFO] - Loading states from /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3000_state
+[2026-07-28 19:18:06,455][accelerate.accelerator][INFO] - Loading DeepSpeed Model and Optimizer
+[2026-07-28 19:18:40,017][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer loaded from input dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3000_state/pytorch_model
+[2026-07-28 19:18:40,017][accelerate.checkpointing][INFO] - All model weights loaded successfully
+[2026-07-28 19:18:40,017][accelerate.checkpointing][INFO] - All optimizer states loaded successfully
+[2026-07-28 19:18:40,017][accelerate.checkpointing][INFO] - All scheduler states loaded successfully
+[2026-07-28 19:18:40,018][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully
+[2026-07-28 19:18:40,029][accelerate.checkpointing][INFO] - All random states loaded successfully
+[2026-07-28 19:18:40,030][accelerate.accelerator][INFO] - Loading in 0 custom states
+[2026-07-28 19:18:42,845][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-07-28 19:18:42,846][starVLA.training.train_starvla][INFO] -   Total optimization steps = 4000
+[2026-07-28 19:18:42,846][starVLA.training.train_starvla][INFO] -   Per device batch size = 8
+[2026-07-28 19:18:42,846][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-07-28 19:18:42,846][starVLA.training.train_starvla][INFO] -   Total batch size = 256
+[2026-07-28 19:20:04,843][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.019282719120383263, 'train/grad_norm_pre_clip': 0.5299254059791565, 'timing/data': 0.0004106108099222183, 'timing/model': 0.6891281958669424, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.55})
+[2026-07-28 19:20:12,280][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.04237920790910721, 'train/grad_norm_pre_clip': 2.3184146881103516, 'timing/data': 0.0003396030515432358, 'timing/model': 0.547906868159771, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.55})
+[2026-07-28 19:20:19,879][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.022324316203594208, 'train/grad_norm_pre_clip': 0.9949431419372559, 'timing/data': 0.0003484915941953659, 'timing/model': 0.5626047141849995, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.55})
+[2026-07-28 19:20:27,190][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.03171294927597046, 'train/grad_norm_pre_clip': 0.909801721572876, 'timing/data': 0.0007149353623390198, 'timing/model': 0.5313371140509844, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.55})
+[2026-07-28 19:20:34,468][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.015307040885090828, 'train/grad_norm_pre_clip': 0.5980374217033386, 'timing/data': 0.0004151761531829834, 'timing/model': 0.5416046641767025, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.55})
+[2026-07-28 19:20:41,775][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.055243439972400665, 'train/grad_norm_pre_clip': 1.2821323871612549, 'timing/data': 0.00046728551387786865, 'timing/model': 0.5391825586557388, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.55})
+[2026-07-28 19:20:49,094][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.027121201157569885, 'train/grad_norm_pre_clip': 1.246240258216858, 'timing/data': 0.0005341209471225739, 'timing/model': 0.5335547234863043, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.55})
+[2026-07-28 19:20:56,433][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.04039693623781204, 'train/grad_norm_pre_clip': 1.6926754713058472, 'timing/data': 0.0004294849932193756, 'timing/model': 0.5439067892730236, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.55})
+[2026-07-28 19:21:03,786][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.03702855110168457, 'train/grad_norm_pre_clip': 0.9897633790969849, 'timing/data': 0.0005248840898275375, 'timing/model': 0.5357393193989992, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.55})
+[2026-07-28 19:21:11,083][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.04068250209093094, 'train/grad_norm_pre_clip': 0.9507504105567932, 'timing/data': 0.00038347020745277405, 'timing/model': 0.5414377637207508, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.55})
+[2026-07-28 19:21:18,360][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.07063671946525574, 'train/grad_norm_pre_clip': 1.8937311172485352, 'timing/data': 0.0005033742636442184, 'timing/model': 0.5372298806905746, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.55})
+[2026-07-28 19:21:25,637][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.03313279151916504, 'train/grad_norm_pre_clip': 0.7044604420661926, 'timing/data': 0.0004706159234046936, 'timing/model': 0.535880571231246, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.55})
+[2026-07-28 19:21:32,877][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.021880436688661575, 'train/grad_norm_pre_clip': 1.165088176727295, 'timing/data': 0.0003835950046777725, 'timing/model': 0.528672069311142, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.55})
+[2026-07-28 19:21:40,106][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.07221923023462296, 'train/grad_norm_pre_clip': 1.3799798488616943, 'timing/data': 0.0003932379186153412, 'timing/model': 0.5318417344242334, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.55})
+[2026-07-28 19:21:47,328][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.08310911059379578, 'train/grad_norm_pre_clip': 1.8197321891784668, 'timing/data': 0.00036323629319667816, 'timing/model': 0.5312827210873365, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.55})
+[2026-07-28 19:21:54,543][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.060441963374614716, 'train/grad_norm_pre_clip': 2.005075693130493, 'timing/data': 0.00036463141441345215, 'timing/model': 0.5293042548000813, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.55})
+[2026-07-28 19:22:01,781][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.04043789952993393, 'train/grad_norm_pre_clip': 1.2415262460708618, 'timing/data': 0.00041993334889411926, 'timing/model': 0.5343614220619202, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.55})
+[2026-07-28 19:22:09,009][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.045816175639629364, 'train/grad_norm_pre_clip': 1.1950567960739136, 'timing/data': 0.0003188531845808029, 'timing/model': 0.537591191008687, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.55})
+[2026-07-28 19:22:16,229][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.030040597543120384, 'train/grad_norm_pre_clip': 1.2793704271316528, 'timing/data': 0.0003184657543897629, 'timing/model': 0.531764404848218, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.55})
+[2026-07-28 19:22:23,453][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.03960849344730377, 'train/grad_norm_pre_clip': 1.0102341175079346, 'timing/data': 0.000489385798573494, 'timing/model': 0.5326356533914804, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.55})
+[2026-07-28 19:22:30,703][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.051247164607048035, 'train/grad_norm_pre_clip': 1.1953551769256592, 'timing/data': 0.0005291569977998734, 'timing/model': 0.5337384920567274, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.55})
+[2026-07-28 19:22:37,960][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.03965763747692108, 'train/grad_norm_pre_clip': 1.8297693729400635, 'timing/data': 0.00041971541941165924, 'timing/model': 0.5348420199006796, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.56})
+[2026-07-28 19:22:45,196][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.029110198840498924, 'train/grad_norm_pre_clip': 1.585187315940857, 'timing/data': 0.0005209539085626602, 'timing/model': 0.5303451213985682, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.56})
+[2026-07-28 19:22:52,452][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.031363144516944885, 'train/grad_norm_pre_clip': 1.4592370986938477, 'timing/data': 0.0005165357142686844, 'timing/model': 0.5316873379051685, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.56})
+[2026-07-28 19:22:59,775][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.04329307749867439, 'train/grad_norm_pre_clip': 1.2284491062164307, 'timing/data': 0.0004603676497936249, 'timing/model': 0.5325006972998381, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.56})
+[2026-07-28 19:23:07,093][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.03325644135475159, 'train/grad_norm_pre_clip': 0.6980854868888855, 'timing/data': 0.0004328060895204544, 'timing/model': 0.5306795723736286, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.56})
+[2026-07-28 19:23:14,383][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.03999297320842743, 'train/grad_norm_pre_clip': 1.6180839538574219, 'timing/data': 0.0004544854164123535, 'timing/model': 0.5362474825233221, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.56})
+[2026-07-28 19:23:22,229][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.021540343761444092, 'train/grad_norm_pre_clip': 0.9319985508918762, 'timing/data': 0.0003304760903120041, 'timing/model': 0.9918206427246332, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.56})
+[2026-07-28 19:23:29,876][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.02287665568292141, 'train/grad_norm_pre_clip': 0.6943028569221497, 'timing/data': 0.00036584585905075073, 'timing/model': 0.5310838092118502, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.56})
+[2026-07-28 19:23:37,110][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.03561479598283768, 'train/grad_norm_pre_clip': 0.9755613207817078, 'timing/data': 0.0003441832959651947, 'timing/model': 0.5337336882948875, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.56})
+[2026-07-28 19:23:44,367][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.02434835210442543, 'train/grad_norm_pre_clip': 0.5769041180610657, 'timing/data': 0.0005719661712646484, 'timing/model': 0.5298257675021887, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.56})
+[2026-07-28 19:23:51,614][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.03853529691696167, 'train/grad_norm_pre_clip': 1.1849669218063354, 'timing/data': 0.00037381239235401154, 'timing/model': 0.5302882790565491, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.56})
+[2026-07-28 19:23:58,839][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.04571586847305298, 'train/grad_norm_pre_clip': 1.4658229351043701, 'timing/data': 0.0004065129905939102, 'timing/model': 0.5306582413613796, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.56})
+[2026-07-28 19:24:06,056][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.02627101168036461, 'train/grad_norm_pre_clip': 1.020153522491455, 'timing/data': 0.00034855492413043976, 'timing/model': 0.5316563509404659, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.56})
+[2026-07-28 19:24:13,271][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.02159702777862549, 'train/grad_norm_pre_clip': 0.6279914975166321, 'timing/data': 0.0004475899040699005, 'timing/model': 0.5315147116780281, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.56})
+[2026-07-28 19:24:20,488][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.018871702253818512, 'train/grad_norm_pre_clip': 0.7212873101234436, 'timing/data': 0.00032228976488113403, 'timing/model': 0.5300405118614435, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.56})
+[2026-07-28 19:24:27,708][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.022500377148389816, 'train/grad_norm_pre_clip': 1.149543285369873, 'timing/data': 0.0003590583801269531, 'timing/model': 0.5307904835790396, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.56})
+[2026-07-28 19:24:34,953][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.016455234959721565, 'train/grad_norm_pre_clip': 0.59885174036026, 'timing/data': 0.0004443507641553879, 'timing/model': 0.5329327341169119, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.56})
+[2026-07-28 19:24:42,177][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.06123654916882515, 'train/grad_norm_pre_clip': 4.568548679351807, 'timing/data': 0.00048109516501426697, 'timing/model': 0.5315616279840469, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.56})
+[2026-07-28 19:24:49,415][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.029995804652571678, 'train/grad_norm_pre_clip': 0.7204329371452332, 'timing/data': 0.00048697739839553833, 'timing/model': 0.534884063526988, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.56})
+[2026-07-28 19:24:56,672][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.015097290277481079, 'train/grad_norm_pre_clip': 0.6463257074356079, 'timing/data': 0.00045737624168395996, 'timing/model': 0.5321826264262199, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.56})
+[2026-07-28 19:25:03,924][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.05362570658326149, 'train/grad_norm_pre_clip': 2.429837703704834, 'timing/data': 0.0004803873598575592, 'timing/model': 0.5360212177038193, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.56})
+[2026-07-28 19:25:11,179][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.032584987580776215, 'train/grad_norm_pre_clip': 1.592928171157837, 'timing/data': 0.0004524160176515579, 'timing/model': 0.5319481994956732, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.56})
+[2026-07-28 19:25:18,451][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.009202498011291027, 'train/grad_norm_pre_clip': 0.6342343688011169, 'timing/data': 0.0004818309098482132, 'timing/model': 0.5308069586753845, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.56})
+[2026-07-28 19:25:25,704][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.01293390616774559, 'train/grad_norm_pre_clip': 0.568828821182251, 'timing/data': 0.00041564926505088806, 'timing/model': 0.5360708571970463, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.56})
+[2026-07-28 19:25:32,964][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.020346861332654953, 'train/grad_norm_pre_clip': 1.3813015222549438, 'timing/data': 0.0004989355802536011, 'timing/model': 0.5308430586010218, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.56})
+[2026-07-28 19:25:40,196][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.08597125113010406, 'train/grad_norm_pre_clip': 2.0468459129333496, 'timing/data': 0.0003217458724975586, 'timing/model': 0.5295381899923086, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.56})
+[2026-07-28 19:25:47,413][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.03945789858698845, 'train/grad_norm_pre_clip': 1.8139532804489136, 'timing/data': 0.000383010134100914, 'timing/model': 0.5329014770686626, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.56})
+[2026-07-28 19:25:54,641][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.04247163236141205, 'train/grad_norm_pre_clip': 1.9319499731063843, 'timing/data': 0.0003323536366224289, 'timing/model': 0.5344592034816742, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.56})
+[2026-07-28 19:26:01,867][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.020852742716670036, 'train/grad_norm_pre_clip': 0.5881521701812744, 'timing/data': 0.00038108229637145996, 'timing/model': 0.5300194323062897, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.56})
+[2026-07-28 19:26:09,134][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.07930797338485718, 'train/grad_norm_pre_clip': 2.3683741092681885, 'timing/data': 0.00036064721643924713, 'timing/model': 0.5303427558392286, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.56})
+[2026-07-28 19:26:16,390][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.012382904067635536, 'train/grad_norm_pre_clip': 0.43119505047798157, 'timing/data': 0.000353911891579628, 'timing/model': 0.5306056439876556, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.56})
+[2026-07-28 19:26:23,611][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.01168003398925066, 'train/grad_norm_pre_clip': 0.6902304887771606, 'timing/data': 0.00033511407673358917, 'timing/model': 0.5310641750693321, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.56})
+[2026-07-28 19:26:30,859][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.03580830991268158, 'train/grad_norm_pre_clip': 13.758354187011719, 'timing/data': 0.000387800857424736, 'timing/model': 0.5285990722477436, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.56})
+[2026-07-28 19:26:38,079][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.013998689129948616, 'train/grad_norm_pre_clip': 0.5992675423622131, 'timing/data': 0.00037822313606739044, 'timing/model': 0.5296066198498011, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.56})
+[2026-07-28 19:26:45,384][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.023584913462400436, 'train/grad_norm_pre_clip': 0.8632457852363586, 'timing/data': 0.0003842804580926895, 'timing/model': 0.5294164251536131, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.56})
+[2026-07-28 19:26:52,609][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.011701608076691628, 'train/grad_norm_pre_clip': 0.48245206475257874, 'timing/data': 0.0003758873790502548, 'timing/model': 0.5303920730948448, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.56})
+[2026-07-28 19:26:59,824][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.06620050966739655, 'train/grad_norm_pre_clip': 2.1149508953094482, 'timing/data': 0.0006414223462343216, 'timing/model': 0.5308340694755316, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.56})
+[2026-07-28 19:27:07,067][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.09156780689954758, 'train/grad_norm_pre_clip': 2.5018296241760254, 'timing/data': 0.00040143541991710663, 'timing/model': 0.5307271145284176, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.56})
+[2026-07-28 19:27:14,286][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.013205979950726032, 'train/grad_norm_pre_clip': 0.5721898674964905, 'timing/data': 0.0004414822906255722, 'timing/model': 0.530446246266365, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.56})
+[2026-07-28 19:27:21,513][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.03711055964231491, 'train/grad_norm_pre_clip': 1.6314340829849243, 'timing/data': 0.00037237443029880524, 'timing/model': 0.5322412978857756, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.56})
+[2026-07-28 19:27:28,731][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.015176134184002876, 'train/grad_norm_pre_clip': 0.9094634056091309, 'timing/data': 0.0003164205700159073, 'timing/model': 0.53003353998065, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.56})
+[2026-07-28 19:27:35,951][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.012785306200385094, 'train/grad_norm_pre_clip': 0.7829889059066772, 'timing/data': 0.0003170035779476166, 'timing/model': 0.5289449468255043, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.56})
+[2026-07-28 19:27:43,172][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.05355660244822502, 'train/grad_norm_pre_clip': 1.7609894275665283, 'timing/data': 0.0003709513694047928, 'timing/model': 0.5305197294801474, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.56})
+[2026-07-28 19:27:50,389][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.028572827577590942, 'train/grad_norm_pre_clip': 1.0943589210510254, 'timing/data': 0.00031052157282829285, 'timing/model': 0.5349923223257065, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.56})
+[2026-07-28 19:27:57,869][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.036925721913576126, 'train/grad_norm_pre_clip': 1.620847225189209, 'timing/data': 0.00037292204797267914, 'timing/model': 0.5294179860502481, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.56})
+[2026-07-28 19:28:05,081][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.013230651617050171, 'train/grad_norm_pre_clip': 0.595267117023468, 'timing/data': 0.00036255083978176117, 'timing/model': 0.5317060612142086, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.56})
+[2026-07-28 19:28:12,592][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.03373570740222931, 'train/grad_norm_pre_clip': 1.7850887775421143, 'timing/data': 0.0003317594528198242, 'timing/model': 0.5319377109408379, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.56})
+[2026-07-28 19:28:19,812][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.023240594193339348, 'train/grad_norm_pre_clip': 0.7088927030563354, 'timing/data': 0.0003207363188266754, 'timing/model': 0.5303169917315245, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.56})
+[2026-07-28 19:28:27,032][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.024721477180719376, 'train/grad_norm_pre_clip': 0.8076308965682983, 'timing/data': 0.00036646798253059387, 'timing/model': 0.5332662407308817, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.56})
+[2026-07-28 19:28:34,258][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.03392510116100311, 'train/grad_norm_pre_clip': 0.9448643922805786, 'timing/data': 0.0003390032798051834, 'timing/model': 0.5291342958807945, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.56})
+[2026-07-28 19:28:41,472][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.0546436533331871, 'train/grad_norm_pre_clip': 1.2245999574661255, 'timing/data': 0.00033642537891864777, 'timing/model': 0.5302329454571009, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.56})
+[2026-07-28 19:28:48,688][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.03475034981966019, 'train/grad_norm_pre_clip': 0.8290149569511414, 'timing/data': 0.0003928299993276596, 'timing/model': 0.5306179840117693, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.56})
+[2026-07-28 19:28:55,932][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.02544979937374592, 'train/grad_norm_pre_clip': 0.695379376411438, 'timing/data': 0.0003710892051458359, 'timing/model': 0.5303777977824211, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.56})
+[2026-07-28 19:29:03,162][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.027922969311475754, 'train/grad_norm_pre_clip': 1.1571043729782104, 'timing/data': 0.000317620113492012, 'timing/model': 0.5282902847975492, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.56})
+[2026-07-28 19:29:10,379][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.04017188772559166, 'train/grad_norm_pre_clip': 1.1710392236709595, 'timing/data': 0.00038591213524341583, 'timing/model': 0.530446745455265, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.56})
+[2026-07-28 19:29:17,604][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.017147071659564972, 'train/grad_norm_pre_clip': 0.8719284534454346, 'timing/data': 0.00036857277154922485, 'timing/model': 0.5308108124881983, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.57})
+[2026-07-28 19:29:24,818][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.019369257614016533, 'train/grad_norm_pre_clip': 0.6844349503517151, 'timing/data': 0.00037237629294395447, 'timing/model': 0.5303487945348024, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.57})
+[2026-07-28 19:29:32,029][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.01170194149017334, 'train/grad_norm_pre_clip': 0.4589882791042328, 'timing/data': 0.0003348942846059799, 'timing/model': 0.5305595509707928, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.57})
+[2026-07-28 19:29:39,241][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.04239977151155472, 'train/grad_norm_pre_clip': 1.0716584920883179, 'timing/data': 0.00038273073732852936, 'timing/model': 0.5318556260317564, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.57})
+[2026-07-28 19:29:46,465][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.03369409590959549, 'train/grad_norm_pre_clip': 1.5798051357269287, 'timing/data': 0.0003180895000696182, 'timing/model': 0.5364906266331673, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.57})
+[2026-07-28 19:29:53,688][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.02589830756187439, 'train/grad_norm_pre_clip': 1.2776540517807007, 'timing/data': 0.0003190748393535614, 'timing/model': 0.5305939633399248, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.57})
+[2026-07-28 19:30:00,921][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.015184530057013035, 'train/grad_norm_pre_clip': 1.4251688718795776, 'timing/data': 0.0003992430865764618, 'timing/model': 0.5303904842585325, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.57})
+[2026-07-28 19:30:08,146][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.019508713856339455, 'train/grad_norm_pre_clip': 0.7282603979110718, 'timing/data': 0.00033312663435935974, 'timing/model': 0.5312775392085314, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.57})
+[2026-07-28 19:30:15,354][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.018796823918819427, 'train/grad_norm_pre_clip': 0.6924747824668884, 'timing/data': 0.0004736725240945816, 'timing/model': 0.5308224502950907, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.57})
+[2026-07-28 19:30:22,579][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.02953556925058365, 'train/grad_norm_pre_clip': 1.0774197578430176, 'timing/data': 0.00033499859273433685, 'timing/model': 0.5314429048448801, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.57})
+[2026-07-28 19:30:29,791][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.03525692597031593, 'train/grad_norm_pre_clip': 1.2370890378952026, 'timing/data': 0.0003053378313779831, 'timing/model': 0.5316748078912497, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.57})
+[2026-07-28 19:30:37,010][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.053332775831222534, 'train/grad_norm_pre_clip': 1.8960731029510498, 'timing/data': 0.0003132447600364685, 'timing/model': 0.5295331720262766, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.57})
+[2026-07-28 19:30:44,223][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.014870420098304749, 'train/grad_norm_pre_clip': 0.9401538372039795, 'timing/data': 0.00032333284616470337, 'timing/model': 0.5305772330611944, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.57})
+[2026-07-28 19:30:51,437][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.037228427827358246, 'train/grad_norm_pre_clip': 1.343504548072815, 'timing/data': 0.0003218725323677063, 'timing/model': 0.5299788285046816, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.57})
+[2026-07-28 19:30:58,659][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.031151410192251205, 'train/grad_norm_pre_clip': 1.06513249874115, 'timing/data': 0.0003165490925312042, 'timing/model': 0.5319608468562365, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.57})
+[2026-07-28 19:31:05,878][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.030805183574557304, 'train/grad_norm_pre_clip': 1.7221969366073608, 'timing/data': 0.00036715343594551086, 'timing/model': 0.5314489398151636, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.57})
+[2026-07-28 19:31:13,101][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.059771034866571426, 'train/grad_norm_pre_clip': 1.3069281578063965, 'timing/data': 0.00037945620715618134, 'timing/model': 0.5345271546393633, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.57})
+[2026-07-28 19:31:20,313][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.03703951835632324, 'train/grad_norm_pre_clip': 1.0741536617279053, 'timing/data': 0.0003167111426591873, 'timing/model': 0.5314394198358059, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.57})
+[2026-07-28 19:31:27,531][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.03533965349197388, 'train/grad_norm_pre_clip': 0.9972966313362122, 'timing/data': 0.0003949068486690521, 'timing/model': 0.5306935086846352, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.57})
+[2026-07-28 19:31:34,751][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.04853173345327377, 'train/grad_norm_pre_clip': 1.1548752784729004, 'timing/data': 0.0003365762531757355, 'timing/model': 0.5311295967549086, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.57})
+[2026-07-28 19:31:41,968][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.019509397447109222, 'train/grad_norm_pre_clip': 0.7393167614936829, 'timing/data': 0.0003400184214115143, 'timing/model': 0.5306283757090569, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.57})
+[2026-07-28 19:31:49,192][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.0371641144156456, 'train/grad_norm_pre_clip': 1.2406730651855469, 'timing/data': 0.0003265161067247391, 'timing/model': 0.5301080755889416, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.57})
+[2026-07-28 19:31:56,420][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.026669269427657127, 'train/grad_norm_pre_clip': 0.6718985438346863, 'timing/data': 0.000343959778547287, 'timing/model': 0.5308119636029005, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.57})
+[2026-07-28 19:32:03,642][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.024948135018348694, 'train/grad_norm_pre_clip': 0.934278130531311, 'timing/data': 0.0003739427775144577, 'timing/model': 0.5310458019375801, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.57})
+[2026-07-28 19:32:10,872][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.019374486058950424, 'train/grad_norm_pre_clip': 1.1220588684082031, 'timing/data': 0.00037021003663539886, 'timing/model': 0.5308097638189793, 'learning_rate/qwen_vl_interface': 9.999998458888285e-06, 'learning_rate/action_model': 9.999998458888285e-05, 'epoch': 0.57})
+[2026-07-28 19:32:18,093][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.0600542277097702, 'train/grad_norm_pre_clip': 1.5694210529327393, 'timing/data': 0.0003363732248544693, 'timing/model': 0.5309015922248363, 'learning_rate/qwen_vl_interface': 9.999993835554141e-06, 'learning_rate/action_model': 9.999993835554141e-05, 'epoch': 0.57})
+[2026-07-28 19:32:25,638][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.05022403597831726, 'train/grad_norm_pre_clip': 0.8714272975921631, 'timing/data': 0.0003456640988588333, 'timing/model': 0.5620808620005846, 'learning_rate/qwen_vl_interface': 9.999986130000566e-06, 'learning_rate/action_model': 9.999986130000566e-05, 'epoch': 0.57})
+[2026-07-28 19:32:32,861][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.02341911755502224, 'train/grad_norm_pre_clip': 1.1201800107955933, 'timing/data': 0.00039307214319705963, 'timing/model': 0.5300996974110603, 'learning_rate/qwen_vl_interface': 9.99997534223256e-06, 'learning_rate/action_model': 9.99997534223256e-05, 'epoch': 0.57})
+[2026-07-28 19:32:40,078][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.029628941789269447, 'train/grad_norm_pre_clip': 1.161430835723877, 'timing/data': 0.0004108697175979614, 'timing/model': 0.5324000623077154, 'learning_rate/qwen_vl_interface': 9.999961472257127e-06, 'learning_rate/action_model': 9.999961472257127e-05, 'epoch': 0.57})
+[2026-07-28 19:32:47,306][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.04033563286066055, 'train/grad_norm_pre_clip': 0.8857515454292297, 'timing/data': 0.00039361976087093353, 'timing/model': 0.5318960305303335, 'learning_rate/qwen_vl_interface': 9.999944520083262e-06, 'learning_rate/action_model': 9.999944520083261e-05, 'epoch': 0.57})
+[2026-07-28 19:32:54,789][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.039382144808769226, 'train/grad_norm_pre_clip': 0.925574779510498, 'timing/data': 0.00037789158523082733, 'timing/model': 0.5301032178103924, 'learning_rate/qwen_vl_interface': 9.999924485721968e-06, 'learning_rate/action_model': 9.999924485721968e-05, 'epoch': 0.57})
+[2026-07-28 19:33:02,008][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.04893864691257477, 'train/grad_norm_pre_clip': 1.7585852146148682, 'timing/data': 0.0003289170563220978, 'timing/model': 0.5321122240275145, 'learning_rate/qwen_vl_interface': 9.999901369186244e-06, 'learning_rate/action_model': 9.999901369186244e-05, 'epoch': 0.57})
+[2026-07-28 19:33:09,226][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.036761172115802765, 'train/grad_norm_pre_clip': 0.8845136165618896, 'timing/data': 0.00032209791243076324, 'timing/model': 0.5311186723411083, 'learning_rate/qwen_vl_interface': 9.999875170491092e-06, 'learning_rate/action_model': 9.999875170491091e-05, 'epoch': 0.57})
+[2026-07-28 19:33:16,448][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.036095183342695236, 'train/grad_norm_pre_clip': 0.9251642823219299, 'timing/data': 0.00036275945603847504, 'timing/model': 0.5324282571673393, 'learning_rate/qwen_vl_interface': 9.999845889653509e-06, 'learning_rate/action_model': 9.999845889653508e-05, 'epoch': 0.57})
+[2026-07-28 19:33:23,678][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.02142200618982315, 'train/grad_norm_pre_clip': 1.0202769041061401, 'timing/data': 0.0004064589738845825, 'timing/model': 0.5319173093885183, 'learning_rate/qwen_vl_interface': 9.999813526692496e-06, 'learning_rate/action_model': 9.999813526692495e-05, 'epoch': 0.57})
+[2026-07-28 19:33:30,900][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.010295011103153229, 'train/grad_norm_pre_clip': 0.36066457629203796, 'timing/data': 0.0007355250418186188, 'timing/model': 0.528721708804369, 'learning_rate/qwen_vl_interface': 9.999778081629055e-06, 'learning_rate/action_model': 9.999778081629054e-05, 'epoch': 0.57})
+[2026-07-28 19:33:38,192][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.026636194437742233, 'train/grad_norm_pre_clip': 0.8961175084114075, 'timing/data': 0.00034940987825393677, 'timing/model': 0.5601854547858238, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.57})
+[2026-07-28 19:33:45,593][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.01764153130352497, 'train/grad_norm_pre_clip': 0.8183702230453491, 'timing/data': 0.0003349222242832184, 'timing/model': 0.5569085534662008, 'learning_rate/qwen_vl_interface': 9.999697945288881e-06, 'learning_rate/action_model': 9.999697945288881e-05, 'epoch': 0.57})
+[2026-07-28 19:33:52,983][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.03040989488363266, 'train/grad_norm_pre_clip': 0.8376421332359314, 'timing/data': 0.00033508986234664917, 'timing/model': 0.564636493101716, 'learning_rate/qwen_vl_interface': 9.999653254064153e-06, 'learning_rate/action_model': 9.999653254064152e-05, 'epoch': 0.57})
+[2026-07-28 19:34:00,292][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.02489144168794155, 'train/grad_norm_pre_clip': 0.6089730262756348, 'timing/data': 0.0003364887088537216, 'timing/model': 0.5296259075403214, 'learning_rate/qwen_vl_interface': 9.99960548084099e-06, 'learning_rate/action_model': 9.99960548084099e-05, 'epoch': 0.57})
+[2026-07-28 19:34:07,517][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.03633499890565872, 'train/grad_norm_pre_clip': 1.3294845819473267, 'timing/data': 0.00032814033329486847, 'timing/model': 0.5307500287890434, 'learning_rate/qwen_vl_interface': 9.999554625650398e-06, 'learning_rate/action_model': 9.999554625650399e-05, 'epoch': 0.57})
+[2026-07-28 19:34:14,924][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.038471877574920654, 'train/grad_norm_pre_clip': 1.6464756727218628, 'timing/data': 0.00031275488436222076, 'timing/model': 0.5353706777095795, 'learning_rate/qwen_vl_interface': 9.999500688525376e-06, 'learning_rate/action_model': 9.999500688525375e-05, 'epoch': 0.57})
+[2026-07-28 19:34:22,169][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.03527205064892769, 'train/grad_norm_pre_clip': 1.2794781923294067, 'timing/data': 0.0003325454890727997, 'timing/model': 0.5375523213297129, 'learning_rate/qwen_vl_interface': 9.99944366950092e-06, 'learning_rate/action_model': 9.999443669500919e-05, 'epoch': 0.57})
+[2026-07-28 19:34:29,417][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.04623682424426079, 'train/grad_norm_pre_clip': 3.0496790409088135, 'timing/data': 0.00034195929765701294, 'timing/model': 0.5400749687105417, 'learning_rate/qwen_vl_interface': 9.99938356861403e-06, 'learning_rate/action_model': 9.99938356861403e-05, 'epoch': 0.57})
+[2026-07-28 19:34:36,659][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.036082517355680466, 'train/grad_norm_pre_clip': 1.9017724990844727, 'timing/data': 0.0003383420407772064, 'timing/model': 0.538415776565671, 'learning_rate/qwen_vl_interface': 9.999320385903709e-06, 'learning_rate/action_model': 9.999320385903708e-05, 'epoch': 0.57})
+[2026-07-28 19:34:43,915][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.024373264983296394, 'train/grad_norm_pre_clip': 1.649425983428955, 'timing/data': 0.00036367960274219513, 'timing/model': 0.5386903341859579, 'learning_rate/qwen_vl_interface': 9.999254121410948e-06, 'learning_rate/action_model': 9.999254121410949e-05, 'epoch': 0.57})
+[2026-07-28 19:34:51,164][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.03920629620552063, 'train/grad_norm_pre_clip': 2.1136183738708496, 'timing/data': 0.00034882500767707825, 'timing/model': 0.537164555862546, 'learning_rate/qwen_vl_interface': 9.999184775178751e-06, 'learning_rate/action_model': 9.999184775178751e-05, 'epoch': 0.57})
+[2026-07-28 19:34:58,421][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.041784871369600296, 'train/grad_norm_pre_clip': 2.0841526985168457, 'timing/data': 0.0003156810998916626, 'timing/model': 0.5342960953712463, 'learning_rate/qwen_vl_interface': 9.999112347252117e-06, 'learning_rate/action_model': 9.999112347252116e-05, 'epoch': 0.57})
+[2026-07-28 19:35:05,667][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.022566955536603928, 'train/grad_norm_pre_clip': 1.394766926765442, 'timing/data': 0.0003852900117635727, 'timing/model': 0.5375737119466066, 'learning_rate/qwen_vl_interface': 9.99903683767804e-06, 'learning_rate/action_model': 9.99903683767804e-05, 'epoch': 0.57})
+[2026-07-28 19:35:12,928][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.007741183042526245, 'train/grad_norm_pre_clip': 0.5203608274459839, 'timing/data': 0.0004082322120666504, 'timing/model': 0.5419268310070038, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.57})
+[2026-07-28 19:35:20,163][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.011814821511507034, 'train/grad_norm_pre_clip': 0.841081976890564, 'timing/data': 0.00032357312738895416, 'timing/model': 0.5331869721412659, 'learning_rate/qwen_vl_interface': 9.998876573785547e-06, 'learning_rate/action_model': 9.998876573785547e-05, 'epoch': 0.57})
+[2026-07-28 19:35:27,403][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.07528925687074661, 'train/grad_norm_pre_clip': 2.1142985820770264, 'timing/data': 0.00032539665699005127, 'timing/model': 0.5391220785677433, 'learning_rate/qwen_vl_interface': 9.998791819571126e-06, 'learning_rate/action_model': 9.998791819571127e-05, 'epoch': 0.57})
+[2026-07-28 19:35:34,642][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.036241624504327774, 'train/grad_norm_pre_clip': 1.5278432369232178, 'timing/data': 0.00035469047725200653, 'timing/model': 0.5357454307377338, 'learning_rate/qwen_vl_interface': 9.998703983917252e-06, 'learning_rate/action_model': 9.998703983917252e-05, 'epoch': 0.57})
+[2026-07-28 19:35:41,848][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.027436096221208572, 'train/grad_norm_pre_clip': 1.0383590459823608, 'timing/data': 0.00031662359833717346, 'timing/model': 0.5287199877202511, 'learning_rate/qwen_vl_interface': 9.998613066880919e-06, 'learning_rate/action_model': 9.998613066880918e-05, 'epoch': 0.57})
+[2026-07-28 19:35:49,055][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.014440401457250118, 'train/grad_norm_pre_clip': 0.677627444267273, 'timing/data': 0.0003275461494922638, 'timing/model': 0.5289258603006601, 'learning_rate/qwen_vl_interface': 9.99851906852112e-06, 'learning_rate/action_model': 9.998519068521119e-05, 'epoch': 0.58})
+[2026-07-28 19:35:56,259][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.08360248804092407, 'train/grad_norm_pre_clip': 2.182178020477295, 'timing/data': 0.0003253631293773651, 'timing/model': 0.5297627691179514, 'learning_rate/qwen_vl_interface': 9.998421988898851e-06, 'learning_rate/action_model': 9.998421988898851e-05, 'epoch': 0.58})
+[2026-07-28 19:36:03,486][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.022395137697458267, 'train/grad_norm_pre_clip': 1.1180005073547363, 'timing/data': 0.00034620054066181183, 'timing/model': 0.5309430006891489, 'learning_rate/qwen_vl_interface': 9.998321828077107e-06, 'learning_rate/action_model': 9.998321828077106e-05, 'epoch': 0.58})
+[2026-07-28 19:36:10,705][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.03780629485845566, 'train/grad_norm_pre_clip': 0.9515622854232788, 'timing/data': 0.00031920336186885834, 'timing/model': 0.5323801096528769, 'learning_rate/qwen_vl_interface': 9.99821858612088e-06, 'learning_rate/action_model': 9.99821858612088e-05, 'epoch': 0.58})
+[2026-07-28 19:36:18,064][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.01842948980629444, 'train/grad_norm_pre_clip': 0.6183863878250122, 'timing/data': 0.0003736913204193115, 'timing/model': 0.5581243243068457, 'learning_rate/qwen_vl_interface': 9.998112263097164e-06, 'learning_rate/action_model': 9.998112263097165e-05, 'epoch': 0.58})
+[2026-07-28 19:36:25,288][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.050619810819625854, 'train/grad_norm_pre_clip': 2.951782464981079, 'timing/data': 0.0003198161721229553, 'timing/model': 0.5323719847947359, 'learning_rate/qwen_vl_interface': 9.99800285907495e-06, 'learning_rate/action_model': 9.99800285907495e-05, 'epoch': 0.58})
+[2026-07-28 19:36:32,517][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.015928151085972786, 'train/grad_norm_pre_clip': 0.6199173927307129, 'timing/data': 0.0003352370113134384, 'timing/model': 0.5318466071039438, 'learning_rate/qwen_vl_interface': 9.997890374125228e-06, 'learning_rate/action_model': 9.997890374125227e-05, 'epoch': 0.58})
+[2026-07-28 19:36:39,742][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.08231334388256073, 'train/grad_norm_pre_clip': 1.1221519708633423, 'timing/data': 0.0003362353891134262, 'timing/model': 0.5295809470117092, 'learning_rate/qwen_vl_interface': 9.997774808320992e-06, 'learning_rate/action_model': 9.997774808320992e-05, 'epoch': 0.58})
+[2026-07-28 19:36:46,966][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.025651058182120323, 'train/grad_norm_pre_clip': 1.3388029336929321, 'timing/data': 0.0007177460938692093, 'timing/model': 0.5291238892823458, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.58})
+[2026-07-28 19:36:54,467][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.01992882788181305, 'train/grad_norm_pre_clip': 0.6181800961494446, 'timing/data': 0.00034112483263015747, 'timing/model': 0.5327872149646282, 'learning_rate/qwen_vl_interface': 9.997534434450922e-06, 'learning_rate/action_model': 9.997534434450921e-05, 'epoch': 0.58})
+[2026-07-28 19:37:01,679][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.039282940328121185, 'train/grad_norm_pre_clip': 1.9863752126693726, 'timing/data': 0.0003941822797060013, 'timing/model': 0.5300326906144619, 'learning_rate/qwen_vl_interface': 9.997409626541065e-06, 'learning_rate/action_model': 9.997409626541066e-05, 'epoch': 0.58})
+[2026-07-28 19:37:08,898][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.02851189859211445, 'train/grad_norm_pre_clip': 0.8161270022392273, 'timing/data': 0.0003147982060909271, 'timing/model': 0.5293180551379919, 'learning_rate/qwen_vl_interface': 9.997281738088647e-06, 'learning_rate/action_model': 9.997281738088646e-05, 'epoch': 0.58})
+[2026-07-28 19:37:16,117][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.02356436476111412, 'train/grad_norm_pre_clip': 1.351481318473816, 'timing/data': 0.00033030472695827484, 'timing/model': 0.5305752418935299, 'learning_rate/qwen_vl_interface': 9.997150769176648e-06, 'learning_rate/action_model': 9.997150769176647e-05, 'epoch': 0.58})
+[2026-07-28 19:37:23,333][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.033876825124025345, 'train/grad_norm_pre_clip': 0.7370008826255798, 'timing/data': 0.0003062598407268524, 'timing/model': 0.5296044796705246, 'learning_rate/qwen_vl_interface': 9.997016719890052e-06, 'learning_rate/action_model': 9.997016719890052e-05, 'epoch': 0.58})
+[2026-07-28 19:37:30,755][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.025431200861930847, 'train/grad_norm_pre_clip': 0.6406704187393188, 'timing/data': 0.00030410289764404297, 'timing/model': 0.7214146628975868, 'learning_rate/qwen_vl_interface': 9.996879590315847e-06, 'learning_rate/action_model': 9.996879590315846e-05, 'epoch': 0.58})
+[2026-07-28 19:37:37,975][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.043350834399461746, 'train/grad_norm_pre_clip': 1.3832621574401855, 'timing/data': 0.00032873451709747314, 'timing/model': 0.5328966304659843, 'learning_rate/qwen_vl_interface': 9.99673938054301e-06, 'learning_rate/action_model': 9.996739380543009e-05, 'epoch': 0.58})
+[2026-07-28 19:37:45,179][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.043995197862386703, 'train/grad_norm_pre_clip': 1.2891976833343506, 'timing/data': 0.0003105495125055313, 'timing/model': 0.5296002831310034, 'learning_rate/qwen_vl_interface': 9.996596090662525e-06, 'learning_rate/action_model': 9.996596090662525e-05, 'epoch': 0.58})
+[2026-07-28 19:37:52,399][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.013065274804830551, 'train/grad_norm_pre_clip': 0.9706422090530396, 'timing/data': 0.00031539425253868103, 'timing/model': 0.5304305497556925, 'learning_rate/qwen_vl_interface': 9.996449720767369e-06, 'learning_rate/action_model': 9.996449720767368e-05, 'epoch': 0.58})
+[2026-07-28 19:37:59,626][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.03311610221862793, 'train/grad_norm_pre_clip': 0.712689995765686, 'timing/data': 0.0003304518759250641, 'timing/model': 0.5305422600358725, 'learning_rate/qwen_vl_interface': 9.99630027095252e-06, 'learning_rate/action_model': 9.99630027095252e-05, 'epoch': 0.58})
+[2026-07-28 19:38:06,848][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.021073829382658005, 'train/grad_norm_pre_clip': 0.9711698889732361, 'timing/data': 0.00035926327109336853, 'timing/model': 0.5306374281644821, 'learning_rate/qwen_vl_interface': 9.996147741314956e-06, 'learning_rate/action_model': 9.996147741314957e-05, 'epoch': 0.58})
+[2026-07-28 19:38:14,092][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.03485524654388428, 'train/grad_norm_pre_clip': 0.9272443056106567, 'timing/data': 0.00034403055906295776, 'timing/model': 0.5325459931045771, 'learning_rate/qwen_vl_interface': 9.99599213195365e-06, 'learning_rate/action_model': 9.99599213195365e-05, 'epoch': 0.58})
+[2026-07-28 19:38:21,327][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.04920452833175659, 'train/grad_norm_pre_clip': 1.4621858596801758, 'timing/data': 0.0003129243850708008, 'timing/model': 0.5339458845555782, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.58})
+[2026-07-28 19:38:28,550][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.018985938280820847, 'train/grad_norm_pre_clip': 0.9238725900650024, 'timing/data': 0.0004630163311958313, 'timing/model': 0.5306380093097687, 'learning_rate/qwen_vl_interface': 9.995671674465709e-06, 'learning_rate/action_model': 9.995671674465709e-05, 'epoch': 0.58})
+[2026-07-28 19:38:35,774][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.01729152351617813, 'train/grad_norm_pre_clip': 0.8858495950698853, 'timing/data': 0.0003152787685394287, 'timing/model': 0.5297717992216349, 'learning_rate/qwen_vl_interface': 9.99550682654701e-06, 'learning_rate/action_model': 9.995506826547011e-05, 'epoch': 0.58})
+[2026-07-28 19:38:43,005][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.07445693016052246, 'train/grad_norm_pre_clip': 3.146834135055542, 'timing/data': 0.00037369504570961, 'timing/model': 0.531117731705308, 'learning_rate/qwen_vl_interface': 9.995338899320457e-06, 'learning_rate/action_model': 9.995338899320456e-05, 'epoch': 0.58})
+[2026-07-28 19:38:50,256][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.0517096221446991, 'train/grad_norm_pre_clip': 1.5363174676895142, 'timing/data': 0.000387703999876976, 'timing/model': 0.5328640751540661, 'learning_rate/qwen_vl_interface': 9.995167892895011e-06, 'learning_rate/action_model': 9.99516789289501e-05, 'epoch': 0.58})
+[2026-07-28 19:38:57,494][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.035920899361371994, 'train/grad_norm_pre_clip': 1.9857996702194214, 'timing/data': 0.0004417840391397476, 'timing/model': 0.5313134733587503, 'learning_rate/qwen_vl_interface': 9.994993807381635e-06, 'learning_rate/action_model': 9.994993807381635e-05, 'epoch': 0.58})
+[2026-07-28 19:39:04,703][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.04090778902173042, 'train/grad_norm_pre_clip': 2.466458320617676, 'timing/data': 0.00036578811705112457, 'timing/model': 0.5329757165163755, 'learning_rate/qwen_vl_interface': 9.994816642893294e-06, 'learning_rate/action_model': 9.994816642893293e-05, 'epoch': 0.58})
+[2026-07-28 19:39:11,933][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.025831226259469986, 'train/grad_norm_pre_clip': 0.7419413328170776, 'timing/data': 0.00033292733132839203, 'timing/model': 0.5286794565618038, 'learning_rate/qwen_vl_interface': 9.994636399544948e-06, 'learning_rate/action_model': 9.994636399544947e-05, 'epoch': 0.58})
+[2026-07-28 19:39:19,139][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.04411061108112335, 'train/grad_norm_pre_clip': 1.6098545789718628, 'timing/data': 0.00038806162774562836, 'timing/model': 0.5314686764031649, 'learning_rate/qwen_vl_interface': 9.994453077453552e-06, 'learning_rate/action_model': 9.994453077453552e-05, 'epoch': 0.58})
+[2026-07-28 19:39:26,355][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.02310228906571865, 'train/grad_norm_pre_clip': 0.7409632802009583, 'timing/data': 0.0003393683582544327, 'timing/model': 0.5309412311762571, 'learning_rate/qwen_vl_interface': 9.994266676738065e-06, 'learning_rate/action_model': 9.994266676738065e-05, 'epoch': 0.58})
+[2026-07-28 19:39:33,575][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.07003941386938095, 'train/grad_norm_pre_clip': 3.2179877758026123, 'timing/data': 0.0003383420407772064, 'timing/model': 0.5311264134943485, 'learning_rate/qwen_vl_interface': 9.99407719751944e-06, 'learning_rate/action_model': 9.994077197519439e-05, 'epoch': 0.58})
+[2026-07-28 19:39:40,796][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.020360995084047318, 'train/grad_norm_pre_clip': 0.5580732822418213, 'timing/data': 0.00040861964225769043, 'timing/model': 0.5311771314591169, 'learning_rate/qwen_vl_interface': 9.993884639920626e-06, 'learning_rate/action_model': 9.993884639920626e-05, 'epoch': 0.58})
+[2026-07-28 19:39:48,013][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.03408707678318024, 'train/grad_norm_pre_clip': 1.3704757690429688, 'timing/data': 0.0003638956695795059, 'timing/model': 0.5311888828873634, 'learning_rate/qwen_vl_interface': 9.993689004066573e-06, 'learning_rate/action_model': 9.993689004066573e-05, 'epoch': 0.58})
+[2026-07-28 19:39:55,236][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.030559539794921875, 'train/grad_norm_pre_clip': 0.7931413054466248, 'timing/data': 0.0003148205578327179, 'timing/model': 0.5304043777287006, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.58})
+[2026-07-28 19:40:02,453][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.023278722539544106, 'train/grad_norm_pre_clip': 1.1476563215255737, 'timing/data': 0.00043277256190776825, 'timing/model': 0.5312762334942818, 'learning_rate/qwen_vl_interface': 9.993288498102531e-06, 'learning_rate/action_model': 9.993288498102532e-05, 'epoch': 0.58})
+[2026-07-28 19:40:09,671][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.054508961737155914, 'train/grad_norm_pre_clip': 1.2481074333190918, 'timing/data': 0.0003237910568714142, 'timing/model': 0.5314209051430225, 'learning_rate/qwen_vl_interface': 9.993083628252424e-06, 'learning_rate/action_model': 9.993083628252424e-05, 'epoch': 0.58})
+[2026-07-28 19:40:16,880][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.028586674481630325, 'train/grad_norm_pre_clip': 1.1795799732208252, 'timing/data': 0.00033671222627162933, 'timing/model': 0.529469657689333, 'learning_rate/qwen_vl_interface': 9.992875680666849e-06, 'learning_rate/action_model': 9.992875680666849e-05, 'epoch': 0.58})
+[2026-07-28 19:40:24,120][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.016036171466112137, 'train/grad_norm_pre_clip': 0.4494108259677887, 'timing/data': 0.0003620721399784088, 'timing/model': 0.5290006436407566, 'learning_rate/qwen_vl_interface': 9.992664655480736e-06, 'learning_rate/action_model': 9.992664655480735e-05, 'epoch': 0.58})
+[2026-07-28 19:40:31,339][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.03146757185459137, 'train/grad_norm_pre_clip': 0.7995524406433105, 'timing/data': 0.000352662056684494, 'timing/model': 0.5290970858186483, 'learning_rate/qwen_vl_interface': 9.992450552831018e-06, 'learning_rate/action_model': 9.992450552831017e-05, 'epoch': 0.58})
+[2026-07-28 19:40:38,563][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.0227142795920372, 'train/grad_norm_pre_clip': 1.1688557863235474, 'timing/data': 0.0003195255994796753, 'timing/model': 0.5300918780267239, 'learning_rate/qwen_vl_interface': 9.992233372856623e-06, 'learning_rate/action_model': 9.992233372856624e-05, 'epoch': 0.58})
+[2026-07-28 19:40:45,826][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.06966002285480499, 'train/grad_norm_pre_clip': 1.2540193796157837, 'timing/data': 0.00039733946323394775, 'timing/model': 0.5296118222177029, 'learning_rate/qwen_vl_interface': 9.99201311569848e-06, 'learning_rate/action_model': 9.99201311569848e-05, 'epoch': 0.58})
+[2026-07-28 19:40:53,039][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.04296048730611801, 'train/grad_norm_pre_clip': 1.250211238861084, 'timing/data': 0.0003331601619720459, 'timing/model': 0.5302948132157326, 'learning_rate/qwen_vl_interface': 9.99178978149951e-06, 'learning_rate/action_model': 9.99178978149951e-05, 'epoch': 0.58})
+[2026-07-28 19:41:00,269][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.007210863754153252, 'train/grad_norm_pre_clip': 0.5809416770935059, 'timing/data': 0.0003292970359325409, 'timing/model': 0.5309864431619644, 'learning_rate/qwen_vl_interface': 9.99156337040463e-06, 'learning_rate/action_model': 9.991563370404629e-05, 'epoch': 0.58})
+[2026-07-28 19:41:07,496][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.04318246245384216, 'train/grad_norm_pre_clip': 1.128166675567627, 'timing/data': 0.0003288891166448593, 'timing/model': 0.5308450311422348, 'learning_rate/qwen_vl_interface': 9.991333882560759e-06, 'learning_rate/action_model': 9.991333882560759e-05, 'epoch': 0.58})
+[2026-07-28 19:41:14,705][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.0615466944873333, 'train/grad_norm_pre_clip': 1.511650800704956, 'timing/data': 0.0003860965371131897, 'timing/model': 0.5299456007778645, 'learning_rate/qwen_vl_interface': 9.991101318116806e-06, 'learning_rate/action_model': 9.991101318116806e-05, 'epoch': 0.58})
+[2026-07-28 19:41:22,203][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.06959299743175507, 'train/grad_norm_pre_clip': 1.7139931917190552, 'timing/data': 0.00035715289413928986, 'timing/model': 0.530890591442585, 'learning_rate/qwen_vl_interface': 9.990865677223684e-06, 'learning_rate/action_model': 9.990865677223684e-05, 'epoch': 0.58})
+[2026-07-28 19:41:29,422][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.048091113567352295, 'train/grad_norm_pre_clip': 1.2276043891906738, 'timing/data': 0.00032144971191883087, 'timing/model': 0.5310892779380083, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.58})
+[2026-07-28 19:41:36,659][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.028933554887771606, 'train/grad_norm_pre_clip': 1.3920921087265015, 'timing/data': 0.0003718007355928421, 'timing/model': 0.5310947559773922, 'learning_rate/qwen_vl_interface': 9.990385166703534e-06, 'learning_rate/action_model': 9.990385166703533e-05, 'epoch': 0.58})
+[2026-07-28 19:41:43,876][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.02617400325834751, 'train/grad_norm_pre_clip': 0.7976472973823547, 'timing/data': 0.00030577555298805237, 'timing/model': 0.529826533049345, 'learning_rate/qwen_vl_interface': 9.990140297388306e-06, 'learning_rate/action_model': 9.990140297388307e-05, 'epoch': 0.58})
+[2026-07-28 19:41:51,131][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.03677774965763092, 'train/grad_norm_pre_clip': 1.131028175354004, 'timing/data': 0.00038853101432323456, 'timing/model': 0.5453311875462532, 'learning_rate/qwen_vl_interface': 9.989892352247505e-06, 'learning_rate/action_model': 9.989892352247504e-05, 'epoch': 0.58})
+[2026-07-28 19:41:58,385][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.036351028829813004, 'train/grad_norm_pre_clip': 1.3839906454086304, 'timing/data': 0.0003645997494459152, 'timing/model': 0.5323925968259573, 'learning_rate/qwen_vl_interface': 9.989641331442014e-06, 'learning_rate/action_model': 9.989641331442013e-05, 'epoch': 0.58})
+[2026-07-28 19:42:05,610][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.047795720398426056, 'train/grad_norm_pre_clip': 1.9568750858306885, 'timing/data': 0.0003489609807729721, 'timing/model': 0.5334060564637184, 'learning_rate/qwen_vl_interface': 9.989387235134722e-06, 'learning_rate/action_model': 9.989387235134722e-05, 'epoch': 0.58})
+[2026-07-28 19:42:13,103][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.03427368029952049, 'train/grad_norm_pre_clip': 1.255192756652832, 'timing/data': 0.0003626439720392227, 'timing/model': 0.5305585414171219, 'learning_rate/qwen_vl_interface': 9.98913006349051e-06, 'learning_rate/action_model': 9.989130063490508e-05, 'epoch': 0.58})
+[2026-07-28 19:42:20,329][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.05862465500831604, 'train/grad_norm_pre_clip': 1.3159743547439575, 'timing/data': 0.0003131255507469177, 'timing/model': 0.5312643349170685, 'learning_rate/qwen_vl_interface': 9.988869816676247e-06, 'learning_rate/action_model': 9.988869816676246e-05, 'epoch': 0.59})
+[2026-07-28 19:42:27,582][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.016997158527374268, 'train/grad_norm_pre_clip': 0.7191210985183716, 'timing/data': 0.0003214683383703232, 'timing/model': 0.5294579360634089, 'learning_rate/qwen_vl_interface': 9.988606494860811e-06, 'learning_rate/action_model': 9.988606494860812e-05, 'epoch': 0.59})
+[2026-07-28 19:42:34,789][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.03820434585213661, 'train/grad_norm_pre_clip': 1.0539554357528687, 'timing/data': 0.00042053498327732086, 'timing/model': 0.5303224213421345, 'learning_rate/qwen_vl_interface': 9.98834009821507e-06, 'learning_rate/action_model': 9.988340098215069e-05, 'epoch': 0.59})
+[2026-07-28 19:42:42,041][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.022533806040883064, 'train/grad_norm_pre_clip': 0.9577446579933167, 'timing/data': 0.00030647963285446167, 'timing/model': 0.5308363810181618, 'learning_rate/qwen_vl_interface': 9.98807062691188e-06, 'learning_rate/action_model': 9.988070626911879e-05, 'epoch': 0.59})
+[2026-07-28 19:42:49,265][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.044214121997356415, 'train/grad_norm_pre_clip': 1.861775517463684, 'timing/data': 0.00040209852159023285, 'timing/model': 0.5320648141205311, 'learning_rate/qwen_vl_interface': 9.987798081126099e-06, 'learning_rate/action_model': 9.987798081126099e-05, 'epoch': 0.59})
+[2026-07-28 19:42:56,491][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.03776243329048157, 'train/grad_norm_pre_clip': 1.1237974166870117, 'timing/data': 0.0005229432135820389, 'timing/model': 0.5356986429542303, 'learning_rate/qwen_vl_interface': 9.987522461034584e-06, 'learning_rate/action_model': 9.987522461034584e-05, 'epoch': 0.59})
+[2026-07-28 19:43:03,726][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.0406835675239563, 'train/grad_norm_pre_clip': 1.93574857711792, 'timing/data': 0.0003317836672067642, 'timing/model': 0.5335099902004004, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.59})
+[2026-07-28 19:43:10,948][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.029015492647886276, 'train/grad_norm_pre_clip': 0.9673151969909668, 'timing/data': 0.00035915151238441467, 'timing/model': 0.5326774101704359, 'learning_rate/qwen_vl_interface': 9.98696199865172e-06, 'learning_rate/action_model': 9.98696199865172e-05, 'epoch': 0.59})
+[2026-07-28 19:43:18,186][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.0768076628446579, 'train/grad_norm_pre_clip': 2.301988363265991, 'timing/data': 0.00040093250572681427, 'timing/model': 0.5332314670085907, 'learning_rate/qwen_vl_interface': 9.986677156724054e-06, 'learning_rate/action_model': 9.986677156724054e-05, 'epoch': 0.59})
+[2026-07-28 19:43:25,412][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.02082822471857071, 'train/grad_norm_pre_clip': 1.0090010166168213, 'timing/data': 0.00038952939212322235, 'timing/model': 0.5352026373147964, 'learning_rate/qwen_vl_interface': 9.986389241218003e-06, 'learning_rate/action_model': 9.986389241218003e-05, 'epoch': 0.59})
+[2026-07-28 19:43:32,636][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.09495025873184204, 'train/grad_norm_pre_clip': 2.294877529144287, 'timing/data': 0.00033822283148765564, 'timing/model': 0.5309165101498365, 'learning_rate/qwen_vl_interface': 9.986098252320397e-06, 'learning_rate/action_model': 9.986098252320397e-05, 'epoch': 0.59})
+[2026-07-28 19:43:39,861][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.04360489547252655, 'train/grad_norm_pre_clip': 1.583136796951294, 'timing/data': 0.0003226008266210556, 'timing/model': 0.5325321480631828, 'learning_rate/qwen_vl_interface': 9.985804190220054e-06, 'learning_rate/action_model': 9.985804190220053e-05, 'epoch': 0.59})
+[2026-07-28 19:43:47,089][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.03820279240608215, 'train/grad_norm_pre_clip': 1.152921199798584, 'timing/data': 0.0003399346023797989, 'timing/model': 0.5317654386162758, 'learning_rate/qwen_vl_interface': 9.985507055107788e-06, 'learning_rate/action_model': 9.985507055107788e-05, 'epoch': 0.59})
+[2026-07-28 19:43:54,337][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.021523673087358475, 'train/grad_norm_pre_clip': 1.1845215559005737, 'timing/data': 0.00035609304904937744, 'timing/model': 0.535706439986825, 'learning_rate/qwen_vl_interface': 9.985206847176406e-06, 'learning_rate/action_model': 9.985206847176406e-05, 'epoch': 0.59})
+[2026-07-28 19:44:01,605][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.030081763863563538, 'train/grad_norm_pre_clip': 0.8703964948654175, 'timing/data': 0.00031878799200057983, 'timing/model': 0.5318929459899664, 'learning_rate/qwen_vl_interface': 9.984903566620711e-06, 'learning_rate/action_model': 9.984903566620712e-05, 'epoch': 0.59})
+[2026-07-28 19:44:08,835][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.02119571343064308, 'train/grad_norm_pre_clip': 0.8890127539634705, 'timing/data': 0.0003557652235031128, 'timing/model': 0.5309335347265005, 'learning_rate/qwen_vl_interface': 9.984597213637497e-06, 'learning_rate/action_model': 9.984597213637497e-05, 'epoch': 0.59})
+[2026-07-28 19:44:16,047][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.02878250554203987, 'train/grad_norm_pre_clip': 1.170839786529541, 'timing/data': 0.0003871731460094452, 'timing/model': 0.5297693740576506, 'learning_rate/qwen_vl_interface': 9.984287788425555e-06, 'learning_rate/action_model': 9.984287788425555e-05, 'epoch': 0.59})
+[2026-07-28 19:44:23,275][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.02322416380047798, 'train/grad_norm_pre_clip': 1.2671432495117188, 'timing/data': 0.0003471300005912781, 'timing/model': 0.5340576954185963, 'learning_rate/qwen_vl_interface': 9.983975291185666e-06, 'learning_rate/action_model': 9.983975291185667e-05, 'epoch': 0.59})
+[2026-07-28 19:44:30,527][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.03825252130627632, 'train/grad_norm_pre_clip': 1.69839608669281, 'timing/data': 0.0003256388008594513, 'timing/model': 0.5305589605122805, 'learning_rate/qwen_vl_interface': 9.983659722120607e-06, 'learning_rate/action_model': 9.983659722120608e-05, 'epoch': 0.59})
+[2026-07-28 19:44:37,772][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.012257428839802742, 'train/grad_norm_pre_clip': 0.43351131677627563, 'timing/data': 0.0003332998603582382, 'timing/model': 0.5301390122622252, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.59})
+[2026-07-28 19:44:45,019][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.02748947963118553, 'train/grad_norm_pre_clip': 0.5978043079376221, 'timing/data': 0.00039106421172618866, 'timing/model': 0.5297485068440437, 'learning_rate/qwen_vl_interface': 9.983019369336048e-06, 'learning_rate/action_model': 9.983019369336048e-05, 'epoch': 0.59})
+[2026-07-28 19:44:52,233][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.05026034638285637, 'train/grad_norm_pre_clip': 2.7956457138061523, 'timing/data': 0.00033376365900039673, 'timing/model': 0.5290480479598045, 'learning_rate/qwen_vl_interface': 9.982694586032066e-06, 'learning_rate/action_model': 9.982694586032065e-05, 'epoch': 0.59})
+[2026-07-28 19:44:59,454][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.02373235486447811, 'train/grad_norm_pre_clip': 0.9671164155006409, 'timing/data': 0.0003300812095403671, 'timing/model': 0.5301440488547087, 'learning_rate/qwen_vl_interface': 9.982366731733948e-06, 'learning_rate/action_model': 9.982366731733947e-05, 'epoch': 0.59})
+[2026-07-28 19:45:06,705][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.027500463649630547, 'train/grad_norm_pre_clip': 1.1783167123794556, 'timing/data': 0.0004325173795223236, 'timing/model': 0.5292857177555561, 'learning_rate/qwen_vl_interface': 9.982035806654437e-06, 'learning_rate/action_model': 9.982035806654437e-05, 'epoch': 0.59})
+[2026-07-28 19:45:13,985][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.030132019892334938, 'train/grad_norm_pre_clip': 1.4307314157485962, 'timing/data': 0.0003361552953720093, 'timing/model': 0.5290130749344826, 'learning_rate/qwen_vl_interface': 9.981701811008267e-06, 'learning_rate/action_model': 9.981701811008267e-05, 'epoch': 0.59})
+[2026-07-28 19:45:21,221][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.023096200078725815, 'train/grad_norm_pre_clip': 0.6507115364074707, 'timing/data': 0.0003227032721042633, 'timing/model': 0.5307592358440161, 'learning_rate/qwen_vl_interface': 9.981364745012163e-06, 'learning_rate/action_model': 9.981364745012163e-05, 'epoch': 0.59})
+[2026-07-28 19:45:28,472][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.07104083150625229, 'train/grad_norm_pre_clip': 2.6683714389801025, 'timing/data': 0.00042478740215301514, 'timing/model': 0.5312071312218904, 'learning_rate/qwen_vl_interface': 9.981024608884842e-06, 'learning_rate/action_model': 9.981024608884843e-05, 'epoch': 0.59})
+[2026-07-28 19:45:35,738][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.06627754867076874, 'train/grad_norm_pre_clip': 1.5668014287948608, 'timing/data': 0.0005124881863594055, 'timing/model': 0.5302020572125912, 'learning_rate/qwen_vl_interface': 9.980681402847017e-06, 'learning_rate/action_model': 9.980681402847016e-05, 'epoch': 0.59})
+[2026-07-28 19:45:42,997][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.03231433779001236, 'train/grad_norm_pre_clip': 1.59807550907135, 'timing/data': 0.0003280807286500931, 'timing/model': 0.5307525601238012, 'learning_rate/qwen_vl_interface': 9.980335127121391e-06, 'learning_rate/action_model': 9.98033512712139e-05, 'epoch': 0.59})
+[2026-07-28 19:45:50,489][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.027128521353006363, 'train/grad_norm_pre_clip': 1.18438720703125, 'timing/data': 0.0003388691693544388, 'timing/model': 0.5302811469882727, 'learning_rate/qwen_vl_interface': 9.979985781932658e-06, 'learning_rate/action_model': 9.979985781932657e-05, 'epoch': 0.59})
+[2026-07-28 19:45:57,729][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.02245221845805645, 'train/grad_norm_pre_clip': 0.689695417881012, 'timing/data': 0.0005177780985832214, 'timing/model': 0.530973594635725, 'learning_rate/qwen_vl_interface': 9.979633367507501e-06, 'learning_rate/action_model': 9.979633367507502e-05, 'epoch': 0.59})
+[2026-07-28 19:46:04,935][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.015256570652127266, 'train/grad_norm_pre_clip': 0.7490767240524292, 'timing/data': 0.0003304034471511841, 'timing/model': 0.5287157408893108, 'learning_rate/qwen_vl_interface': 9.979277884074604e-06, 'learning_rate/action_model': 9.979277884074604e-05, 'epoch': 0.59})
+[2026-07-28 19:46:12,203][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.04222961887717247, 'train/grad_norm_pre_clip': 1.4829895496368408, 'timing/data': 0.0004183407872915268, 'timing/model': 0.5304461475461721, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.59})
+[2026-07-28 19:46:19,447][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.04032488167285919, 'train/grad_norm_pre_clip': 1.5773613452911377, 'timing/data': 0.00033518485724925995, 'timing/model': 0.5317352954298258, 'learning_rate/qwen_vl_interface': 9.978557711110245e-06, 'learning_rate/action_model': 9.978557711110244e-05, 'epoch': 0.59})
+[2026-07-28 19:46:26,679][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.02313046157360077, 'train/grad_norm_pre_clip': 0.6998468637466431, 'timing/data': 0.0004321392625570297, 'timing/model': 0.5323327276855707, 'learning_rate/qwen_vl_interface': 9.978193022046097e-06, 'learning_rate/action_model': 9.978193022046097e-05, 'epoch': 0.59})
+[2026-07-28 19:46:33,911][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.05267181247472763, 'train/grad_norm_pre_clip': 1.3018702268600464, 'timing/data': 0.0006471630185842514, 'timing/model': 0.5315897352993488, 'learning_rate/qwen_vl_interface': 9.977825264908832e-06, 'learning_rate/action_model': 9.977825264908832e-05, 'epoch': 0.59})
+[2026-07-28 19:46:41,164][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.03149387985467911, 'train/grad_norm_pre_clip': 1.2607165575027466, 'timing/data': 0.00041021406650543213, 'timing/model': 0.5328042842447758, 'learning_rate/qwen_vl_interface': 9.97745443993708e-06, 'learning_rate/action_model': 9.97745443993708e-05, 'epoch': 0.59})
+[2026-07-28 19:46:48,406][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.01038607768714428, 'train/grad_norm_pre_clip': 0.9641194939613342, 'timing/data': 0.00048794597387313843, 'timing/model': 0.5335026886314154, 'learning_rate/qwen_vl_interface': 9.977080547371468e-06, 'learning_rate/action_model': 9.977080547371468e-05, 'epoch': 0.59})
+[2026-07-28 19:46:55,616][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.015852365642786026, 'train/grad_norm_pre_clip': 1.0464855432510376, 'timing/data': 0.0003932025283575058, 'timing/model': 0.5296860095113516, 'learning_rate/qwen_vl_interface': 9.97670358745461e-06, 'learning_rate/action_model': 9.97670358745461e-05, 'epoch': 0.59})
+[2026-07-28 19:47:03,109][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.032424867153167725, 'train/grad_norm_pre_clip': 1.085934042930603, 'timing/data': 0.00036627426743507385, 'timing/model': 0.8097895104438066, 'learning_rate/qwen_vl_interface': 9.976323560431112e-06, 'learning_rate/action_model': 9.976323560431111e-05, 'epoch': 0.59})
+[2026-07-28 19:47:10,341][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.0299852192401886, 'train/grad_norm_pre_clip': 0.8650460839271545, 'timing/data': 0.0003954097628593445, 'timing/model': 0.529696024954319, 'learning_rate/qwen_vl_interface': 9.975940466547567e-06, 'learning_rate/action_model': 9.975940466547567e-05, 'epoch': 0.59})
+[2026-07-28 19:47:17,570][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.03696414828300476, 'train/grad_norm_pre_clip': 1.17062509059906, 'timing/data': 0.00033890828490257263, 'timing/model': 0.5295001063495874, 'learning_rate/qwen_vl_interface': 9.975554306052563e-06, 'learning_rate/action_model': 9.975554306052562e-05, 'epoch': 0.59})
+[2026-07-28 19:47:24,799][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.029779233038425446, 'train/grad_norm_pre_clip': 0.8463326096534729, 'timing/data': 0.0003600623458623886, 'timing/model': 0.5296317245811224, 'learning_rate/qwen_vl_interface': 9.975165079196674e-06, 'learning_rate/action_model': 9.975165079196673e-05, 'epoch': 0.59})
+[2026-07-28 19:47:32,029][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.02256445214152336, 'train/grad_norm_pre_clip': 0.5802955627441406, 'timing/data': 0.00038787350058555603, 'timing/model': 0.5307703018188477, 'learning_rate/qwen_vl_interface': 9.974772786232466e-06, 'learning_rate/action_model': 9.974772786232466e-05, 'epoch': 0.59})
+[2026-07-28 19:47:39,274][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.03824734687805176, 'train/grad_norm_pre_clip': 0.9314444661140442, 'timing/data': 0.0003555305302143097, 'timing/model': 0.5303475186228752, 'learning_rate/qwen_vl_interface': 9.974377427414493e-06, 'learning_rate/action_model': 9.974377427414493e-05, 'epoch': 0.59})
+[2026-07-28 19:47:46,547][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.0177658312022686, 'train/grad_norm_pre_clip': 0.9426897764205933, 'timing/data': 0.0003394130617380142, 'timing/model': 0.530585253611207, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.59})
+[2026-07-28 19:47:53,772][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.043650105595588684, 'train/grad_norm_pre_clip': 1.407518744468689, 'timing/data': 0.00037593021988868713, 'timing/model': 0.5297715403139591, 'learning_rate/qwen_vl_interface': 9.973577513245417e-06, 'learning_rate/action_model': 9.973577513245417e-05, 'epoch': 0.59})
+[2026-07-28 19:48:01,004][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.02054159715771675, 'train/grad_norm_pre_clip': 0.8368255496025085, 'timing/data': 0.0004101581871509552, 'timing/model': 0.530773913487792, 'learning_rate/qwen_vl_interface': 9.973172958413372e-06, 'learning_rate/action_model': 9.97317295841337e-05, 'epoch': 0.59})
+[2026-07-28 19:48:08,230][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.019775375723838806, 'train/grad_norm_pre_clip': 0.7671636939048767, 'timing/data': 0.00037737004458904266, 'timing/model': 0.5360123105347157, 'learning_rate/qwen_vl_interface': 9.97276533876567e-06, 'learning_rate/action_model': 9.97276533876567e-05, 'epoch': 0.59})
+[2026-07-28 19:48:15,465][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.03652849420905113, 'train/grad_norm_pre_clip': 1.3656086921691895, 'timing/data': 0.0003499593585729599, 'timing/model': 0.5302222371101379, 'learning_rate/qwen_vl_interface': 9.972354654566816e-06, 'learning_rate/action_model': 9.972354654566816e-05, 'epoch': 0.59})
+[2026-07-28 19:48:22,709][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.022329211235046387, 'train/grad_norm_pre_clip': 1.376875877380371, 'timing/data': 0.0004100929945707321, 'timing/model': 0.5294683501124382, 'learning_rate/qwen_vl_interface': 9.971940906083295e-06, 'learning_rate/action_model': 9.971940906083295e-05, 'epoch': 0.59})
+[2026-07-28 19:48:29,956][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.028674528002738953, 'train/grad_norm_pre_clip': 0.7004112601280212, 'timing/data': 0.0003884360194206238, 'timing/model': 0.5315103344619274, 'learning_rate/qwen_vl_interface': 9.971524093583586e-06, 'learning_rate/action_model': 9.971524093583586e-05, 'epoch': 0.59})
+[2026-07-28 19:48:37,196][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.05685817450284958, 'train/grad_norm_pre_clip': 1.5467122793197632, 'timing/data': 0.0003968067467212677, 'timing/model': 0.5284949243068695, 'learning_rate/qwen_vl_interface': 9.971104217338154e-06, 'learning_rate/action_model': 9.971104217338153e-05, 'epoch': 0.59})
+[2026-07-28 19:48:44,487][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.02618078514933586, 'train/grad_norm_pre_clip': 0.7671858668327332, 'timing/data': 0.0003470480442047119, 'timing/model': 0.5301606766879559, 'learning_rate/qwen_vl_interface': 9.970681277619452e-06, 'learning_rate/action_model': 9.970681277619452e-05, 'epoch': 0.59})
+[2026-07-28 19:48:51,743][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.027205906808376312, 'train/grad_norm_pre_clip': 0.9150741696357727, 'timing/data': 0.0003484077751636505, 'timing/model': 0.5303514078259468, 'learning_rate/qwen_vl_interface': 9.97025527470192e-06, 'learning_rate/action_model': 9.970255274701919e-05, 'epoch': 0.59})
+[2026-07-28 19:48:58,965][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.023329032585024834, 'train/grad_norm_pre_clip': 0.8775104880332947, 'timing/data': 0.0005403310060501099, 'timing/model': 0.5316621400415897, 'learning_rate/qwen_vl_interface': 9.969826208861987e-06, 'learning_rate/action_model': 9.969826208861987e-05, 'epoch': 0.6})
+[2026-07-28 19:49:06,196][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.039593007415533066, 'train/grad_norm_pre_clip': 1.6613261699676514, 'timing/data': 0.00041889213025569916, 'timing/model': 0.5301687512546778, 'learning_rate/qwen_vl_interface': 9.96939408037807e-06, 'learning_rate/action_model': 9.96939408037807e-05, 'epoch': 0.6})
+[2026-07-28 19:49:13,428][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.024219712242484093, 'train/grad_norm_pre_clip': 1.0301165580749512, 'timing/data': 0.00041469745337963104, 'timing/model': 0.5330484993755817, 'learning_rate/qwen_vl_interface': 9.968958889530573e-06, 'learning_rate/action_model': 9.968958889530573e-05, 'epoch': 0.6})
+[2026-07-28 19:49:20,668][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.020543191581964493, 'train/grad_norm_pre_clip': 0.6140017509460449, 'timing/data': 0.00039000436663627625, 'timing/model': 0.5295219328254461, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.6})
+[2026-07-28 19:49:27,880][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.03400476276874542, 'train/grad_norm_pre_clip': 1.5914592742919922, 'timing/data': 0.0003933403640985489, 'timing/model': 0.53078493475914, 'learning_rate/qwen_vl_interface': 9.968079321876382e-06, 'learning_rate/action_model': 9.968079321876382e-05, 'epoch': 0.6})
+[2026-07-28 19:49:35,099][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.026595339179039, 'train/grad_norm_pre_clip': 0.7926523685455322, 'timing/data': 0.00032001733779907227, 'timing/model': 0.5314134191721678, 'learning_rate/qwen_vl_interface': 9.967634945640432e-06, 'learning_rate/action_model': 9.967634945640431e-05, 'epoch': 0.6})
+[2026-07-28 19:49:42,354][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.09776845574378967, 'train/grad_norm_pre_clip': 3.1034560203552246, 'timing/data': 0.00034196674823760986, 'timing/model': 0.543678404763341, 'learning_rate/qwen_vl_interface': 9.967187508182385e-06, 'learning_rate/action_model': 9.967187508182385e-05, 'epoch': 0.6})
+[2026-07-28 19:49:49,678][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.037668850272893906, 'train/grad_norm_pre_clip': 1.305371642112732, 'timing/data': 0.0006052553653717041, 'timing/model': 0.5480630565434694, 'learning_rate/qwen_vl_interface': 9.966737009792577e-06, 'learning_rate/action_model': 9.966737009792577e-05, 'epoch': 0.6})
+[2026-07-28 19:49:56,910][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.06870778650045395, 'train/grad_norm_pre_clip': 2.591012477874756, 'timing/data': 0.00038932450115680695, 'timing/model': 0.530586015433073, 'learning_rate/qwen_vl_interface': 9.966283450763333e-06, 'learning_rate/action_model': 9.966283450763333e-05, 'epoch': 0.6})
+[2026-07-28 19:50:04,133][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.09081493318080902, 'train/grad_norm_pre_clip': 2.990856170654297, 'timing/data': 0.0004153568297624588, 'timing/model': 0.5292966291308403, 'learning_rate/qwen_vl_interface': 9.965826831388962e-06, 'learning_rate/action_model': 9.965826831388961e-05, 'epoch': 0.6})
+[2026-07-28 19:50:11,399][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3250
+[2026-07-28 19:51:29,656][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3250
+[2026-07-28 20:08:11,548][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3250
+[2026-07-28 20:08:11,549][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.02633362077176571, 'train/grad_norm_pre_clip': 1.097044587135315, 'eval/action_loss/seconds': 78.25372889265418, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.04410348832607269, 'eval/latency_3/loss': 0.04410348832607269, 'eval/flappy/loss': 0.04410348832607269, 'eval/flappy/latency_3/loss': 0.04410348832607269, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 376.70500443093476, 'rl_games_eval/mid_train/mean_length': 3053.6, 'rl_games_eval/mid_train/std_reward': 120.41256086491829, 'rl_games_eval/mid_train/std_length': 969.9987835043918, 'rl_games_eval/mid_train/macro_mean_reward': 376.70500443093476, 'rl_games_eval/mid_train/macro_mean_length': 3053.6, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 376.70500443093476, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 3053.6, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 120.41256086491829, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 969.9987835043918, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 376.70500443093476, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003179963678121567, 'timing/model': 0.5505000781267881, 'learning_rate/qwen_vl_interface': 9.965367151965757e-06, 'learning_rate/action_model': 9.965367151965756e-05, 'epoch': 0.6})
+[2026-07-28 20:08:20,890][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.026350947096943855, 'train/grad_norm_pre_clip': 0.7394227981567383, 'timing/data': 0.00033569522202014923, 'timing/model': 0.6227922588586807, 'learning_rate/qwen_vl_interface': 9.964904412792002e-06, 'learning_rate/action_model': 9.964904412792002e-05, 'epoch': 0.6})
+[2026-07-28 20:08:30,289][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.010866045951843262, 'train/grad_norm_pre_clip': 0.5542160272598267, 'timing/data': 0.0003890041261911392, 'timing/model': 0.624745849519968, 'learning_rate/qwen_vl_interface': 9.964438614167964e-06, 'learning_rate/action_model': 9.964438614167964e-05, 'epoch': 0.6})
+[2026-07-28 20:08:39,765][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.03219124674797058, 'train/grad_norm_pre_clip': 1.34186851978302, 'timing/data': 0.00038761459290981293, 'timing/model': 0.6195922438055277, 'learning_rate/qwen_vl_interface': 9.963969756395892e-06, 'learning_rate/action_model': 9.96396975639589e-05, 'epoch': 0.6})
+[2026-07-28 20:08:48,961][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.06510940939188004, 'train/grad_norm_pre_clip': 2.211047649383545, 'timing/data': 0.00039894506335258484, 'timing/model': 0.6198311969637871, 'learning_rate/qwen_vl_interface': 9.963497839780025e-06, 'learning_rate/action_model': 9.963497839780024e-05, 'epoch': 0.6})
+[2026-07-28 20:08:57,865][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.022936711087822914, 'train/grad_norm_pre_clip': 1.1320263147354126, 'timing/data': 0.00047005899250507355, 'timing/model': 0.6193205323070288, 'learning_rate/qwen_vl_interface': 9.963022864626582e-06, 'learning_rate/action_model': 9.963022864626581e-05, 'epoch': 0.6})
+[2026-07-28 20:09:06,788][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.030268844217061996, 'train/grad_norm_pre_clip': 1.364858627319336, 'timing/data': 0.0003282073885202408, 'timing/model': 0.6281879525631666, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.6})
+[2026-07-28 20:09:16,173][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.05060863494873047, 'train/grad_norm_pre_clip': 1.9674547910690308, 'timing/data': 0.0004224870353937149, 'timing/model': 0.8430657368153334, 'learning_rate/qwen_vl_interface': 9.962063739941783e-06, 'learning_rate/action_model': 9.962063739941783e-05, 'epoch': 0.6})
+[2026-07-28 20:09:25,406][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.0539296492934227, 'train/grad_norm_pre_clip': 2.201070785522461, 'timing/data': 0.0005542635917663574, 'timing/model': 0.6530733332037926, 'learning_rate/qwen_vl_interface': 9.961579591032791e-06, 'learning_rate/action_model': 9.96157959103279e-05, 'epoch': 0.6})
+[2026-07-28 20:09:34,647][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.024013083428144455, 'train/grad_norm_pre_clip': 1.3870892524719238, 'timing/data': 0.00039111077785491943, 'timing/model': 0.626493263989687, 'learning_rate/qwen_vl_interface': 9.961092384830956e-06, 'learning_rate/action_model': 9.961092384830956e-05, 'epoch': 0.6})
+[2026-07-28 20:09:43,793][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.06177597865462303, 'train/grad_norm_pre_clip': 1.522038221359253, 'timing/data': 0.00042013823986053467, 'timing/model': 0.6212056335061789, 'learning_rate/qwen_vl_interface': 9.96060212165242e-06, 'learning_rate/action_model': 9.96060212165242e-05, 'epoch': 0.6})
+[2026-07-28 20:09:53,187][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.02141287550330162, 'train/grad_norm_pre_clip': 0.6068724989891052, 'timing/data': 0.00038333237171173096, 'timing/model': 0.6212233155965805, 'learning_rate/qwen_vl_interface': 9.96010880181531e-06, 'learning_rate/action_model': 9.960108801815309e-05, 'epoch': 0.6})
+[2026-07-28 20:10:02,299][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.026069611310958862, 'train/grad_norm_pre_clip': 0.7755225896835327, 'timing/data': 0.00037796609103679657, 'timing/model': 0.6194064822047949, 'learning_rate/qwen_vl_interface': 9.959612425639736e-06, 'learning_rate/action_model': 9.959612425639735e-05, 'epoch': 0.6})
+[2026-07-28 20:10:11,690][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.023422259837388992, 'train/grad_norm_pre_clip': 1.292561650276184, 'timing/data': 0.00037901662290096283, 'timing/model': 0.6197947412729263, 'learning_rate/qwen_vl_interface': 9.95911299344779e-06, 'learning_rate/action_model': 9.959112993447789e-05, 'epoch': 0.6})
+[2026-07-28 20:10:20,807][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.028003808110952377, 'train/grad_norm_pre_clip': 1.2139496803283691, 'timing/data': 0.00040096603333950043, 'timing/model': 0.619988588616252, 'learning_rate/qwen_vl_interface': 9.958610505563546e-06, 'learning_rate/action_model': 9.958610505563546e-05, 'epoch': 0.6})
+[2026-07-28 20:10:30,160][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.03827318176627159, 'train/grad_norm_pre_clip': 1.2321034669876099, 'timing/data': 0.0003695916384458542, 'timing/model': 0.6201159376651049, 'learning_rate/qwen_vl_interface': 9.95810496231307e-06, 'learning_rate/action_model': 9.95810496231307e-05, 'epoch': 0.6})
+[2026-07-28 20:10:39,286][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.008944513276219368, 'train/grad_norm_pre_clip': 0.5912946462631226, 'timing/data': 0.0006659906357526779, 'timing/model': 0.6226984336972237, 'learning_rate/qwen_vl_interface': 9.957596364024396e-06, 'learning_rate/action_model': 9.957596364024396e-05, 'epoch': 0.6})
+[2026-07-28 20:10:48,680][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.024361005052924156, 'train/grad_norm_pre_clip': 0.921567976474762, 'timing/data': 0.0005480833351612091, 'timing/model': 0.8393639698624611, 'learning_rate/qwen_vl_interface': 9.957084711027553e-06, 'learning_rate/action_model': 9.957084711027552e-05, 'epoch': 0.6})
+[2026-07-28 20:10:58,023][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.023628629744052887, 'train/grad_norm_pre_clip': 0.7193193435668945, 'timing/data': 0.0004025697708129883, 'timing/model': 0.8038716372102499, 'learning_rate/qwen_vl_interface': 9.956570003654544e-06, 'learning_rate/action_model': 9.956570003654544e-05, 'epoch': 0.6})
+[2026-07-28 20:11:07,149][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.060966186225414276, 'train/grad_norm_pre_clip': 1.7502082586288452, 'timing/data': 0.0003659147769212723, 'timing/model': 0.6202738601714373, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.6})
+[2026-07-28 20:11:16,263][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.024480853229761124, 'train/grad_norm_pre_clip': 1.327017068862915, 'timing/data': 0.0004353616386651993, 'timing/model': 0.6216180697083473, 'learning_rate/qwen_vl_interface': 9.955531427117964e-06, 'learning_rate/action_model': 9.955531427117964e-05, 'epoch': 0.6})
+[2026-07-28 20:11:25,624][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.009997420012950897, 'train/grad_norm_pre_clip': 0.9568922519683838, 'timing/data': 0.00032316334545612335, 'timing/model': 0.6208725422620773, 'learning_rate/qwen_vl_interface': 9.955007558628316e-06, 'learning_rate/action_model': 9.955007558628316e-05, 'epoch': 0.6})
+[2026-07-28 20:11:34,998][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.03313314542174339, 'train/grad_norm_pre_clip': 1.5825895071029663, 'timing/data': 0.00037047266960144043, 'timing/model': 0.8613732885569334, 'learning_rate/qwen_vl_interface': 9.954480637110346e-06, 'learning_rate/action_model': 9.954480637110344e-05, 'epoch': 0.6})
+[2026-07-28 20:11:44,126][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.0943048745393753, 'train/grad_norm_pre_clip': 8.150837898254395, 'timing/data': 0.00036516226828098297, 'timing/model': 0.6197082195430994, 'learning_rate/qwen_vl_interface': 9.953950662905963e-06, 'learning_rate/action_model': 9.953950662905964e-05, 'epoch': 0.6})
+[2026-07-28 20:11:53,503][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.02148617058992386, 'train/grad_norm_pre_clip': 0.9383730292320251, 'timing/data': 0.00041542015969753265, 'timing/model': 0.8418649639934301, 'learning_rate/qwen_vl_interface': 9.953417636359067e-06, 'learning_rate/action_model': 9.953417636359067e-05, 'epoch': 0.6})
+[2026-07-28 20:12:02,692][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.013696854002773762, 'train/grad_norm_pre_clip': 0.7500405311584473, 'timing/data': 0.00034167803823947906, 'timing/model': 0.6199919916689396, 'learning_rate/qwen_vl_interface': 9.95288155781553e-06, 'learning_rate/action_model': 9.95288155781553e-05, 'epoch': 0.6})
+[2026-07-28 20:12:12,129][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.050197575241327286, 'train/grad_norm_pre_clip': 1.687972903251648, 'timing/data': 0.000361759215593338, 'timing/model': 0.6198568847030401, 'learning_rate/qwen_vl_interface': 9.952342427623209e-06, 'learning_rate/action_model': 9.952342427623208e-05, 'epoch': 0.6})
+[2026-07-28 20:12:21,251][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.0399266816675663, 'train/grad_norm_pre_clip': 1.1672618389129639, 'timing/data': 0.00032357871532440186, 'timing/model': 0.6204769480973482, 'learning_rate/qwen_vl_interface': 9.95180024613194e-06, 'learning_rate/action_model': 9.951800246131939e-05, 'epoch': 0.6})
+[2026-07-28 20:12:30,392][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.05489800497889519, 'train/grad_norm_pre_clip': 1.9906641244888306, 'timing/data': 0.000339386984705925, 'timing/model': 0.6197445224970579, 'learning_rate/qwen_vl_interface': 9.951255013693537e-06, 'learning_rate/action_model': 9.951255013693536e-05, 'epoch': 0.6})
+[2026-07-28 20:12:39,761][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.022587886080145836, 'train/grad_norm_pre_clip': 1.2089020013809204, 'timing/data': 0.0003440678119659424, 'timing/model': 0.6245326362550259, 'learning_rate/qwen_vl_interface': 9.950706730661797e-06, 'learning_rate/action_model': 9.950706730661797e-05, 'epoch': 0.6})
+[2026-07-28 20:12:49,125][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.00940365344285965, 'train/grad_norm_pre_clip': 1.0044307708740234, 'timing/data': 0.0003406573086977005, 'timing/model': 0.6195566952228546, 'learning_rate/qwen_vl_interface': 9.950155397392494e-06, 'learning_rate/action_model': 9.950155397392493e-05, 'epoch': 0.6})
+[2026-07-28 20:12:58,016][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.03251509368419647, 'train/grad_norm_pre_clip': 2.098752737045288, 'timing/data': 0.0003309529274702072, 'timing/model': 0.6205138750374317, 'learning_rate/qwen_vl_interface': 9.949601014243382e-06, 'learning_rate/action_model': 9.949601014243382e-05, 'epoch': 0.6})
+[2026-07-28 20:13:06,952][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.04538078233599663, 'train/grad_norm_pre_clip': 1.9577633142471313, 'timing/data': 0.0003518238663673401, 'timing/model': 0.6213422603905201, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.6})
+[2026-07-28 20:13:15,917][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.0678899884223938, 'train/grad_norm_pre_clip': 1.467301368713379, 'timing/data': 0.0004742685705423355, 'timing/model': 0.6224608272314072, 'learning_rate/qwen_vl_interface': 9.948483099746645e-06, 'learning_rate/action_model': 9.948483099746645e-05, 'epoch': 0.6})
+[2026-07-28 20:13:25,130][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.06368669867515564, 'train/grad_norm_pre_clip': 1.4331510066986084, 'timing/data': 0.0003374628722667694, 'timing/model': 0.8666874077171087, 'learning_rate/qwen_vl_interface': 9.947919569124423e-06, 'learning_rate/action_model': 9.947919569124422e-05, 'epoch': 0.6})
+[2026-07-28 20:13:34,532][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.049345165491104126, 'train/grad_norm_pre_clip': 1.4778224229812622, 'timing/data': 0.0003416668623685837, 'timing/model': 0.620023412629962, 'learning_rate/qwen_vl_interface': 9.947352990073196e-06, 'learning_rate/action_model': 9.947352990073195e-05, 'epoch': 0.6})
+[2026-07-28 20:13:43,688][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.02913348749279976, 'train/grad_norm_pre_clip': 0.8767344355583191, 'timing/data': 0.0003432314842939377, 'timing/model': 0.6213771104812622, 'learning_rate/qwen_vl_interface': 9.946783362960612e-06, 'learning_rate/action_model': 9.946783362960612e-05, 'epoch': 0.6})
+[2026-07-28 20:13:52,894][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.01995140314102173, 'train/grad_norm_pre_clip': 1.264196515083313, 'timing/data': 0.0003411862999200821, 'timing/model': 0.6211515609174967, 'learning_rate/qwen_vl_interface': 9.946210688156297e-06, 'learning_rate/action_model': 9.946210688156297e-05, 'epoch': 0.6})
+[2026-07-28 20:14:02,157][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.0525374598801136, 'train/grad_norm_pre_clip': 1.8239779472351074, 'timing/data': 0.00032638199627399445, 'timing/model': 0.6194290146231651, 'learning_rate/qwen_vl_interface': 9.945634966031852e-06, 'learning_rate/action_model': 9.945634966031851e-05, 'epoch': 0.6})
+[2026-07-28 20:14:11,607][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.05650123581290245, 'train/grad_norm_pre_clip': 1.6304457187652588, 'timing/data': 0.0003231428563594818, 'timing/model': 0.6222415398806334, 'learning_rate/qwen_vl_interface': 9.945056196960858e-06, 'learning_rate/action_model': 9.945056196960858e-05, 'epoch': 0.6})
+[2026-07-28 20:14:21,071][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.03066958300769329, 'train/grad_norm_pre_clip': 1.4315574169158936, 'timing/data': 0.0003571901470422745, 'timing/model': 0.6191389262676239, 'learning_rate/qwen_vl_interface': 9.944474381318871e-06, 'learning_rate/action_model': 9.94447438131887e-05, 'epoch': 0.6})
+[2026-07-28 20:14:30,232][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.05087605118751526, 'train/grad_norm_pre_clip': 2.3069076538085938, 'timing/data': 0.00034120865166187286, 'timing/model': 0.62129120901227, 'learning_rate/qwen_vl_interface': 9.943889519483424e-06, 'learning_rate/action_model': 9.943889519483424e-05, 'epoch': 0.6})
+[2026-07-28 20:14:39,709][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.04222578555345535, 'train/grad_norm_pre_clip': 0.9103832244873047, 'timing/data': 0.00047070905566215515, 'timing/model': 0.623061042279005, 'learning_rate/qwen_vl_interface': 9.94330161183403e-06, 'learning_rate/action_model': 9.94330161183403e-05, 'epoch': 0.6})
+[2026-07-28 20:14:49,128][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.042289577424526215, 'train/grad_norm_pre_clip': 0.9061720967292786, 'timing/data': 0.0003394223749637604, 'timing/model': 0.8607572130858898, 'learning_rate/qwen_vl_interface': 9.942710658752175e-06, 'learning_rate/action_model': 9.942710658752174e-05, 'epoch': 0.6})
+[2026-07-28 20:14:58,363][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.014080598950386047, 'train/grad_norm_pre_clip': 0.65218585729599, 'timing/data': 0.0003282669931650162, 'timing/model': 0.6204357240349054, 'learning_rate/qwen_vl_interface': 9.942116660621319e-06, 'learning_rate/action_model': 9.94211666062132e-05, 'epoch': 0.61})
+[2026-07-28 20:15:07,510][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.015090509317815304, 'train/grad_norm_pre_clip': 0.5198051929473877, 'timing/data': 0.00035966746509075165, 'timing/model': 0.6205317918211222, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.61})
+[2026-07-28 20:15:16,915][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.04040958732366562, 'train/grad_norm_pre_clip': 1.4525057077407837, 'timing/data': 0.0003879964351654053, 'timing/model': 0.6198632270097733, 'learning_rate/qwen_vl_interface': 9.940919530756347e-06, 'learning_rate/action_model': 9.940919530756347e-05, 'epoch': 0.61})
+[2026-07-28 20:15:26,060][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.021202389150857925, 'train/grad_norm_pre_clip': 1.0537158250808716, 'timing/data': 0.00033976510167121887, 'timing/model': 0.6209263354539871, 'learning_rate/qwen_vl_interface': 9.940316399799032e-06, 'learning_rate/action_model': 9.940316399799033e-05, 'epoch': 0.61})
+[2026-07-28 20:15:35,458][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.017251580953598022, 'train/grad_norm_pre_clip': 0.6521386504173279, 'timing/data': 0.00042623840272426605, 'timing/model': 0.8562575485557318, 'learning_rate/qwen_vl_interface': 9.939710225346329e-06, 'learning_rate/action_model': 9.939710225346329e-05, 'epoch': 0.61})
+[2026-07-28 20:15:44,650][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.059990569949150085, 'train/grad_norm_pre_clip': 2.3440465927124023, 'timing/data': 0.0004208255559206009, 'timing/model': 0.6199618075042963, 'learning_rate/qwen_vl_interface': 9.939101007791575e-06, 'learning_rate/action_model': 9.939101007791575e-05, 'epoch': 0.61})
+[2026-07-28 20:15:54,052][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.013314371928572655, 'train/grad_norm_pre_clip': 0.7460259199142456, 'timing/data': 0.00032611191272735596, 'timing/model': 0.6195656564086676, 'learning_rate/qwen_vl_interface': 9.938488747530087e-06, 'learning_rate/action_model': 9.938488747530086e-05, 'epoch': 0.61})
+[2026-07-28 20:16:03,094][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.02428043819963932, 'train/grad_norm_pre_clip': 0.9031009674072266, 'timing/data': 0.0004587750881910324, 'timing/model': 0.620672820135951, 'learning_rate/qwen_vl_interface': 9.937873444959151e-06, 'learning_rate/action_model': 9.937873444959151e-05, 'epoch': 0.61})
+[2026-07-28 20:16:12,288][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.027523325756192207, 'train/grad_norm_pre_clip': 1.3013490438461304, 'timing/data': 0.0003468915820121765, 'timing/model': 0.6191646363586187, 'learning_rate/qwen_vl_interface': 9.937255100478036e-06, 'learning_rate/action_model': 9.937255100478036e-05, 'epoch': 0.61})
+[2026-07-28 20:16:21,413][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.08439093828201294, 'train/grad_norm_pre_clip': 3.793560266494751, 'timing/data': 0.0003431476652622223, 'timing/model': 0.6195024661719799, 'learning_rate/qwen_vl_interface': 9.93663371448797e-06, 'learning_rate/action_model': 9.93663371448797e-05, 'epoch': 0.61})
+[2026-07-28 20:16:30,724][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.050190165638923645, 'train/grad_norm_pre_clip': 2.015329122543335, 'timing/data': 0.00039608776569366455, 'timing/model': 0.839042292907834, 'learning_rate/qwen_vl_interface': 9.936009287392173e-06, 'learning_rate/action_model': 9.936009287392171e-05, 'epoch': 0.61})
+[2026-07-28 20:16:40,089][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.01176008116453886, 'train/grad_norm_pre_clip': 0.8106555342674255, 'timing/data': 0.00031767599284648895, 'timing/model': 0.6226633787155151, 'learning_rate/qwen_vl_interface': 9.935381819595822e-06, 'learning_rate/action_model': 9.935381819595822e-05, 'epoch': 0.61})
+[2026-07-28 20:16:49,307][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.017078792676329613, 'train/grad_norm_pre_clip': 1.0409115552902222, 'timing/data': 0.000365307554602623, 'timing/model': 0.6202441491186619, 'learning_rate/qwen_vl_interface': 9.934751311506078e-06, 'learning_rate/action_model': 9.934751311506077e-05, 'epoch': 0.61})
+[2026-07-28 20:16:58,480][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.06553065776824951, 'train/grad_norm_pre_clip': 1.931384801864624, 'timing/data': 0.0004031732678413391, 'timing/model': 0.6205735951662064, 'learning_rate/qwen_vl_interface': 9.93411776353207e-06, 'learning_rate/action_model': 9.934117763532069e-05, 'epoch': 0.61})
+[2026-07-28 20:17:07,658][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.01738952100276947, 'train/grad_norm_pre_clip': 0.6890851259231567, 'timing/data': 0.00040765106678009033, 'timing/model': 0.620669849216938, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.61})
+[2026-07-28 20:17:16,867][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.01597478613257408, 'train/grad_norm_pre_clip': 1.3543531894683838, 'timing/data': 0.00033864937722682953, 'timing/model': 0.6209566723555326, 'learning_rate/qwen_vl_interface': 9.932841549577642e-06, 'learning_rate/action_model': 9.932841549577641e-05, 'epoch': 0.61})
+[2026-07-28 20:17:26,265][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.029511703178286552, 'train/grad_norm_pre_clip': 1.5873568058013916, 'timing/data': 0.0003431178629398346, 'timing/model': 0.6183995027095079, 'learning_rate/qwen_vl_interface': 9.932198884425343e-06, 'learning_rate/action_model': 9.932198884425342e-05, 'epoch': 0.61})
+[2026-07-28 20:17:35,650][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.015787210315465927, 'train/grad_norm_pre_clip': 1.0533076524734497, 'timing/data': 0.0003437027335166931, 'timing/model': 0.6206490527838469, 'learning_rate/qwen_vl_interface': 9.931553181045022e-06, 'learning_rate/action_model': 9.931553181045022e-05, 'epoch': 0.61})
+[2026-07-28 20:17:44,846][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.018640510737895966, 'train/grad_norm_pre_clip': 1.2089773416519165, 'timing/data': 0.0003826506435871124, 'timing/model': 0.6409444604068995, 'learning_rate/qwen_vl_interface': 9.930904439855667e-06, 'learning_rate/action_model': 9.930904439855667e-05, 'epoch': 0.61})
+[2026-07-28 20:17:54,016][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.051771149039268494, 'train/grad_norm_pre_clip': 2.3541431427001953, 'timing/data': 0.0003536343574523926, 'timing/model': 0.6203502882272005, 'learning_rate/qwen_vl_interface': 9.930252661278245e-06, 'learning_rate/action_model': 9.930252661278244e-05, 'epoch': 0.61})
+[2026-07-28 20:18:03,387][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.012793586589396, 'train/grad_norm_pre_clip': 0.6985288858413696, 'timing/data': 0.0003981012850999832, 'timing/model': 0.849053705111146, 'learning_rate/qwen_vl_interface': 9.92959784573568e-06, 'learning_rate/action_model': 9.92959784573568e-05, 'epoch': 0.61})
+[2026-07-28 20:18:12,566][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.03780023381114006, 'train/grad_norm_pre_clip': 0.8140566349029541, 'timing/data': 0.0003694649785757065, 'timing/model': 0.6206050049513578, 'learning_rate/qwen_vl_interface': 9.92893999365288e-06, 'learning_rate/action_model': 9.92893999365288e-05, 'epoch': 0.61})
+[2026-07-28 20:18:21,508][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.01338009350001812, 'train/grad_norm_pre_clip': 0.9856293797492981, 'timing/data': 0.00033378228545188904, 'timing/model': 0.6206480115652084, 'learning_rate/qwen_vl_interface': 9.928279105456717e-06, 'learning_rate/action_model': 9.928279105456718e-05, 'epoch': 0.61})
+[2026-07-28 20:18:30,640][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.018589504063129425, 'train/grad_norm_pre_clip': 1.033560872077942, 'timing/data': 0.0003197360783815384, 'timing/model': 0.6199963130056858, 'learning_rate/qwen_vl_interface': 9.927615181576034e-06, 'learning_rate/action_model': 9.927615181576034e-05, 'epoch': 0.61})
+[2026-07-28 20:18:40,009][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.044898636639118195, 'train/grad_norm_pre_clip': 1.338326334953308, 'timing/data': 0.00032059289515018463, 'timing/model': 0.6208312474191189, 'learning_rate/qwen_vl_interface': 9.926948222441644e-06, 'learning_rate/action_model': 9.926948222441643e-05, 'epoch': 0.61})
+[2026-07-28 20:18:49,381][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.009578866884112358, 'train/grad_norm_pre_clip': 0.5731133222579956, 'timing/data': 0.00034610554575920105, 'timing/model': 0.8556443676352501, 'learning_rate/qwen_vl_interface': 9.926278228486328e-06, 'learning_rate/action_model': 9.926278228486327e-05, 'epoch': 0.61})
+[2026-07-28 20:18:58,546][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.04886315017938614, 'train/grad_norm_pre_clip': 1.5098971128463745, 'timing/data': 0.00034816376864910126, 'timing/model': 0.6215360127389431, 'learning_rate/qwen_vl_interface': 9.92560520014484e-06, 'learning_rate/action_model': 9.92560520014484e-05, 'epoch': 0.61})
+[2026-07-28 20:19:07,934][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.03520497679710388, 'train/grad_norm_pre_clip': 1.4003193378448486, 'timing/data': 0.00037273019552230835, 'timing/model': 0.8534410446882248, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.61})
+[2026-07-28 20:19:17,093][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.033805057406425476, 'train/grad_norm_pre_clip': 1.167694330215454, 'timing/data': 0.0003455188125371933, 'timing/model': 0.6204870771616697, 'learning_rate/qwen_vl_interface': 9.924250042052197e-06, 'learning_rate/action_model': 9.924250042052197e-05, 'epoch': 0.61})
+[2026-07-28 20:19:26,018][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.04483278840780258, 'train/grad_norm_pre_clip': 1.3490023612976074, 'timing/data': 0.00036115385591983795, 'timing/model': 0.6218824181705713, 'learning_rate/qwen_vl_interface': 9.92356791318039e-06, 'learning_rate/action_model': 9.92356791318039e-05, 'epoch': 0.61})
+[2026-07-28 20:19:34,974][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.04955582320690155, 'train/grad_norm_pre_clip': 0.9789308905601501, 'timing/data': 0.0004340875893831253, 'timing/model': 0.6203155182301998, 'learning_rate/qwen_vl_interface': 9.922882751681104e-06, 'learning_rate/action_model': 9.922882751681104e-05, 'epoch': 0.61})
+[2026-07-28 20:19:43,933][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.03316890820860863, 'train/grad_norm_pre_clip': 1.048156976699829, 'timing/data': 0.000327499583363533, 'timing/model': 0.6197175569832325, 'learning_rate/qwen_vl_interface': 9.922194557998932e-06, 'learning_rate/action_model': 9.922194557998933e-05, 'epoch': 0.61})
+[2026-07-28 20:19:53,109][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.025840196758508682, 'train/grad_norm_pre_clip': 0.8726410269737244, 'timing/data': 0.0003467593342065811, 'timing/model': 0.6221541184931993, 'learning_rate/qwen_vl_interface': 9.921503332580439e-06, 'learning_rate/action_model': 9.921503332580439e-05, 'epoch': 0.61})
+[2026-07-28 20:20:02,080][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.042456403374671936, 'train/grad_norm_pre_clip': 1.43844735622406, 'timing/data': 0.00039475783705711365, 'timing/model': 0.6216972023248672, 'learning_rate/qwen_vl_interface': 9.92080907587415e-06, 'learning_rate/action_model': 9.92080907587415e-05, 'epoch': 0.61})
+[2026-07-28 20:20:11,299][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.033391159027814865, 'train/grad_norm_pre_clip': 0.8905609250068665, 'timing/data': 0.0005546212196350098, 'timing/model': 0.6196381635963917, 'learning_rate/qwen_vl_interface': 9.92011178833056e-06, 'learning_rate/action_model': 9.92011178833056e-05, 'epoch': 0.61})
+[2026-07-28 20:20:20,493][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.04812300577759743, 'train/grad_norm_pre_clip': 1.1024343967437744, 'timing/data': 0.00035573355853557587, 'timing/model': 0.6202703695744276, 'learning_rate/qwen_vl_interface': 9.919411470402136e-06, 'learning_rate/action_model': 9.919411470402136e-05, 'epoch': 0.61})
+[2026-07-28 20:20:29,642][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.03340558707714081, 'train/grad_norm_pre_clip': 0.8756622076034546, 'timing/data': 0.00040246546268463135, 'timing/model': 0.621658805757761, 'learning_rate/qwen_vl_interface': 9.918708122543305e-06, 'learning_rate/action_model': 9.918708122543305e-05, 'epoch': 0.61})
+[2026-07-28 20:20:39,074][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.0392114594578743, 'train/grad_norm_pre_clip': 1.7057383060455322, 'timing/data': 0.00039015524089336395, 'timing/model': 0.8452371619641781, 'learning_rate/qwen_vl_interface': 9.918001745210458e-06, 'learning_rate/action_model': 9.918001745210456e-05, 'epoch': 0.61})
+[2026-07-28 20:20:48,530][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.0811317041516304, 'train/grad_norm_pre_clip': 2.0805835723876953, 'timing/data': 0.00039874203503131866, 'timing/model': 0.6219330094754696, 'learning_rate/qwen_vl_interface': 9.917292338861958e-06, 'learning_rate/action_model': 9.917292338861957e-05, 'epoch': 0.61})
+[2026-07-28 20:20:57,728][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.03786896914243698, 'train/grad_norm_pre_clip': 1.0523438453674316, 'timing/data': 0.0005509499460458755, 'timing/model': 0.6199862621724606, 'learning_rate/qwen_vl_interface': 9.916579903958134e-06, 'learning_rate/action_model': 9.916579903958133e-05, 'epoch': 0.61})
+[2026-07-28 20:21:06,917][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.034154098480939865, 'train/grad_norm_pre_clip': 0.8116636276245117, 'timing/data': 0.0003986787050962448, 'timing/model': 0.6201938576996326, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.61})
+[2026-07-28 20:21:16,032][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.04000966250896454, 'train/grad_norm_pre_clip': 0.910168468952179, 'timing/data': 0.00039848871529102325, 'timing/model': 0.6198639068752527, 'learning_rate/qwen_vl_interface': 9.915145950335631e-06, 'learning_rate/action_model': 9.915145950335632e-05, 'epoch': 0.61})
+[2026-07-28 20:21:25,484][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.0346260629594326, 'train/grad_norm_pre_clip': 0.7208448648452759, 'timing/data': 0.0003470294177532196, 'timing/model': 0.6199235767126083, 'learning_rate/qwen_vl_interface': 9.914424432547434e-06, 'learning_rate/action_model': 9.914424432547433e-05, 'epoch': 0.61})
+[2026-07-28 20:21:35,048][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.03295617550611496, 'train/grad_norm_pre_clip': 1.1989939212799072, 'timing/data': 0.0004721079021692276, 'timing/model': 0.6203174702823162, 'learning_rate/qwen_vl_interface': 9.913699888064862e-06, 'learning_rate/action_model': 9.913699888064861e-05, 'epoch': 0.61})
+[2026-07-28 20:21:44,393][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.04255705326795578, 'train/grad_norm_pre_clip': 1.1780000925064087, 'timing/data': 0.0003565549850463867, 'timing/model': 0.6211078967899084, 'learning_rate/qwen_vl_interface': 9.912972317358066e-06, 'learning_rate/action_model': 9.912972317358066e-05, 'epoch': 0.61})
+[2026-07-28 20:21:53,899][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.035433024168014526, 'train/grad_norm_pre_clip': 1.11983060836792, 'timing/data': 0.00041811540722846985, 'timing/model': 0.6244392972439528, 'learning_rate/qwen_vl_interface': 9.912241720899157e-06, 'learning_rate/action_model': 9.912241720899157e-05, 'epoch': 0.61})
+[2026-07-28 20:22:03,135][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.02315542846918106, 'train/grad_norm_pre_clip': 1.1742116212844849, 'timing/data': 0.00040469877421855927, 'timing/model': 0.6208246797323227, 'learning_rate/qwen_vl_interface': 9.911508099162215e-06, 'learning_rate/action_model': 9.911508099162214e-05, 'epoch': 0.61})
+[2026-07-28 20:22:12,597][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.042946211993694305, 'train/grad_norm_pre_clip': 1.587602972984314, 'timing/data': 0.00035593099892139435, 'timing/model': 0.6200910620391369, 'learning_rate/qwen_vl_interface': 9.910771452623276e-06, 'learning_rate/action_model': 9.910771452623275e-05, 'epoch': 0.61})
+[2026-07-28 20:22:21,793][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.05735429376363754, 'train/grad_norm_pre_clip': 1.6892715692520142, 'timing/data': 0.000453755259513855, 'timing/model': 0.6194704473018646, 'learning_rate/qwen_vl_interface': 9.910031781760342e-06, 'learning_rate/action_model': 9.910031781760342e-05, 'epoch': 0.61})
+[2026-07-28 20:22:31,190][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.0328981876373291, 'train/grad_norm_pre_clip': 1.0183252096176147, 'timing/data': 0.00038534030318260193, 'timing/model': 0.6200700048357248, 'learning_rate/qwen_vl_interface': 9.909289087053378e-06, 'learning_rate/action_model': 9.909289087053378e-05, 'epoch': 0.61})
+[2026-07-28 20:22:40,368][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.0402766577899456, 'train/grad_norm_pre_clip': 1.6655901670455933, 'timing/data': 0.00040606968104839325, 'timing/model': 0.6196777708828449, 'learning_rate/qwen_vl_interface': 9.90854336898431e-06, 'learning_rate/action_model': 9.90854336898431e-05, 'epoch': 0.61})
+[2026-07-28 20:22:49,499][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.027671903371810913, 'train/grad_norm_pre_clip': 1.1514031887054443, 'timing/data': 0.00048824213445186615, 'timing/model': 0.6198487542569637, 'learning_rate/qwen_vl_interface': 9.907794628037028e-06, 'learning_rate/action_model': 9.907794628037028e-05, 'epoch': 0.61})
+[2026-07-28 20:22:58,872][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.026824206113815308, 'train/grad_norm_pre_clip': 1.0888490676879883, 'timing/data': 0.0003903694450855255, 'timing/model': 0.6254679672420025, 'learning_rate/qwen_vl_interface': 9.90704286469738e-06, 'learning_rate/action_model': 9.90704286469738e-05, 'epoch': 0.61})
+[2026-07-28 20:23:08,266][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.04181612655520439, 'train/grad_norm_pre_clip': 1.2229094505310059, 'timing/data': 0.0005720760673284531, 'timing/model': 0.6198664717376232, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.61})
+[2026-07-28 20:23:17,513][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.054504767060279846, 'train/grad_norm_pre_clip': 1.4426250457763672, 'timing/data': 0.00039241276681423187, 'timing/model': 0.6217014789581299, 'learning_rate/qwen_vl_interface': 9.905530272794195e-06, 'learning_rate/action_model': 9.905530272794195e-05, 'epoch': 0.61})
+[2026-07-28 20:23:26,740][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.0185028538107872, 'train/grad_norm_pre_clip': 0.8209801912307739, 'timing/data': 0.0003965906798839569, 'timing/model': 0.6201159618794918, 'learning_rate/qwen_vl_interface': 9.904769445212163e-06, 'learning_rate/action_model': 9.904769445212162e-05, 'epoch': 0.62})
+[2026-07-28 20:23:35,886][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.016678206622600555, 'train/grad_norm_pre_clip': 0.6924426555633545, 'timing/data': 0.00040745921432971954, 'timing/model': 0.6202418729662895, 'learning_rate/qwen_vl_interface': 9.904005597200772e-06, 'learning_rate/action_model': 9.904005597200772e-05, 'epoch': 0.62})
+[2026-07-28 20:23:45,275][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.02532540261745453, 'train/grad_norm_pre_clip': 1.5238124132156372, 'timing/data': 0.00039546750485897064, 'timing/model': 0.6185008306056261, 'learning_rate/qwen_vl_interface': 9.903238729255678e-06, 'learning_rate/action_model': 9.903238729255677e-05, 'epoch': 0.62})
+[2026-07-28 20:23:54,717][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.022874556481838226, 'train/grad_norm_pre_clip': 0.8935849666595459, 'timing/data': 0.0003137011080980301, 'timing/model': 0.8502804525196552, 'learning_rate/qwen_vl_interface': 9.902468841874492e-06, 'learning_rate/action_model': 9.902468841874492e-05, 'epoch': 0.62})
+[2026-07-28 20:24:03,886][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.04764558747410774, 'train/grad_norm_pre_clip': 1.143343210220337, 'timing/data': 0.000532960519194603, 'timing/model': 0.6219621375203133, 'learning_rate/qwen_vl_interface': 9.901695935556784e-06, 'learning_rate/action_model': 9.901695935556783e-05, 'epoch': 0.62})
+[2026-07-28 20:24:13,510][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.02340148389339447, 'train/grad_norm_pre_clip': 1.4173033237457275, 'timing/data': 0.0003621745854616165, 'timing/model': 0.8516666237264872, 'learning_rate/qwen_vl_interface': 9.900920010804088e-06, 'learning_rate/action_model': 9.900920010804088e-05, 'epoch': 0.62})
+[2026-07-28 20:24:22,809][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.051341354846954346, 'train/grad_norm_pre_clip': 1.1992833614349365, 'timing/data': 0.0005153976380825043, 'timing/model': 0.6214958820492029, 'learning_rate/qwen_vl_interface': 9.900141068119891e-06, 'learning_rate/action_model': 9.900141068119891e-05, 'epoch': 0.62})
+[2026-07-28 20:24:32,026][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.0063499752432107925, 'train/grad_norm_pre_clip': 0.5427053570747375, 'timing/data': 0.00040375255048274994, 'timing/model': 0.6195373348891735, 'learning_rate/qwen_vl_interface': 9.89935910800964e-06, 'learning_rate/action_model': 9.89935910800964e-05, 'epoch': 0.62})
+[2026-07-28 20:24:41,204][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.036815568804740906, 'train/grad_norm_pre_clip': 1.5793644189834595, 'timing/data': 0.00040284544229507446, 'timing/model': 0.6191507931798697, 'learning_rate/qwen_vl_interface': 9.898574130980743e-06, 'learning_rate/action_model': 9.898574130980742e-05, 'epoch': 0.62})
+[2026-07-28 20:24:50,614][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.02035660669207573, 'train/grad_norm_pre_clip': 1.0117462873458862, 'timing/data': 0.00041962601244449615, 'timing/model': 0.6221502609550953, 'learning_rate/qwen_vl_interface': 9.89778613754256e-06, 'learning_rate/action_model': 9.89778613754256e-05, 'epoch': 0.62})
+[2026-07-28 20:24:59,998][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.009861395694315434, 'train/grad_norm_pre_clip': 0.5499008297920227, 'timing/data': 0.00039809197187423706, 'timing/model': 0.8534799329936504, 'learning_rate/qwen_vl_interface': 9.896995128206412e-06, 'learning_rate/action_model': 9.896995128206411e-05, 'epoch': 0.62})
+[2026-07-28 20:25:09,181][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.04403969645500183, 'train/grad_norm_pre_clip': 1.5558441877365112, 'timing/data': 0.0004011504352092743, 'timing/model': 0.6192509792745113, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.62})
+[2026-07-28 20:25:18,396][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.02909363992512226, 'train/grad_norm_pre_clip': 1.1763789653778076, 'timing/data': 0.0004305243492126465, 'timing/model': 0.6198775377124548, 'learning_rate/qwen_vl_interface': 9.89540406389529e-06, 'learning_rate/action_model': 9.89540406389529e-05, 'epoch': 0.62})
+[2026-07-28 20:25:27,840][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.053804680705070496, 'train/grad_norm_pre_clip': 1.9007415771484375, 'timing/data': 0.0004031118005514145, 'timing/model': 0.9032922126352787, 'learning_rate/qwen_vl_interface': 9.894604009952742e-06, 'learning_rate/action_model': 9.89460400995274e-05, 'epoch': 0.62})
+[2026-07-28 20:25:37,004][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.01231150422245264, 'train/grad_norm_pre_clip': 0.5832743644714355, 'timing/data': 0.0004585571587085724, 'timing/model': 0.6202264875173569, 'learning_rate/qwen_vl_interface': 9.893800942177073e-06, 'learning_rate/action_model': 9.893800942177073e-05, 'epoch': 0.62})
+[2026-07-28 20:25:46,159][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.047030478715896606, 'train/grad_norm_pre_clip': 1.2950490713119507, 'timing/data': 0.0005598552525043488, 'timing/model': 0.6202677860856056, 'learning_rate/qwen_vl_interface': 9.892994861089393e-06, 'learning_rate/action_model': 9.892994861089393e-05, 'epoch': 0.62})
+[2026-07-28 20:25:55,325][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.02469383180141449, 'train/grad_norm_pre_clip': 0.8608254790306091, 'timing/data': 0.00039205141365528107, 'timing/model': 0.6202329806983471, 'learning_rate/qwen_vl_interface': 9.892185767212756e-06, 'learning_rate/action_model': 9.892185767212756e-05, 'epoch': 0.62})
+[2026-07-28 20:26:04,734][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.022605054080486298, 'train/grad_norm_pre_clip': 1.3160396814346313, 'timing/data': 0.00034836120903491974, 'timing/model': 0.619241451844573, 'learning_rate/qwen_vl_interface': 9.891373661072172e-06, 'learning_rate/action_model': 9.891373661072173e-05, 'epoch': 0.62})
+[2026-07-28 20:26:14,122][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.020251182839274406, 'train/grad_norm_pre_clip': 1.6944364309310913, 'timing/data': 0.00039952993392944336, 'timing/model': 0.6227176431566477, 'learning_rate/qwen_vl_interface': 9.89055854319461e-06, 'learning_rate/action_model': 9.890558543194611e-05, 'epoch': 0.62})
+[2026-07-28 20:26:23,231][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.039685796946287155, 'train/grad_norm_pre_clip': 1.1578035354614258, 'timing/data': 0.00033039413392543793, 'timing/model': 0.6197747774422169, 'learning_rate/qwen_vl_interface': 9.889740414108994e-06, 'learning_rate/action_model': 9.889740414108995e-05, 'epoch': 0.62})
+[2026-07-28 20:26:32,381][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.03081916645169258, 'train/grad_norm_pre_clip': 0.9800947904586792, 'timing/data': 0.00041762925684452057, 'timing/model': 0.6191954202950001, 'learning_rate/qwen_vl_interface': 9.888919274346196e-06, 'learning_rate/action_model': 9.888919274346195e-05, 'epoch': 0.62})
+[2026-07-28 20:26:41,830][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.03802253305912018, 'train/grad_norm_pre_clip': 1.547250509262085, 'timing/data': 0.0004035830497741699, 'timing/model': 0.6219657007604837, 'learning_rate/qwen_vl_interface': 9.888095124439043e-06, 'learning_rate/action_model': 9.888095124439043e-05, 'epoch': 0.62})
+[2026-07-28 20:26:51,001][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.048565931618213654, 'train/grad_norm_pre_clip': 2.8544445037841797, 'timing/data': 0.0004196278750896454, 'timing/model': 0.6201563309878111, 'learning_rate/qwen_vl_interface': 9.88726796492232e-06, 'learning_rate/action_model': 9.88726796492232e-05, 'epoch': 0.62})
+[2026-07-28 20:27:00,404][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.04197809845209122, 'train/grad_norm_pre_clip': 1.597548484802246, 'timing/data': 0.0003773532807826996, 'timing/model': 0.8616790268570185, 'learning_rate/qwen_vl_interface': 9.886437796332759e-06, 'learning_rate/action_model': 9.88643779633276e-05, 'epoch': 0.62})
+[2026-07-28 20:27:09,549][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.03351224213838577, 'train/grad_norm_pre_clip': 0.9597777724266052, 'timing/data': 0.0004002191126346588, 'timing/model': 0.6188401412218809, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.62})
+[2026-07-28 20:27:18,720][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.037193477153778076, 'train/grad_norm_pre_clip': 1.2107408046722412, 'timing/data': 0.0003889352083206177, 'timing/model': 0.6198431197553873, 'learning_rate/qwen_vl_interface': 9.884768434091832e-06, 'learning_rate/action_model': 9.884768434091832e-05, 'epoch': 0.62})
+[2026-07-28 20:27:27,874][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.01300618052482605, 'train/grad_norm_pre_clip': 0.338954895734787, 'timing/data': 0.00038833729922771454, 'timing/model': 0.619221979752183, 'learning_rate/qwen_vl_interface': 9.883929241523698e-06, 'learning_rate/action_model': 9.883929241523697e-05, 'epoch': 0.62})
+[2026-07-28 20:27:36,778][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.03085188940167427, 'train/grad_norm_pre_clip': 0.905236542224884, 'timing/data': 0.0004003811627626419, 'timing/model': 0.6204934362322092, 'learning_rate/qwen_vl_interface': 9.883087042049185e-06, 'learning_rate/action_model': 9.883087042049186e-05, 'epoch': 0.62})
+[2026-07-28 20:27:46,169][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.023372946307063103, 'train/grad_norm_pre_clip': 0.872815728187561, 'timing/data': 0.0004094243049621582, 'timing/model': 0.6194274947047234, 'learning_rate/qwen_vl_interface': 9.882241836214794e-06, 'learning_rate/action_model': 9.882241836214794e-05, 'epoch': 0.62})
+[2026-07-28 20:27:55,520][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.030424706637859344, 'train/grad_norm_pre_clip': 0.7969114184379578, 'timing/data': 0.0003532432019710541, 'timing/model': 0.6211999543011189, 'learning_rate/qwen_vl_interface': 9.881393624568965e-06, 'learning_rate/action_model': 9.881393624568965e-05, 'epoch': 0.62})
+[2026-07-28 20:28:04,705][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.05333705246448517, 'train/grad_norm_pre_clip': 2.33836030960083, 'timing/data': 0.00035061873495578766, 'timing/model': 0.6199292596429586, 'learning_rate/qwen_vl_interface': 9.880542407662098e-06, 'learning_rate/action_model': 9.880542407662097e-05, 'epoch': 0.62})
+[2026-07-28 20:28:13,924][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.03497079014778137, 'train/grad_norm_pre_clip': 2.029106616973877, 'timing/data': 0.00044313259422779083, 'timing/model': 0.619920838624239, 'learning_rate/qwen_vl_interface': 9.879688186046535e-06, 'learning_rate/action_model': 9.879688186046534e-05, 'epoch': 0.62})
+[2026-07-28 20:28:23,314][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.04153990000486374, 'train/grad_norm_pre_clip': 2.8204829692840576, 'timing/data': 0.0005397163331508636, 'timing/model': 0.6185990236699581, 'learning_rate/qwen_vl_interface': 9.87883096027657e-06, 'learning_rate/action_model': 9.87883096027657e-05, 'epoch': 0.62})
+[2026-07-28 20:28:32,494][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.059888970106840134, 'train/grad_norm_pre_clip': 2.153078556060791, 'timing/data': 0.0005896128714084625, 'timing/model': 0.6197266951203346, 'learning_rate/qwen_vl_interface': 9.87797073090845e-06, 'learning_rate/action_model': 9.87797073090845e-05, 'epoch': 0.62})
+[2026-07-28 20:28:41,849][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.07188977301120758, 'train/grad_norm_pre_clip': 3.9880106449127197, 'timing/data': 0.0003321561962366104, 'timing/model': 0.8193354178220034, 'learning_rate/qwen_vl_interface': 9.877107498500368e-06, 'learning_rate/action_model': 9.877107498500367e-05, 'epoch': 0.62})
+[2026-07-28 20:28:51,239][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.037448495626449585, 'train/grad_norm_pre_clip': 0.9701632261276245, 'timing/data': 0.0003908257931470871, 'timing/model': 0.8489407338202, 'learning_rate/qwen_vl_interface': 9.876241263612467e-06, 'learning_rate/action_model': 9.876241263612465e-05, 'epoch': 0.62})
+[2026-07-28 20:29:00,445][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.03647972270846367, 'train/grad_norm_pre_clip': 2.230954885482788, 'timing/data': 0.0005556531250476837, 'timing/model': 0.6191422715783119, 'learning_rate/qwen_vl_interface': 9.875372026806835e-06, 'learning_rate/action_model': 9.875372026806835e-05, 'epoch': 0.62})
+[2026-07-28 20:29:09,586][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.033637817949056625, 'train/grad_norm_pre_clip': 1.2001233100891113, 'timing/data': 0.0004943292587995529, 'timing/model': 0.6209790203720331, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.62})
+[2026-07-28 20:29:18,759][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.0379304438829422, 'train/grad_norm_pre_clip': 1.2891128063201904, 'timing/data': 0.0004973430186510086, 'timing/model': 0.6208549961447716, 'learning_rate/qwen_vl_interface': 9.873624549700485e-06, 'learning_rate/action_model': 9.873624549700484e-05, 'epoch': 0.62})
+[2026-07-28 20:29:28,158][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.045208416879177094, 'train/grad_norm_pre_clip': 2.049669027328491, 'timing/data': 0.0005563907325267792, 'timing/model': 0.6202106680721045, 'learning_rate/qwen_vl_interface': 9.872746310533683e-06, 'learning_rate/action_model': 9.872746310533684e-05, 'epoch': 0.62})
+[2026-07-28 20:29:37,663][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.03158096969127655, 'train/grad_norm_pre_clip': 0.7560999393463135, 'timing/data': 0.0004587620496749878, 'timing/model': 0.8799202051013708, 'learning_rate/qwen_vl_interface': 9.871865071716991e-06, 'learning_rate/action_model': 9.87186507171699e-05, 'epoch': 0.62})
+[2026-07-28 20:29:46,934][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.06235847622156143, 'train/grad_norm_pre_clip': 5.692343711853027, 'timing/data': 0.0003407523036003113, 'timing/model': 0.6239958591759205, 'learning_rate/qwen_vl_interface': 9.870980833822231e-06, 'learning_rate/action_model': 9.87098083382223e-05, 'epoch': 0.62})
+[2026-07-28 20:29:56,404][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.037606868892908096, 'train/grad_norm_pre_clip': 1.5864733457565308, 'timing/data': 0.00044535286724567413, 'timing/model': 0.6228896956890821, 'learning_rate/qwen_vl_interface': 9.87009359742318e-06, 'learning_rate/action_model': 9.87009359742318e-05, 'epoch': 0.62})
+[2026-07-28 20:30:05,573][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.05127281695604324, 'train/grad_norm_pre_clip': 1.7721422910690308, 'timing/data': 0.0005654729902744293, 'timing/model': 0.6195947304368019, 'learning_rate/qwen_vl_interface': 9.86920336309555e-06, 'learning_rate/action_model': 9.86920336309555e-05, 'epoch': 0.62})
+[2026-07-28 20:30:14,488][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.050135329365730286, 'train/grad_norm_pre_clip': 2.7994117736816406, 'timing/data': 0.00036591291427612305, 'timing/model': 0.6190771646797657, 'learning_rate/qwen_vl_interface': 9.86831013141701e-06, 'learning_rate/action_model': 9.86831013141701e-05, 'epoch': 0.62})
+[2026-07-28 20:30:23,616][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.03496731445193291, 'train/grad_norm_pre_clip': 1.3005924224853516, 'timing/data': 0.0006848089396953583, 'timing/model': 0.6217647939920425, 'learning_rate/qwen_vl_interface': 9.867413902967166e-06, 'learning_rate/action_model': 9.867413902967166e-05, 'epoch': 0.62})
+[2026-07-28 20:30:33,019][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.03664251044392586, 'train/grad_norm_pre_clip': 1.765036940574646, 'timing/data': 0.0003383271396160126, 'timing/model': 0.8659923765808344, 'learning_rate/qwen_vl_interface': 9.86651467832757e-06, 'learning_rate/action_model': 9.86651467832757e-05, 'epoch': 0.62})
+[2026-07-28 20:30:41,933][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.045140862464904785, 'train/grad_norm_pre_clip': 2.445263624191284, 'timing/data': 0.00039343908429145813, 'timing/model': 0.6179674305021763, 'learning_rate/qwen_vl_interface': 9.865612458081718e-06, 'learning_rate/action_model': 9.865612458081719e-05, 'epoch': 0.62})
+[2026-07-28 20:30:51,208][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.07984621822834015, 'train/grad_norm_pre_clip': 3.2701756954193115, 'timing/data': 0.00032979249954223633, 'timing/model': 0.6230744514614344, 'learning_rate/qwen_vl_interface': 9.864707242815057e-06, 'learning_rate/action_model': 9.864707242815056e-05, 'epoch': 0.62})
+[2026-07-28 20:31:00,352][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.10132165998220444, 'train/grad_norm_pre_clip': 2.323862314224243, 'timing/data': 0.000336620956659317, 'timing/model': 0.6196384765207767, 'learning_rate/qwen_vl_interface': 9.863799033114966e-06, 'learning_rate/action_model': 9.863799033114966e-05, 'epoch': 0.62})
+[2026-07-28 20:31:09,539][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.030288146808743477, 'train/grad_norm_pre_clip': 2.0329720973968506, 'timing/data': 0.0004040561616420746, 'timing/model': 0.6218633521348238, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.62})
+[2026-07-28 20:31:18,913][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.02990780398249626, 'train/grad_norm_pre_clip': 1.0360203981399536, 'timing/data': 0.0004123486578464508, 'timing/model': 0.6189785487949848, 'learning_rate/qwen_vl_interface': 9.861973632773748e-06, 'learning_rate/action_model': 9.861973632773748e-05, 'epoch': 0.62})
+[2026-07-28 20:31:28,284][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.02985616773366928, 'train/grad_norm_pre_clip': 0.9994741082191467, 'timing/data': 0.000352710485458374, 'timing/model': 0.619591286405921, 'learning_rate/qwen_vl_interface': 9.861056443317105e-06, 'learning_rate/action_model': 9.861056443317105e-05, 'epoch': 0.62})
+[2026-07-28 20:31:37,453][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.04805292934179306, 'train/grad_norm_pre_clip': 1.6845375299453735, 'timing/data': 0.00043478794395923615, 'timing/model': 0.6207911800593138, 'learning_rate/qwen_vl_interface': 9.860136261796e-06, 'learning_rate/action_model': 9.860136261795998e-05, 'epoch': 0.62})
+[2026-07-28 20:31:46,671][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.03900828957557678, 'train/grad_norm_pre_clip': 0.9809978604316711, 'timing/data': 0.00042089633643627167, 'timing/model': 0.6212636101990938, 'learning_rate/qwen_vl_interface': 9.859213088807521e-06, 'learning_rate/action_model': 9.85921308880752e-05, 'epoch': 0.63})
+[2026-07-28 20:31:56,062][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.029151804745197296, 'train/grad_norm_pre_clip': 0.8084215521812439, 'timing/data': 0.00032784603536129, 'timing/model': 0.620199179276824, 'learning_rate/qwen_vl_interface': 9.858286924950714e-06, 'learning_rate/action_model': 9.858286924950713e-05, 'epoch': 0.63})
+[2026-07-28 20:32:05,391][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.01756317913532257, 'train/grad_norm_pre_clip': 0.6517876982688904, 'timing/data': 0.0003304537385702133, 'timing/model': 0.837952584028244, 'learning_rate/qwen_vl_interface': 9.85735777082655e-06, 'learning_rate/action_model': 9.857357770826551e-05, 'epoch': 0.63})
+[2026-07-28 20:32:14,548][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.038808535784482956, 'train/grad_norm_pre_clip': 1.8026989698410034, 'timing/data': 0.0003764983266592026, 'timing/model': 0.6201854832470417, 'learning_rate/qwen_vl_interface': 9.856425627037951e-06, 'learning_rate/action_model': 9.856425627037952e-05, 'epoch': 0.63})
+[2026-07-28 20:32:23,702][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.016031119972467422, 'train/grad_norm_pre_clip': 0.8395956754684448, 'timing/data': 0.0004403498023748398, 'timing/model': 0.619873084127903, 'learning_rate/qwen_vl_interface': 9.855490494189777e-06, 'learning_rate/action_model': 9.855490494189776e-05, 'epoch': 0.63})
+[2026-07-28 20:32:33,082][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.028276631608605385, 'train/grad_norm_pre_clip': 2.266655445098877, 'timing/data': 0.0004192143678665161, 'timing/model': 0.622860811650753, 'learning_rate/qwen_vl_interface': 9.854552372888819e-06, 'learning_rate/action_model': 9.854552372888819e-05, 'epoch': 0.63})
+[2026-07-28 20:32:42,480][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.06994734704494476, 'train/grad_norm_pre_clip': 2.3113198280334473, 'timing/data': 0.0006263330578804016, 'timing/model': 0.620916111394763, 'learning_rate/qwen_vl_interface': 9.853611263743817e-06, 'learning_rate/action_model': 9.853611263743817e-05, 'epoch': 0.63})
+[2026-07-28 20:32:51,638][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.030900035053491592, 'train/grad_norm_pre_clip': 1.44376540184021, 'timing/data': 0.0004127137362957001, 'timing/model': 0.6203329730778933, 'learning_rate/qwen_vl_interface': 9.85266716736545e-06, 'learning_rate/action_model': 9.852667167365449e-05, 'epoch': 0.63})
+[2026-07-28 20:33:00,985][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.06008496135473251, 'train/grad_norm_pre_clip': 1.672683835029602, 'timing/data': 0.0004014279693365097, 'timing/model': 0.7988036405295134, 'learning_rate/qwen_vl_interface': 9.851720084366324e-06, 'learning_rate/action_model': 9.851720084366324e-05, 'epoch': 0.63})
+[2026-07-28 20:33:10,147][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.04637311398983002, 'train/grad_norm_pre_clip': 1.4227405786514282, 'timing/data': 0.0003822706639766693, 'timing/model': 0.6236834060400724, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.63})
+[2026-07-28 20:33:19,513][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.08966019004583359, 'train/grad_norm_pre_clip': 2.9506940841674805, 'timing/data': 0.00040346570312976837, 'timing/model': 0.8444861136376858, 'learning_rate/qwen_vl_interface': 9.849816960965959e-06, 'learning_rate/action_model': 9.849816960965959e-05, 'epoch': 0.63})
+[2026-07-28 20:33:28,650][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.027367182075977325, 'train/grad_norm_pre_clip': 1.2824807167053223, 'timing/data': 0.00039790384471416473, 'timing/model': 0.6191725600510836, 'learning_rate/qwen_vl_interface': 9.848860921799633e-06, 'learning_rate/action_model': 9.848860921799632e-05, 'epoch': 0.63})
+[2026-07-28 20:33:38,027][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.017398793250322342, 'train/grad_norm_pre_clip': 0.6199262142181396, 'timing/data': 0.0003557596355676651, 'timing/model': 0.859487259760499, 'learning_rate/qwen_vl_interface': 9.847901898482385e-06, 'learning_rate/action_model': 9.847901898482385e-05, 'epoch': 0.63})
+[2026-07-28 20:33:47,188][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.08269098401069641, 'train/grad_norm_pre_clip': 2.3980958461761475, 'timing/data': 0.0004223734140396118, 'timing/model': 0.6197963617742062, 'learning_rate/qwen_vl_interface': 9.846939891636512e-06, 'learning_rate/action_model': 9.846939891636512e-05, 'epoch': 0.63})
+[2026-07-28 20:33:56,364][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.06657078862190247, 'train/grad_norm_pre_clip': 1.512290358543396, 'timing/data': 0.0005728546530008316, 'timing/model': 0.6204640585929155, 'learning_rate/qwen_vl_interface': 9.845974901886253e-06, 'learning_rate/action_model': 9.845974901886252e-05, 'epoch': 0.63})
+[2026-07-28 20:34:05,567][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.05993057042360306, 'train/grad_norm_pre_clip': 2.17075252532959, 'timing/data': 0.000332878902554512, 'timing/model': 0.6218121014535427, 'learning_rate/qwen_vl_interface': 9.845006929857777e-06, 'learning_rate/action_model': 9.845006929857776e-05, 'epoch': 0.63})
+[2026-07-28 20:34:14,955][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.030232416465878487, 'train/grad_norm_pre_clip': 0.8312033414840698, 'timing/data': 0.00032023340463638306, 'timing/model': 0.6212123818695545, 'learning_rate/qwen_vl_interface': 9.844035976179193e-06, 'learning_rate/action_model': 9.844035976179192e-05, 'epoch': 0.63})
+[2026-07-28 20:34:24,051][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.026336833834648132, 'train/grad_norm_pre_clip': 0.5543492436408997, 'timing/data': 0.0003778468817472458, 'timing/model': 0.6216760724782944, 'learning_rate/qwen_vl_interface': 9.843062041480538e-06, 'learning_rate/action_model': 9.843062041480539e-05, 'epoch': 0.63})
+[2026-07-28 20:34:33,181][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.04860180988907814, 'train/grad_norm_pre_clip': 1.3265682458877563, 'timing/data': 0.00039840489625930786, 'timing/model': 0.6243261974304914, 'learning_rate/qwen_vl_interface': 9.842085126393793e-06, 'learning_rate/action_model': 9.842085126393792e-05, 'epoch': 0.63})
+[2026-07-28 20:34:42,326][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.031802572309970856, 'train/grad_norm_pre_clip': 1.4636547565460205, 'timing/data': 0.00046739354729652405, 'timing/model': 0.6191026885062456, 'learning_rate/qwen_vl_interface': 9.841105231552863e-06, 'learning_rate/action_model': 9.841105231552864e-05, 'epoch': 0.63})
+[2026-07-28 20:34:51,875][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.05403190106153488, 'train/grad_norm_pre_clip': 1.0836317539215088, 'timing/data': 0.000398905947804451, 'timing/model': 0.8418775051832199, 'learning_rate/qwen_vl_interface': 9.840122357593594e-06, 'learning_rate/action_model': 9.840122357593594e-05, 'epoch': 0.63})
+[2026-07-28 20:35:01,022][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.033973000943660736, 'train/grad_norm_pre_clip': 1.3051238059997559, 'timing/data': 0.0003996770828962326, 'timing/model': 0.6245708800852299, 'learning_rate/qwen_vl_interface': 9.839136505153762e-06, 'learning_rate/action_model': 9.839136505153762e-05, 'epoch': 0.63})
+[2026-07-28 20:35:10,166][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.018230730667710304, 'train/grad_norm_pre_clip': 0.6606720685958862, 'timing/data': 0.0005035698413848877, 'timing/model': 0.6210349779576063, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.63})
+[2026-07-28 20:35:19,609][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.03748733550310135, 'train/grad_norm_pre_clip': 0.8447743058204651, 'timing/data': 0.000394551083445549, 'timing/model': 0.6187193691730499, 'learning_rate/qwen_vl_interface': 9.837155867393176e-06, 'learning_rate/action_model': 9.837155867393175e-05, 'epoch': 0.63})
+[2026-07-28 20:35:28,999][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.0748225748538971, 'train/grad_norm_pre_clip': 2.8280510902404785, 'timing/data': 0.00040329620242118835, 'timing/model': 0.8576341327279806, 'learning_rate/qwen_vl_interface': 9.836161083357635e-06, 'learning_rate/action_model': 9.836161083357635e-05, 'epoch': 0.63})
+[2026-07-28 20:35:38,159][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.04719462990760803, 'train/grad_norm_pre_clip': 0.7275788187980652, 'timing/data': 0.00034741126000881195, 'timing/model': 0.6185374837368727, 'learning_rate/qwen_vl_interface': 9.835163323411957e-06, 'learning_rate/action_model': 9.835163323411958e-05, 'epoch': 0.63})
+[2026-07-28 20:35:47,318][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.049032218754291534, 'train/grad_norm_pre_clip': 1.0368077754974365, 'timing/data': 0.00033993087708950043, 'timing/model': 0.6209652461111546, 'learning_rate/qwen_vl_interface': 9.834162588203581e-06, 'learning_rate/action_model': 9.83416258820358e-05, 'epoch': 0.63})
+[2026-07-28 20:35:56,694][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.03218371421098709, 'train/grad_norm_pre_clip': 1.0432372093200684, 'timing/data': 0.00034099072217941284, 'timing/model': 0.8590985778719187, 'learning_rate/qwen_vl_interface': 9.833158878381872e-06, 'learning_rate/action_model': 9.83315887838187e-05, 'epoch': 0.63})
+[2026-07-28 20:36:06,055][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.023810163140296936, 'train/grad_norm_pre_clip': 1.2710102796554565, 'timing/data': 0.0003945492208003998, 'timing/model': 0.618812957778573, 'learning_rate/qwen_vl_interface': 9.832152194598123e-06, 'learning_rate/action_model': 9.832152194598122e-05, 'epoch': 0.63})
+[2026-07-28 20:36:15,160][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.03256836533546448, 'train/grad_norm_pre_clip': 0.7283591628074646, 'timing/data': 0.0003177281469106674, 'timing/model': 0.6209467519074678, 'learning_rate/qwen_vl_interface': 9.831142537505563e-06, 'learning_rate/action_model': 9.831142537505562e-05, 'epoch': 0.63})
+[2026-07-28 20:36:24,083][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.022327197715640068, 'train/grad_norm_pre_clip': 0.6968917846679688, 'timing/data': 0.00033389031887054443, 'timing/model': 0.6206337325274944, 'learning_rate/qwen_vl_interface': 9.830129907759347e-06, 'learning_rate/action_model': 9.830129907759347e-05, 'epoch': 0.63})
+[2026-07-28 20:36:33,023][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.022069234400987625, 'train/grad_norm_pre_clip': 0.7104883790016174, 'timing/data': 0.0004929490387439728, 'timing/model': 0.6210400350391865, 'learning_rate/qwen_vl_interface': 9.82911430601656e-06, 'learning_rate/action_model': 9.82911430601656e-05, 'epoch': 0.63})
+[2026-07-28 20:36:42,210][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.030126556754112244, 'train/grad_norm_pre_clip': 0.7756884098052979, 'timing/data': 0.00036130286753177643, 'timing/model': 0.8730423003435135, 'learning_rate/qwen_vl_interface': 9.828095732936216e-06, 'learning_rate/action_model': 9.828095732936216e-05, 'epoch': 0.63})
+[2026-07-28 20:36:51,137][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.01730271801352501, 'train/grad_norm_pre_clip': 1.005043864250183, 'timing/data': 0.00035476312041282654, 'timing/model': 0.6194445639848709, 'learning_rate/qwen_vl_interface': 9.827074189179254e-06, 'learning_rate/action_model': 9.827074189179253e-05, 'epoch': 0.63})
+[2026-07-28 20:37:00,267][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.06811395287513733, 'train/grad_norm_pre_clip': 1.8753753900527954, 'timing/data': 0.0003878399729728699, 'timing/model': 0.6243737451732159, 'learning_rate/qwen_vl_interface': 9.82604967540854e-06, 'learning_rate/action_model': 9.82604967540854e-05, 'epoch': 0.63})
+[2026-07-28 20:37:09,405][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.06066330522298813, 'train/grad_norm_pre_clip': 1.6941618919372559, 'timing/data': 0.00034576840698719025, 'timing/model': 0.6191705949604511, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.63})
+[2026-07-28 20:37:18,557][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.028608787804841995, 'train/grad_norm_pre_clip': 1.2275482416152954, 'timing/data': 0.0003326907753944397, 'timing/model': 0.6192209478467703, 'learning_rate/qwen_vl_interface': 9.823991740486982e-06, 'learning_rate/action_model': 9.823991740486982e-05, 'epoch': 0.63})
+[2026-07-28 20:37:27,722][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.008978459984064102, 'train/grad_norm_pre_clip': 0.5255653858184814, 'timing/data': 0.00036259181797504425, 'timing/model': 0.6216280553489923, 'learning_rate/qwen_vl_interface': 9.822958320671506e-06, 'learning_rate/action_model': 9.822958320671506e-05, 'epoch': 0.63})
+[2026-07-28 20:37:37,095][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.02646297588944435, 'train/grad_norm_pre_clip': 1.3488188982009888, 'timing/data': 0.0004215296357870102, 'timing/model': 0.623797569423914, 'learning_rate/qwen_vl_interface': 9.821921933513023e-06, 'learning_rate/action_model': 9.821921933513023e-05, 'epoch': 0.63})
+[2026-07-28 20:37:46,505][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.017561174929142, 'train/grad_norm_pre_clip': 0.6628040075302124, 'timing/data': 0.00039051100611686707, 'timing/model': 0.6206918507814407, 'learning_rate/qwen_vl_interface': 9.820882579684037e-06, 'learning_rate/action_model': 9.820882579684037e-05, 'epoch': 0.63})
+[2026-07-28 20:37:55,671][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.027646586298942566, 'train/grad_norm_pre_clip': 1.5193601846694946, 'timing/data': 0.00031773000955581665, 'timing/model': 0.6197581719607115, 'learning_rate/qwen_vl_interface': 9.819840259858968e-06, 'learning_rate/action_model': 9.819840259858968e-05, 'epoch': 0.63})
+[2026-07-28 20:38:04,794][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.023173779249191284, 'train/grad_norm_pre_clip': 1.2020312547683716, 'timing/data': 0.0003828220069408417, 'timing/model': 0.6208633817732334, 'learning_rate/qwen_vl_interface': 9.818794974714167e-06, 'learning_rate/action_model': 9.818794974714168e-05, 'epoch': 0.63})
+[2026-07-28 20:38:13,924][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.01112175639718771, 'train/grad_norm_pre_clip': 1.0772281885147095, 'timing/data': 0.00037444010376930237, 'timing/model': 0.6193112339824438, 'learning_rate/qwen_vl_interface': 9.817746724927912e-06, 'learning_rate/action_model': 9.817746724927912e-05, 'epoch': 0.63})
+[2026-07-28 20:38:23,298][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.06348837912082672, 'train/grad_norm_pre_clip': 1.9839575290679932, 'timing/data': 0.0003841295838356018, 'timing/model': 0.6202699206769466, 'learning_rate/qwen_vl_interface': 9.8166955111804e-06, 'learning_rate/action_model': 9.816695511180399e-05, 'epoch': 0.63})
+[2026-07-28 20:38:32,455][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.047163210809230804, 'train/grad_norm_pre_clip': 8.038662910461426, 'timing/data': 0.00040389224886894226, 'timing/model': 0.620597593486309, 'learning_rate/qwen_vl_interface': 9.815641334153746e-06, 'learning_rate/action_model': 9.815641334153746e-05, 'epoch': 0.63})
+[2026-07-28 20:38:41,819][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.011332840658724308, 'train/grad_norm_pre_clip': 0.666242778301239, 'timing/data': 0.0004023127257823944, 'timing/model': 0.6209753919392824, 'learning_rate/qwen_vl_interface': 9.814584194532001e-06, 'learning_rate/action_model': 9.814584194532001e-05, 'epoch': 0.63})
+[2026-07-28 20:38:50,951][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.053501278162002563, 'train/grad_norm_pre_clip': 2.501675605773926, 'timing/data': 0.0004313420504331589, 'timing/model': 0.6207049116492271, 'learning_rate/qwen_vl_interface': 9.813524093001128e-06, 'learning_rate/action_model': 9.813524093001129e-05, 'epoch': 0.63})
+[2026-07-28 20:39:00,339][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.046028368175029755, 'train/grad_norm_pre_clip': 2.189439296722412, 'timing/data': 0.00037843361496925354, 'timing/model': 0.6209217254072428, 'learning_rate/qwen_vl_interface': 9.81246103024902e-06, 'learning_rate/action_model': 9.81246103024902e-05, 'epoch': 0.63})
+[2026-07-28 20:39:09,524][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.035926006734371185, 'train/grad_norm_pre_clip': 1.3403398990631104, 'timing/data': 0.00037328340113162994, 'timing/model': 0.6214221026748419, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.63})
+[2026-07-28 20:39:18,682][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.05006735771894455, 'train/grad_norm_pre_clip': 1.759973168373108, 'timing/data': 0.00039141811430454254, 'timing/model': 0.6205668542534113, 'learning_rate/qwen_vl_interface': 9.810326023842245e-06, 'learning_rate/action_model': 9.810326023842245e-05, 'epoch': 0.63})
+[2026-07-28 20:39:28,278][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.039503276348114014, 'train/grad_norm_pre_clip': 1.3912016153335571, 'timing/data': 0.0003344584256410599, 'timing/model': 0.8507949113845825, 'learning_rate/qwen_vl_interface': 9.809254081572963e-06, 'learning_rate/action_model': 9.809254081572961e-05, 'epoch': 0.63})
+[2026-07-28 20:39:37,375][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.04294443875551224, 'train/grad_norm_pre_clip': 1.9198030233383179, 'timing/data': 0.0003891400992870331, 'timing/model': 0.6205791980028152, 'learning_rate/qwen_vl_interface': 9.808179180853204e-06, 'learning_rate/action_model': 9.808179180853204e-05, 'epoch': 0.63})
+[2026-07-28 20:39:46,501][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.03561212867498398, 'train/grad_norm_pre_clip': 1.0141514539718628, 'timing/data': 0.0003926791250705719, 'timing/model': 0.6195958834141493, 'learning_rate/qwen_vl_interface': 9.807101322380464e-06, 'learning_rate/action_model': 9.807101322380464e-05, 'epoch': 0.63})
+[2026-07-28 20:39:55,645][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.0623457208275795, 'train/grad_norm_pre_clip': 1.7514132261276245, 'timing/data': 0.0005321372300386429, 'timing/model': 0.6194839049130678, 'learning_rate/qwen_vl_interface': 9.806020506854152e-06, 'learning_rate/action_model': 9.806020506854153e-05, 'epoch': 0.63})
+[2026-07-28 20:40:05,031][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.023442136123776436, 'train/grad_norm_pre_clip': 1.1400521993637085, 'timing/data': 0.00038455426692962646, 'timing/model': 0.6188079155981541, 'learning_rate/qwen_vl_interface': 9.804936734975596e-06, 'learning_rate/action_model': 9.804936734975596e-05, 'epoch': 0.63})
+[2026-07-28 20:40:14,389][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.03451240062713623, 'train/grad_norm_pre_clip': 0.7936941385269165, 'timing/data': 0.0003325473517179489, 'timing/model': 0.6188549436628819, 'learning_rate/qwen_vl_interface': 9.803850007448046e-06, 'learning_rate/action_model': 9.803850007448045e-05, 'epoch': 0.64})
+[2026-07-28 20:40:23,552][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.04527512565255165, 'train/grad_norm_pre_clip': 1.7667100429534912, 'timing/data': 0.00040366873145103455, 'timing/model': 0.6206036116927862, 'learning_rate/qwen_vl_interface': 9.802760324976666e-06, 'learning_rate/action_model': 9.802760324976667e-05, 'epoch': 0.64})
+[2026-07-28 20:40:32,897][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.018776141107082367, 'train/grad_norm_pre_clip': 1.0839601755142212, 'timing/data': 0.0006378386169672012, 'timing/model': 0.8268646188080311, 'learning_rate/qwen_vl_interface': 9.80166768826854e-06, 'learning_rate/action_model': 9.80166768826854e-05, 'epoch': 0.64})
+[2026-07-28 20:40:42,041][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.051065657287836075, 'train/grad_norm_pre_clip': 2.007728338241577, 'timing/data': 0.00032838620245456696, 'timing/model': 0.6194851510226727, 'learning_rate/qwen_vl_interface': 9.800572098032668e-06, 'learning_rate/action_model': 9.800572098032668e-05, 'epoch': 0.64})
+[2026-07-28 20:40:51,164][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.019231978803873062, 'train/grad_norm_pre_clip': 0.5277105569839478, 'timing/data': 0.0003299340605735779, 'timing/model': 0.6203794460743666, 'learning_rate/qwen_vl_interface': 9.799473554979966e-06, 'learning_rate/action_model': 9.799473554979965e-05, 'epoch': 0.64})
+[2026-07-28 20:41:00,519][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.02209930494427681, 'train/grad_norm_pre_clip': 0.9275872707366943, 'timing/data': 0.00041580386459827423, 'timing/model': 0.8442996609956026, 'learning_rate/qwen_vl_interface': 9.798372059823268e-06, 'learning_rate/action_model': 9.798372059823268e-05, 'epoch': 0.64})
+[2026-07-28 20:41:09,890][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.02431822009384632, 'train/grad_norm_pre_clip': 0.704126238822937, 'timing/data': 0.00046928785741329193, 'timing/model': 0.861936654895544, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.64})
+[2026-07-28 20:41:18,988][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.051154181361198425, 'train/grad_norm_pre_clip': 1.4133118391036987, 'timing/data': 0.0004039052873849869, 'timing/model': 0.619857294484973, 'learning_rate/qwen_vl_interface': 9.79616021605879e-06, 'learning_rate/action_model': 9.796160216058788e-05, 'epoch': 0.64})
+[2026-07-28 20:41:28,122][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.021494992077350616, 'train/grad_norm_pre_clip': 0.7723424434661865, 'timing/data': 0.0004114583134651184, 'timing/model': 0.6207827366888523, 'learning_rate/qwen_vl_interface': 9.795049868886247e-06, 'learning_rate/action_model': 9.795049868886247e-05, 'epoch': 0.64})
+[2026-07-28 20:41:37,254][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.039435453712940216, 'train/grad_norm_pre_clip': 1.7880455255508423, 'timing/data': 0.00048072077333927155, 'timing/model': 0.620841983705759, 'learning_rate/qwen_vl_interface': 9.793936572480192e-06, 'learning_rate/action_model': 9.793936572480192e-05, 'epoch': 0.64})
+[2026-07-28 20:41:46,885][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.01358325406908989, 'train/grad_norm_pre_clip': 0.6546611785888672, 'timing/data': 0.00041303597390651703, 'timing/model': 0.6235822271555662, 'learning_rate/qwen_vl_interface': 9.792820327563028e-06, 'learning_rate/action_model': 9.792820327563028e-05, 'epoch': 0.64})
+[2026-07-28 20:41:55,998][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.02781103178858757, 'train/grad_norm_pre_clip': 1.2938098907470703, 'timing/data': 0.000393155962228775, 'timing/model': 0.6222354769706726, 'learning_rate/qwen_vl_interface': 9.791701134859076e-06, 'learning_rate/action_model': 9.791701134859075e-05, 'epoch': 0.64})
+[2026-07-28 20:42:05,152][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.03323030844330788, 'train/grad_norm_pre_clip': 1.1303471326828003, 'timing/data': 0.00038329511880874634, 'timing/model': 0.6192562244832516, 'learning_rate/qwen_vl_interface': 9.790578995094565e-06, 'learning_rate/action_model': 9.790578995094563e-05, 'epoch': 0.64})
+[2026-07-28 20:42:14,560][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.024338223040103912, 'train/grad_norm_pre_clip': 0.8885124325752258, 'timing/data': 0.0004070643335580826, 'timing/model': 0.6200926303863525, 'learning_rate/qwen_vl_interface': 9.789453908997639e-06, 'learning_rate/action_model': 9.789453908997638e-05, 'epoch': 0.64})
+[2026-07-28 20:42:23,969][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.017104072496294975, 'train/grad_norm_pre_clip': 0.577804684638977, 'timing/data': 0.00033463165163993835, 'timing/model': 0.6198455579578876, 'learning_rate/qwen_vl_interface': 9.788325877298355e-06, 'learning_rate/action_model': 9.788325877298355e-05, 'epoch': 0.64})
+[2026-07-28 20:42:33,092][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.02187928557395935, 'train/grad_norm_pre_clip': 1.2155781984329224, 'timing/data': 0.0004276670515537262, 'timing/model': 0.6201263330876827, 'learning_rate/qwen_vl_interface': 9.787194900728682e-06, 'learning_rate/action_model': 9.787194900728683e-05, 'epoch': 0.64})
+[2026-07-28 20:42:42,237][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.031686749309301376, 'train/grad_norm_pre_clip': 0.8943350911140442, 'timing/data': 0.0003658663481473923, 'timing/model': 0.6205595955252647, 'learning_rate/qwen_vl_interface': 9.786060980022499e-06, 'learning_rate/action_model': 9.786060980022498e-05, 'epoch': 0.64})
+[2026-07-28 20:42:51,656][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.01958722062408924, 'train/grad_norm_pre_clip': 1.4611228704452515, 'timing/data': 0.00039855949580669403, 'timing/model': 0.620604844763875, 'learning_rate/qwen_vl_interface': 9.784924115915592e-06, 'learning_rate/action_model': 9.784924115915592e-05, 'epoch': 0.64})
+[2026-07-28 20:43:00,807][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.025642935186624527, 'train/grad_norm_pre_clip': 1.5934126377105713, 'timing/data': 0.00040122494101524353, 'timing/model': 0.6208989676088095, 'learning_rate/qwen_vl_interface': 9.783784309145663e-06, 'learning_rate/action_model': 9.783784309145661e-05, 'epoch': 0.64})
+[2026-07-28 20:43:10,143][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.031387925148010254, 'train/grad_norm_pre_clip': 2.380648612976074, 'timing/data': 0.00039197318255901337, 'timing/model': 0.620003106072545, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.64})
+[2026-07-28 20:43:19,258][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.035528939217329025, 'train/grad_norm_pre_clip': 1.5266810655593872, 'timing/data': 0.000534864142537117, 'timing/model': 0.6198108289390802, 'learning_rate/qwen_vl_interface': 9.78149587057707e-06, 'learning_rate/action_model': 9.78149587057707e-05, 'epoch': 0.64})
+[2026-07-28 20:43:28,407][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.08930352330207825, 'train/grad_norm_pre_clip': 2.8568270206451416, 'timing/data': 0.0004032142460346222, 'timing/model': 0.6194158401340246, 'learning_rate/qwen_vl_interface': 9.780347240263354e-06, 'learning_rate/action_model': 9.780347240263353e-05, 'epoch': 0.64})
+[2026-07-28 20:43:37,779][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.011872565373778343, 'train/grad_norm_pre_clip': 0.7307072877883911, 'timing/data': 0.0004028007388114929, 'timing/model': 0.6193922683596611, 'learning_rate/qwen_vl_interface': 9.779195670256498e-06, 'learning_rate/action_model': 9.779195670256497e-05, 'epoch': 0.64})
+[2026-07-28 20:43:46,917][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.041563861072063446, 'train/grad_norm_pre_clip': 1.7498042583465576, 'timing/data': 0.0003999602049589157, 'timing/model': 0.6197097655385733, 'learning_rate/qwen_vl_interface': 9.778041161303746e-06, 'learning_rate/action_model': 9.778041161303745e-05, 'epoch': 0.64})
+[2026-07-28 20:43:56,281][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.03036363795399666, 'train/grad_norm_pre_clip': 0.8137646913528442, 'timing/data': 0.0005087703466415405, 'timing/model': 0.6196478772908449, 'learning_rate/qwen_vl_interface': 9.776883714154242e-06, 'learning_rate/action_model': 9.776883714154241e-05, 'epoch': 0.64})
+[2026-07-28 20:44:05,653][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.010822403244674206, 'train/grad_norm_pre_clip': 0.7243347764015198, 'timing/data': 0.0005493015050888062, 'timing/model': 0.6216413583606482, 'learning_rate/qwen_vl_interface': 9.775723329559044e-06, 'learning_rate/action_model': 9.775723329559044e-05, 'epoch': 0.64})
+[2026-07-28 20:44:14,815][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.019066713750362396, 'train/grad_norm_pre_clip': 1.1107008457183838, 'timing/data': 0.0003795512020587921, 'timing/model': 0.6203401759266853, 'learning_rate/qwen_vl_interface': 9.774560008271114e-06, 'learning_rate/action_model': 9.774560008271114e-05, 'epoch': 0.64})
+[2026-07-28 20:44:24,216][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.03238885849714279, 'train/grad_norm_pre_clip': 2.185358762741089, 'timing/data': 0.00032429397106170654, 'timing/model': 0.8793067820370197, 'learning_rate/qwen_vl_interface': 9.773393751045315e-06, 'learning_rate/action_model': 9.773393751045315e-05, 'epoch': 0.64})
+[2026-07-28 20:44:33,505][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.05766242370009422, 'train/grad_norm_pre_clip': 3.50718355178833, 'timing/data': 0.00038778968155384064, 'timing/model': 0.7575274538248777, 'learning_rate/qwen_vl_interface': 9.772224558638422e-06, 'learning_rate/action_model': 9.772224558638422e-05, 'epoch': 0.64})
+[2026-07-28 20:44:42,642][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.039774805307388306, 'train/grad_norm_pre_clip': 1.2149577140808105, 'timing/data': 0.00038187578320503235, 'timing/model': 0.6242042798548937, 'learning_rate/qwen_vl_interface': 9.77105243180911e-06, 'learning_rate/action_model': 9.771052431809109e-05, 'epoch': 0.64})
+[2026-07-28 20:44:51,796][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.03358028084039688, 'train/grad_norm_pre_clip': 1.966893196105957, 'timing/data': 0.000401470810174942, 'timing/model': 0.6200828962028027, 'learning_rate/qwen_vl_interface': 9.76987737131796e-06, 'learning_rate/action_model': 9.76987737131796e-05, 'epoch': 0.64})
+[2026-07-28 20:45:00,947][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.02265780419111252, 'train/grad_norm_pre_clip': 0.6733420491218567, 'timing/data': 0.0004057474434375763, 'timing/model': 0.6200616378337145, 'learning_rate/qwen_vl_interface': 9.768699377927454e-06, 'learning_rate/action_model': 9.768699377927453e-05, 'epoch': 0.64})
+[2026-07-28 20:45:10,334][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.0320170521736145, 'train/grad_norm_pre_clip': 1.4393614530563354, 'timing/data': 0.0005739051848649979, 'timing/model': 0.8521119374781847, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.64})
+[2026-07-28 20:45:19,461][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.023618213832378387, 'train/grad_norm_pre_clip': 0.7059051394462585, 'timing/data': 0.00039943866431713104, 'timing/model': 0.6213505435734987, 'learning_rate/qwen_vl_interface': 9.76633459550783e-06, 'learning_rate/action_model': 9.76633459550783e-05, 'epoch': 0.64})
+[2026-07-28 20:45:28,580][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.029843328520655632, 'train/grad_norm_pre_clip': 0.6976194977760315, 'timing/data': 0.0004532281309366226, 'timing/model': 0.6220008432865143, 'learning_rate/qwen_vl_interface': 9.765147808013193e-06, 'learning_rate/action_model': 9.765147808013192e-05, 'epoch': 0.64})
+[2026-07-28 20:45:37,946][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.07318738102912903, 'train/grad_norm_pre_clip': 2.3742542266845703, 'timing/data': 0.0003454722464084625, 'timing/model': 0.8308036476373672, 'learning_rate/qwen_vl_interface': 9.763958090688165e-06, 'learning_rate/action_model': 9.763958090688164e-05, 'epoch': 0.64})
+[2026-07-28 20:45:47,101][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.04628370329737663, 'train/grad_norm_pre_clip': 1.0381639003753662, 'timing/data': 0.00039374642074108124, 'timing/model': 0.6203370410948992, 'learning_rate/qwen_vl_interface': 9.762765444304737e-06, 'learning_rate/action_model': 9.762765444304737e-05, 'epoch': 0.64})
+[2026-07-28 20:45:56,235][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.02976180799305439, 'train/grad_norm_pre_clip': 0.7709885239601135, 'timing/data': 0.000369025394320488, 'timing/model': 0.6193774193525314, 'learning_rate/qwen_vl_interface': 9.761569869636808e-06, 'learning_rate/action_model': 9.761569869636808e-05, 'epoch': 0.64})
+[2026-07-28 20:46:05,368][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.027872741222381592, 'train/grad_norm_pre_clip': 1.0435192584991455, 'timing/data': 0.00033789873123168945, 'timing/model': 0.619699751958251, 'learning_rate/qwen_vl_interface': 9.760371367460172e-06, 'learning_rate/action_model': 9.760371367460171e-05, 'epoch': 0.64})
+[2026-07-28 20:46:14,735][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.006766809616237879, 'train/grad_norm_pre_clip': 0.3182358741760254, 'timing/data': 0.00038886629045009613, 'timing/model': 0.6221738643944263, 'learning_rate/qwen_vl_interface': 9.759169938552524e-06, 'learning_rate/action_model': 9.759169938552524e-05, 'epoch': 0.64})
+[2026-07-28 20:46:24,088][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.039669640362262726, 'train/grad_norm_pre_clip': 1.4013314247131348, 'timing/data': 0.0003206860274076462, 'timing/model': 0.6186795104295015, 'learning_rate/qwen_vl_interface': 9.757965583693458e-06, 'learning_rate/action_model': 9.757965583693457e-05, 'epoch': 0.64})
+[2026-07-28 20:46:33,189][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.02956811711192131, 'train/grad_norm_pre_clip': 1.244969129562378, 'timing/data': 0.0003928598016500473, 'timing/model': 0.6215303465723991, 'learning_rate/qwen_vl_interface': 9.756758303664467e-06, 'learning_rate/action_model': 9.756758303664468e-05, 'epoch': 0.64})
+[2026-07-28 20:46:42,292][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3500
+[2026-07-28 20:46:52,651][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3500
+[2026-07-28 21:04:16,841][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3500
+[2026-07-28 21:04:16,846][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.018188705667853355, 'train/grad_norm_pre_clip': 0.5791876316070557, 'eval/action_loss/seconds': 10.35595003142953, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.03657408803701401, 'eval/latency_3/loss': 0.03657408803701401, 'eval/flappy/loss': 0.03657408803701401, 'eval/flappy/latency_3/loss': 0.03657408803701401, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 416.78500490002335, 'rl_games_eval/mid_train/mean_length': 3376.7, 'rl_games_eval/mid_train/std_reward': 63.41565555134738, 'rl_games_eval/mid_train/std_length': 510.38819539640605, 'rl_games_eval/mid_train/macro_mean_reward': 416.78500490002335, 'rl_games_eval/mid_train/macro_mean_length': 3376.7, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 416.78500490002335, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 3376.7, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 63.41565555134738, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 510.38819539640605, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 416.78500490002335, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0004184246063232422, 'timing/model': 0.6201035678386688, 'learning_rate/qwen_vl_interface': 9.755548099248942e-06, 'learning_rate/action_model': 9.75554809924894e-05, 'epoch': 0.64})
+[2026-07-28 21:04:16,848][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3500_state
+[2026-07-28 21:04:16,848][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 21:04:54,756][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3500_state/pytorch_model
+[2026-07-28 21:04:54,756][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3500_state/sampler.bin
+[2026-07-28 21:04:54,757][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3500_state/sampler_1.bin
+[2026-07-28 21:04:54,760][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_3500_state/random_states_0.pkl
+[2026-07-28 21:05:02,848][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500
+[2026-07-28 21:05:02,848][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 21:05:02,869][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 21:05:12,515][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.028406275436282158, 'train/grad_norm_pre_clip': 1.1482826471328735, 'timing/data': 0.00046338140964508057, 'timing/model': 0.6316848248243332, 'learning_rate/qwen_vl_interface': 9.75433497123217e-06, 'learning_rate/action_model': 9.754334971232171e-05, 'epoch': 0.64})
+[2026-07-28 21:05:21,713][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.037178341299295425, 'train/grad_norm_pre_clip': 1.1236307621002197, 'timing/data': 0.0004484802484512329, 'timing/model': 0.6260432563722134, 'learning_rate/qwen_vl_interface': 9.753118920401341e-06, 'learning_rate/action_model': 9.75311892040134e-05, 'epoch': 0.64})
+[2026-07-28 21:05:30,799][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.021948982030153275, 'train/grad_norm_pre_clip': 0.8847002983093262, 'timing/data': 0.0003672465682029724, 'timing/model': 0.6436514183878899, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.64})
+[2026-07-28 21:05:39,737][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.010548793710768223, 'train/grad_norm_pre_clip': 0.5322908759117126, 'timing/data': 0.00035137496888637543, 'timing/model': 0.6241002827882767, 'learning_rate/qwen_vl_interface': 9.750678053455725e-06, 'learning_rate/action_model': 9.750678053455725e-05, 'epoch': 0.64})
+[2026-07-28 21:05:49,123][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.009228793904185295, 'train/grad_norm_pre_clip': 0.5707805752754211, 'timing/data': 0.0003513917326927185, 'timing/model': 0.860901340842247, 'learning_rate/qwen_vl_interface': 9.749453238924793e-06, 'learning_rate/action_model': 9.749453238924792e-05, 'epoch': 0.64})
+[2026-07-28 21:05:58,311][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.046361975371837616, 'train/grad_norm_pre_clip': 2.474433422088623, 'timing/data': 0.00042258016765117645, 'timing/model': 0.6365134604275227, 'learning_rate/qwen_vl_interface': 9.7482255047475e-06, 'learning_rate/action_model': 9.7482255047475e-05, 'epoch': 0.64})
+[2026-07-28 21:06:07,488][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.05116608366370201, 'train/grad_norm_pre_clip': 1.7876306772232056, 'timing/data': 0.00038279592990875244, 'timing/model': 0.6225849650800228, 'learning_rate/qwen_vl_interface': 9.746994851720518e-06, 'learning_rate/action_model': 9.746994851720517e-05, 'epoch': 0.64})
+[2026-07-28 21:06:16,681][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.0223393552005291, 'train/grad_norm_pre_clip': 0.951590895652771, 'timing/data': 0.0004800688475370407, 'timing/model': 0.621727954596281, 'learning_rate/qwen_vl_interface': 9.745761280642397e-06, 'learning_rate/action_model': 9.745761280642397e-05, 'epoch': 0.64})
+[2026-07-28 21:06:25,914][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.015832286328077316, 'train/grad_norm_pre_clip': 0.7035467028617859, 'timing/data': 0.0004986114799976349, 'timing/model': 0.6238574106246233, 'learning_rate/qwen_vl_interface': 9.744524792313591e-06, 'learning_rate/action_model': 9.74452479231359e-05, 'epoch': 0.64})
+[2026-07-28 21:06:35,329][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.03232496231794357, 'train/grad_norm_pre_clip': 1.0186353921890259, 'timing/data': 0.0003728587180376053, 'timing/model': 0.8494750317186117, 'learning_rate/qwen_vl_interface': 9.743285387536445e-06, 'learning_rate/action_model': 9.743285387536444e-05, 'epoch': 0.64})
+[2026-07-28 21:06:44,506][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.04924123361706734, 'train/grad_norm_pre_clip': 1.766412377357483, 'timing/data': 0.00035701319575309753, 'timing/model': 0.6221179533749819, 'learning_rate/qwen_vl_interface': 9.742043067115193e-06, 'learning_rate/action_model': 9.742043067115192e-05, 'epoch': 0.64})
+[2026-07-28 21:06:53,897][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.026104368269443512, 'train/grad_norm_pre_clip': 0.8300052285194397, 'timing/data': 0.00036049075424671173, 'timing/model': 0.621434273198247, 'learning_rate/qwen_vl_interface': 9.740797831855965e-06, 'learning_rate/action_model': 9.740797831855965e-05, 'epoch': 0.65})
+[2026-07-28 21:07:03,066][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.05687941610813141, 'train/grad_norm_pre_clip': 1.415794014930725, 'timing/data': 0.000362202525138855, 'timing/model': 0.6213073469698429, 'learning_rate/qwen_vl_interface': 9.73954968256678e-06, 'learning_rate/action_model': 9.73954968256678e-05, 'epoch': 0.65})
+[2026-07-28 21:07:12,226][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.02579810470342636, 'train/grad_norm_pre_clip': 1.1549838781356812, 'timing/data': 0.000375356525182724, 'timing/model': 0.6216662880033255, 'learning_rate/qwen_vl_interface': 9.738298620057547e-06, 'learning_rate/action_model': 9.738298620057546e-05, 'epoch': 0.65})
+[2026-07-28 21:07:21,435][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.046991586685180664, 'train/grad_norm_pre_clip': 1.1075650453567505, 'timing/data': 0.00038733892142772675, 'timing/model': 0.623522786423564, 'learning_rate/qwen_vl_interface': 9.73704464514007e-06, 'learning_rate/action_model': 9.737044645140068e-05, 'epoch': 0.65})
+[2026-07-28 21:07:31,142][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.021875161677598953, 'train/grad_norm_pre_clip': 1.0006906986236572, 'timing/data': 0.00036393478512763977, 'timing/model': 0.8822247534990311, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.65})
+[2026-07-28 21:07:40,324][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.042386963963508606, 'train/grad_norm_pre_clip': 1.267127513885498, 'timing/data': 0.0007627159357070923, 'timing/model': 0.621406827121973, 'learning_rate/qwen_vl_interface': 9.734527961337028e-06, 'learning_rate/action_model': 9.734527961337028e-05, 'epoch': 0.65})
+[2026-07-28 21:07:49,493][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.025959521532058716, 'train/grad_norm_pre_clip': 0.7848978638648987, 'timing/data': 0.0003636647015810013, 'timing/model': 0.6229111086577177, 'learning_rate/qwen_vl_interface': 9.733265254084514e-06, 'learning_rate/action_model': 9.733265254084513e-05, 'epoch': 0.65})
+[2026-07-28 21:07:58,945][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.0474068745970726, 'train/grad_norm_pre_clip': 1.5740612745285034, 'timing/data': 0.00035269372165203094, 'timing/model': 0.6383206583559513, 'learning_rate/qwen_vl_interface': 9.73199963768985e-06, 'learning_rate/action_model': 9.731999637689849e-05, 'epoch': 0.65})
+[2026-07-28 21:08:08,178][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.0617268905043602, 'train/grad_norm_pre_clip': 1.8495259284973145, 'timing/data': 0.0003694836050271988, 'timing/model': 0.6217441018670797, 'learning_rate/qwen_vl_interface': 9.73073111297428e-06, 'learning_rate/action_model': 9.73073111297428e-05, 'epoch': 0.65})
+[2026-07-28 21:08:17,648][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.044814109802246094, 'train/grad_norm_pre_clip': 0.88355952501297, 'timing/data': 0.0003709085285663605, 'timing/model': 0.6240076255053282, 'learning_rate/qwen_vl_interface': 9.72945968076094e-06, 'learning_rate/action_model': 9.729459680760939e-05, 'epoch': 0.65})
+[2026-07-28 21:08:27,023][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.010955357924103737, 'train/grad_norm_pre_clip': 0.6149879693984985, 'timing/data': 0.0004968941211700439, 'timing/model': 0.7949536610394716, 'learning_rate/qwen_vl_interface': 9.728185341874844e-06, 'learning_rate/action_model': 9.728185341874844e-05, 'epoch': 0.65})
+[2026-07-28 21:08:36,246][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.044328927993774414, 'train/grad_norm_pre_clip': 1.303694725036621, 'timing/data': 0.0003701932728290558, 'timing/model': 0.6216977089643478, 'learning_rate/qwen_vl_interface': 9.7269080971429e-06, 'learning_rate/action_model': 9.726908097142899e-05, 'epoch': 0.65})
+[2026-07-28 21:08:45,448][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.06780858337879181, 'train/grad_norm_pre_clip': 1.9056460857391357, 'timing/data': 0.00035286135971546173, 'timing/model': 0.6212332844734192, 'learning_rate/qwen_vl_interface': 9.725627947393896e-06, 'learning_rate/action_model': 9.725627947393896e-05, 'epoch': 0.65})
+[2026-07-28 21:08:54,922][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.027214009314775467, 'train/grad_norm_pre_clip': 0.7198732495307922, 'timing/data': 0.0003617890179157257, 'timing/model': 0.8604654259979725, 'learning_rate/qwen_vl_interface': 9.72434489345851e-06, 'learning_rate/action_model': 9.72434489345851e-05, 'epoch': 0.65})
+[2026-07-28 21:09:04,350][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.028932610526680946, 'train/grad_norm_pre_clip': 0.7500515580177307, 'timing/data': 0.0003712158650159836, 'timing/model': 0.6302051488310099, 'learning_rate/qwen_vl_interface': 9.723058936169298e-06, 'learning_rate/action_model': 9.723058936169297e-05, 'epoch': 0.65})
+[2026-07-28 21:09:13,525][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.04370056092739105, 'train/grad_norm_pre_clip': 1.5172793865203857, 'timing/data': 0.0003459993749856949, 'timing/model': 0.6218118499964476, 'learning_rate/qwen_vl_interface': 9.721770076360707e-06, 'learning_rate/action_model': 9.721770076360706e-05, 'epoch': 0.65})
+[2026-07-28 21:09:22,696][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.020787829533219337, 'train/grad_norm_pre_clip': 0.6979948282241821, 'timing/data': 0.0003436971455812454, 'timing/model': 0.6217277850955725, 'learning_rate/qwen_vl_interface': 9.720478314869061e-06, 'learning_rate/action_model': 9.72047831486906e-05, 'epoch': 0.65})
+[2026-07-28 21:09:32,036][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.044103678315877914, 'train/grad_norm_pre_clip': 1.3430070877075195, 'timing/data': 0.000365380197763443, 'timing/model': 0.9955615159124136, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.65})
+[2026-07-28 21:09:41,239][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.048991166055202484, 'train/grad_norm_pre_clip': 0.924461305141449, 'timing/data': 0.0004708096385002136, 'timing/model': 0.6238399464637041, 'learning_rate/qwen_vl_interface': 9.717886090191334e-06, 'learning_rate/action_model': 9.717886090191333e-05, 'epoch': 0.65})
+[2026-07-28 21:09:50,487][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.024069778621196747, 'train/grad_norm_pre_clip': 1.014634132385254, 'timing/data': 0.00036374107003211975, 'timing/model': 0.6223010085523129, 'learning_rate/qwen_vl_interface': 9.716585628687317e-06, 'learning_rate/action_model': 9.716585628687316e-05, 'epoch': 0.65})
+[2026-07-28 21:09:59,665][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.025091810151934624, 'train/grad_norm_pre_clip': 0.8157221674919128, 'timing/data': 0.00037193670868873596, 'timing/model': 0.6243244390934706, 'learning_rate/qwen_vl_interface': 9.715282268864379e-06, 'learning_rate/action_model': 9.715282268864378e-05, 'epoch': 0.65})
+[2026-07-28 21:10:09,522][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.02005786821246147, 'train/grad_norm_pre_clip': 0.5796322822570801, 'timing/data': 0.0004374980926513672, 'timing/model': 1.0388378165662289, 'learning_rate/qwen_vl_interface': 9.713976011568256e-06, 'learning_rate/action_model': 9.713976011568256e-05, 'epoch': 0.65})
+[2026-07-28 21:10:19,144][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.013612342067062855, 'train/grad_norm_pre_clip': 0.8096933960914612, 'timing/data': 0.0005851667374372482, 'timing/model': 0.6342308931052685, 'learning_rate/qwen_vl_interface': 9.712666857646565e-06, 'learning_rate/action_model': 9.712666857646564e-05, 'epoch': 0.65})
+[2026-07-28 21:10:28,732][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.026219572871923447, 'train/grad_norm_pre_clip': 0.9094300866127014, 'timing/data': 0.0004624258726835251, 'timing/model': 0.6454762797802687, 'learning_rate/qwen_vl_interface': 9.711354807948798e-06, 'learning_rate/action_model': 9.711354807948798e-05, 'epoch': 0.65})
+[2026-07-28 21:10:38,114][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.019994204863905907, 'train/grad_norm_pre_clip': 0.7505273222923279, 'timing/data': 0.00040605850517749786, 'timing/model': 0.626307487487793, 'learning_rate/qwen_vl_interface': 9.710039863326334e-06, 'learning_rate/action_model': 9.710039863326334e-05, 'epoch': 0.65})
+[2026-07-28 21:10:47,389][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.030284255743026733, 'train/grad_norm_pre_clip': 1.6912871599197388, 'timing/data': 0.0006522387266159058, 'timing/model': 0.6296205427497625, 'learning_rate/qwen_vl_interface': 9.708722024632425e-06, 'learning_rate/action_model': 9.708722024632424e-05, 'epoch': 0.65})
+[2026-07-28 21:10:57,056][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.036708440631628036, 'train/grad_norm_pre_clip': 1.2499847412109375, 'timing/data': 0.0005082916468381882, 'timing/model': 0.9452612083405256, 'learning_rate/qwen_vl_interface': 9.7074012927222e-06, 'learning_rate/action_model': 9.7074012927222e-05, 'epoch': 0.65})
+[2026-07-28 21:11:06,775][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.008785163052380085, 'train/grad_norm_pre_clip': 0.8322905898094177, 'timing/data': 0.0004698745906352997, 'timing/model': 0.8574553225189447, 'learning_rate/qwen_vl_interface': 9.706077668452671e-06, 'learning_rate/action_model': 9.706077668452671e-05, 'epoch': 0.65})
+[2026-07-28 21:11:16,204][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.045573458075523376, 'train/grad_norm_pre_clip': 1.0840644836425781, 'timing/data': 0.000465957447886467, 'timing/model': 0.6252177823334932, 'learning_rate/qwen_vl_interface': 9.704751152682722e-06, 'learning_rate/action_model': 9.704751152682721e-05, 'epoch': 0.65})
+[2026-07-28 21:11:25,243][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.03129899501800537, 'train/grad_norm_pre_clip': 2.1336584091186523, 'timing/data': 0.0004297439008951187, 'timing/model': 0.6210571378469467, 'learning_rate/qwen_vl_interface': 9.703421746273115e-06, 'learning_rate/action_model': 9.703421746273114e-05, 'epoch': 0.65})
+[2026-07-28 21:11:34,483][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.03345317021012306, 'train/grad_norm_pre_clip': 0.9374871850013733, 'timing/data': 0.00041082315146923065, 'timing/model': 0.6228806748986244, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.65})
+[2026-07-28 21:11:43,789][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.025877144187688828, 'train/grad_norm_pre_clip': 1.3759554624557495, 'timing/data': 0.0004986170679330826, 'timing/model': 0.6311550885438919, 'learning_rate/qwen_vl_interface': 9.700754264987347e-06, 'learning_rate/action_model': 9.700754264987347e-05, 'epoch': 0.65})
+[2026-07-28 21:11:53,312][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.03474611043930054, 'train/grad_norm_pre_clip': 1.3930708169937134, 'timing/data': 0.0005553737282752991, 'timing/model': 0.6245700158178806, 'learning_rate/qwen_vl_interface': 9.69941619184209e-06, 'learning_rate/action_model': 9.699416191842089e-05, 'epoch': 0.65})
+[2026-07-28 21:12:02,922][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.04735732823610306, 'train/grad_norm_pre_clip': 1.6656996011734009, 'timing/data': 0.0005106981843709946, 'timing/model': 0.6278343796730042, 'learning_rate/qwen_vl_interface': 9.69807523151897e-06, 'learning_rate/action_model': 9.698075231518969e-05, 'epoch': 0.65})
+[2026-07-28 21:12:12,225][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.03235209360718727, 'train/grad_norm_pre_clip': 1.2799880504608154, 'timing/data': 0.0003802906721830368, 'timing/model': 0.6288050953298807, 'learning_rate/qwen_vl_interface': 9.696731384888125e-06, 'learning_rate/action_model': 9.696731384888125e-05, 'epoch': 0.65})
+[2026-07-28 21:12:21,756][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.01503564789891243, 'train/grad_norm_pre_clip': 0.6240215301513672, 'timing/data': 0.0005482863634824753, 'timing/model': 0.6268320642411709, 'learning_rate/qwen_vl_interface': 9.695384652821562e-06, 'learning_rate/action_model': 9.695384652821561e-05, 'epoch': 0.65})
+[2026-07-28 21:12:31,082][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.04527537524700165, 'train/grad_norm_pre_clip': 1.2780506610870361, 'timing/data': 0.0005615521222352982, 'timing/model': 0.6385887246578932, 'learning_rate/qwen_vl_interface': 9.694035036193162e-06, 'learning_rate/action_model': 9.694035036193161e-05, 'epoch': 0.65})
+[2026-07-28 21:12:40,536][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.02356424741446972, 'train/grad_norm_pre_clip': 0.6593958139419556, 'timing/data': 0.00037384033203125, 'timing/model': 0.6330783609300852, 'learning_rate/qwen_vl_interface': 9.692682535878674e-06, 'learning_rate/action_model': 9.692682535878673e-05, 'epoch': 0.65})
+[2026-07-28 21:12:50,170][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.017315881326794624, 'train/grad_norm_pre_clip': 0.9101271033287048, 'timing/data': 0.0005651712417602539, 'timing/model': 0.6343777999281883, 'learning_rate/qwen_vl_interface': 9.691327152755721e-06, 'learning_rate/action_model': 9.691327152755722e-05, 'epoch': 0.65})
+[2026-07-28 21:12:59,802][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.029316378757357597, 'train/grad_norm_pre_clip': 0.9369004964828491, 'timing/data': 0.0003695636987686157, 'timing/model': 0.6234611831605434, 'learning_rate/qwen_vl_interface': 9.6899688877038e-06, 'learning_rate/action_model': 9.689968887703798e-05, 'epoch': 0.65})
+[2026-07-28 21:13:09,071][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.02115543931722641, 'train/grad_norm_pre_clip': 0.9506431818008423, 'timing/data': 0.0006019286811351776, 'timing/model': 0.6250570565462112, 'learning_rate/qwen_vl_interface': 9.688607741604269e-06, 'learning_rate/action_model': 9.688607741604269e-05, 'epoch': 0.65})
+[2026-07-28 21:13:18,522][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.04759715497493744, 'train/grad_norm_pre_clip': 1.509399652481079, 'timing/data': 0.0003443453460931778, 'timing/model': 0.6223361622542143, 'learning_rate/qwen_vl_interface': 9.687243715340365e-06, 'learning_rate/action_model': 9.687243715340364e-05, 'epoch': 0.65})
+[2026-07-28 21:13:28,290][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.025293026119470596, 'train/grad_norm_pre_clip': 0.8357651233673096, 'timing/data': 0.00037183240056037903, 'timing/model': 0.6179515272378922, 'learning_rate/qwen_vl_interface': 9.685876809797188e-06, 'learning_rate/action_model': 9.685876809797188e-05, 'epoch': 0.65})
+[2026-07-28 21:13:37,768][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.02135249227285385, 'train/grad_norm_pre_clip': 0.9864588975906372, 'timing/data': 0.0003843773156404495, 'timing/model': 0.6218667887151241, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.65})
+[2026-07-28 21:13:46,974][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.03351679816842079, 'train/grad_norm_pre_clip': 1.3762586116790771, 'timing/data': 0.0004203617572784424, 'timing/model': 0.6199240200221539, 'learning_rate/qwen_vl_interface': 9.683134364422766e-06, 'learning_rate/action_model': 9.683134364422765e-05, 'epoch': 0.65})
+[2026-07-28 21:13:55,947][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.022605981677770615, 'train/grad_norm_pre_clip': 0.7437821626663208, 'timing/data': 0.00042195431888103485, 'timing/model': 0.619694322347641, 'learning_rate/qwen_vl_interface': 9.681758826371064e-06, 'learning_rate/action_model': 9.681758826371063e-05, 'epoch': 0.65})
+[2026-07-28 21:14:05,189][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.032996803522109985, 'train/grad_norm_pre_clip': 1.6781712770462036, 'timing/data': 0.00041837990283966064, 'timing/model': 0.6209964528679848, 'learning_rate/qwen_vl_interface': 9.680380412599174e-06, 'learning_rate/action_model': 9.680380412599173e-05, 'epoch': 0.65})
+[2026-07-28 21:14:14,961][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.026583002880215645, 'train/grad_norm_pre_clip': 0.7228108644485474, 'timing/data': 0.0003617517650127411, 'timing/model': 0.8731389474123716, 'learning_rate/qwen_vl_interface': 9.678999124001533e-06, 'learning_rate/action_model': 9.678999124001533e-05, 'epoch': 0.65})
+[2026-07-28 21:14:24,171][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.051293663680553436, 'train/grad_norm_pre_clip': 1.1222537755966187, 'timing/data': 0.0004310011863708496, 'timing/model': 0.6189411077648401, 'learning_rate/qwen_vl_interface': 9.67761496147445e-06, 'learning_rate/action_model': 9.677614961474449e-05, 'epoch': 0.65})
+[2026-07-28 21:14:33,462][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.05262012779712677, 'train/grad_norm_pre_clip': 1.4508863687515259, 'timing/data': 0.00034863129258155823, 'timing/model': 0.6192430760711432, 'learning_rate/qwen_vl_interface': 9.676227925916082e-06, 'learning_rate/action_model': 9.676227925916082e-05, 'epoch': 0.65})
+[2026-07-28 21:14:42,693][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.012040680274367332, 'train/grad_norm_pre_clip': 0.7299342155456543, 'timing/data': 0.00045256875455379486, 'timing/model': 0.6206335425376892, 'learning_rate/qwen_vl_interface': 9.674838018226472e-06, 'learning_rate/action_model': 9.674838018226471e-05, 'epoch': 0.65})
+[2026-07-28 21:14:52,154][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.08843282610177994, 'train/grad_norm_pre_clip': 1.4032750129699707, 'timing/data': 0.0003553684800863266, 'timing/model': 0.62027857452631, 'learning_rate/qwen_vl_interface': 9.67344523930751e-06, 'learning_rate/action_model': 9.673445239307511e-05, 'epoch': 0.65})
+[2026-07-28 21:15:01,321][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.02371928095817566, 'train/grad_norm_pre_clip': 1.5294170379638672, 'timing/data': 0.0004515126347541809, 'timing/model': 0.6209620125591755, 'learning_rate/qwen_vl_interface': 9.672049590062959e-06, 'learning_rate/action_model': 9.672049590062958e-05, 'epoch': 0.65})
+[2026-07-28 21:15:10,743][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.05539565905928612, 'train/grad_norm_pre_clip': 1.8151980638504028, 'timing/data': 0.00034080445766448975, 'timing/model': 0.6266760099679232, 'learning_rate/qwen_vl_interface': 9.670651071398436e-06, 'learning_rate/action_model': 9.670651071398436e-05, 'epoch': 0.65})
+[2026-07-28 21:15:19,883][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.09720723330974579, 'train/grad_norm_pre_clip': 2.273883819580078, 'timing/data': 0.0003710538148880005, 'timing/model': 0.6193202678114176, 'learning_rate/qwen_vl_interface': 9.669249684221427e-06, 'learning_rate/action_model': 9.669249684221428e-05, 'epoch': 0.65})
+[2026-07-28 21:15:29,074][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.0056715719401836395, 'train/grad_norm_pre_clip': 0.49608466029167175, 'timing/data': 0.0003562718629837036, 'timing/model': 0.6242198906838894, 'learning_rate/qwen_vl_interface': 9.66784542944128e-06, 'learning_rate/action_model': 9.66784542944128e-05, 'epoch': 0.66})
+[2026-07-28 21:15:38,437][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.029901959002017975, 'train/grad_norm_pre_clip': 2.389078378677368, 'timing/data': 0.0003511980175971985, 'timing/model': 0.618519889190793, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.66})
+[2026-07-28 21:15:47,822][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.017843417823314667, 'train/grad_norm_pre_clip': 0.7993015646934509, 'timing/data': 0.0004245098680257797, 'timing/model': 0.8502132929861546, 'learning_rate/qwen_vl_interface': 9.66502832071824e-06, 'learning_rate/action_model': 9.66502832071824e-05, 'epoch': 0.66})
+[2026-07-28 21:15:57,023][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.061219699680805206, 'train/grad_norm_pre_clip': 2.6414053440093994, 'timing/data': 0.00034373626112937927, 'timing/model': 0.6205231882631779, 'learning_rate/qwen_vl_interface': 9.663615468603344e-06, 'learning_rate/action_model': 9.663615468603343e-05, 'epoch': 0.66})
+[2026-07-28 21:16:06,174][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.029618699103593826, 'train/grad_norm_pre_clip': 0.873842179775238, 'timing/data': 0.00037005357444286346, 'timing/model': 0.6266042925417423, 'learning_rate/qwen_vl_interface': 9.662199752541287e-06, 'learning_rate/action_model': 9.662199752541287e-05, 'epoch': 0.66})
+[2026-07-28 21:16:15,323][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.035430729389190674, 'train/grad_norm_pre_clip': 1.2041950225830078, 'timing/data': 0.00036824867129325867, 'timing/model': 0.6226243656128645, 'learning_rate/qwen_vl_interface': 9.660781173450711e-06, 'learning_rate/action_model': 9.660781173450712e-05, 'epoch': 0.66})
+[2026-07-28 21:16:24,743][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.04077466204762459, 'train/grad_norm_pre_clip': 1.6190987825393677, 'timing/data': 0.0003514457494020462, 'timing/model': 0.620094396173954, 'learning_rate/qwen_vl_interface': 9.659359732252121e-06, 'learning_rate/action_model': 9.659359732252121e-05, 'epoch': 0.66})
+[2026-07-28 21:16:33,950][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.024010363966226578, 'train/grad_norm_pre_clip': 0.7795193195343018, 'timing/data': 0.0003375578671693802, 'timing/model': 0.622523732483387, 'learning_rate/qwen_vl_interface': 9.657935429867872e-06, 'learning_rate/action_model': 9.657935429867873e-05, 'epoch': 0.66})
+[2026-07-28 21:16:43,233][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.028016824275255203, 'train/grad_norm_pre_clip': 1.9005557298660278, 'timing/data': 0.00035427697002887726, 'timing/model': 0.6226111222058535, 'learning_rate/qwen_vl_interface': 9.65650826722218e-06, 'learning_rate/action_model': 9.656508267222179e-05, 'epoch': 0.66})
+[2026-07-28 21:16:52,425][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.03304169699549675, 'train/grad_norm_pre_clip': 2.5948448181152344, 'timing/data': 0.00043344125151634216, 'timing/model': 0.623574698343873, 'learning_rate/qwen_vl_interface': 9.655078245241114e-06, 'learning_rate/action_model': 9.655078245241112e-05, 'epoch': 0.66})
+[2026-07-28 21:17:01,910][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.07528501749038696, 'train/grad_norm_pre_clip': 2.0696403980255127, 'timing/data': 0.00036127492785453796, 'timing/model': 0.8741006478667259, 'learning_rate/qwen_vl_interface': 9.653645364852598e-06, 'learning_rate/action_model': 9.653645364852597e-05, 'epoch': 0.66})
+[2026-07-28 21:17:11,583][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.04082479327917099, 'train/grad_norm_pre_clip': 1.4930436611175537, 'timing/data': 0.00038331374526023865, 'timing/model': 0.637605944648385, 'learning_rate/qwen_vl_interface': 9.652209626986414e-06, 'learning_rate/action_model': 9.652209626986413e-05, 'epoch': 0.66})
+[2026-07-28 21:17:20,862][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.012274493463337421, 'train/grad_norm_pre_clip': 0.46183082461357117, 'timing/data': 0.00033495761454105377, 'timing/model': 0.6258625444024801, 'learning_rate/qwen_vl_interface': 9.6507710325742e-06, 'learning_rate/action_model': 9.6507710325742e-05, 'epoch': 0.66})
+[2026-07-28 21:17:30,086][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.012430170550942421, 'train/grad_norm_pre_clip': 0.9130709171295166, 'timing/data': 0.00041443854570388794, 'timing/model': 0.6261110603809357, 'learning_rate/qwen_vl_interface': 9.649329582549439e-06, 'learning_rate/action_model': 9.649329582549439e-05, 'epoch': 0.66})
+[2026-07-28 21:17:39,589][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.01666746847331524, 'train/grad_norm_pre_clip': 0.7729865312576294, 'timing/data': 0.0004959572106599808, 'timing/model': 0.6510509140789509, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.66})
+[2026-07-28 21:17:49,561][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.018957844004034996, 'train/grad_norm_pre_clip': 1.3816179037094116, 'timing/data': 0.0004635807126760483, 'timing/model': 0.6301299575716257, 'learning_rate/qwen_vl_interface': 9.646438119405503e-06, 'learning_rate/action_model': 9.646438119405503e-05, 'epoch': 0.66})
+[2026-07-28 21:17:59,097][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.013345958665013313, 'train/grad_norm_pre_clip': 0.8131404519081116, 'timing/data': 0.0005363281816244125, 'timing/model': 0.6298744976520538, 'learning_rate/qwen_vl_interface': 9.644988108162566e-06, 'learning_rate/action_model': 9.644988108162567e-05, 'epoch': 0.66})
+[2026-07-28 21:18:08,490][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.050767265260219574, 'train/grad_norm_pre_clip': 2.5148539543151855, 'timing/data': 0.00048227980732917786, 'timing/model': 0.6263470984995365, 'learning_rate/qwen_vl_interface': 9.643535245059562e-06, 'learning_rate/action_model': 9.643535245059561e-05, 'epoch': 0.66})
+[2026-07-28 21:18:17,831][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.01900050789117813, 'train/grad_norm_pre_clip': 1.1013599634170532, 'timing/data': 0.00037434324622154236, 'timing/model': 0.6260343603789806, 'learning_rate/qwen_vl_interface': 9.642079531039235e-06, 'learning_rate/action_model': 9.642079531039235e-05, 'epoch': 0.66})
+[2026-07-28 21:18:27,469][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.0075519997626543045, 'train/grad_norm_pre_clip': 0.6597808003425598, 'timing/data': 0.000493885949254036, 'timing/model': 0.934859674423933, 'learning_rate/qwen_vl_interface': 9.640620967046186e-06, 'learning_rate/action_model': 9.640620967046186e-05, 'epoch': 0.66})
+[2026-07-28 21:18:36,740][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.0016912801656872034, 'train/grad_norm_pre_clip': 0.17812564969062805, 'timing/data': 0.00042948126792907715, 'timing/model': 0.6220641639083624, 'learning_rate/qwen_vl_interface': 9.63915955402686e-06, 'learning_rate/action_model': 9.63915955402686e-05, 'epoch': 0.66})
+[2026-07-28 21:18:46,223][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.01954852044582367, 'train/grad_norm_pre_clip': 1.4203190803527832, 'timing/data': 0.000434044748544693, 'timing/model': 0.6233421135693789, 'learning_rate/qwen_vl_interface': 9.637695292929552e-06, 'learning_rate/action_model': 9.637695292929552e-05, 'epoch': 0.66})
+[2026-07-28 21:18:55,473][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.013937829062342644, 'train/grad_norm_pre_clip': 0.9445067048072815, 'timing/data': 0.0003954283893108368, 'timing/model': 0.6222365070134401, 'learning_rate/qwen_vl_interface': 9.636228184704405e-06, 'learning_rate/action_model': 9.636228184704405e-05, 'epoch': 0.66})
+[2026-07-28 21:19:04,692][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.032686952501535416, 'train/grad_norm_pre_clip': 1.8580670356750488, 'timing/data': 0.00042030587792396545, 'timing/model': 0.6312755346298218, 'learning_rate/qwen_vl_interface': 9.634758230303408e-06, 'learning_rate/action_model': 9.634758230303408e-05, 'epoch': 0.66})
+[2026-07-28 21:19:14,486][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.07197301089763641, 'train/grad_norm_pre_clip': 4.692596912384033, 'timing/data': 0.0004052557051181793, 'timing/model': 0.6416531912982464, 'learning_rate/qwen_vl_interface': 9.633285430680401e-06, 'learning_rate/action_model': 9.633285430680401e-05, 'epoch': 0.66})
+[2026-07-28 21:19:24,025][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.04947374016046524, 'train/grad_norm_pre_clip': 1.5772429704666138, 'timing/data': 0.0004491601139307022, 'timing/model': 0.6279582194983959, 'learning_rate/qwen_vl_interface': 9.631809786791066e-06, 'learning_rate/action_model': 9.631809786791066e-05, 'epoch': 0.66})
+[2026-07-28 21:19:33,325][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.020885225385427475, 'train/grad_norm_pre_clip': 0.8190801739692688, 'timing/data': 0.00041220150887966156, 'timing/model': 0.6217477545142174, 'learning_rate/qwen_vl_interface': 9.630331299592932e-06, 'learning_rate/action_model': 9.630331299592933e-05, 'epoch': 0.66})
+[2026-07-28 21:19:42,593][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.039279185235500336, 'train/grad_norm_pre_clip': 1.8790377378463745, 'timing/data': 0.00041938386857509613, 'timing/model': 0.6257502771914005, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.66})
+[2026-07-28 21:19:52,108][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.023676693439483643, 'train/grad_norm_pre_clip': 1.103364109992981, 'timing/data': 0.00033342838287353516, 'timing/model': 0.8677474986761808, 'learning_rate/qwen_vl_interface': 9.627365799109617e-06, 'learning_rate/action_model': 9.627365799109617e-05, 'epoch': 0.66})
+[2026-07-28 21:20:01,500][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.05156201869249344, 'train/grad_norm_pre_clip': 2.2235333919525146, 'timing/data': 0.00037647970020771027, 'timing/model': 0.6231709904968739, 'learning_rate/qwen_vl_interface': 9.625878787748712e-06, 'learning_rate/action_model': 9.625878787748713e-05, 'epoch': 0.66})
+[2026-07-28 21:20:10,747][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.04894281178712845, 'train/grad_norm_pre_clip': 1.8777885437011719, 'timing/data': 0.0003314204514026642, 'timing/model': 0.6221284084022045, 'learning_rate/qwen_vl_interface': 9.624388936927572e-06, 'learning_rate/action_model': 9.624388936927571e-05, 'epoch': 0.66})
+[2026-07-28 21:20:19,999][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.008939879015088081, 'train/grad_norm_pre_clip': 0.5646393299102783, 'timing/data': 0.0003959089517593384, 'timing/model': 0.6225554663687944, 'learning_rate/qwen_vl_interface': 9.622896247612943e-06, 'learning_rate/action_model': 9.622896247612942e-05, 'epoch': 0.66})
+[2026-07-28 21:20:29,445][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.05161654204130173, 'train/grad_norm_pre_clip': 2.069423198699951, 'timing/data': 0.0003772098571062088, 'timing/model': 0.799126461148262, 'learning_rate/qwen_vl_interface': 9.621400720773414e-06, 'learning_rate/action_model': 9.621400720773414e-05, 'epoch': 0.66})
+[2026-07-28 21:20:38,952][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.042982377111911774, 'train/grad_norm_pre_clip': 1.1473441123962402, 'timing/data': 0.00042039528489112854, 'timing/model': 0.622786482796073, 'learning_rate/qwen_vl_interface': 9.619902357379416e-06, 'learning_rate/action_model': 9.619902357379415e-05, 'epoch': 0.66})
+[2026-07-28 21:20:48,211][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.05818168818950653, 'train/grad_norm_pre_clip': 2.3971922397613525, 'timing/data': 0.00034331344068050385, 'timing/model': 0.6213049087673426, 'learning_rate/qwen_vl_interface': 9.618401158403222e-06, 'learning_rate/action_model': 9.618401158403222e-05, 'epoch': 0.66})
+[2026-07-28 21:20:57,461][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.05429837107658386, 'train/grad_norm_pre_clip': 1.5051562786102295, 'timing/data': 0.00037123076617717743, 'timing/model': 0.6231906767934561, 'learning_rate/qwen_vl_interface': 9.616897124818944e-06, 'learning_rate/action_model': 9.616897124818945e-05, 'epoch': 0.66})
+[2026-07-28 21:21:06,980][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.05014915391802788, 'train/grad_norm_pre_clip': 2.2070751190185547, 'timing/data': 0.00035322830080986023, 'timing/model': 0.863512834534049, 'learning_rate/qwen_vl_interface': 9.615390257602532e-06, 'learning_rate/action_model': 9.615390257602531e-05, 'epoch': 0.66})
+[2026-07-28 21:21:16,225][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.009460401721298695, 'train/grad_norm_pre_clip': 0.6289004683494568, 'timing/data': 0.00032756850123405457, 'timing/model': 0.6231048423796892, 'learning_rate/qwen_vl_interface': 9.613880557731776e-06, 'learning_rate/action_model': 9.613880557731776e-05, 'epoch': 0.66})
+[2026-07-28 21:21:25,739][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.03140869736671448, 'train/grad_norm_pre_clip': 0.7119185328483582, 'timing/data': 0.0004146937280893326, 'timing/model': 0.8529555965214968, 'learning_rate/qwen_vl_interface': 9.612368026186304e-06, 'learning_rate/action_model': 9.612368026186304e-05, 'epoch': 0.66})
+[2026-07-28 21:21:34,976][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.053489573299884796, 'train/grad_norm_pre_clip': 1.9295456409454346, 'timing/data': 0.0003885868936777115, 'timing/model': 0.6219009831547737, 'learning_rate/qwen_vl_interface': 9.610852663947582e-06, 'learning_rate/action_model': 9.610852663947583e-05, 'epoch': 0.66})
+[2026-07-28 21:21:44,209][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.05404866486787796, 'train/grad_norm_pre_clip': 2.093285322189331, 'timing/data': 0.0003864802420139313, 'timing/model': 0.6220355033874512, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.66})
+[2026-07-28 21:21:53,462][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.03627074882388115, 'train/grad_norm_pre_clip': 0.5107072591781616, 'timing/data': 0.00037914328277111053, 'timing/model': 0.6228159498423338, 'learning_rate/qwen_vl_interface': 9.607813451325431e-06, 'learning_rate/action_model': 9.60781345132543e-05, 'epoch': 0.66})
+[2026-07-28 21:22:02,921][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.05403969809412956, 'train/grad_norm_pre_clip': 0.9220117330551147, 'timing/data': 0.000391373410820961, 'timing/model': 0.6218371726572514, 'learning_rate/qwen_vl_interface': 9.606289602914113e-06, 'learning_rate/action_model': 9.606289602914112e-05, 'epoch': 0.66})
+[2026-07-28 21:22:12,215][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.07961653918027878, 'train/grad_norm_pre_clip': 2.344245433807373, 'timing/data': 0.0003893468528985977, 'timing/model': 0.6232036147266626, 'learning_rate/qwen_vl_interface': 9.604762927753768e-06, 'learning_rate/action_model': 9.604762927753768e-05, 'epoch': 0.66})
+[2026-07-28 21:22:21,490][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.03556087240576744, 'train/grad_norm_pre_clip': 0.8386498689651489, 'timing/data': 0.0003457162529230118, 'timing/model': 0.6301951203495264, 'learning_rate/qwen_vl_interface': 9.603233426835037e-06, 'learning_rate/action_model': 9.603233426835036e-05, 'epoch': 0.66})
+[2026-07-28 21:22:30,864][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.022785894572734833, 'train/grad_norm_pre_clip': 0.4952383041381836, 'timing/data': 0.0003279242664575577, 'timing/model': 0.6209682989865541, 'learning_rate/qwen_vl_interface': 9.601701101150399e-06, 'learning_rate/action_model': 9.601701101150398e-05, 'epoch': 0.66})
+[2026-07-28 21:22:40,391][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.03980129957199097, 'train/grad_norm_pre_clip': 0.8234016299247742, 'timing/data': 0.00034076161682605743, 'timing/model': 0.6244426127523184, 'learning_rate/qwen_vl_interface': 9.600165951694162e-06, 'learning_rate/action_model': 9.60016595169416e-05, 'epoch': 0.66})
+[2026-07-28 21:22:49,647][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.04004737734794617, 'train/grad_norm_pre_clip': 1.0832887887954712, 'timing/data': 0.0003232192248106003, 'timing/model': 0.6210854053497314, 'learning_rate/qwen_vl_interface': 9.598627979462466e-06, 'learning_rate/action_model': 9.598627979462467e-05, 'epoch': 0.66})
+[2026-07-28 21:22:59,109][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.009058305993676186, 'train/grad_norm_pre_clip': 0.3973316550254822, 'timing/data': 0.00038024410605430603, 'timing/model': 0.8472475837916136, 'learning_rate/qwen_vl_interface': 9.59708718545329e-06, 'learning_rate/action_model': 9.597087185453288e-05, 'epoch': 0.66})
+[2026-07-28 21:23:08,380][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.021567516028881073, 'train/grad_norm_pre_clip': 0.5832030177116394, 'timing/data': 0.0004028845578432083, 'timing/model': 0.6237675473093987, 'learning_rate/qwen_vl_interface': 9.595543570666433e-06, 'learning_rate/action_model': 9.595543570666432e-05, 'epoch': 0.66})
+[2026-07-28 21:23:17,624][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.04202638566493988, 'train/grad_norm_pre_clip': 0.7103355526924133, 'timing/data': 0.0003258008509874344, 'timing/model': 0.6215907074511051, 'learning_rate/qwen_vl_interface': 9.593997136103533e-06, 'learning_rate/action_model': 9.593997136103533e-05, 'epoch': 0.66})
+[2026-07-28 21:23:26,843][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.034931961447000504, 'train/grad_norm_pre_clip': 0.7154964804649353, 'timing/data': 0.000405728816986084, 'timing/model': 0.6214774735271931, 'learning_rate/qwen_vl_interface': 9.592447882768053e-06, 'learning_rate/action_model': 9.592447882768052e-05, 'epoch': 0.66})
+[2026-07-28 21:23:36,069][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.015545211732387543, 'train/grad_norm_pre_clip': 0.5911709070205688, 'timing/data': 0.0003912895917892456, 'timing/model': 0.6210941094905138, 'learning_rate/qwen_vl_interface': 9.590895811665284e-06, 'learning_rate/action_model': 9.590895811665285e-05, 'epoch': 0.66})
+[2026-07-28 21:23:45,317][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.038129933178424835, 'train/grad_norm_pre_clip': 1.7007911205291748, 'timing/data': 0.0005094483494758606, 'timing/model': 0.6207487341016531, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.66})
+[2026-07-28 21:23:54,862][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.03536020219326019, 'train/grad_norm_pre_clip': 1.1375466585159302, 'timing/data': 0.0003449786454439163, 'timing/model': 0.6221741810441017, 'learning_rate/qwen_vl_interface': 9.587783220188212e-06, 'learning_rate/action_model': 9.587783220188212e-05, 'epoch': 0.67})
+[2026-07-28 21:24:04,354][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.021592218428850174, 'train/grad_norm_pre_clip': 1.1997238397598267, 'timing/data': 0.00034964270889759064, 'timing/model': 0.8806227799504995, 'learning_rate/qwen_vl_interface': 9.586222701833631e-06, 'learning_rate/action_model': 9.58622270183363e-05, 'epoch': 0.67})
+[2026-07-28 21:24:13,560][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.032289303839206696, 'train/grad_norm_pre_clip': 1.6184718608856201, 'timing/data': 0.0004009772092103958, 'timing/model': 0.6231822483241558, 'learning_rate/qwen_vl_interface': 9.584659369751216e-06, 'learning_rate/action_model': 9.584659369751215e-05, 'epoch': 0.67})
+[2026-07-28 21:24:22,799][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.01938110589981079, 'train/grad_norm_pre_clip': 0.9794424176216125, 'timing/data': 0.0004002656787633896, 'timing/model': 0.6216829214245081, 'learning_rate/qwen_vl_interface': 9.583093224955397e-06, 'learning_rate/action_model': 9.583093224955397e-05, 'epoch': 0.67})
+[2026-07-28 21:24:32,038][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.04165232554078102, 'train/grad_norm_pre_clip': 1.198930025100708, 'timing/data': 0.0005783569067716599, 'timing/model': 0.622032780200243, 'learning_rate/qwen_vl_interface': 9.58152426846243e-06, 'learning_rate/action_model': 9.581524268462429e-05, 'epoch': 0.67})
+[2026-07-28 21:24:41,521][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.029316555708646774, 'train/grad_norm_pre_clip': 1.6506584882736206, 'timing/data': 0.00037156976759433746, 'timing/model': 0.8431724589318037, 'learning_rate/qwen_vl_interface': 9.579952501290388e-06, 'learning_rate/action_model': 9.579952501290387e-05, 'epoch': 0.67})
+[2026-07-28 21:24:51,163][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.01908114366233349, 'train/grad_norm_pre_clip': 0.8164514899253845, 'timing/data': 0.00037011876702308655, 'timing/model': 0.6226156987249851, 'learning_rate/qwen_vl_interface': 9.57837792445918e-06, 'learning_rate/action_model': 9.578377924459181e-05, 'epoch': 0.67})
+[2026-07-28 21:25:00,481][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.04511179402470589, 'train/grad_norm_pre_clip': 3.8942713737487793, 'timing/data': 0.00040593743324279785, 'timing/model': 0.6219819039106369, 'learning_rate/qwen_vl_interface': 9.57680053899053e-06, 'learning_rate/action_model': 9.576800538990529e-05, 'epoch': 0.67})
+[2026-07-28 21:25:09,757][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.029352670535445213, 'train/grad_norm_pre_clip': 1.09530770778656, 'timing/data': 0.00039408542215824127, 'timing/model': 0.6226439401507378, 'learning_rate/qwen_vl_interface': 9.575220345907984e-06, 'learning_rate/action_model': 9.575220345907984e-05, 'epoch': 0.67})
+[2026-07-28 21:25:19,023][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.044517625123262405, 'train/grad_norm_pre_clip': 1.6262235641479492, 'timing/data': 0.0004251822829246521, 'timing/model': 0.6239096913486719, 'learning_rate/qwen_vl_interface': 9.573637346236915e-06, 'learning_rate/action_model': 9.573637346236915e-05, 'epoch': 0.67})
+[2026-07-28 21:25:28,506][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.015036683529615402, 'train/grad_norm_pre_clip': 0.6777108311653137, 'timing/data': 0.00033244118094444275, 'timing/model': 0.622401250526309, 'learning_rate/qwen_vl_interface': 9.572051541004513e-06, 'learning_rate/action_model': 9.572051541004513e-05, 'epoch': 0.67})
+[2026-07-28 21:25:37,967][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.013714885339140892, 'train/grad_norm_pre_clip': 1.5437285900115967, 'timing/data': 0.000370025634765625, 'timing/model': 0.624885831028223, 'learning_rate/qwen_vl_interface': 9.57046293123979e-06, 'learning_rate/action_model': 9.570462931239789e-05, 'epoch': 0.67})
+[2026-07-28 21:25:47,153][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.005799285136163235, 'train/grad_norm_pre_clip': 0.5270473957061768, 'timing/data': 0.0003683231770992279, 'timing/model': 0.6218065228313208, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.67})
+[2026-07-28 21:25:56,396][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.01558394730091095, 'train/grad_norm_pre_clip': 0.8703929781913757, 'timing/data': 0.0004308987408876419, 'timing/model': 0.6212937999516726, 'learning_rate/qwen_vl_interface': 9.567277302238524e-06, 'learning_rate/action_model': 9.567277302238524e-05, 'epoch': 0.67})
+[2026-07-28 21:26:05,914][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.02833280898630619, 'train/grad_norm_pre_clip': 1.16400146484375, 'timing/data': 0.0004036352038383484, 'timing/model': 0.8485485445708036, 'learning_rate/qwen_vl_interface': 9.565680285069104e-06, 'learning_rate/action_model': 9.565680285069103e-05, 'epoch': 0.67})
+[2026-07-28 21:26:15,448][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.034323982894420624, 'train/grad_norm_pre_clip': 1.0369807481765747, 'timing/data': 0.0003193262964487076, 'timing/model': 0.6243229676038027, 'learning_rate/qwen_vl_interface': 9.564080467501601e-06, 'learning_rate/action_model': 9.5640804675016e-05, 'epoch': 0.67})
+[2026-07-28 21:26:24,729][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.016844216734170914, 'train/grad_norm_pre_clip': 0.5641102194786072, 'timing/data': 0.0003868415951728821, 'timing/model': 0.6206196025013924, 'learning_rate/qwen_vl_interface': 9.562477850574118e-06, 'learning_rate/action_model': 9.562477850574119e-05, 'epoch': 0.67})
+[2026-07-28 21:26:33,990][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.05276751518249512, 'train/grad_norm_pre_clip': 5.634420871734619, 'timing/data': 0.00032966211438179016, 'timing/model': 0.6211773212999105, 'learning_rate/qwen_vl_interface': 9.56087243532658e-06, 'learning_rate/action_model': 9.56087243532658e-05, 'epoch': 0.67})
+[2026-07-28 21:26:43,415][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.024570154026150703, 'train/grad_norm_pre_clip': 1.0438240766525269, 'timing/data': 0.00036600418388843536, 'timing/model': 0.6203141938894987, 'learning_rate/qwen_vl_interface': 9.55926422280072e-06, 'learning_rate/action_model': 9.55926422280072e-05, 'epoch': 0.67})
+[2026-07-28 21:26:52,900][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.0329684317111969, 'train/grad_norm_pre_clip': 1.3314335346221924, 'timing/data': 0.0003077276051044464, 'timing/model': 0.6219589915126562, 'learning_rate/qwen_vl_interface': 9.55765321404009e-06, 'learning_rate/action_model': 9.557653214040091e-05, 'epoch': 0.67})
+[2026-07-28 21:27:02,149][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.011027111671864986, 'train/grad_norm_pre_clip': 0.5446917414665222, 'timing/data': 0.0003816932439804077, 'timing/model': 0.6215116735547781, 'learning_rate/qwen_vl_interface': 9.556039410090061e-06, 'learning_rate/action_model': 9.556039410090061e-05, 'epoch': 0.67})
+[2026-07-28 21:27:11,631][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.0913049727678299, 'train/grad_norm_pre_clip': 1.527004361152649, 'timing/data': 0.0003226622939109802, 'timing/model': 0.8601603657007217, 'learning_rate/qwen_vl_interface': 9.554422811997804e-06, 'learning_rate/action_model': 9.554422811997805e-05, 'epoch': 0.67})
+[2026-07-28 21:27:20,880][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.06878509372472763, 'train/grad_norm_pre_clip': 2.594573736190796, 'timing/data': 0.0003939922899007797, 'timing/model': 0.6212248206138611, 'learning_rate/qwen_vl_interface': 9.55280342081232e-06, 'learning_rate/action_model': 9.55280342081232e-05, 'epoch': 0.67})
+[2026-07-28 21:27:30,356][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.01056507509201765, 'train/grad_norm_pre_clip': 0.42403605580329895, 'timing/data': 0.0003204233944416046, 'timing/model': 0.6242524031549692, 'learning_rate/qwen_vl_interface': 9.55118123758441e-06, 'learning_rate/action_model': 9.551181237584409e-05, 'epoch': 0.67})
+[2026-07-28 21:27:39,555][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.014608710072934628, 'train/grad_norm_pre_clip': 0.8562584519386292, 'timing/data': 0.0003527868539094925, 'timing/model': 0.6217298377305269, 'learning_rate/qwen_vl_interface': 9.549556263366691e-06, 'learning_rate/action_model': 9.549556263366692e-05, 'epoch': 0.67})
+[2026-07-28 21:27:48,769][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.08021022379398346, 'train/grad_norm_pre_clip': 1.7350012063980103, 'timing/data': 0.00031897611916065216, 'timing/model': 0.6210082583129406, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.67})
+[2026-07-28 21:27:58,194][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.014475155621767044, 'train/grad_norm_pre_clip': 0.5768773555755615, 'timing/data': 0.00031364522874355316, 'timing/model': 0.6202639788389206, 'learning_rate/qwen_vl_interface': 9.546297946181354e-06, 'learning_rate/action_model': 9.546297946181354e-05, 'epoch': 0.67})
+[2026-07-28 21:28:07,653][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.07240070402622223, 'train/grad_norm_pre_clip': 4.071506023406982, 'timing/data': 0.00033809803426265717, 'timing/model': 0.6204660758376122, 'learning_rate/qwen_vl_interface': 9.544664605328022e-06, 'learning_rate/action_model': 9.544664605328022e-05, 'epoch': 0.67})
+[2026-07-28 21:28:16,674][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.037840791046619415, 'train/grad_norm_pre_clip': 1.5632225275039673, 'timing/data': 0.00032342784106731415, 'timing/model': 0.6204292010515928, 'learning_rate/qwen_vl_interface': 9.543028477713455e-06, 'learning_rate/action_model': 9.543028477713453e-05, 'epoch': 0.67})
+[2026-07-28 21:28:25,687][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.03840222209692001, 'train/grad_norm_pre_clip': 2.1901583671569824, 'timing/data': 0.0003148224204778671, 'timing/model': 0.6240674126893282, 'learning_rate/qwen_vl_interface': 9.541389564399316e-06, 'learning_rate/action_model': 9.541389564399315e-05, 'epoch': 0.67})
+[2026-07-28 21:28:35,061][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.02297460101544857, 'train/grad_norm_pre_clip': 1.1368045806884766, 'timing/data': 0.0003278236836194992, 'timing/model': 0.7869489584118128, 'learning_rate/qwen_vl_interface': 9.53974786644908e-06, 'learning_rate/action_model': 9.53974786644908e-05, 'epoch': 0.67})
+[2026-07-28 21:28:44,518][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.020336616784334183, 'train/grad_norm_pre_clip': 0.8692371249198914, 'timing/data': 0.0005760453641414642, 'timing/model': 0.8526474833488464, 'learning_rate/qwen_vl_interface': 9.538103384928029e-06, 'learning_rate/action_model': 9.538103384928027e-05, 'epoch': 0.67})
+[2026-07-28 21:28:53,716][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.030834490433335304, 'train/grad_norm_pre_clip': 1.3519744873046875, 'timing/data': 0.0003246534615755081, 'timing/model': 0.6253274362534285, 'learning_rate/qwen_vl_interface': 9.536456120903244e-06, 'learning_rate/action_model': 9.536456120903244e-05, 'epoch': 0.67})
+[2026-07-28 21:29:02,895][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.047776803374290466, 'train/grad_norm_pre_clip': 2.784788131713867, 'timing/data': 0.0003096871078014374, 'timing/model': 0.6213530786335468, 'learning_rate/qwen_vl_interface': 9.534806075443619e-06, 'learning_rate/action_model': 9.534806075443619e-05, 'epoch': 0.67})
+[2026-07-28 21:29:12,341][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.009814566001296043, 'train/grad_norm_pre_clip': 0.6438453197479248, 'timing/data': 0.00034429505467414856, 'timing/model': 0.8541939668357372, 'learning_rate/qwen_vl_interface': 9.533153249619853e-06, 'learning_rate/action_model': 9.533153249619853e-05, 'epoch': 0.67})
+[2026-07-28 21:29:21,597][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.0480736568570137, 'train/grad_norm_pre_clip': 1.9403494596481323, 'timing/data': 0.0003199819475412369, 'timing/model': 0.620769651606679, 'learning_rate/qwen_vl_interface': 9.531497644504444e-06, 'learning_rate/action_model': 9.531497644504445e-05, 'epoch': 0.67})
+[2026-07-28 21:29:31,063][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.01984664425253868, 'train/grad_norm_pre_clip': 0.8665574789047241, 'timing/data': 0.0003172103315591812, 'timing/model': 0.6216553617268801, 'learning_rate/qwen_vl_interface': 9.529839261171698e-06, 'learning_rate/action_model': 9.529839261171696e-05, 'epoch': 0.67})
+[2026-07-28 21:29:40,280][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.025106491521000862, 'train/grad_norm_pre_clip': 0.9090991020202637, 'timing/data': 0.00033578649163246155, 'timing/model': 0.6196892373263836, 'learning_rate/qwen_vl_interface': 9.52817810069772e-06, 'learning_rate/action_model': 9.52817810069772e-05, 'epoch': 0.67})
+[2026-07-28 21:29:49,480][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.03552376106381416, 'train/grad_norm_pre_clip': 1.0137574672698975, 'timing/data': 0.00039595551788806915, 'timing/model': 0.6205393653362989, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.67})
+[2026-07-28 21:29:59,195][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.02616075798869133, 'train/grad_norm_pre_clip': 1.0144444704055786, 'timing/data': 0.0003594961017370224, 'timing/model': 0.8506447691470385, 'learning_rate/qwen_vl_interface': 9.524847452639509e-06, 'learning_rate/action_model': 9.524847452639509e-05, 'epoch': 0.67})
+[2026-07-28 21:30:08,402][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.022910797968506813, 'train/grad_norm_pre_clip': 1.290273666381836, 'timing/data': 0.00031612440943717957, 'timing/model': 0.6207999680191278, 'learning_rate/qwen_vl_interface': 9.523177967216496e-06, 'learning_rate/action_model': 9.523177967216496e-05, 'epoch': 0.67})
+[2026-07-28 21:30:17,619][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.052665021270513535, 'train/grad_norm_pre_clip': 2.0161285400390625, 'timing/data': 0.0003524869680404663, 'timing/model': 0.6205650437623262, 'learning_rate/qwen_vl_interface': 9.521505708974694e-06, 'learning_rate/action_model': 9.521505708974694e-05, 'epoch': 0.67})
+[2026-07-28 21:30:26,887][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.034853994846343994, 'train/grad_norm_pre_clip': 1.253991723060608, 'timing/data': 0.00031578727066516876, 'timing/model': 0.6208448838442564, 'learning_rate/qwen_vl_interface': 9.519830678999214e-06, 'learning_rate/action_model': 9.519830678999213e-05, 'epoch': 0.67})
+[2026-07-28 21:30:36,368][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.030345378443598747, 'train/grad_norm_pre_clip': 1.269830584526062, 'timing/data': 0.00034490786492824554, 'timing/model': 0.6219616178423166, 'learning_rate/qwen_vl_interface': 9.518152878376961e-06, 'learning_rate/action_model': 9.51815287837696e-05, 'epoch': 0.67})
+[2026-07-28 21:30:45,667][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.019779080525040627, 'train/grad_norm_pre_clip': 0.7718112468719482, 'timing/data': 0.0003317762166261673, 'timing/model': 0.620964091271162, 'learning_rate/qwen_vl_interface': 9.516472308196646e-06, 'learning_rate/action_model': 9.516472308196645e-05, 'epoch': 0.67})
+[2026-07-28 21:30:54,910][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.023162826895713806, 'train/grad_norm_pre_clip': 0.663981556892395, 'timing/data': 0.00031505338847637177, 'timing/model': 0.6216801889240742, 'learning_rate/qwen_vl_interface': 9.51478896954877e-06, 'learning_rate/action_model': 9.51478896954877e-05, 'epoch': 0.67})
+[2026-07-28 21:31:04,585][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.02865324169397354, 'train/grad_norm_pre_clip': 0.7849911451339722, 'timing/data': 0.00030561909079551697, 'timing/model': 0.6221008356660604, 'learning_rate/qwen_vl_interface': 9.51310286352563e-06, 'learning_rate/action_model': 9.51310286352563e-05, 'epoch': 0.67})
+[2026-07-28 21:31:14,021][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.027577098459005356, 'train/grad_norm_pre_clip': 1.1505131721496582, 'timing/data': 0.00034688226878643036, 'timing/model': 0.6207149047404528, 'learning_rate/qwen_vl_interface': 9.51141399122133e-06, 'learning_rate/action_model': 9.511413991221331e-05, 'epoch': 0.67})
+[2026-07-28 21:31:23,383][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.04512617737054825, 'train/grad_norm_pre_clip': 1.7052134275436401, 'timing/data': 0.00031217746436595917, 'timing/model': 0.6202813591808081, 'learning_rate/qwen_vl_interface': 9.509722353731757e-06, 'learning_rate/action_model': 9.509722353731756e-05, 'epoch': 0.67})
+[2026-07-28 21:31:32,653][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.018530771136283875, 'train/grad_norm_pre_clip': 1.2852442264556885, 'timing/data': 0.00031216442584991455, 'timing/model': 0.6212302278727293, 'learning_rate/qwen_vl_interface': 9.508027952154595e-06, 'learning_rate/action_model': 9.508027952154593e-05, 'epoch': 0.67})
+[2026-07-28 21:31:42,135][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.024056077003479004, 'train/grad_norm_pre_clip': 1.1222561597824097, 'timing/data': 0.00039682723581790924, 'timing/model': 0.622283985838294, 'learning_rate/qwen_vl_interface': 9.506330787589321e-06, 'learning_rate/action_model': 9.506330787589322e-05, 'epoch': 0.67})
+[2026-07-28 21:31:51,791][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.05483464151620865, 'train/grad_norm_pre_clip': 1.3130450248718262, 'timing/data': 0.0003043990582227707, 'timing/model': 0.8553217072039843, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.67})
+[2026-07-28 21:32:01,430][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.018011795356869698, 'train/grad_norm_pre_clip': 3.2001953125, 'timing/data': 0.00037494488060474396, 'timing/model': 0.8410992212593555, 'learning_rate/qwen_vl_interface': 9.50292817390133e-06, 'learning_rate/action_model': 9.502928173901329e-05, 'epoch': 0.67})
+[2026-07-28 21:32:10,784][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.036218978464603424, 'train/grad_norm_pre_clip': 1.1798241138458252, 'timing/data': 0.0003218669444322586, 'timing/model': 0.6258087120950222, 'learning_rate/qwen_vl_interface': 9.501222726986525e-06, 'learning_rate/action_model': 9.501222726986524e-05, 'epoch': 0.67})
+[2026-07-28 21:32:20,062][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.026895716786384583, 'train/grad_norm_pre_clip': 0.7694621682167053, 'timing/data': 0.0004656407982110977, 'timing/model': 0.6212785914540291, 'learning_rate/qwen_vl_interface': 9.49951452149945e-06, 'learning_rate/action_model': 9.499514521499451e-05, 'epoch': 0.68})
+[2026-07-28 21:32:29,308][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.020931463688611984, 'train/grad_norm_pre_clip': 0.6213601231575012, 'timing/data': 0.00033781304955482483, 'timing/model': 0.6205318626016378, 'learning_rate/qwen_vl_interface': 9.497803558548539e-06, 'learning_rate/action_model': 9.497803558548539e-05, 'epoch': 0.68})
+[2026-07-28 21:32:38,685][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.08346399664878845, 'train/grad_norm_pre_clip': 2.6076152324676514, 'timing/data': 0.00031571462750434875, 'timing/model': 0.6221669409424067, 'learning_rate/qwen_vl_interface': 9.496089839244015e-06, 'learning_rate/action_model': 9.496089839244015e-05, 'epoch': 0.68})
+[2026-07-28 21:32:48,353][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.0391426719725132, 'train/grad_norm_pre_clip': 0.9827311038970947, 'timing/data': 0.00030124373733997345, 'timing/model': 0.625637236982584, 'learning_rate/qwen_vl_interface': 9.494373364697895e-06, 'learning_rate/action_model': 9.494373364697894e-05, 'epoch': 0.68})
+[2026-07-28 21:32:57,568][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.028363779187202454, 'train/grad_norm_pre_clip': 0.7690744996070862, 'timing/data': 0.0007177870720624924, 'timing/model': 0.6194063667207956, 'learning_rate/qwen_vl_interface': 9.492654136023978e-06, 'learning_rate/action_model': 9.492654136023979e-05, 'epoch': 0.68})
+[2026-07-28 21:33:06,796][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.018859989941120148, 'train/grad_norm_pre_clip': 0.7243203520774841, 'timing/data': 0.0003005173057317734, 'timing/model': 0.6210898105055094, 'learning_rate/qwen_vl_interface': 9.490932154337855e-06, 'learning_rate/action_model': 9.490932154337855e-05, 'epoch': 0.68})
+[2026-07-28 21:33:16,282][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.01168813370168209, 'train/grad_norm_pre_clip': 0.5559678673744202, 'timing/data': 0.0007780883461236954, 'timing/model': 0.6220432985574007, 'learning_rate/qwen_vl_interface': 9.4892074207569e-06, 'learning_rate/action_model': 9.489207420756899e-05, 'epoch': 0.68})
+[2026-07-28 21:33:25,730][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.025898300111293793, 'train/grad_norm_pre_clip': 0.884852409362793, 'timing/data': 0.00030985474586486816, 'timing/model': 0.6210693921893835, 'learning_rate/qwen_vl_interface': 9.487479936400274e-06, 'learning_rate/action_model': 9.487479936400272e-05, 'epoch': 0.68})
+[2026-07-28 21:33:34,960][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.044877246022224426, 'train/grad_norm_pre_clip': 2.056084632873535, 'timing/data': 0.0003268606960773468, 'timing/model': 0.6280395239591599, 'learning_rate/qwen_vl_interface': 9.485749702388923e-06, 'learning_rate/action_model': 9.485749702388923e-05, 'epoch': 0.68})
+[2026-07-28 21:33:44,206][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.026952413842082024, 'train/grad_norm_pre_clip': 1.8008438348770142, 'timing/data': 0.00031253136694431305, 'timing/model': 0.620142538100481, 'learning_rate/qwen_vl_interface': 9.484016719845576e-06, 'learning_rate/action_model': 9.484016719845576e-05, 'epoch': 0.68})
+[2026-07-28 21:33:53,659][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.022507887333631516, 'train/grad_norm_pre_clip': 1.6226378679275513, 'timing/data': 0.00037266314029693604, 'timing/model': 0.8556961640715599, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.68})
+[2026-07-28 21:34:02,915][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.022477179765701294, 'train/grad_norm_pre_clip': 1.3908089399337769, 'timing/data': 0.00030631572008132935, 'timing/model': 0.6215586606413126, 'learning_rate/qwen_vl_interface': 9.480542513662734e-06, 'learning_rate/action_model': 9.480542513662734e-05, 'epoch': 0.68})
+[2026-07-28 21:34:12,094][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.007963775657117367, 'train/grad_norm_pre_clip': 0.6652145981788635, 'timing/data': 0.00031520985066890717, 'timing/model': 0.62165960483253, 'learning_rate/qwen_vl_interface': 9.478801292277617e-06, 'learning_rate/action_model': 9.478801292277616e-05, 'epoch': 0.68})
+[2026-07-28 21:34:21,065][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.036842554807662964, 'train/grad_norm_pre_clip': 0.7046845555305481, 'timing/data': 0.00032480061054229736, 'timing/model': 0.6210814584046602, 'learning_rate/qwen_vl_interface': 9.477057326869249e-06, 'learning_rate/action_model': 9.477057326869249e-05, 'epoch': 0.68})
+[2026-07-28 21:34:30,228][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.02976451814174652, 'train/grad_norm_pre_clip': 1.2084051370620728, 'timing/data': 0.00031876377761363983, 'timing/model': 0.6215927619487047, 'learning_rate/qwen_vl_interface': 9.475310618569274e-06, 'learning_rate/action_model': 9.475310618569274e-05, 'epoch': 0.68})
+[2026-07-28 21:34:39,432][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.04499536007642746, 'train/grad_norm_pre_clip': 1.6796104907989502, 'timing/data': 0.00031398236751556396, 'timing/model': 0.8671412449330091, 'learning_rate/qwen_vl_interface': 9.473561168511115e-06, 'learning_rate/action_model': 9.473561168511115e-05, 'epoch': 0.68})
+[2026-07-28 21:34:48,373][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.04075847566127777, 'train/grad_norm_pre_clip': 1.607092022895813, 'timing/data': 0.0003165900707244873, 'timing/model': 0.6198720037937164, 'learning_rate/qwen_vl_interface': 9.471808977829967e-06, 'learning_rate/action_model': 9.471808977829967e-05, 'epoch': 0.68})
+[2026-07-28 21:34:57,379][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.042730096727609634, 'train/grad_norm_pre_clip': 1.1139066219329834, 'timing/data': 0.0003076493740081787, 'timing/model': 0.6209142450243235, 'learning_rate/qwen_vl_interface': 9.470054047662808e-06, 'learning_rate/action_model': 9.470054047662808e-05, 'epoch': 0.68})
+[2026-07-28 21:35:06,626][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.051575470715761185, 'train/grad_norm_pre_clip': 1.7037020921707153, 'timing/data': 0.0003086067736148834, 'timing/model': 0.6213020328432322, 'learning_rate/qwen_vl_interface': 9.468296379148394e-06, 'learning_rate/action_model': 9.468296379148393e-05, 'epoch': 0.68})
+[2026-07-28 21:35:15,835][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.0339893139898777, 'train/grad_norm_pre_clip': 0.8257870078086853, 'timing/data': 0.0003163013607263565, 'timing/model': 0.6199485156685114, 'learning_rate/qwen_vl_interface': 9.466535973427259e-06, 'learning_rate/action_model': 9.466535973427258e-05, 'epoch': 0.68})
+[2026-07-28 21:35:25,038][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.03200820833444595, 'train/grad_norm_pre_clip': 2.1743814945220947, 'timing/data': 0.000327233225107193, 'timing/model': 0.6233036145567894, 'learning_rate/qwen_vl_interface': 9.464772831641706e-06, 'learning_rate/action_model': 9.464772831641707e-05, 'epoch': 0.68})
+[2026-07-28 21:35:34,459][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.03880259022116661, 'train/grad_norm_pre_clip': 1.2421122789382935, 'timing/data': 0.0003130864351987839, 'timing/model': 0.6210073940455914, 'learning_rate/qwen_vl_interface': 9.463006954935823e-06, 'learning_rate/action_model': 9.463006954935822e-05, 'epoch': 0.68})
+[2026-07-28 21:35:43,897][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.03571752831339836, 'train/grad_norm_pre_clip': 1.6793156862258911, 'timing/data': 0.0003737024962902069, 'timing/model': 0.6203917842358351, 'learning_rate/qwen_vl_interface': 9.461238344455468e-06, 'learning_rate/action_model': 9.461238344455468e-05, 'epoch': 0.68})
+[2026-07-28 21:35:53,157][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.04899836331605911, 'train/grad_norm_pre_clip': 2.2837331295013428, 'timing/data': 0.00039484165608882904, 'timing/model': 0.6598124001175165, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.68})
+[2026-07-28 21:36:02,743][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.0319848507642746, 'train/grad_norm_pre_clip': 0.7697884440422058, 'timing/data': 0.0003232862800359726, 'timing/model': 0.622173298150301, 'learning_rate/qwen_vl_interface': 9.457692926763638e-06, 'learning_rate/action_model': 9.457692926763638e-05, 'epoch': 0.68})
+[2026-07-28 21:36:11,956][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.025025829672813416, 'train/grad_norm_pre_clip': 1.4686466455459595, 'timing/data': 0.00030923448503017426, 'timing/model': 0.6223361156880856, 'learning_rate/qwen_vl_interface': 9.455916121852748e-06, 'learning_rate/action_model': 9.455916121852747e-05, 'epoch': 0.68})
+[2026-07-28 21:36:21,242][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.035074494779109955, 'train/grad_norm_pre_clip': 1.1042673587799072, 'timing/data': 0.0003879126161336899, 'timing/model': 0.6251373626291752, 'learning_rate/qwen_vl_interface': 9.454136587768549e-06, 'learning_rate/action_model': 9.454136587768548e-05, 'epoch': 0.68})
+[2026-07-28 21:36:30,984][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.017922692000865936, 'train/grad_norm_pre_clip': 0.7480890154838562, 'timing/data': 0.0003308746963739395, 'timing/model': 0.8472115658223629, 'learning_rate/qwen_vl_interface': 9.452354325665763e-06, 'learning_rate/action_model': 9.452354325665763e-05, 'epoch': 0.68})
+[2026-07-28 21:36:40,213][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.01310537476092577, 'train/grad_norm_pre_clip': 0.7850131988525391, 'timing/data': 0.0003105662763118744, 'timing/model': 0.6214749999344349, 'learning_rate/qwen_vl_interface': 9.450569336700878e-06, 'learning_rate/action_model': 9.450569336700878e-05, 'epoch': 0.68})
+[2026-07-28 21:36:49,500][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.04113170504570007, 'train/grad_norm_pre_clip': 1.3135381937026978, 'timing/data': 0.00032375194132328033, 'timing/model': 0.6213450990617275, 'learning_rate/qwen_vl_interface': 9.448781622032155e-06, 'learning_rate/action_model': 9.448781622032155e-05, 'epoch': 0.68})
+[2026-07-28 21:36:58,760][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.023432835936546326, 'train/grad_norm_pre_clip': 0.6445086598396301, 'timing/data': 0.0003863181918859482, 'timing/model': 0.6230979468673468, 'learning_rate/qwen_vl_interface': 9.446991182819624e-06, 'learning_rate/action_model': 9.446991182819623e-05, 'epoch': 0.68})
+[2026-07-28 21:37:08,059][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.023448891937732697, 'train/grad_norm_pre_clip': 1.5735840797424316, 'timing/data': 0.0005067028105258942, 'timing/model': 0.6246023084968328, 'learning_rate/qwen_vl_interface': 9.44519802022508e-06, 'learning_rate/action_model': 9.44519802022508e-05, 'epoch': 0.68})
+[2026-07-28 21:37:17,290][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.007325660437345505, 'train/grad_norm_pre_clip': 0.5110005736351013, 'timing/data': 0.0003270823508501053, 'timing/model': 0.6223591305315495, 'learning_rate/qwen_vl_interface': 9.443402135412088e-06, 'learning_rate/action_model': 9.443402135412088e-05, 'epoch': 0.68})
+[2026-07-28 21:37:26,549][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.024491073563694954, 'train/grad_norm_pre_clip': 0.7479907274246216, 'timing/data': 0.00031320005655288696, 'timing/model': 0.6212384887039661, 'learning_rate/qwen_vl_interface': 9.441603529545977e-06, 'learning_rate/action_model': 9.441603529545976e-05, 'epoch': 0.68})
+[2026-07-28 21:37:35,977][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.03963671252131462, 'train/grad_norm_pre_clip': 1.6760976314544678, 'timing/data': 0.0003648214042186737, 'timing/model': 0.8449570015072823, 'learning_rate/qwen_vl_interface': 9.439802203793843e-06, 'learning_rate/action_model': 9.439802203793842e-05, 'epoch': 0.68})
+[2026-07-28 21:37:45,233][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.034539468586444855, 'train/grad_norm_pre_clip': 1.0163390636444092, 'timing/data': 0.0003628525882959366, 'timing/model': 0.6220164783298969, 'learning_rate/qwen_vl_interface': 9.437998159324547e-06, 'learning_rate/action_model': 9.437998159324547e-05, 'epoch': 0.68})
+[2026-07-28 21:37:54,716][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.036925990134477615, 'train/grad_norm_pre_clip': 1.4669238328933716, 'timing/data': 0.00037432461977005005, 'timing/model': 0.6208639666438103, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.68})
+[2026-07-28 21:38:03,913][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.0330621674656868, 'train/grad_norm_pre_clip': 2.3976454734802246, 'timing/data': 0.00036122649908065796, 'timing/model': 0.6202688924968243, 'learning_rate/qwen_vl_interface': 9.434381918918731e-06, 'learning_rate/action_model': 9.434381918918731e-05, 'epoch': 0.68})
+[2026-07-28 21:38:13,123][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.019330311566591263, 'train/grad_norm_pre_clip': 0.7884346842765808, 'timing/data': 0.00042723678052425385, 'timing/model': 0.6242322660982609, 'learning_rate/qwen_vl_interface': 9.432569725328753e-06, 'learning_rate/action_model': 9.432569725328753e-05, 'epoch': 0.68})
+[2026-07-28 21:38:22,639][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.010647398419678211, 'train/grad_norm_pre_clip': 0.4176754057407379, 'timing/data': 0.0003301762044429779, 'timing/model': 0.6217971332371235, 'learning_rate/qwen_vl_interface': 9.430754817714686e-06, 'learning_rate/action_model': 9.430754817714687e-05, 'epoch': 0.68})
+[2026-07-28 21:38:32,131][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.038641393184661865, 'train/grad_norm_pre_clip': 0.9762765169143677, 'timing/data': 0.0003571435809135437, 'timing/model': 0.6191038582473993, 'learning_rate/qwen_vl_interface': 9.42893719725421e-06, 'learning_rate/action_model': 9.42893719725421e-05, 'epoch': 0.68})
+[2026-07-28 21:38:41,486][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.03997839614748955, 'train/grad_norm_pre_clip': 1.76809823513031, 'timing/data': 0.0003019217401742935, 'timing/model': 0.6200307514518499, 'learning_rate/qwen_vl_interface': 9.427116865126757e-06, 'learning_rate/action_model': 9.427116865126756e-05, 'epoch': 0.68})
+[2026-07-28 21:38:50,458][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.017165757715702057, 'train/grad_norm_pre_clip': 0.9591280817985535, 'timing/data': 0.00039857812225818634, 'timing/model': 0.620486993342638, 'learning_rate/qwen_vl_interface': 9.42529382251352e-06, 'learning_rate/action_model': 9.42529382251352e-05, 'epoch': 0.68})
+[2026-07-28 21:38:59,712][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.050310343503952026, 'train/grad_norm_pre_clip': 1.843993902206421, 'timing/data': 0.00031064264476299286, 'timing/model': 0.6213269587606192, 'learning_rate/qwen_vl_interface': 9.423468070597453e-06, 'learning_rate/action_model': 9.423468070597452e-05, 'epoch': 0.68})
+[2026-07-28 21:39:09,473][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.027977539226412773, 'train/grad_norm_pre_clip': 1.1711002588272095, 'timing/data': 0.00040633417665958405, 'timing/model': 0.8563217613846064, 'learning_rate/qwen_vl_interface': 9.421639610563264e-06, 'learning_rate/action_model': 9.421639610563264e-05, 'epoch': 0.68})
+[2026-07-28 21:39:18,722][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.02359367534518242, 'train/grad_norm_pre_clip': 1.222493290901184, 'timing/data': 0.0003902427852153778, 'timing/model': 0.6206301376223564, 'learning_rate/qwen_vl_interface': 9.419808443597422e-06, 'learning_rate/action_model': 9.419808443597423e-05, 'epoch': 0.68})
+[2026-07-28 21:39:27,970][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.052584752440452576, 'train/grad_norm_pre_clip': 1.6286450624465942, 'timing/data': 0.00040164031088352203, 'timing/model': 0.621635427698493, 'learning_rate/qwen_vl_interface': 9.417974570888152e-06, 'learning_rate/action_model': 9.417974570888153e-05, 'epoch': 0.68})
+[2026-07-28 21:39:37,459][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.029929697513580322, 'train/grad_norm_pre_clip': 0.7622261047363281, 'timing/data': 0.0003897789865732193, 'timing/model': 0.8446014560759068, 'learning_rate/qwen_vl_interface': 9.416137993625437e-06, 'learning_rate/action_model': 9.416137993625437e-05, 'epoch': 0.68})
+[2026-07-28 21:39:46,688][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.02465439960360527, 'train/grad_norm_pre_clip': 0.9393530488014221, 'timing/data': 0.00032407231628894806, 'timing/model': 0.6206712741404772, 'learning_rate/qwen_vl_interface': 9.414298713001007e-06, 'learning_rate/action_model': 9.414298713001006e-05, 'epoch': 0.68})
+[2026-07-28 21:39:56,120][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.016784610226750374, 'train/grad_norm_pre_clip': 0.7542503476142883, 'timing/data': 0.0003936588764190674, 'timing/model': 0.6221255268901587, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.68})
+[2026-07-28 21:40:05,334][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.03317571431398392, 'train/grad_norm_pre_clip': 0.9380253553390503, 'timing/data': 0.00031608156859874725, 'timing/model': 0.620452281087637, 'learning_rate/qwen_vl_interface': 9.410612046442719e-06, 'learning_rate/action_model': 9.410612046442718e-05, 'epoch': 0.68})
+[2026-07-28 21:40:14,525][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.030126051977276802, 'train/grad_norm_pre_clip': 0.95460045337677, 'timing/data': 0.0003386419266462326, 'timing/model': 0.6224306840449572, 'learning_rate/qwen_vl_interface': 9.408764662901097e-06, 'learning_rate/action_model': 9.408764662901098e-05, 'epoch': 0.68})
+[2026-07-28 21:40:24,211][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.02258145809173584, 'train/grad_norm_pre_clip': 0.750174880027771, 'timing/data': 0.0003345143049955368, 'timing/model': 0.8480466511100531, 'learning_rate/qwen_vl_interface': 9.406914580782238e-06, 'learning_rate/action_model': 9.406914580782238e-05, 'epoch': 0.68})
+[2026-07-28 21:40:33,469][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.018403105437755585, 'train/grad_norm_pre_clip': 0.9657856225967407, 'timing/data': 0.00032093562185764313, 'timing/model': 0.6256445813924074, 'learning_rate/qwen_vl_interface': 9.405061801286638e-06, 'learning_rate/action_model': 9.405061801286638e-05, 'epoch': 0.68})
+[2026-07-28 21:40:42,755][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.046292368322610855, 'train/grad_norm_pre_clip': 1.2614082098007202, 'timing/data': 0.0003388896584510803, 'timing/model': 0.6276774480938911, 'learning_rate/qwen_vl_interface': 9.403206325616546e-06, 'learning_rate/action_model': 9.403206325616545e-05, 'epoch': 0.68})
+[2026-07-28 21:40:52,026][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.006996883545070887, 'train/grad_norm_pre_clip': 0.5747576951980591, 'timing/data': 0.00033939629793167114, 'timing/model': 0.6299952212721109, 'learning_rate/qwen_vl_interface': 9.40134815497596e-06, 'learning_rate/action_model': 9.40134815497596e-05, 'epoch': 0.69})
+[2026-07-28 21:41:01,622][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.048271436244249344, 'train/grad_norm_pre_clip': 1.7304489612579346, 'timing/data': 0.0004025232046842575, 'timing/model': 0.6255862582474947, 'learning_rate/qwen_vl_interface': 9.399487290570629e-06, 'learning_rate/action_model': 9.399487290570629e-05, 'epoch': 0.69})
+[2026-07-28 21:41:10,981][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.05317874997854233, 'train/grad_norm_pre_clip': 2.351238489151001, 'timing/data': 0.0003155563026666641, 'timing/model': 0.6270412839949131, 'learning_rate/qwen_vl_interface': 9.397623733608042e-06, 'learning_rate/action_model': 9.397623733608042e-05, 'epoch': 0.69})
+[2026-07-28 21:41:20,254][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.028617512434720993, 'train/grad_norm_pre_clip': 1.2000539302825928, 'timing/data': 0.0006819590926170349, 'timing/model': 0.621300708502531, 'learning_rate/qwen_vl_interface': 9.395757485297447e-06, 'learning_rate/action_model': 9.395757485297446e-05, 'epoch': 0.69})
+[2026-07-28 21:41:29,727][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.039808887988328934, 'train/grad_norm_pre_clip': 1.3828232288360596, 'timing/data': 0.00039638951420783997, 'timing/model': 0.6225200276821852, 'learning_rate/qwen_vl_interface': 9.393888546849828e-06, 'learning_rate/action_model': 9.393888546849828e-05, 'epoch': 0.69})
+[2026-07-28 21:41:38,973][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.05230293050408363, 'train/grad_norm_pre_clip': 1.9770443439483643, 'timing/data': 0.00030929967761039734, 'timing/model': 0.6216489057987928, 'learning_rate/qwen_vl_interface': 9.392016919477922e-06, 'learning_rate/action_model': 9.392016919477921e-05, 'epoch': 0.69})
+[2026-07-28 21:41:48,199][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.010671893134713173, 'train/grad_norm_pre_clip': 0.5834430456161499, 'timing/data': 0.0004150606691837311, 'timing/model': 0.6210534684360027, 'learning_rate/qwen_vl_interface': 9.390142604396206e-06, 'learning_rate/action_model': 9.390142604396205e-05, 'epoch': 0.69})
+[2026-07-28 21:41:57,425][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.07024411857128143, 'train/grad_norm_pre_clip': 3.9551267623901367, 'timing/data': 0.00032144784927368164, 'timing/model': 0.6218520645052195, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.69})
+[2026-07-28 21:42:06,670][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.022914011031389236, 'train/grad_norm_pre_clip': 0.7952282428741455, 'timing/data': 0.0003421194851398468, 'timing/model': 0.6243772506713867, 'learning_rate/qwen_vl_interface': 9.386385915969975e-06, 'learning_rate/action_model': 9.386385915969976e-05, 'epoch': 0.69})
+[2026-07-28 21:42:16,249][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.041425738483667374, 'train/grad_norm_pre_clip': 1.2880767583847046, 'timing/data': 0.0004189852625131607, 'timing/model': 0.889026740565896, 'learning_rate/qwen_vl_interface': 9.384503545063138e-06, 'learning_rate/action_model': 9.384503545063137e-05, 'epoch': 0.69})
+[2026-07-28 21:42:25,698][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.009162059985101223, 'train/grad_norm_pre_clip': 0.3776171803474426, 'timing/data': 0.0003149360418319702, 'timing/model': 0.8442278131842613, 'learning_rate/qwen_vl_interface': 9.382618491321837e-06, 'learning_rate/action_model': 9.382618491321837e-05, 'epoch': 0.69})
+[2026-07-28 21:42:34,947][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.01857934705913067, 'train/grad_norm_pre_clip': 0.7378485202789307, 'timing/data': 0.0003803577274084091, 'timing/model': 0.6250832267105579, 'learning_rate/qwen_vl_interface': 9.380730755969265e-06, 'learning_rate/action_model': 9.380730755969264e-05, 'epoch': 0.69})
+[2026-07-28 21:42:44,176][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.031361330300569534, 'train/grad_norm_pre_clip': 1.322291374206543, 'timing/data': 0.0003399942070245743, 'timing/model': 0.6239979714155197, 'learning_rate/qwen_vl_interface': 9.378840340230353e-06, 'learning_rate/action_model': 9.378840340230352e-05, 'epoch': 0.69})
+[2026-07-28 21:42:53,467][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.035806722939014435, 'train/grad_norm_pre_clip': 2.706183910369873, 'timing/data': 0.0003183837980031967, 'timing/model': 0.6218222938477993, 'learning_rate/qwen_vl_interface': 9.376947245331769e-06, 'learning_rate/action_model': 9.376947245331767e-05, 'epoch': 0.69})
+[2026-07-28 21:43:02,769][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.02995474264025688, 'train/grad_norm_pre_clip': 0.8011565804481506, 'timing/data': 0.0003948155790567398, 'timing/model': 0.6220465172082186, 'learning_rate/qwen_vl_interface': 9.375051472501922e-06, 'learning_rate/action_model': 9.375051472501923e-05, 'epoch': 0.69})
+[2026-07-28 21:43:12,198][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.0164275411516428, 'train/grad_norm_pre_clip': 0.7377836108207703, 'timing/data': 0.0004059039056301117, 'timing/model': 0.6266967486590147, 'learning_rate/qwen_vl_interface': 9.373153022970961e-06, 'learning_rate/action_model': 9.373153022970961e-05, 'epoch': 0.69})
+[2026-07-28 21:43:21,530][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.022946935147047043, 'train/grad_norm_pre_clip': 0.9371814727783203, 'timing/data': 0.0004188753664493561, 'timing/model': 0.631733475252986, 'learning_rate/qwen_vl_interface': 9.371251897970768e-06, 'learning_rate/action_model': 9.371251897970768e-05, 'epoch': 0.69})
+[2026-07-28 21:43:30,559][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.009659120813012123, 'train/grad_norm_pre_clip': 0.5210957527160645, 'timing/data': 0.00036022625863552094, 'timing/model': 0.6263626478612423, 'learning_rate/qwen_vl_interface': 9.369348098734961e-06, 'learning_rate/action_model': 9.369348098734961e-05, 'epoch': 0.69})
+[2026-07-28 21:43:39,876][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.017552904784679413, 'train/grad_norm_pre_clip': 0.7890427708625793, 'timing/data': 0.00034239329397678375, 'timing/model': 0.623035654425621, 'learning_rate/qwen_vl_interface': 9.367441626498897e-06, 'learning_rate/action_model': 9.367441626498896e-05, 'epoch': 0.69})
+[2026-07-28 21:43:49,643][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.011258352547883987, 'train/grad_norm_pre_clip': 0.45819225907325745, 'timing/data': 0.00045561790466308594, 'timing/model': 0.8837345782667398, 'learning_rate/qwen_vl_interface': 9.365532482499663e-06, 'learning_rate/action_model': 9.365532482499662e-05, 'epoch': 0.69})
+[2026-07-28 21:43:58,900][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3750
+[2026-07-28 21:44:09,187][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3750
+[2026-07-28 22:01:37,741][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3750
+[2026-07-28 22:01:37,746][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.046809546649456024, 'train/grad_norm_pre_clip': 2.3545830249786377, 'eval/action_loss/seconds': 10.283450270071626, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.030658388510346413, 'eval/latency_3/loss': 0.030658388510346413, 'eval/flappy/loss': 0.030658388510346413, 'eval/flappy/latency_3/loss': 0.030658388510346413, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 411.52000483796, 'rl_games_eval/mid_train/mean_length': 3334.25, 'rl_games_eval/mid_train/std_reward': 55.36169859795965, 'rl_games_eval/mid_train/std_length': 445.4333704382733, 'rl_games_eval/mid_train/macro_mean_reward': 411.52000483796, 'rl_games_eval/mid_train/macro_mean_length': 3334.25, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 411.52000483796, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 3334.25, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 55.36169859795965, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 445.4333704382733, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 411.52000483796, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003133229911327362, 'timing/model': 0.6220023613423109, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.69})
+[2026-07-28 22:01:47,567][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.04333765432238579, 'train/grad_norm_pre_clip': 1.2669111490249634, 'timing/data': 0.0003880690783262253, 'timing/model': 0.6282333843410015, 'learning_rate/qwen_vl_interface': 9.361706184168714e-06, 'learning_rate/action_model': 9.361706184168713e-05, 'epoch': 0.69})
+[2026-07-28 22:01:57,003][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.029043225571513176, 'train/grad_norm_pre_clip': 2.0151398181915283, 'timing/data': 0.00031352974474430084, 'timing/model': 0.6226823758333921, 'learning_rate/qwen_vl_interface': 9.359789032319843e-06, 'learning_rate/action_model': 9.359789032319842e-05, 'epoch': 0.69})
+[2026-07-28 22:02:06,238][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.008856777101755142, 'train/grad_norm_pre_clip': 0.954797089099884, 'timing/data': 0.00030609220266342163, 'timing/model': 0.622620789334178, 'learning_rate/qwen_vl_interface': 9.357869213673484e-06, 'learning_rate/action_model': 9.357869213673485e-05, 'epoch': 0.69})
+[2026-07-28 22:02:15,475][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.0922061949968338, 'train/grad_norm_pre_clip': 3.1616482734680176, 'timing/data': 0.00042511336505413055, 'timing/model': 0.62181281670928, 'learning_rate/qwen_vl_interface': 9.355946729475393e-06, 'learning_rate/action_model': 9.355946729475393e-05, 'epoch': 0.69})
+[2026-07-28 22:02:24,963][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.02893403358757496, 'train/grad_norm_pre_clip': 2.363372802734375, 'timing/data': 0.0003204699605703354, 'timing/model': 0.8582028839737177, 'learning_rate/qwen_vl_interface': 9.354021580973049e-06, 'learning_rate/action_model': 9.354021580973048e-05, 'epoch': 0.69})
+[2026-07-28 22:02:34,386][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.025004955008625984, 'train/grad_norm_pre_clip': 1.9727200269699097, 'timing/data': 0.0003645867109298706, 'timing/model': 0.8408933524042368, 'learning_rate/qwen_vl_interface': 9.352093769415655e-06, 'learning_rate/action_model': 9.352093769415655e-05, 'epoch': 0.69})
+[2026-07-28 22:02:43,577][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.01583872176706791, 'train/grad_norm_pre_clip': 1.1373695135116577, 'timing/data': 0.00037588179111480713, 'timing/model': 0.6225091814994812, 'learning_rate/qwen_vl_interface': 9.35016329605415e-06, 'learning_rate/action_model': 9.350163296054151e-05, 'epoch': 0.69})
+[2026-07-28 22:02:52,777][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.01247769221663475, 'train/grad_norm_pre_clip': 0.6561160087585449, 'timing/data': 0.0003708302974700928, 'timing/model': 0.6208570804446936, 'learning_rate/qwen_vl_interface': 9.348230162141198e-06, 'learning_rate/action_model': 9.348230162141198e-05, 'epoch': 0.69})
+[2026-07-28 22:03:01,982][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.013288349844515324, 'train/grad_norm_pre_clip': 0.8091486096382141, 'timing/data': 0.00038206763565540314, 'timing/model': 0.6214809641242027, 'learning_rate/qwen_vl_interface': 9.346294368931188e-06, 'learning_rate/action_model': 9.346294368931188e-05, 'epoch': 0.69})
+[2026-07-28 22:03:11,416][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.03826815262436867, 'train/grad_norm_pre_clip': 1.9633640050888062, 'timing/data': 0.0003699846565723419, 'timing/model': 0.8679943233728409, 'learning_rate/qwen_vl_interface': 9.344355917680235e-06, 'learning_rate/action_model': 9.344355917680235e-05, 'epoch': 0.69})
+[2026-07-28 22:03:20,876][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.036250919103622437, 'train/grad_norm_pre_clip': 2.234733819961548, 'timing/data': 0.00039600953459739685, 'timing/model': 0.622055996209383, 'learning_rate/qwen_vl_interface': 9.342414809646179e-06, 'learning_rate/action_model': 9.342414809646179e-05, 'epoch': 0.69})
+[2026-07-28 22:03:29,843][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.07307387888431549, 'train/grad_norm_pre_clip': 2.365950345993042, 'timing/data': 0.0003214646130800247, 'timing/model': 0.6219126712530851, 'learning_rate/qwen_vl_interface': 9.340471046088581e-06, 'learning_rate/action_model': 9.34047104608858e-05, 'epoch': 0.69})
+[2026-07-28 22:03:39,058][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.06299979984760284, 'train/grad_norm_pre_clip': 2.2719368934631348, 'timing/data': 0.00040173716843128204, 'timing/model': 0.6205561552196741, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.69})
+[2026-07-28 22:03:48,582][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.026659108698368073, 'train/grad_norm_pre_clip': 1.3385906219482422, 'timing/data': 0.00032554566860198975, 'timing/model': 0.6221802514046431, 'learning_rate/qwen_vl_interface': 9.336575557449641e-06, 'learning_rate/action_model': 9.336575557449641e-05, 'epoch': 0.69})
+[2026-07-28 22:03:58,145][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.020005805417895317, 'train/grad_norm_pre_clip': 1.0354615449905396, 'timing/data': 0.0003301873803138733, 'timing/model': 0.6233970243483782, 'learning_rate/qwen_vl_interface': 9.334623834896039e-06, 'learning_rate/action_model': 9.334623834896038e-05, 'epoch': 0.69})
+[2026-07-28 22:04:07,407][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.094352126121521, 'train/grad_norm_pre_clip': 2.210714340209961, 'timing/data': 0.0004439596086740494, 'timing/model': 0.6216831225901842, 'learning_rate/qwen_vl_interface': 9.332669461874376e-06, 'learning_rate/action_model': 9.332669461874374e-05, 'epoch': 0.69})
+[2026-07-28 22:04:16,655][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.048819735646247864, 'train/grad_norm_pre_clip': 1.5637662410736084, 'timing/data': 0.00038170628249645233, 'timing/model': 0.6218277774751186, 'learning_rate/qwen_vl_interface': 9.330712439652824e-06, 'learning_rate/action_model': 9.330712439652823e-05, 'epoch': 0.69})
+[2026-07-28 22:04:26,147][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.03991860896348953, 'train/grad_norm_pre_clip': 0.957983136177063, 'timing/data': 0.0003197621554136276, 'timing/model': 0.8533544652163982, 'learning_rate/qwen_vl_interface': 9.328752769501272e-06, 'learning_rate/action_model': 9.328752769501272e-05, 'epoch': 0.69})
+[2026-07-28 22:04:35,581][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.046529315412044525, 'train/grad_norm_pre_clip': 2.5667643547058105, 'timing/data': 0.0003118906170129776, 'timing/model': 0.6241213530302048, 'learning_rate/qwen_vl_interface': 9.326790452691332e-06, 'learning_rate/action_model': 9.326790452691332e-05, 'epoch': 0.69})
+[2026-07-28 22:04:44,573][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.054964639246463776, 'train/grad_norm_pre_clip': 1.5742988586425781, 'timing/data': 0.00033479370176792145, 'timing/model': 0.6245483346283436, 'learning_rate/qwen_vl_interface': 9.324825490496327e-06, 'learning_rate/action_model': 9.324825490496327e-05, 'epoch': 0.69})
+[2026-07-28 22:04:53,774][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.015158170834183693, 'train/grad_norm_pre_clip': 0.5770000219345093, 'timing/data': 0.0003193691372871399, 'timing/model': 0.6232329942286015, 'learning_rate/qwen_vl_interface': 9.322857884191303e-06, 'learning_rate/action_model': 9.322857884191302e-05, 'epoch': 0.69})
+[2026-07-28 22:05:03,005][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.03153672069311142, 'train/grad_norm_pre_clip': 1.745302438735962, 'timing/data': 0.00032280758023262024, 'timing/model': 0.6208412740379572, 'learning_rate/qwen_vl_interface': 9.320887635053013e-06, 'learning_rate/action_model': 9.320887635053014e-05, 'epoch': 0.69})
+[2026-07-28 22:05:12,444][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.0785169005393982, 'train/grad_norm_pre_clip': 2.175417423248291, 'timing/data': 0.0003244113177061081, 'timing/model': 0.6245018690824509, 'learning_rate/qwen_vl_interface': 9.318914744359936e-06, 'learning_rate/action_model': 9.318914744359937e-05, 'epoch': 0.69})
+[2026-07-28 22:05:21,877][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.03441016376018524, 'train/grad_norm_pre_clip': 1.5411077737808228, 'timing/data': 0.00039102137088775635, 'timing/model': 0.6210358813405037, 'learning_rate/qwen_vl_interface': 9.316939213392257e-06, 'learning_rate/action_model': 9.316939213392257e-05, 'epoch': 0.69})
+[2026-07-28 22:05:31,231][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.03609897196292877, 'train/grad_norm_pre_clip': 1.5461174249649048, 'timing/data': 0.0003444179892539978, 'timing/model': 0.7281917352229357, 'learning_rate/qwen_vl_interface': 9.314961043431877e-06, 'learning_rate/action_model': 9.314961043431876e-05, 'epoch': 0.69})
+[2026-07-28 22:05:40,199][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.013697447255253792, 'train/grad_norm_pre_clip': 0.4774426221847534, 'timing/data': 0.00032797642052173615, 'timing/model': 0.6205648332834244, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.69})
+[2026-07-28 22:05:49,652][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.02312197908759117, 'train/grad_norm_pre_clip': 0.6792194843292236, 'timing/data': 0.0003282073885202408, 'timing/model': 0.6207264829427004, 'learning_rate/qwen_vl_interface': 9.310996791669175e-06, 'learning_rate/action_model': 9.310996791669175e-05, 'epoch': 0.69})
+[2026-07-28 22:05:58,623][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.021738482639193535, 'train/grad_norm_pre_clip': 1.140799641609192, 'timing/data': 0.0003249999135732651, 'timing/model': 0.6245896089822054, 'learning_rate/qwen_vl_interface': 9.309010712439213e-06, 'learning_rate/action_model': 9.309010712439213e-05, 'epoch': 0.69})
+[2026-07-28 22:06:07,827][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.027130048722028732, 'train/grad_norm_pre_clip': 1.283162236213684, 'timing/data': 0.0003177933394908905, 'timing/model': 0.6247092857956886, 'learning_rate/qwen_vl_interface': 9.307021999361267e-06, 'learning_rate/action_model': 9.307021999361267e-05, 'epoch': 0.69})
+[2026-07-28 22:06:16,986][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.043078795075416565, 'train/grad_norm_pre_clip': 1.5057255029678345, 'timing/data': 0.0003205053508281708, 'timing/model': 0.7921914998441935, 'learning_rate/qwen_vl_interface': 9.305030653725794e-06, 'learning_rate/action_model': 9.305030653725794e-05, 'epoch': 0.69})
+[2026-07-28 22:06:26,176][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.026395371183753014, 'train/grad_norm_pre_clip': 0.8916826248168945, 'timing/data': 0.0003779008984565735, 'timing/model': 0.6228424049913883, 'learning_rate/qwen_vl_interface': 9.303036676824954e-06, 'learning_rate/action_model': 9.303036676824953e-05, 'epoch': 0.69})
+[2026-07-28 22:06:35,593][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.022789765149354935, 'train/grad_norm_pre_clip': 0.7418280839920044, 'timing/data': 0.0003200843930244446, 'timing/model': 0.620853878557682, 'learning_rate/qwen_vl_interface': 9.301040069952616e-06, 'learning_rate/action_model': 9.301040069952616e-05, 'epoch': 0.69})
+[2026-07-28 22:06:44,802][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.011707572266459465, 'train/grad_norm_pre_clip': 0.9270134568214417, 'timing/data': 0.00032753124833106995, 'timing/model': 0.6211497709155083, 'learning_rate/qwen_vl_interface': 9.299040834404362e-06, 'learning_rate/action_model': 9.299040834404361e-05, 'epoch': 0.69})
+[2026-07-28 22:06:54,272][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.027242843061685562, 'train/grad_norm_pre_clip': 0.7543554306030273, 'timing/data': 0.0003368258476257324, 'timing/model': 0.6215222347527742, 'learning_rate/qwen_vl_interface': 9.297038971477468e-06, 'learning_rate/action_model': 9.297038971477466e-05, 'epoch': 0.7})
+[2026-07-28 22:07:03,470][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.03297816216945648, 'train/grad_norm_pre_clip': 1.537976622581482, 'timing/data': 0.0003971625119447708, 'timing/model': 0.6212160345166922, 'learning_rate/qwen_vl_interface': 9.295034482470922e-06, 'learning_rate/action_model': 9.295034482470923e-05, 'epoch': 0.7})
+[2026-07-28 22:07:12,687][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.019099397584795952, 'train/grad_norm_pre_clip': 0.9155712127685547, 'timing/data': 0.0003264043480157852, 'timing/model': 0.62155482172966, 'learning_rate/qwen_vl_interface': 9.293027368685418e-06, 'learning_rate/action_model': 9.293027368685419e-05, 'epoch': 0.7})
+[2026-07-28 22:07:22,363][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.06151658669114113, 'train/grad_norm_pre_clip': 2.581120014190674, 'timing/data': 0.0005820319056510925, 'timing/model': 0.8551409300416708, 'learning_rate/qwen_vl_interface': 9.29101763142335e-06, 'learning_rate/action_model': 9.29101763142335e-05, 'epoch': 0.7})
+[2026-07-28 22:07:31,617][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.03557176887989044, 'train/grad_norm_pre_clip': 1.8539910316467285, 'timing/data': 0.00032564811408519745, 'timing/model': 0.6210212931036949, 'learning_rate/qwen_vl_interface': 9.28900527198881e-06, 'learning_rate/action_model': 9.289005271988811e-05, 'epoch': 0.7})
+[2026-07-28 22:07:40,634][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.04943430423736572, 'train/grad_norm_pre_clip': 2.1569721698760986, 'timing/data': 0.00031188875436782837, 'timing/model': 0.6194242686033249, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.7})
+[2026-07-28 22:07:50,058][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.033572107553482056, 'train/grad_norm_pre_clip': 1.1143648624420166, 'timing/data': 0.000396067276597023, 'timing/model': 0.8525327444076538, 'learning_rate/qwen_vl_interface': 9.284972691827226e-06, 'learning_rate/action_model': 9.284972691827224e-05, 'epoch': 0.7})
+[2026-07-28 22:07:59,256][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.0584423691034317, 'train/grad_norm_pre_clip': 2.0237772464752197, 'timing/data': 0.0003244541585445404, 'timing/model': 0.6204980853945017, 'learning_rate/qwen_vl_interface': 9.282952473716873e-06, 'learning_rate/action_model': 9.282952473716874e-05, 'epoch': 0.7})
+[2026-07-28 22:08:08,704][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.041686058044433594, 'train/grad_norm_pre_clip': 1.182993769645691, 'timing/data': 0.00039790384471416473, 'timing/model': 0.621649632230401, 'learning_rate/qwen_vl_interface': 9.280929638667447e-06, 'learning_rate/action_model': 9.280929638667447e-05, 'epoch': 0.7})
+[2026-07-28 22:08:17,908][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.03486320376396179, 'train/grad_norm_pre_clip': 1.7076117992401123, 'timing/data': 0.00042528286576271057, 'timing/model': 0.6203132346272469, 'learning_rate/qwen_vl_interface': 9.278904187991542e-06, 'learning_rate/action_model': 9.278904187991541e-05, 'epoch': 0.7})
+[2026-07-28 22:08:27,149][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.020098449662327766, 'train/grad_norm_pre_clip': 1.2826236486434937, 'timing/data': 0.0004030466079711914, 'timing/model': 0.6207639593631029, 'learning_rate/qwen_vl_interface': 9.27687612300345e-06, 'learning_rate/action_model': 9.27687612300345e-05, 'epoch': 0.7})
+[2026-07-28 22:08:36,612][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.039116811007261276, 'train/grad_norm_pre_clip': 1.7108285427093506, 'timing/data': 0.000333661213517189, 'timing/model': 0.8475891184061766, 'learning_rate/qwen_vl_interface': 9.27484544501916e-06, 'learning_rate/action_model': 9.274845445019159e-05, 'epoch': 0.7})
+[2026-07-28 22:08:45,820][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.04536547511816025, 'train/grad_norm_pre_clip': 2.5893399715423584, 'timing/data': 0.00035975687205791473, 'timing/model': 0.621682183817029, 'learning_rate/qwen_vl_interface': 9.27281215535636e-06, 'learning_rate/action_model': 9.272812155356358e-05, 'epoch': 0.7})
+[2026-07-28 22:08:55,295][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.02882625162601471, 'train/grad_norm_pre_clip': 0.773821234703064, 'timing/data': 0.0003173407167196274, 'timing/model': 0.8554450795054436, 'learning_rate/qwen_vl_interface': 9.270776255334425e-06, 'learning_rate/action_model': 9.270776255334425e-05, 'epoch': 0.7})
+[2026-07-28 22:09:04,499][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.036340825259685516, 'train/grad_norm_pre_clip': 1.8139381408691406, 'timing/data': 0.00038802996277809143, 'timing/model': 0.6208256147801876, 'learning_rate/qwen_vl_interface': 9.268737746274432e-06, 'learning_rate/action_model': 9.268737746274432e-05, 'epoch': 0.7})
+[2026-07-28 22:09:13,710][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.02669360116124153, 'train/grad_norm_pre_clip': 0.8355572819709778, 'timing/data': 0.0004008915275335312, 'timing/model': 0.6206721588969231, 'learning_rate/qwen_vl_interface': 9.266696629499146e-06, 'learning_rate/action_model': 9.266696629499145e-05, 'epoch': 0.7})
+[2026-07-28 22:09:22,929][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.02629627101123333, 'train/grad_norm_pre_clip': 0.6029757261276245, 'timing/data': 0.000346209853887558, 'timing/model': 0.6188537441194057, 'learning_rate/qwen_vl_interface': 9.264652906333026e-06, 'learning_rate/action_model': 9.264652906333026e-05, 'epoch': 0.7})
+[2026-07-28 22:09:32,224][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.022040728479623795, 'train/grad_norm_pre_clip': 0.5017358064651489, 'timing/data': 0.000340290367603302, 'timing/model': 0.6235679276287556, 'learning_rate/qwen_vl_interface': 9.26260657810222e-06, 'learning_rate/action_model': 9.262606578102221e-05, 'epoch': 0.7})
+[2026-07-28 22:09:41,688][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.009048259817063808, 'train/grad_norm_pre_clip': 0.4418085217475891, 'timing/data': 0.000307328999042511, 'timing/model': 0.6190639641135931, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.7})
+[2026-07-28 22:09:50,886][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.02163391374051571, 'train/grad_norm_pre_clip': 0.9082098007202148, 'timing/data': 0.0003292970359325409, 'timing/model': 0.6212994437664747, 'learning_rate/qwen_vl_interface': 9.258506111759608e-06, 'learning_rate/action_model': 9.258506111759609e-05, 'epoch': 0.7})
+[2026-07-28 22:10:00,063][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.011477628722786903, 'train/grad_norm_pre_clip': 0.893486738204956, 'timing/data': 0.00033046677708625793, 'timing/model': 0.6200846713036299, 'learning_rate/qwen_vl_interface': 9.256451976308548e-06, 'learning_rate/action_model': 9.25645197630855e-05, 'epoch': 0.7})
+[2026-07-28 22:10:09,598][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.03576631471514702, 'train/grad_norm_pre_clip': 1.5326497554779053, 'timing/data': 0.00038318708539009094, 'timing/model': 0.6250054556876421, 'learning_rate/qwen_vl_interface': 9.2543952411143e-06, 'learning_rate/action_model': 9.254395241114299e-05, 'epoch': 0.7})
+[2026-07-28 22:10:18,840][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.056390441954135895, 'train/grad_norm_pre_clip': 1.5723265409469604, 'timing/data': 0.0003190618008375168, 'timing/model': 0.6243705321103334, 'learning_rate/qwen_vl_interface': 9.252335907511452e-06, 'learning_rate/action_model': 9.252335907511453e-05, 'epoch': 0.7})
+[2026-07-28 22:10:28,043][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.028247101232409477, 'train/grad_norm_pre_clip': 1.0670174360275269, 'timing/data': 0.00032730773091316223, 'timing/model': 0.6201502438634634, 'learning_rate/qwen_vl_interface': 9.250273976836289e-06, 'learning_rate/action_model': 9.250273976836289e-05, 'epoch': 0.7})
+[2026-07-28 22:10:37,608][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.009779086336493492, 'train/grad_norm_pre_clip': 0.6318597197532654, 'timing/data': 0.00033108144998550415, 'timing/model': 0.6200926601886749, 'learning_rate/qwen_vl_interface': 9.248209450426772e-06, 'learning_rate/action_model': 9.248209450426772e-05, 'epoch': 0.7})
+[2026-07-28 22:10:46,965][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.014901025220751762, 'train/grad_norm_pre_clip': 0.5989506244659424, 'timing/data': 0.0003177151083946228, 'timing/model': 0.6193411964923143, 'learning_rate/qwen_vl_interface': 9.24614232962255e-06, 'learning_rate/action_model': 9.24614232962255e-05, 'epoch': 0.7})
+[2026-07-28 22:10:56,134][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.02811792679131031, 'train/grad_norm_pre_clip': 1.6111202239990234, 'timing/data': 0.0003296658396720886, 'timing/model': 0.6189724914729595, 'learning_rate/qwen_vl_interface': 9.244072615764957e-06, 'learning_rate/action_model': 9.244072615764956e-05, 'epoch': 0.7})
+[2026-07-28 22:11:05,320][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.032900337129831314, 'train/grad_norm_pre_clip': 1.5317811965942383, 'timing/data': 0.00032572075724601746, 'timing/model': 0.6192150507122278, 'learning_rate/qwen_vl_interface': 9.242000310197005e-06, 'learning_rate/action_model': 9.242000310197004e-05, 'epoch': 0.7})
+[2026-07-28 22:11:14,348][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.01000276766717434, 'train/grad_norm_pre_clip': 1.4179412126541138, 'timing/data': 0.00033197738230228424, 'timing/model': 0.6224821656942368, 'learning_rate/qwen_vl_interface': 9.239925414263392e-06, 'learning_rate/action_model': 9.239925414263392e-05, 'epoch': 0.7})
+[2026-07-28 22:11:24,076][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.06051347404718399, 'train/grad_norm_pre_clip': 2.1490321159362793, 'timing/data': 0.00033671967685222626, 'timing/model': 0.6208746489137411, 'learning_rate/qwen_vl_interface': 9.237847929310497e-06, 'learning_rate/action_model': 9.237847929310496e-05, 'epoch': 0.7})
+[2026-07-28 22:11:33,455][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.013562600128352642, 'train/grad_norm_pre_clip': 0.780502438545227, 'timing/data': 0.0005351342260837555, 'timing/model': 0.6202756278216839, 'learning_rate/qwen_vl_interface': 9.235767856686375e-06, 'learning_rate/action_model': 9.235767856686375e-05, 'epoch': 0.7})
+[2026-07-28 22:11:42,690][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.009848954156041145, 'train/grad_norm_pre_clip': 0.49961212277412415, 'timing/data': 0.00031878985464572906, 'timing/model': 0.6188802067190409, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.7})
+[2026-07-28 22:11:51,666][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.026200678199529648, 'train/grad_norm_pre_clip': 0.7416682243347168, 'timing/data': 0.00039476901292800903, 'timing/model': 0.6214585062116385, 'learning_rate/qwen_vl_interface': 9.23159995382508e-06, 'learning_rate/action_model': 9.231599953825079e-05, 'epoch': 0.7})
+[2026-07-28 22:12:01,312][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.024254050105810165, 'train/grad_norm_pre_clip': 1.564095377922058, 'timing/data': 0.00038878805935382843, 'timing/model': 0.8425268027931452, 'learning_rate/qwen_vl_interface': 9.22951212629241e-06, 'learning_rate/action_model': 9.229512126292411e-05, 'epoch': 0.7})
+[2026-07-28 22:12:10,488][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.07422348111867905, 'train/grad_norm_pre_clip': 3.5842738151550293, 'timing/data': 0.0003122277557849884, 'timing/model': 0.6203667614609003, 'learning_rate/qwen_vl_interface': 9.227421716497531e-06, 'learning_rate/action_model': 9.227421716497531e-05, 'epoch': 0.7})
+[2026-07-28 22:12:19,671][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.04808695614337921, 'train/grad_norm_pre_clip': 2.1449732780456543, 'timing/data': 0.00031919777393341064, 'timing/model': 0.6214450243860483, 'learning_rate/qwen_vl_interface': 9.225328725796879e-06, 'learning_rate/action_model': 9.22532872579688e-05, 'epoch': 0.7})
+[2026-07-28 22:12:29,066][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.02245994284749031, 'train/grad_norm_pre_clip': 1.4601190090179443, 'timing/data': 0.0003380049020051956, 'timing/model': 0.8159653730690479, 'learning_rate/qwen_vl_interface': 9.223233155548579e-06, 'learning_rate/action_model': 9.223233155548578e-05, 'epoch': 0.7})
+[2026-07-28 22:12:38,287][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.01112976111471653, 'train/grad_norm_pre_clip': 0.9590379595756531, 'timing/data': 0.00031300634145736694, 'timing/model': 0.6247723214328289, 'learning_rate/qwen_vl_interface': 9.22113500711242e-06, 'learning_rate/action_model': 9.221135007112419e-05, 'epoch': 0.7})
+[2026-07-28 22:12:47,512][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.06114829331636429, 'train/grad_norm_pre_clip': 2.587209463119507, 'timing/data': 0.0003038104623556137, 'timing/model': 0.6252319347113371, 'learning_rate/qwen_vl_interface': 9.219034281849867e-06, 'learning_rate/action_model': 9.219034281849867e-05, 'epoch': 0.7})
+[2026-07-28 22:12:56,915][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.029363542795181274, 'train/grad_norm_pre_clip': 1.551563024520874, 'timing/data': 0.00033369287848472595, 'timing/model': 0.6220411323010921, 'learning_rate/qwen_vl_interface': 9.216930981124062e-06, 'learning_rate/action_model': 9.216930981124062e-05, 'epoch': 0.7})
+[2026-07-28 22:13:06,367][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.03784555941820145, 'train/grad_norm_pre_clip': 0.7141618728637695, 'timing/data': 0.00031789951026439667, 'timing/model': 0.8560424894094467, 'learning_rate/qwen_vl_interface': 9.214825106299808e-06, 'learning_rate/action_model': 9.214825106299808e-05, 'epoch': 0.7})
+[2026-07-28 22:13:15,529][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.061508722603321075, 'train/grad_norm_pre_clip': 1.4610778093338013, 'timing/data': 0.000301925465464592, 'timing/model': 0.6209196224808693, 'learning_rate/qwen_vl_interface': 9.212716658743592e-06, 'learning_rate/action_model': 9.212716658743592e-05, 'epoch': 0.7})
+[2026-07-28 22:13:24,679][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.022886980324983597, 'train/grad_norm_pre_clip': 0.9484222531318665, 'timing/data': 0.0003655720502138138, 'timing/model': 0.6217775531113148, 'learning_rate/qwen_vl_interface': 9.210605639823558e-06, 'learning_rate/action_model': 9.210605639823557e-05, 'epoch': 0.7})
+[2026-07-28 22:13:34,071][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.03963866084814072, 'train/grad_norm_pre_clip': 2.217550277709961, 'timing/data': 0.00036969780921936035, 'timing/model': 0.8374657649546862, 'learning_rate/qwen_vl_interface': 9.208492050909523e-06, 'learning_rate/action_model': 9.208492050909523e-05, 'epoch': 0.7})
+[2026-07-28 22:13:43,230][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.0401763841509819, 'train/grad_norm_pre_clip': 2.0199270248413086, 'timing/data': 0.00035617873072624207, 'timing/model': 0.6191443428397179, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.7})
+[2026-07-28 22:13:52,664][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.017382092773914337, 'train/grad_norm_pre_clip': 0.9146376252174377, 'timing/data': 0.0003172531723976135, 'timing/model': 0.8465469218790531, 'learning_rate/qwen_vl_interface': 9.204257168587061e-06, 'learning_rate/action_model': 9.204257168587061e-05, 'epoch': 0.7})
+[2026-07-28 22:14:01,810][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.05123002454638481, 'train/grad_norm_pre_clip': 1.7229962348937988, 'timing/data': 0.0003153625875711441, 'timing/model': 0.6225448697805405, 'learning_rate/qwen_vl_interface': 9.2021358779266e-06, 'learning_rate/action_model': 9.202135877926601e-05, 'epoch': 0.7})
+[2026-07-28 22:14:11,186][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.04064326733350754, 'train/grad_norm_pre_clip': 2.1137590408325195, 'timing/data': 0.0003777220845222473, 'timing/model': 0.6210468206554651, 'learning_rate/qwen_vl_interface': 9.200012022768077e-06, 'learning_rate/action_model': 9.200012022768077e-05, 'epoch': 0.7})
+[2026-07-28 22:14:20,363][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.053020525723695755, 'train/grad_norm_pre_clip': 2.0741539001464844, 'timing/data': 0.000326450914144516, 'timing/model': 0.6194424200803041, 'learning_rate/qwen_vl_interface': 9.19788560448964e-06, 'learning_rate/action_model': 9.197885604489639e-05, 'epoch': 0.7})
+[2026-07-28 22:14:29,552][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.03042720817029476, 'train/grad_norm_pre_clip': 1.3061720132827759, 'timing/data': 0.0003340672701597214, 'timing/model': 0.6187269017100334, 'learning_rate/qwen_vl_interface': 9.195756624471094e-06, 'learning_rate/action_model': 9.195756624471094e-05, 'epoch': 0.7})
+[2026-07-28 22:14:38,980][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.04804763197898865, 'train/grad_norm_pre_clip': 1.4423882961273193, 'timing/data': 0.00034124962985515594, 'timing/model': 0.8473267164081335, 'learning_rate/qwen_vl_interface': 9.193625084093912e-06, 'learning_rate/action_model': 9.193625084093911e-05, 'epoch': 0.7})
+[2026-07-28 22:14:48,414][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.02486177533864975, 'train/grad_norm_pre_clip': 0.9061101675033569, 'timing/data': 0.000377558171749115, 'timing/model': 0.8469585422426462, 'learning_rate/qwen_vl_interface': 9.191490984741227e-06, 'learning_rate/action_model': 9.191490984741227e-05, 'epoch': 0.7})
+[2026-07-28 22:14:57,606][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.013355327770113945, 'train/grad_norm_pre_clip': 0.7752047777175903, 'timing/data': 0.0003281068056821823, 'timing/model': 0.6191122941672802, 'learning_rate/qwen_vl_interface': 9.189354327797835e-06, 'learning_rate/action_model': 9.189354327797836e-05, 'epoch': 0.7})
+[2026-07-28 22:15:06,768][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.024208206683397293, 'train/grad_norm_pre_clip': 1.250800371170044, 'timing/data': 0.0003947429358959198, 'timing/model': 0.6195983774960041, 'learning_rate/qwen_vl_interface': 9.18721511465019e-06, 'learning_rate/action_model': 9.187215114650191e-05, 'epoch': 0.7})
+[2026-07-28 22:15:15,926][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.026791077107191086, 'train/grad_norm_pre_clip': 0.9527576565742493, 'timing/data': 0.00039291754364967346, 'timing/model': 0.6207853630185127, 'learning_rate/qwen_vl_interface': 9.185073346686402e-06, 'learning_rate/action_model': 9.185073346686401e-05, 'epoch': 0.7})
+[2026-07-28 22:15:25,339][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.06768734753131866, 'train/grad_norm_pre_clip': 3.0006704330444336, 'timing/data': 0.00032576918601989746, 'timing/model': 0.6239556819200516, 'learning_rate/qwen_vl_interface': 9.182929025296242e-06, 'learning_rate/action_model': 9.182929025296242e-05, 'epoch': 0.71})
+[2026-07-28 22:15:34,757][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.04227888584136963, 'train/grad_norm_pre_clip': 1.5127060413360596, 'timing/data': 0.00039847008883953094, 'timing/model': 0.6208130549639463, 'learning_rate/qwen_vl_interface': 9.180782151871138e-06, 'learning_rate/action_model': 9.180782151871137e-05, 'epoch': 0.71})
+[2026-07-28 22:15:43,939][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.04260582476854324, 'train/grad_norm_pre_clip': 1.0423067808151245, 'timing/data': 0.0005949158221483231, 'timing/model': 0.6191377248615026, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.71})
+[2026-07-28 22:15:53,136][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.030678488314151764, 'train/grad_norm_pre_clip': 0.8589981198310852, 'timing/data': 0.000315934419631958, 'timing/model': 0.6196021866053343, 'learning_rate/qwen_vl_interface': 9.176480754490078e-06, 'learning_rate/action_model': 9.176480754490078e-05, 'epoch': 0.71})
+[2026-07-28 22:16:02,341][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.012837475165724754, 'train/grad_norm_pre_clip': 0.7317163944244385, 'timing/data': 0.000458715483546257, 'timing/model': 0.6198284607380629, 'learning_rate/qwen_vl_interface': 9.174326233325253e-06, 'learning_rate/action_model': 9.174326233325253e-05, 'epoch': 0.71})
+[2026-07-28 22:16:12,049][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.023653948679566383, 'train/grad_norm_pre_clip': 1.1211990118026733, 'timing/data': 0.0003710053861141205, 'timing/model': 0.8356528840959072, 'learning_rate/qwen_vl_interface': 9.17216916570774e-06, 'learning_rate/action_model': 9.172169165707741e-05, 'epoch': 0.71})
+[2026-07-28 22:16:21,235][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.054850056767463684, 'train/grad_norm_pre_clip': 1.7366621494293213, 'timing/data': 0.00037130340933799744, 'timing/model': 0.6173346098512411, 'learning_rate/qwen_vl_interface': 9.170009553037239e-06, 'learning_rate/action_model': 9.170009553037238e-05, 'epoch': 0.71})
+[2026-07-28 22:16:30,382][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.04395546764135361, 'train/grad_norm_pre_clip': 1.0729575157165527, 'timing/data': 0.00037845782935619354, 'timing/model': 0.6189548019319773, 'learning_rate/qwen_vl_interface': 9.167847396715098e-06, 'learning_rate/action_model': 9.167847396715098e-05, 'epoch': 0.71})
+[2026-07-28 22:16:39,546][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.02752457745373249, 'train/grad_norm_pre_clip': 0.9004694819450378, 'timing/data': 0.00036994926631450653, 'timing/model': 0.6189664099365473, 'learning_rate/qwen_vl_interface': 9.165682698144317e-06, 'learning_rate/action_model': 9.165682698144317e-05, 'epoch': 0.71})
+[2026-07-28 22:16:48,972][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.019287757575511932, 'train/grad_norm_pre_clip': 1.216200828552246, 'timing/data': 0.00035431794822216034, 'timing/model': 0.6198299527168274, 'learning_rate/qwen_vl_interface': 9.163515458729545e-06, 'learning_rate/action_model': 9.163515458729545e-05, 'epoch': 0.71})
+[2026-07-28 22:16:58,384][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.04981639236211777, 'train/grad_norm_pre_clip': 1.220678687095642, 'timing/data': 0.00036452896893024445, 'timing/model': 0.6199900023639202, 'learning_rate/qwen_vl_interface': 9.16134567987708e-06, 'learning_rate/action_model': 9.16134567987708e-05, 'epoch': 0.71})
+[2026-07-28 22:17:07,586][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.03341077268123627, 'train/grad_norm_pre_clip': 0.8956151008605957, 'timing/data': 0.0004029609262943268, 'timing/model': 0.6202332880347967, 'learning_rate/qwen_vl_interface': 9.159173362994868e-06, 'learning_rate/action_model': 9.159173362994869e-05, 'epoch': 0.71})
+[2026-07-28 22:17:16,940][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.0213584303855896, 'train/grad_norm_pre_clip': 0.7874531745910645, 'timing/data': 0.00037087686359882355, 'timing/model': 0.7861294262111187, 'learning_rate/qwen_vl_interface': 9.156998509492503e-06, 'learning_rate/action_model': 9.156998509492503e-05, 'epoch': 0.71})
+[2026-07-28 22:17:26,386][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.02238277718424797, 'train/grad_norm_pre_clip': 0.836430013179779, 'timing/data': 0.00030996091663837433, 'timing/model': 0.6201635599136353, 'learning_rate/qwen_vl_interface': 9.154821120781221e-06, 'learning_rate/action_model': 9.154821120781222e-05, 'epoch': 0.71})
+[2026-07-28 22:17:35,610][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.03388352692127228, 'train/grad_norm_pre_clip': 1.6740096807479858, 'timing/data': 0.0003681723028421402, 'timing/model': 0.619374256581068, 'learning_rate/qwen_vl_interface': 9.152641198273911e-06, 'learning_rate/action_model': 9.15264119827391e-05, 'epoch': 0.71})
+[2026-07-28 22:17:44,830][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.031855154782533646, 'train/grad_norm_pre_clip': 1.3658238649368286, 'timing/data': 0.00033520162105560303, 'timing/model': 0.620739059522748, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.71})
+[2026-07-28 22:17:54,225][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.03258272260427475, 'train/grad_norm_pre_clip': 1.129982352256775, 'timing/data': 0.000339290127158165, 'timing/model': 0.8621597122400999, 'learning_rate/qwen_vl_interface': 9.148273757530948e-06, 'learning_rate/action_model': 9.148273757530948e-05, 'epoch': 0.71})
+[2026-07-28 22:18:03,620][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.04227167367935181, 'train/grad_norm_pre_clip': 1.552211046218872, 'timing/data': 0.00033286213874816895, 'timing/model': 0.6188057400286198, 'learning_rate/qwen_vl_interface': 9.146086242129283e-06, 'learning_rate/action_model': 9.146086242129282e-05, 'epoch': 0.71})
+[2026-07-28 22:18:12,569][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.012718584388494492, 'train/grad_norm_pre_clip': 0.6319894194602966, 'timing/data': 0.0003328882157802582, 'timing/model': 0.6207812838256359, 'learning_rate/qwen_vl_interface': 9.143896198599553e-06, 'learning_rate/action_model': 9.143896198599553e-05, 'epoch': 0.71})
+[2026-07-28 22:18:21,559][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.012739941477775574, 'train/grad_norm_pre_clip': 0.7568367123603821, 'timing/data': 0.00032012350857257843, 'timing/model': 0.6223609354346991, 'learning_rate/qwen_vl_interface': 9.141703628362858e-06, 'learning_rate/action_model': 9.141703628362858e-05, 'epoch': 0.71})
+[2026-07-28 22:18:31,006][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.025692017748951912, 'train/grad_norm_pre_clip': 4.358061790466309, 'timing/data': 0.0003902968019247055, 'timing/model': 0.6202645823359489, 'learning_rate/qwen_vl_interface': 9.139508532841928e-06, 'learning_rate/action_model': 9.139508532841927e-05, 'epoch': 0.71})
+[2026-07-28 22:18:40,416][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.052788522094488144, 'train/grad_norm_pre_clip': 1.3701800107955933, 'timing/data': 0.00033872202038764954, 'timing/model': 0.6198046207427979, 'learning_rate/qwen_vl_interface': 9.137310913461139e-06, 'learning_rate/action_model': 9.137310913461139e-05, 'epoch': 0.71})
+[2026-07-28 22:18:49,623][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.022157924249768257, 'train/grad_norm_pre_clip': 1.484100580215454, 'timing/data': 0.00033449381589889526, 'timing/model': 0.6189281065016985, 'learning_rate/qwen_vl_interface': 9.135110771646502e-06, 'learning_rate/action_model': 9.135110771646502e-05, 'epoch': 0.71})
+[2026-07-28 22:18:58,822][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.019716357812285423, 'train/grad_norm_pre_clip': 1.2233116626739502, 'timing/data': 0.00035189278423786163, 'timing/model': 0.6213527265936136, 'learning_rate/qwen_vl_interface': 9.132908108825667e-06, 'learning_rate/action_model': 9.132908108825667e-05, 'epoch': 0.71})
+[2026-07-28 22:19:08,275][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.01619880460202694, 'train/grad_norm_pre_clip': 0.7459167242050171, 'timing/data': 0.0003297440707683563, 'timing/model': 0.6183912865817547, 'learning_rate/qwen_vl_interface': 9.130702926427915e-06, 'learning_rate/action_model': 9.130702926427914e-05, 'epoch': 0.71})
+[2026-07-28 22:19:17,667][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.012270499020814896, 'train/grad_norm_pre_clip': 0.7802963256835938, 'timing/data': 0.00033267587423324585, 'timing/model': 0.8405629992485046, 'learning_rate/qwen_vl_interface': 9.128495225884167e-06, 'learning_rate/action_model': 9.128495225884166e-05, 'epoch': 0.71})
+[2026-07-28 22:19:26,854][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.01933809369802475, 'train/grad_norm_pre_clip': 0.9799089431762695, 'timing/data': 0.00031466782093048096, 'timing/model': 0.6188334301114082, 'learning_rate/qwen_vl_interface': 9.126285008626975e-06, 'learning_rate/action_model': 9.126285008626974e-05, 'epoch': 0.71})
+[2026-07-28 22:19:36,187][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.06945102661848068, 'train/grad_norm_pre_clip': 2.5943973064422607, 'timing/data': 0.0004137754440307617, 'timing/model': 0.7870712671428919, 'learning_rate/qwen_vl_interface': 9.124072276090524e-06, 'learning_rate/action_model': 9.124072276090524e-05, 'epoch': 0.71})
+[2026-07-28 22:19:45,352][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.05075758695602417, 'train/grad_norm_pre_clip': 1.8849961757659912, 'timing/data': 0.00035043805837631226, 'timing/model': 0.6209530886262655, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.71})
+[2026-07-28 22:19:54,512][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.011726528406143188, 'train/grad_norm_pre_clip': 0.6980541944503784, 'timing/data': 0.00032423995435237885, 'timing/model': 0.6201132219284773, 'learning_rate/qwen_vl_interface': 9.119639270924754e-06, 'learning_rate/action_model': 9.119639270924754e-05, 'epoch': 0.71})
+[2026-07-28 22:20:03,682][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.06155909225344658, 'train/grad_norm_pre_clip': 4.795238494873047, 'timing/data': 0.00032674893736839294, 'timing/model': 0.6204538904130459, 'learning_rate/qwen_vl_interface': 9.11741900117196e-06, 'learning_rate/action_model': 9.11741900117196e-05, 'epoch': 0.71})
+[2026-07-28 22:20:13,320][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.05864044278860092, 'train/grad_norm_pre_clip': 3.0270211696624756, 'timing/data': 0.00031685642898082733, 'timing/model': 0.8528241217136383, 'learning_rate/qwen_vl_interface': 9.115196221892969e-06, 'learning_rate/action_model': 9.115196221892969e-05, 'epoch': 0.71})
+[2026-07-28 22:20:22,477][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.020174618810415268, 'train/grad_norm_pre_clip': 1.371298909187317, 'timing/data': 0.0003047119826078415, 'timing/model': 0.6191979646682739, 'learning_rate/qwen_vl_interface': 9.112970934530111e-06, 'learning_rate/action_model': 9.112970934530111e-05, 'epoch': 0.71})
+[2026-07-28 22:20:31,636][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.04341074824333191, 'train/grad_norm_pre_clip': 2.3905463218688965, 'timing/data': 0.00038651376962661743, 'timing/model': 0.6194060612469912, 'learning_rate/qwen_vl_interface': 9.110743140527353e-06, 'learning_rate/action_model': 9.110743140527352e-05, 'epoch': 0.71})
+[2026-07-28 22:20:40,826][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.021107954904437065, 'train/grad_norm_pre_clip': 1.3308767080307007, 'timing/data': 0.0003901086747646332, 'timing/model': 0.6227332428097725, 'learning_rate/qwen_vl_interface': 9.108512841330288e-06, 'learning_rate/action_model': 9.108512841330287e-05, 'epoch': 0.71})
+[2026-07-28 22:20:50,207][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.028667666018009186, 'train/grad_norm_pre_clip': 1.4159561395645142, 'timing/data': 0.0004216674715280533, 'timing/model': 0.619576808065176, 'learning_rate/qwen_vl_interface': 9.106280038386129e-06, 'learning_rate/action_model': 9.10628003838613e-05, 'epoch': 0.71})
+[2026-07-28 22:20:59,678][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.046286799013614655, 'train/grad_norm_pre_clip': 2.098588705062866, 'timing/data': 0.0003127641975879669, 'timing/model': 0.8520498834550381, 'learning_rate/qwen_vl_interface': 9.104044733143723e-06, 'learning_rate/action_model': 9.104044733143723e-05, 'epoch': 0.71})
+[2026-07-28 22:21:09,038][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.02355073019862175, 'train/grad_norm_pre_clip': 0.7210916876792908, 'timing/data': 0.0003321748226881027, 'timing/model': 0.6174893621355295, 'learning_rate/qwen_vl_interface': 9.10180692705353e-06, 'learning_rate/action_model': 9.10180692705353e-05, 'epoch': 0.71})
+[2026-07-28 22:21:18,305][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.028097642585635185, 'train/grad_norm_pre_clip': 1.2729856967926025, 'timing/data': 0.00031246989965438843, 'timing/model': 0.632408345118165, 'learning_rate/qwen_vl_interface': 9.099566621567643e-06, 'learning_rate/action_model': 9.099566621567642e-05, 'epoch': 0.71})
+[2026-07-28 22:21:27,548][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.04688100516796112, 'train/grad_norm_pre_clip': 1.8045233488082886, 'timing/data': 0.0003740377724170685, 'timing/model': 0.6324939373880625, 'learning_rate/qwen_vl_interface': 9.097323818139767e-06, 'learning_rate/action_model': 9.097323818139767e-05, 'epoch': 0.71})
+[2026-07-28 22:21:37,015][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.028410673141479492, 'train/grad_norm_pre_clip': 1.0853697061538696, 'timing/data': 0.0003239847719669342, 'timing/model': 0.8514844980090857, 'learning_rate/qwen_vl_interface': 9.095078518225237e-06, 'learning_rate/action_model': 9.095078518225236e-05, 'epoch': 0.71})
+[2026-07-28 22:21:46,256][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.03993555158376694, 'train/grad_norm_pre_clip': 1.3850550651550293, 'timing/data': 0.0003204401582479477, 'timing/model': 0.6331021506339312, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.71})
+[2026-07-28 22:21:55,493][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.012965986505150795, 'train/grad_norm_pre_clip': 0.6783137917518616, 'timing/data': 0.0003237500786781311, 'timing/model': 0.6351829748600721, 'learning_rate/qwen_vl_interface': 9.09058043476563e-06, 'learning_rate/action_model': 9.09058043476563e-05, 'epoch': 0.71})
+[2026-07-28 22:22:04,900][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.02442404441535473, 'train/grad_norm_pre_clip': 1.1211259365081787, 'timing/data': 0.0003124848008155823, 'timing/model': 0.6326377019286156, 'learning_rate/qwen_vl_interface': 9.088327654139311e-06, 'learning_rate/action_model': 9.08832765413931e-05, 'epoch': 0.71})
+[2026-07-28 22:22:14,063][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.011306430213153362, 'train/grad_norm_pre_clip': 0.6082023978233337, 'timing/data': 0.00032121501863002777, 'timing/model': 0.6341038681566715, 'learning_rate/qwen_vl_interface': 9.08607238286385e-06, 'learning_rate/action_model': 9.08607238286385e-05, 'epoch': 0.71})
+[2026-07-28 22:22:23,359][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.02942493185400963, 'train/grad_norm_pre_clip': 0.7911571860313416, 'timing/data': 0.00031048431992530823, 'timing/model': 0.6695064958184958, 'learning_rate/qwen_vl_interface': 9.083814622402667e-06, 'learning_rate/action_model': 9.083814622402666e-05, 'epoch': 0.71})
+[2026-07-28 22:22:32,897][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.025832567363977432, 'train/grad_norm_pre_clip': 0.7172331213951111, 'timing/data': 0.00031519122421741486, 'timing/model': 0.6208390351384878, 'learning_rate/qwen_vl_interface': 9.0815543742208e-06, 'learning_rate/action_model': 9.081554374220799e-05, 'epoch': 0.71})
+[2026-07-28 22:22:42,090][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.07552842795848846, 'train/grad_norm_pre_clip': 1.7636646032333374, 'timing/data': 0.0003616847097873688, 'timing/model': 0.6202837228775024, 'learning_rate/qwen_vl_interface': 9.079291639784897e-06, 'learning_rate/action_model': 9.079291639784896e-05, 'epoch': 0.71})
+[2026-07-28 22:22:51,286][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.014378940686583519, 'train/grad_norm_pre_clip': 1.1088913679122925, 'timing/data': 0.00032260268926620483, 'timing/model': 0.6196051575243473, 'learning_rate/qwen_vl_interface': 9.077026420563223e-06, 'learning_rate/action_model': 9.077026420563223e-05, 'epoch': 0.71})
+[2026-07-28 22:23:00,709][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.05610115826129913, 'train/grad_norm_pre_clip': 2.169438600540161, 'timing/data': 0.00031223706901073456, 'timing/model': 0.616737587377429, 'learning_rate/qwen_vl_interface': 9.074758718025652e-06, 'learning_rate/action_model': 9.074758718025652e-05, 'epoch': 0.71})
+[2026-07-28 22:23:10,098][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.030945824459195137, 'train/grad_norm_pre_clip': 0.9737566113471985, 'timing/data': 0.0003689136356115341, 'timing/model': 0.8385265935212374, 'learning_rate/qwen_vl_interface': 9.072488533643678e-06, 'learning_rate/action_model': 9.072488533643677e-05, 'epoch': 0.71})
+[2026-07-28 22:23:19,263][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.019960949197411537, 'train/grad_norm_pre_clip': 0.6649893522262573, 'timing/data': 0.0003518592566251755, 'timing/model': 0.619586793705821, 'learning_rate/qwen_vl_interface': 9.070215868890391e-06, 'learning_rate/action_model': 9.070215868890391e-05, 'epoch': 0.71})
+[2026-07-28 22:23:28,405][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.02830551750957966, 'train/grad_norm_pre_clip': 0.8328803777694702, 'timing/data': 0.00035038217902183533, 'timing/model': 0.6179547868669033, 'learning_rate/qwen_vl_interface': 9.067940725240503e-06, 'learning_rate/action_model': 9.067940725240503e-05, 'epoch': 0.71})
+[2026-07-28 22:23:37,545][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.04898928850889206, 'train/grad_norm_pre_clip': 1.7771034240722656, 'timing/data': 0.00030961446464061737, 'timing/model': 0.618795670568943, 'learning_rate/qwen_vl_interface': 9.06566310417033e-06, 'learning_rate/action_model': 9.06566310417033e-05, 'epoch': 0.71})
+[2026-07-28 22:23:47,015][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.04560604318976402, 'train/grad_norm_pre_clip': 2.5312490463256836, 'timing/data': 0.0003228578716516495, 'timing/model': 0.6191518679261208, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.72})
+[2026-07-28 22:23:55,979][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.03827425464987755, 'train/grad_norm_pre_clip': 1.1214091777801514, 'timing/data': 0.0003692079335451126, 'timing/model': 0.6218664199113846, 'learning_rate/qwen_vl_interface': 9.061100435682429e-06, 'learning_rate/action_model': 9.061100435682429e-05, 'epoch': 0.72})
+[2026-07-28 22:24:05,154][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.03657369688153267, 'train/grad_norm_pre_clip': 0.867791473865509, 'timing/data': 0.0006250441074371338, 'timing/model': 0.6204217746853828, 'learning_rate/qwen_vl_interface': 9.058815391225363e-06, 'learning_rate/action_model': 9.058815391225362e-05, 'epoch': 0.72})
+[2026-07-28 22:24:14,590][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.01781562902033329, 'train/grad_norm_pre_clip': 0.8575895428657532, 'timing/data': 0.00038770027458667755, 'timing/model': 0.8698536269366741, 'learning_rate/qwen_vl_interface': 9.056527875269344e-06, 'learning_rate/action_model': 9.056527875269344e-05, 'epoch': 0.72})
+[2026-07-28 22:24:24,035][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.028318529948592186, 'train/grad_norm_pre_clip': 0.5567948818206787, 'timing/data': 0.0003290679305791855, 'timing/model': 0.6445647142827511, 'learning_rate/qwen_vl_interface': 9.054237889298714e-06, 'learning_rate/action_model': 9.054237889298714e-05, 'epoch': 0.72})
+[2026-07-28 22:24:33,231][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.025383392348885536, 'train/grad_norm_pre_clip': 0.8704405426979065, 'timing/data': 0.0003222152590751648, 'timing/model': 0.6361279729753733, 'learning_rate/qwen_vl_interface': 9.05194543479942e-06, 'learning_rate/action_model': 9.05194543479942e-05, 'epoch': 0.72})
+[2026-07-28 22:24:42,404][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.02277432568371296, 'train/grad_norm_pre_clip': 0.5663299560546875, 'timing/data': 0.00037369877099990845, 'timing/model': 0.6187922302633524, 'learning_rate/qwen_vl_interface': 9.049650513259007e-06, 'learning_rate/action_model': 9.049650513259008e-05, 'epoch': 0.72})
+[2026-07-28 22:24:51,749][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.035195864737033844, 'train/grad_norm_pre_clip': 0.8438628315925598, 'timing/data': 0.00031555257737636566, 'timing/model': 0.796846728771925, 'learning_rate/qwen_vl_interface': 9.047353126166632e-06, 'learning_rate/action_model': 9.047353126166632e-05, 'epoch': 0.72})
+[2026-07-28 22:25:01,135][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.010731082409620285, 'train/grad_norm_pre_clip': 0.405479371547699, 'timing/data': 0.00031056255102157593, 'timing/model': 0.618132658302784, 'learning_rate/qwen_vl_interface': 9.045053275013038e-06, 'learning_rate/action_model': 9.045053275013038e-05, 'epoch': 0.72})
+[2026-07-28 22:25:10,291][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.03695955127477646, 'train/grad_norm_pre_clip': 0.9354010224342346, 'timing/data': 0.0003259647637605667, 'timing/model': 0.6182963065803051, 'learning_rate/qwen_vl_interface': 9.042750961290578e-06, 'learning_rate/action_model': 9.042750961290578e-05, 'epoch': 0.72})
+[2026-07-28 22:25:19,445][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.011451082304120064, 'train/grad_norm_pre_clip': 0.39617449045181274, 'timing/data': 0.0003744475543498993, 'timing/model': 0.619126096367836, 'learning_rate/qwen_vl_interface': 9.040446186493195e-06, 'learning_rate/action_model': 9.040446186493194e-05, 'epoch': 0.72})
+[2026-07-28 22:25:28,854][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.020615367218852043, 'train/grad_norm_pre_clip': 1.113952875137329, 'timing/data': 0.0003642141819000244, 'timing/model': 0.8606106210500002, 'learning_rate/qwen_vl_interface': 9.038138952116435e-06, 'learning_rate/action_model': 9.038138952116435e-05, 'epoch': 0.72})
+[2026-07-28 22:25:38,051][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.0190106313675642, 'train/grad_norm_pre_clip': 0.6223781704902649, 'timing/data': 0.00031156279146671295, 'timing/model': 0.6225043330341578, 'learning_rate/qwen_vl_interface': 9.035829259657437e-06, 'learning_rate/action_model': 9.035829259657436e-05, 'epoch': 0.72})
+[2026-07-28 22:25:47,464][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.014255142770707607, 'train/grad_norm_pre_clip': 0.571488082408905, 'timing/data': 0.0003771483898162842, 'timing/model': 0.8704065475612879, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.72})
+[2026-07-28 22:25:56,645][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.014487171545624733, 'train/grad_norm_pre_clip': 1.1193084716796875, 'timing/data': 0.00031420402228832245, 'timing/model': 0.6197104752063751, 'learning_rate/qwen_vl_interface': 9.031202506489254e-06, 'learning_rate/action_model': 9.031202506489253e-05, 'epoch': 0.72})
+[2026-07-28 22:26:06,046][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.025627397000789642, 'train/grad_norm_pre_clip': 1.4319508075714111, 'timing/data': 0.00037091784179210663, 'timing/model': 0.6198045313358307, 'learning_rate/qwen_vl_interface': 9.02888544878232e-06, 'learning_rate/action_model': 9.028885448782319e-05, 'epoch': 0.72})
+[2026-07-28 22:26:15,211][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.029825057834386826, 'train/grad_norm_pre_clip': 1.3534619808197021, 'timing/data': 0.0003290250897407532, 'timing/model': 0.6205958910286427, 'learning_rate/qwen_vl_interface': 9.026565938997644e-06, 'learning_rate/action_model': 9.026565938997645e-05, 'epoch': 0.72})
+[2026-07-28 22:26:24,368][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.027213234454393387, 'train/grad_norm_pre_clip': 1.043840765953064, 'timing/data': 0.0003754720091819763, 'timing/model': 0.6190444398671389, 'learning_rate/qwen_vl_interface': 9.024243978640334e-06, 'learning_rate/action_model': 9.024243978640333e-05, 'epoch': 0.72})
+[2026-07-28 22:26:33,768][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.008771214634180069, 'train/grad_norm_pre_clip': 0.6850455403327942, 'timing/data': 0.00035735592246055603, 'timing/model': 0.6196437496691942, 'learning_rate/qwen_vl_interface': 9.02191956921708e-06, 'learning_rate/action_model': 9.02191956921708e-05, 'epoch': 0.72})
+[2026-07-28 22:26:43,210][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.01638176292181015, 'train/grad_norm_pre_clip': 1.1802057027816772, 'timing/data': 0.00032839737832546234, 'timing/model': 0.6204542275518179, 'learning_rate/qwen_vl_interface': 9.019592712236172e-06, 'learning_rate/action_model': 9.019592712236171e-05, 'epoch': 0.72})
+[2026-07-28 22:26:52,484][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.047216229140758514, 'train/grad_norm_pre_clip': 1.4856982231140137, 'timing/data': 0.0003427695482969284, 'timing/model': 0.6197794880717993, 'learning_rate/qwen_vl_interface': 9.017263409207478e-06, 'learning_rate/action_model': 9.017263409207478e-05, 'epoch': 0.72})
+[2026-07-28 22:27:02,087][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.031975578516721725, 'train/grad_norm_pre_clip': 1.0115247964859009, 'timing/data': 0.00029588304460048676, 'timing/model': 0.6208742447197437, 'learning_rate/qwen_vl_interface': 9.014931661642455e-06, 'learning_rate/action_model': 9.014931661642455e-05, 'epoch': 0.72})
+[2026-07-28 22:27:11,492][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.014641273766756058, 'train/grad_norm_pre_clip': 1.1052378416061401, 'timing/data': 0.000376066192984581, 'timing/model': 0.8607598710805178, 'learning_rate/qwen_vl_interface': 9.012597471054154e-06, 'learning_rate/action_model': 9.012597471054154e-05, 'epoch': 0.72})
+[2026-07-28 22:27:20,710][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.030370347201824188, 'train/grad_norm_pre_clip': 1.3139426708221436, 'timing/data': 0.0003389548510313034, 'timing/model': 0.6225090734660625, 'learning_rate/qwen_vl_interface': 9.010260838957206e-06, 'learning_rate/action_model': 9.010260838957205e-05, 'epoch': 0.72})
+[2026-07-28 22:27:29,891][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.0381787084043026, 'train/grad_norm_pre_clip': 0.9607139229774475, 'timing/data': 0.0003261435776948929, 'timing/model': 0.6276920568197966, 'learning_rate/qwen_vl_interface': 9.007921766867823e-06, 'learning_rate/action_model': 9.007921766867822e-05, 'epoch': 0.72})
+[2026-07-28 22:27:39,285][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.018111510202288628, 'train/grad_norm_pre_clip': 1.196415662765503, 'timing/data': 0.00039762817323207855, 'timing/model': 0.8291919976472855, 'learning_rate/qwen_vl_interface': 9.005580256303804e-06, 'learning_rate/action_model': 9.005580256303803e-05, 'epoch': 0.72})
+[2026-07-28 22:27:48,485][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.03456190973520279, 'train/grad_norm_pre_clip': 6.369165420532227, 'timing/data': 0.00032823532819747925, 'timing/model': 0.6197492126375437, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.72})
+[2026-07-28 22:27:57,912][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.022915318608283997, 'train/grad_norm_pre_clip': 1.6525146961212158, 'timing/data': 0.00031685642898082733, 'timing/model': 0.6200676783919334, 'learning_rate/qwen_vl_interface': 9.000889925830964e-06, 'learning_rate/action_model': 9.000889925830965e-05, 'epoch': 0.72})
+[2026-07-28 22:28:07,068][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.06256645917892456, 'train/grad_norm_pre_clip': 3.5545668601989746, 'timing/data': 0.00037564896047115326, 'timing/model': 0.6195607855916023, 'learning_rate/qwen_vl_interface': 8.998541108965648e-06, 'learning_rate/action_model': 8.998541108965648e-05, 'epoch': 0.72})
+[2026-07-28 22:28:16,245][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.006675867363810539, 'train/grad_norm_pre_clip': 0.5617417693138123, 'timing/data': 0.0007588639855384827, 'timing/model': 0.6200793143361807, 'learning_rate/qwen_vl_interface': 8.996189859712704e-06, 'learning_rate/action_model': 8.996189859712703e-05, 'epoch': 0.72})
+[2026-07-28 22:28:25,411][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.028164437040686607, 'train/grad_norm_pre_clip': 2.5108070373535156, 'timing/data': 0.00031455233693122864, 'timing/model': 0.6196182835847139, 'learning_rate/qwen_vl_interface': 8.993836179597833e-06, 'learning_rate/action_model': 8.993836179597833e-05, 'epoch': 0.72})
+[2026-07-28 22:28:34,826][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.0672786682844162, 'train/grad_norm_pre_clip': 1.0859581232070923, 'timing/data': 0.00038078799843788147, 'timing/model': 0.6182005126029253, 'learning_rate/qwen_vl_interface': 8.99148007014831e-06, 'learning_rate/action_model': 8.99148007014831e-05, 'epoch': 0.72})
+[2026-07-28 22:28:43,988][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.026377419009804726, 'train/grad_norm_pre_clip': 1.1067644357681274, 'timing/data': 0.000383647158741951, 'timing/model': 0.6198988724499941, 'learning_rate/qwen_vl_interface': 8.989121532892989e-06, 'learning_rate/action_model': 8.989121532892988e-05, 'epoch': 0.72})
+[2026-07-28 22:28:53,200][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.009160188026726246, 'train/grad_norm_pre_clip': 0.5261096954345703, 'timing/data': 0.0003723651170730591, 'timing/model': 0.6183940637856722, 'learning_rate/qwen_vl_interface': 8.986760569362302e-06, 'learning_rate/action_model': 8.986760569362301e-05, 'epoch': 0.72})
+[2026-07-28 22:29:02,553][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.022940106689929962, 'train/grad_norm_pre_clip': 0.9867254495620728, 'timing/data': 0.0003138575702905655, 'timing/model': 0.7866695187985897, 'learning_rate/qwen_vl_interface': 8.984397181088251e-06, 'learning_rate/action_model': 8.984397181088251e-05, 'epoch': 0.72})
+[2026-07-28 22:29:11,976][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.02935643680393696, 'train/grad_norm_pre_clip': 1.2784695625305176, 'timing/data': 0.00036372989416122437, 'timing/model': 0.6241985186934471, 'learning_rate/qwen_vl_interface': 8.982031369604412e-06, 'learning_rate/action_model': 8.982031369604411e-05, 'epoch': 0.72})
+[2026-07-28 22:29:21,195][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.023014677688479424, 'train/grad_norm_pre_clip': 1.1908701658248901, 'timing/data': 0.0003806650638580322, 'timing/model': 0.619179155677557, 'learning_rate/qwen_vl_interface': 8.979663136445936e-06, 'learning_rate/action_model': 8.979663136445935e-05, 'epoch': 0.72})
+[2026-07-28 22:29:30,422][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.02719336748123169, 'train/grad_norm_pre_clip': 0.8235551714897156, 'timing/data': 0.00037344545125961304, 'timing/model': 0.6209332086145878, 'learning_rate/qwen_vl_interface': 8.977292483149542e-06, 'learning_rate/action_model': 8.977292483149542e-05, 'epoch': 0.72})
+[2026-07-28 22:29:39,868][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.01646227017045021, 'train/grad_norm_pre_clip': 0.7890195846557617, 'timing/data': 0.0003842003643512726, 'timing/model': 0.8544569555670023, 'learning_rate/qwen_vl_interface': 8.974919411253523e-06, 'learning_rate/action_model': 8.974919411253523e-05, 'epoch': 0.72})
+[2026-07-28 22:29:48,809][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.017682412639260292, 'train/grad_norm_pre_clip': 0.9990382790565491, 'timing/data': 0.0003740154206752777, 'timing/model': 0.6202531717717648, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.72})
+[2026-07-28 22:29:57,759][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.014567523263394833, 'train/grad_norm_pre_clip': 0.9792371392250061, 'timing/data': 0.00040206126868724823, 'timing/model': 0.6236223634332418, 'learning_rate/qwen_vl_interface': 8.970166017823621e-06, 'learning_rate/action_model': 8.97016601782362e-05, 'epoch': 0.72})
+[2026-07-28 22:30:06,726][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.04629688709974289, 'train/grad_norm_pre_clip': 1.280795931816101, 'timing/data': 0.00030180998146533966, 'timing/model': 0.6192991640418768, 'learning_rate/qwen_vl_interface': 8.96778569937416e-06, 'learning_rate/action_model': 8.96778569937416e-05, 'epoch': 0.72})
+[2026-07-28 22:30:16,138][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.023339565843343735, 'train/grad_norm_pre_clip': 1.2097948789596558, 'timing/data': 0.0003646276891231537, 'timing/model': 0.6214229073375463, 'learning_rate/qwen_vl_interface': 8.965402968493922e-06, 'learning_rate/action_model': 8.965402968493922e-05, 'epoch': 0.72})
+[2026-07-28 22:30:25,553][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.04587375372648239, 'train/grad_norm_pre_clip': 1.9648996591567993, 'timing/data': 0.00032662227749824524, 'timing/model': 0.6211185250431299, 'learning_rate/qwen_vl_interface': 8.96301782672904e-06, 'learning_rate/action_model': 8.963017826729038e-05, 'epoch': 0.72})
+[2026-07-28 22:30:34,723][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.016559265553951263, 'train/grad_norm_pre_clip': 0.9325820207595825, 'timing/data': 0.0003107879310846329, 'timing/model': 0.6187779419124126, 'learning_rate/qwen_vl_interface': 8.960630275627197e-06, 'learning_rate/action_model': 8.960630275627197e-05, 'epoch': 0.72})
+[2026-07-28 22:30:43,881][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.02037438005208969, 'train/grad_norm_pre_clip': 1.0228397846221924, 'timing/data': 0.0003219936043024063, 'timing/model': 0.6205882001668215, 'learning_rate/qwen_vl_interface': 8.958240316737655e-06, 'learning_rate/action_model': 8.958240316737656e-05, 'epoch': 0.72})
+[2026-07-28 22:30:53,353][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.048692137002944946, 'train/grad_norm_pre_clip': 1.787725806236267, 'timing/data': 0.00041232816874980927, 'timing/model': 0.6204509474337101, 'learning_rate/qwen_vl_interface': 8.955847951611233e-06, 'learning_rate/action_model': 8.955847951611232e-05, 'epoch': 0.72})
+[2026-07-28 22:31:02,508][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.05870157480239868, 'train/grad_norm_pre_clip': 1.596374273300171, 'timing/data': 0.0003763679414987564, 'timing/model': 0.6210784520953894, 'learning_rate/qwen_vl_interface': 8.953453181800307e-06, 'learning_rate/action_model': 8.953453181800306e-05, 'epoch': 0.72})
+[2026-07-28 22:31:11,920][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.028691424056887627, 'train/grad_norm_pre_clip': 0.8103492856025696, 'timing/data': 0.0003249943256378174, 'timing/model': 0.8435537908226252, 'learning_rate/qwen_vl_interface': 8.951056008858819e-06, 'learning_rate/action_model': 8.95105600885882e-05, 'epoch': 0.72})
+[2026-07-28 22:31:21,247][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.04108954593539238, 'train/grad_norm_pre_clip': 0.972578227519989, 'timing/data': 0.00043023936450481415, 'timing/model': 0.7887231316417456, 'learning_rate/qwen_vl_interface': 8.94865643434227e-06, 'learning_rate/action_model': 8.948656434342269e-05, 'epoch': 0.72})
+[2026-07-28 22:31:30,407][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.021492760628461838, 'train/grad_norm_pre_clip': 0.4907348155975342, 'timing/data': 0.00036942027509212494, 'timing/model': 0.6189667712897062, 'learning_rate/qwen_vl_interface': 8.946254459807717e-06, 'learning_rate/action_model': 8.946254459807717e-05, 'epoch': 0.72})
+[2026-07-28 22:31:39,538][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.04111243784427643, 'train/grad_norm_pre_clip': 1.0616931915283203, 'timing/data': 0.0003249961882829666, 'timing/model': 0.6186328567564487, 'learning_rate/qwen_vl_interface': 8.943850086813773e-06, 'learning_rate/action_model': 8.943850086813773e-05, 'epoch': 0.72})
+[2026-07-28 22:31:48,684][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.06350544840097427, 'train/grad_norm_pre_clip': 1.5990087985992432, 'timing/data': 0.0003929324448108673, 'timing/model': 0.6200410965830088, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.72})
+[2026-07-28 22:31:58,335][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.0243507232517004, 'train/grad_norm_pre_clip': 0.8877690434455872, 'timing/data': 0.00033710338175296783, 'timing/model': 0.8542547505348921, 'learning_rate/qwen_vl_interface': 8.93903415168996e-06, 'learning_rate/action_model': 8.93903415168996e-05, 'epoch': 0.72})
+[2026-07-28 22:32:07,517][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.018692871555685997, 'train/grad_norm_pre_clip': 0.8071836233139038, 'timing/data': 0.0003451649099588394, 'timing/model': 0.619345435872674, 'learning_rate/qwen_vl_interface': 8.9366225926851e-06, 'learning_rate/action_model': 8.936622592685098e-05, 'epoch': 0.72})
+[2026-07-28 22:32:16,693][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.05119055509567261, 'train/grad_norm_pre_clip': 1.436002254486084, 'timing/data': 0.0003286581486463547, 'timing/model': 0.627786360681057, 'learning_rate/qwen_vl_interface': 8.934208641470863e-06, 'learning_rate/action_model': 8.934208641470863e-05, 'epoch': 0.73})
+[2026-07-28 22:32:26,151][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.03543589264154434, 'train/grad_norm_pre_clip': 0.7858939170837402, 'timing/data': 0.0005137957632541656, 'timing/model': 0.6189598385244608, 'learning_rate/qwen_vl_interface': 8.931792299613637e-06, 'learning_rate/action_model': 8.931792299613636e-05, 'epoch': 0.73})
+[2026-07-28 22:32:35,607][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.016791721805930138, 'train/grad_norm_pre_clip': 1.039610743522644, 'timing/data': 0.0003778897225856781, 'timing/model': 0.6219043359160423, 'learning_rate/qwen_vl_interface': 8.92937356868136e-06, 'learning_rate/action_model': 8.92937356868136e-05, 'epoch': 0.73})
+[2026-07-28 22:32:44,780][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.027314908802509308, 'train/grad_norm_pre_clip': 0.8222529888153076, 'timing/data': 0.00033735670149326324, 'timing/model': 0.6206864286214113, 'learning_rate/qwen_vl_interface': 8.926952450243522e-06, 'learning_rate/action_model': 8.926952450243522e-05, 'epoch': 0.73})
+[2026-07-28 22:32:54,033][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.02847352623939514, 'train/grad_norm_pre_clip': 1.1628096103668213, 'timing/data': 0.0004119165241718292, 'timing/model': 0.6187487877905369, 'learning_rate/qwen_vl_interface': 8.924528945871157e-06, 'learning_rate/action_model': 8.924528945871156e-05, 'epoch': 0.73})
+[2026-07-28 22:33:03,550][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.04742459952831268, 'train/grad_norm_pre_clip': 1.380925178527832, 'timing/data': 0.00036971643567085266, 'timing/model': 0.6184334922581911, 'learning_rate/qwen_vl_interface': 8.922103057136857e-06, 'learning_rate/action_model': 8.922103057136856e-05, 'epoch': 0.73})
+[2026-07-28 22:33:12,988][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.024072915315628052, 'train/grad_norm_pre_clip': 1.3135602474212646, 'timing/data': 0.0003113783895969391, 'timing/model': 0.620084073394537, 'learning_rate/qwen_vl_interface': 8.919674785614747e-06, 'learning_rate/action_model': 8.919674785614747e-05, 'epoch': 0.73})
+[2026-07-28 22:33:22,133][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.04392532631754875, 'train/grad_norm_pre_clip': 1.1896499395370483, 'timing/data': 0.00031698495149612427, 'timing/model': 0.6179361436516047, 'learning_rate/qwen_vl_interface': 8.91724413288051e-06, 'learning_rate/action_model': 8.91724413288051e-05, 'epoch': 0.73})
+[2026-07-28 22:33:31,487][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.02065795473754406, 'train/grad_norm_pre_clip': 0.8709526658058167, 'timing/data': 0.00033199600875377655, 'timing/model': 0.6219074204564095, 'learning_rate/qwen_vl_interface': 8.91481110051137e-06, 'learning_rate/action_model': 8.91481110051137e-05, 'epoch': 0.73})
+[2026-07-28 22:33:41,095][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.006862933281809092, 'train/grad_norm_pre_clip': 0.40756911039352417, 'timing/data': 0.0003226678818464279, 'timing/model': 0.7905014045536518, 'learning_rate/qwen_vl_interface': 8.912375690086097e-06, 'learning_rate/action_model': 8.912375690086096e-05, 'epoch': 0.73})
+[2026-07-28 22:33:50,309][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.03360180929303169, 'train/grad_norm_pre_clip': 0.9125657081604004, 'timing/data': 0.00038760900497436523, 'timing/model': 0.6202562395483255, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.73})
+[2026-07-28 22:33:59,484][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.023421920835971832, 'train/grad_norm_pre_clip': 0.8262100219726562, 'timing/data': 0.00033291056752204895, 'timing/model': 0.6335643194615841, 'learning_rate/qwen_vl_interface': 8.907497741389932e-06, 'learning_rate/action_model': 8.907497741389932e-05, 'epoch': 0.73})
+[2026-07-28 22:34:08,694][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.028406091034412384, 'train/grad_norm_pre_clip': 0.8531169891357422, 'timing/data': 0.00033793970942497253, 'timing/model': 0.6209427844733, 'learning_rate/qwen_vl_interface': 8.905055206284289e-06, 'learning_rate/action_model': 8.905055206284288e-05, 'epoch': 0.73})
+[2026-07-28 22:34:18,113][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.01729516312479973, 'train/grad_norm_pre_clip': 0.5982450842857361, 'timing/data': 0.0003198441118001938, 'timing/model': 0.8500991128385067, 'learning_rate/qwen_vl_interface': 8.902610299453005e-06, 'learning_rate/action_model': 8.902610299453004e-05, 'epoch': 0.73})
+[2026-07-28 22:34:27,408][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.02407430298626423, 'train/grad_norm_pre_clip': 1.141624927520752, 'timing/data': 0.00039581209421157837, 'timing/model': 0.6211568806320429, 'learning_rate/qwen_vl_interface': 8.900163022482554e-06, 'learning_rate/action_model': 8.900163022482553e-05, 'epoch': 0.73})
+[2026-07-28 22:34:36,661][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.014818653464317322, 'train/grad_norm_pre_clip': 0.6655237078666687, 'timing/data': 0.0003820769488811493, 'timing/model': 0.6197296213358641, 'learning_rate/qwen_vl_interface': 8.897713376960946e-06, 'learning_rate/action_model': 8.897713376960946e-05, 'epoch': 0.73})
+[2026-07-28 22:34:45,896][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.02674344927072525, 'train/grad_norm_pre_clip': 2.6391208171844482, 'timing/data': 0.00041887909173965454, 'timing/model': 0.6198810301721096, 'learning_rate/qwen_vl_interface': 8.895261364477731e-06, 'learning_rate/action_model': 8.895261364477731e-05, 'epoch': 0.73})
+[2026-07-28 22:34:55,378][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.01436705607920885, 'train/grad_norm_pre_clip': 0.8500477075576782, 'timing/data': 0.0003275536000728607, 'timing/model': 0.6217216737568378, 'learning_rate/qwen_vl_interface': 8.892806986623993e-06, 'learning_rate/action_model': 8.892806986623992e-05, 'epoch': 0.73})
+[2026-07-28 22:35:04,582][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.03797859698534012, 'train/grad_norm_pre_clip': 1.5607047080993652, 'timing/data': 0.00031638704240322113, 'timing/model': 0.6210940796881914, 'learning_rate/qwen_vl_interface': 8.890350244992349e-06, 'learning_rate/action_model': 8.890350244992348e-05, 'epoch': 0.73})
+[2026-07-28 22:35:14,054][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.010000083595514297, 'train/grad_norm_pre_clip': 0.4238298833370209, 'timing/data': 0.00037463195621967316, 'timing/model': 0.6193779185414314, 'learning_rate/qwen_vl_interface': 8.887891141176954e-06, 'learning_rate/action_model': 8.887891141176954e-05, 'epoch': 0.73})
+[2026-07-28 22:35:23,259][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.04372109845280647, 'train/grad_norm_pre_clip': 1.7127965688705444, 'timing/data': 0.00032821111381053925, 'timing/model': 0.6222029458731413, 'learning_rate/qwen_vl_interface': 8.885429676773495e-06, 'learning_rate/action_model': 8.885429676773495e-05, 'epoch': 0.73})
+[2026-07-28 22:35:32,683][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.05965482443571091, 'train/grad_norm_pre_clip': 2.245549440383911, 'timing/data': 0.00031531043350696564, 'timing/model': 0.6199329905211926, 'learning_rate/qwen_vl_interface': 8.882965853379188e-06, 'learning_rate/action_model': 8.882965853379187e-05, 'epoch': 0.73})
+[2026-07-28 22:35:41,853][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.0157326627522707, 'train/grad_norm_pre_clip': 1.2085541486740112, 'timing/data': 0.0003173071891069412, 'timing/model': 0.6207870598882437, 'learning_rate/qwen_vl_interface': 8.88049967259278e-06, 'learning_rate/action_model': 8.880499672592778e-05, 'epoch': 0.73})
+[2026-07-28 22:35:51,248][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.010244377888739109, 'train/grad_norm_pre_clip': 0.6927377581596375, 'timing/data': 0.0003765169531106949, 'timing/model': 0.8490614201873541, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.73})
+[2026-07-28 22:36:00,596][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.022070443257689476, 'train/grad_norm_pre_clip': 1.2936815023422241, 'timing/data': 0.00039842724800109863, 'timing/model': 0.7905954699963331, 'learning_rate/qwen_vl_interface': 8.875560245246307e-06, 'learning_rate/action_model': 8.875560245246307e-05, 'epoch': 0.73})
+[2026-07-28 22:36:09,836][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.03228707239031792, 'train/grad_norm_pre_clip': 1.8988980054855347, 'timing/data': 0.00037015415728092194, 'timing/model': 0.6204756516963243, 'learning_rate/qwen_vl_interface': 8.873087001891381e-06, 'learning_rate/action_model': 8.87308700189138e-05, 'epoch': 0.73})
+[2026-07-28 22:36:19,023][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.014711592346429825, 'train/grad_norm_pre_clip': 0.5607218146324158, 'timing/data': 0.00038653798401355743, 'timing/model': 0.6213579215109348, 'learning_rate/qwen_vl_interface': 8.870611407554635e-06, 'learning_rate/action_model': 8.870611407554636e-05, 'epoch': 0.73})
+[2026-07-28 22:36:28,460][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.014058751054108143, 'train/grad_norm_pre_clip': 1.102698802947998, 'timing/data': 0.0003825407475233078, 'timing/model': 0.6196142118424177, 'learning_rate/qwen_vl_interface': 8.868133463842456e-06, 'learning_rate/action_model': 8.868133463842457e-05, 'epoch': 0.73})
+[2026-07-28 22:36:37,878][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.04232045263051987, 'train/grad_norm_pre_clip': 1.5458933115005493, 'timing/data': 0.0006851945072412491, 'timing/model': 0.8525538463145494, 'learning_rate/qwen_vl_interface': 8.865653172362755e-06, 'learning_rate/action_model': 8.865653172362754e-05, 'epoch': 0.73})
+[2026-07-28 22:36:47,073][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.030079126358032227, 'train/grad_norm_pre_clip': 1.3113441467285156, 'timing/data': 0.00037078000605106354, 'timing/model': 0.6243200320750475, 'learning_rate/qwen_vl_interface': 8.863170534724962e-06, 'learning_rate/action_model': 8.863170534724961e-05, 'epoch': 0.73})
+[2026-07-28 22:36:56,496][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.02996782585978508, 'train/grad_norm_pre_clip': 1.0849553346633911, 'timing/data': 0.0003762226551771164, 'timing/model': 0.8683807589113712, 'learning_rate/qwen_vl_interface': 8.860685552540038e-06, 'learning_rate/action_model': 8.860685552540037e-05, 'epoch': 0.73})
+[2026-07-28 22:37:05,796][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.011766137555241585, 'train/grad_norm_pre_clip': 0.6786105632781982, 'timing/data': 0.0003262162208557129, 'timing/model': 0.6267350576817989, 'learning_rate/qwen_vl_interface': 8.858198227420461e-06, 'learning_rate/action_model': 8.85819822742046e-05, 'epoch': 0.73})
+[2026-07-28 22:37:15,272][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.02340022101998329, 'train/grad_norm_pre_clip': 1.403381586074829, 'timing/data': 0.00035645999014377594, 'timing/model': 0.621988620609045, 'learning_rate/qwen_vl_interface': 8.855708560980225e-06, 'learning_rate/action_model': 8.855708560980225e-05, 'epoch': 0.73})
+[2026-07-28 22:37:24,496][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.018688634037971497, 'train/grad_norm_pre_clip': 1.8636354207992554, 'timing/data': 0.0004034508019685745, 'timing/model': 0.6202173661440611, 'learning_rate/qwen_vl_interface': 8.853216554834852e-06, 'learning_rate/action_model': 8.85321655483485e-05, 'epoch': 0.73})
+[2026-07-28 22:37:33,785][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.026538338512182236, 'train/grad_norm_pre_clip': 0.8463502526283264, 'timing/data': 0.0003881137818098068, 'timing/model': 0.635397532954812, 'learning_rate/qwen_vl_interface': 8.850722210601376e-06, 'learning_rate/action_model': 8.850722210601377e-05, 'epoch': 0.73})
+[2026-07-28 22:37:43,223][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.04091790318489075, 'train/grad_norm_pre_clip': 4.207153797149658, 'timing/data': 0.00032876618206501007, 'timing/model': 0.622726321220398, 'learning_rate/qwen_vl_interface': 8.848225529898351e-06, 'learning_rate/action_model': 8.84822552989835e-05, 'epoch': 0.73})
+[2026-07-28 22:37:52,640][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.01721358858048916, 'train/grad_norm_pre_clip': 1.064738392829895, 'timing/data': 0.00037050433456897736, 'timing/model': 0.6202579513192177, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.73})
+[2026-07-28 22:38:01,582][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.04754456877708435, 'train/grad_norm_pre_clip': 4.0157647132873535, 'timing/data': 0.00037035346031188965, 'timing/model': 0.6208119112998247, 'learning_rate/qwen_vl_interface': 8.84322516556544e-06, 'learning_rate/action_model': 8.84322516556544e-05, 'epoch': 0.73})
+[2026-07-28 22:38:10,533][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.09564466774463654, 'train/grad_norm_pre_clip': 2.2751646041870117, 'timing/data': 0.00040297769010066986, 'timing/model': 0.6209290903061628, 'learning_rate/qwen_vl_interface': 8.840721485180237e-06, 'learning_rate/action_model': 8.840721485180236e-05, 'epoch': 0.73})
+[2026-07-28 22:38:19,874][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.008911894634366035, 'train/grad_norm_pre_clip': 0.870441198348999, 'timing/data': 0.0003391522914171219, 'timing/model': 0.79874537140131, 'learning_rate/qwen_vl_interface': 8.838215474814848e-06, 'learning_rate/action_model': 8.838215474814847e-05, 'epoch': 0.73})
+[2026-07-28 22:38:29,326][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.03441174328327179, 'train/grad_norm_pre_clip': 1.143375277519226, 'timing/data': 0.0003873184323310852, 'timing/model': 0.8468456547707319, 'learning_rate/qwen_vl_interface': 8.835707136095392e-06, 'learning_rate/action_model': 8.835707136095391e-05, 'epoch': 0.73})
+[2026-07-28 22:38:38,488][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.012530623003840446, 'train/grad_norm_pre_clip': 1.116929292678833, 'timing/data': 0.00032876431941986084, 'timing/model': 0.6200248654931784, 'learning_rate/qwen_vl_interface': 8.833196470649505e-06, 'learning_rate/action_model': 8.833196470649505e-05, 'epoch': 0.73})
+[2026-07-28 22:38:47,672][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.07613930106163025, 'train/grad_norm_pre_clip': 2.445327043533325, 'timing/data': 0.0003687143325805664, 'timing/model': 0.619020938873291, 'learning_rate/qwen_vl_interface': 8.83068348010633e-06, 'learning_rate/action_model': 8.83068348010633e-05, 'epoch': 0.73})
+[2026-07-28 22:38:57,137][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.053989801555871964, 'train/grad_norm_pre_clip': 2.189182996749878, 'timing/data': 0.00031588226556777954, 'timing/model': 0.8523786142468452, 'learning_rate/qwen_vl_interface': 8.82816816609652e-06, 'learning_rate/action_model': 8.828168166096521e-05, 'epoch': 0.73})
+[2026-07-28 22:39:06,331][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.04910042881965637, 'train/grad_norm_pre_clip': 2.6007561683654785, 'timing/data': 0.0003868546336889267, 'timing/model': 0.6200263071805239, 'learning_rate/qwen_vl_interface': 8.825650530252238e-06, 'learning_rate/action_model': 8.825650530252238e-05, 'epoch': 0.73})
+[2026-07-28 22:39:15,729][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.04057975858449936, 'train/grad_norm_pre_clip': 1.591639757156372, 'timing/data': 0.0003137625753879547, 'timing/model': 0.6197574362158775, 'learning_rate/qwen_vl_interface': 8.823130574207144e-06, 'learning_rate/action_model': 8.823130574207143e-05, 'epoch': 0.73})
+[2026-07-28 22:39:24,914][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.061001501977443695, 'train/grad_norm_pre_clip': 2.442129611968994, 'timing/data': 0.0003149006515741348, 'timing/model': 0.6199630778282881, 'learning_rate/qwen_vl_interface': 8.82060829959641e-06, 'learning_rate/action_model': 8.820608299596411e-05, 'epoch': 0.73})
+[2026-07-28 22:39:34,302][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.027976039797067642, 'train/grad_norm_pre_clip': 2.06258487701416, 'timing/data': 0.00042081065475940704, 'timing/model': 0.6256586872041225, 'learning_rate/qwen_vl_interface': 8.81808370805672e-06, 'learning_rate/action_model': 8.818083708056721e-05, 'epoch': 0.73})
+[2026-07-28 22:39:43,719][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.03646422177553177, 'train/grad_norm_pre_clip': 0.6894313097000122, 'timing/data': 0.00039221346378326416, 'timing/model': 0.6200352981686592, 'learning_rate/qwen_vl_interface': 8.815556801226251e-06, 'learning_rate/action_model': 8.815556801226251e-05, 'epoch': 0.73})
+[2026-07-28 22:39:52,963][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.02234325185418129, 'train/grad_norm_pre_clip': 1.2729723453521729, 'timing/data': 0.0005120113492012024, 'timing/model': 0.6322603635489941, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.73})
+[2026-07-28 22:40:02,180][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.05674758553504944, 'train/grad_norm_pre_clip': 1.221558690071106, 'timing/data': 0.00032635778188705444, 'timing/model': 0.625394344329834, 'learning_rate/qwen_vl_interface': 8.810496048253203e-06, 'learning_rate/action_model': 8.810496048253202e-05, 'epoch': 0.73})
+[2026-07-28 22:40:11,590][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.01379297487437725, 'train/grad_norm_pre_clip': 0.6916541457176208, 'timing/data': 0.0003229808062314987, 'timing/model': 0.6197173558175564, 'learning_rate/qwen_vl_interface': 8.80796220539449e-06, 'learning_rate/action_model': 8.80796220539449e-05, 'epoch': 0.73})
+[2026-07-28 22:40:20,988][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4000
+[2026-07-28 22:40:31,355][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4000
+[2026-07-28 22:56:58,743][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000
+[2026-07-28 22:56:58,744][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.042078275233507156, 'train/grad_norm_pre_clip': 1.4446320533752441, 'eval/action_loss/seconds': 10.363419380038977, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.03596002981066704, 'eval/latency_3/loss': 0.03596002981066704, 'eval/flappy/loss': 0.03596002981066704, 'eval/flappy/latency_3/loss': 0.03596002981066704, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 377.76000444367526, 'rl_games_eval/mid_train/mean_length': 3062.25, 'rl_games_eval/mid_train/std_reward': 126.19024436197301, 'rl_games_eval/mid_train/std_length': 1016.0166767824237, 'rl_games_eval/mid_train/macro_mean_reward': 377.76000444367526, 'rl_games_eval/mid_train/macro_mean_length': 3062.25, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 377.76000444367526, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 3062.25, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 126.19024436197301, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1016.0166767824237, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 377.76000444367526, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003125816583633423, 'timing/model': 0.6183772627264261, 'learning_rate/qwen_vl_interface': 8.805426053812731e-06, 'learning_rate/action_model': 8.805426053812731e-05, 'epoch': 0.73})
+[2026-07-28 22:56:58,747][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state
+[2026-07-28 22:56:58,747][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 22:57:36,373][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state/pytorch_model
+[2026-07-28 22:57:36,373][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state/sampler.bin
+[2026-07-28 22:57:36,373][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state/sampler_1.bin
+[2026-07-28 22:57:36,375][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state/random_states_0.pkl
+[2026-07-28 22:57:44,392][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000
+[2026-07-28 22:57:44,393][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 22:57:44,413][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 22:57:44,414][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints
+[2026-07-28 23:19:11,231][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 23:19:11,232][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21'
+[2026-07-28 23:19:11,232][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 23:19:11,451][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 23:19:11,451][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 23:19:11,453][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 23:19:11,453][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 23:19:12,198][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 23:19:12,214][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 23:19:12,552][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 23:19:12,568][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 23:19:14,024][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 23:19:14,298][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 23:19:14,975][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 23:19:14,976][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 23:19:14,976][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 23:19:14,976][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 23:19:14,977][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 23:19:14,977][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 23:19:14,977][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 23:19:14,978][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 23:19:14,978][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 23:19:14,978][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 23:19:14,979][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 23:19:14,979][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 23:19:14,979][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 23:19:14,980][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 23:19:14,980][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 23:19:14,980][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 23:19:14,981][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 23:19:14,981][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 23:19:14,981][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 23:19:14,982][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 23:19:14,982][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 23:19:14,982][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 23:19:14,983][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 23:19:14,983][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 23:19:14,983][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 23:19:14,984][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 23:19:14,984][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 23:19:14,984][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 23:19:14,984][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 23:19:14,984][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 23:19:14,984][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 23:19:15,453][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 23:19:15,454][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 23:19:15,454][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 23:19:15,454][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 23:19:15,454][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 23:19:15,455][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 23:19:15,455][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 23:19:15,455][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 23:19:15,455][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 23:19:15,455][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 23:19:15,456][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 23:19:15,456][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 23:19:15,456][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 23:19:15,457][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 23:19:15,458][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 23:19:15,458][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 23:19:15,458][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 23:19:15,458][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 23:19:15,459][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 23:19:15,459][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 23:19:15,459][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 23:19:15,459][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 23:19:15,460][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 23:19:15,460][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 23:19:15,460][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 23:19:15,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 23:19:15,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 23:19:15,461][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 23:19:15,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 23:19:15,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 23:19:15,462][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 23:19:16,688][starVLA.dataloader][INFO] - Created train VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 23:19:16,691][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 23:19:16,700][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 23:19:16,716][starVLA.dataloader][INFO] - Created eval VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 23:19:16,723][starVLA.training.train_starvla][INFO] - LR Group qwen_vl_interface: lr=1e-05, num_params=713
+[2026-07-28 23:19:16,723][starVLA.training.train_starvla][INFO] - LR Group action_model: lr=0.0001, num_params=16
+[2026-07-28 23:19:17,438][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state, steps: 4000
+[2026-07-28 23:19:17,450][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/config.full.yaml
+[2026-07-28 23:19:17,463][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/config.yaml
+[2026-07-28 23:19:17,464][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state, steps: 4000
+[2026-07-28 23:19:17,469][accelerate.accelerator][INFO] - Since you passed both train and evaluation dataloader, `is_train_batch_min` (here True will decide the `train_batch_size` (8).
+[2026-07-28 23:19:29,334][accelerate.accelerator][INFO] - Loading states from /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state
+[2026-07-28 23:19:29,334][accelerate.accelerator][INFO] - Loading DeepSpeed Model and Optimizer
+[2026-07-28 23:20:06,102][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer loaded from input dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints/steps_4000_state/pytorch_model
+[2026-07-28 23:20:06,102][accelerate.checkpointing][INFO] - All model weights loaded successfully
+[2026-07-28 23:20:06,102][accelerate.checkpointing][INFO] - All optimizer states loaded successfully
+[2026-07-28 23:20:06,102][accelerate.checkpointing][INFO] - All scheduler states loaded successfully
+[2026-07-28 23:20:06,102][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully
+[2026-07-28 23:20:06,114][accelerate.checkpointing][INFO] - All random states loaded successfully
+[2026-07-28 23:20:06,115][accelerate.accelerator][INFO] - Loading in 0 custom states
+[2026-07-28 23:20:08,266][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-07-28 23:20:08,266][starVLA.training.train_starvla][INFO] -   Total optimization steps = 4000
+[2026-07-28 23:20:08,266][starVLA.training.train_starvla][INFO] -   Per device batch size = 8
+[2026-07-28 23:20:08,267][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-07-28 23:20:08,267][starVLA.training.train_starvla][INFO] -   Total batch size = 256
+[2026-07-28 23:20:08,268][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/checkpoints
+[2026-07-28 23:58:32,014][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000
+[2026-07-28 23:58:33,647][starVLA.training.train_starvla][INFO] - ... and that's all, folks!
+[2026-07-28 23:58:33,647][starVLA.training.train_starvla][INFO] - ... and that's all, folks!
+[2026-07-28 23:58:34,211][starVLA.training.train_starvla][INFO] - Destroyed distributed process group
+[2026-07-28 23:58:34,235][starVLA.training.train_starvla][INFO] - Destroyed distributed process group
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..be2f104ab6b1b1f13fb764ff66918587f79556f5
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 32.80000041425228, "episode_return_env": 32.80000041425228, "game_score": 6.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 285, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 285}
+{"episode_id": 5, "episode_return": 314.8000037074089, "episode_return_env": 314.8000037074089, "game_score": 67.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2556, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2556}
+{"episode_id": 7, "episode_return": 144.80000172555447, "episode_return_env": 144.80000172555447, "game_score": 30.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1189, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1189}
+{"episode_id": 9, "episode_return": 242.80000287294388, "episode_return_env": 242.80000287294388, "game_score": 51.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1980, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1980}
+{"episode_id": 11, "episode_return": 327.00000385940075, "episode_return_env": 327.00000385940075, "game_score": 69.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2660, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2660}
+{"episode_id": 13, "episode_return": 23.40000030398369, "episode_return_env": 23.40000030398369, "game_score": 4.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 209, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 209}
+{"episode_id": 15, "episode_return": 336.100003965199, "episode_return_env": 336.100003965199, "game_score": 72.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2733, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2733}
+{"episode_id": 17, "episode_return": 97.8000011742115, "episode_return_env": 97.8000011742115, "game_score": 20.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 809, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 809}
+{"episode_id": 19, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..e4aff3360d379f2c9f29a6535110afea44ef7e8d
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 291.3000034317374, "episode_return_env": 291.3000034317374, "game_score": 62.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2366, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2366}
+{"episode_id": 3, "episode_return": 6.800000101327896, "episode_return_env": 6.800000101327896, "game_score": 1.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 70, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 70}
+{"episode_id": 5, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 7, "episode_return": 230.80000272393227, "episode_return_env": 230.80000272393227, "game_score": 49.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1878, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1878}
+{"episode_id": 9, "episode_return": 252.00000298023224, "episode_return_env": 252.00000298023224, "game_score": 53.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2054, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2054}
+{"episode_id": 11, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 19, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..09d36d9f943f1a1ea0a9e3e51f57f9f3783263be
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 7, "episode_return": 23.40000030398369, "episode_return_env": 23.40000030398369, "game_score": 4.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 209, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 209}
+{"episode_id": 9, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 11, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 13, "episode_return": 81.30000097304583, "episode_return_env": 81.30000097304583, "game_score": 17.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 671, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 671}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 19, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..e8df2cb000266a477451c9c38a2915c7d181121d
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 261.40000309050083, "episode_return_env": 261.40000309050083, "game_score": 55.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2130, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2130}
+{"episode_id": 2, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 4, "episode_return": 230.80000272393227, "episode_return_env": 230.80000272393227, "game_score": 49.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1878, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1878}
+{"episode_id": 6, "episode_return": 158.70000188797712, "episode_return_env": 158.70000188797712, "game_score": 33.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1301, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1301}
+{"episode_id": 8, "episode_return": 146.80000174045563, "episode_return_env": 146.80000174045563, "game_score": 31.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1200, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1200}
+{"episode_id": 10, "episode_return": 286.80000337958336, "episode_return_env": 286.80000337958336, "game_score": 61.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2330, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2330}
+{"episode_id": 12, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 14, "episode_return": 146.80000174045563, "episode_return_env": 146.80000174045563, "game_score": 31.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1200, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1200}
+{"episode_id": 16, "episode_return": 356.8000041991472, "episode_return_env": 356.8000041991472, "game_score": 76.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2895, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2895}
+{"episode_id": 18, "episode_return": 228.2000027000904, "episode_return_env": 228.2000027000904, "game_score": 48.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1861, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1861}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..11430cd6f33710cad24f01130ae955b8affa1563
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 67.30000080913305, "episode_return_env": 67.30000080913305, "game_score": 14.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 558, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 558}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 34.80000042915344, "episode_return_env": 34.80000042915344, "game_score": 7.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 296, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 296}
+{"episode_id": 7, "episode_return": 328.8000038713217, "episode_return_env": 328.8000038713217, "game_score": 70.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2669, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2669}
+{"episode_id": 9, "episode_return": 421.80000495910645, "episode_return_env": 421.80000495910645, "game_score": 90.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3419, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3419}
+{"episode_id": 11, "episode_return": 67.30000080913305, "episode_return_env": 67.30000080913305, "game_score": 14.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 558, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 558}
+{"episode_id": 13, "episode_return": 393.8000046312809, "episode_return_env": 393.8000046312809, "game_score": 84.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3193, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3193}
+{"episode_id": 15, "episode_return": 221.30000261217356, "episode_return_env": 221.30000261217356, "game_score": 47.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1801, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1801}
+{"episode_id": 17, "episode_return": 305.3000035956502, "episode_return_env": 305.3000035956502, "game_score": 65.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2479, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2479}
+{"episode_id": 19, "episode_return": 244.80000288784504, "episode_return_env": 244.80000288784504, "game_score": 52.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1991, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1991}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..5db8a89c2e2f89a378315c71a2f22e6fbf69d022
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 113.40000134706497, "episode_return_env": 113.40000134706497, "game_score": 24.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 929, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 929}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 7, "episode_return": 158.0000018775463, "episode_return_env": 158.0000018775463, "game_score": 33.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1294, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1294}
+{"episode_id": 9, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 11, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 409.9000048264861, "episode_return_env": 409.9000048264861, "game_score": 87.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3327, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3327}
+{"episode_id": 19, "episode_return": 193.300002284348, "episode_return_env": 193.300002284348, "game_score": 41.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1575, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1575}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..aecdda50b612f4adcf58330523c4df966ad24578
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 8.400000125169754, "episode_return_env": 8.400000125169754, "game_score": 1.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 86, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 86}
+{"episode_id": 3, "episode_return": 13.700000189244747, "episode_return_env": 13.700000189244747, "game_score": 2.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 130, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 130}
+{"episode_id": 5, "episode_return": 9.200000137090683, "episode_return_env": 9.200000137090683, "game_score": 1.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 94, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 94}
+{"episode_id": 7, "episode_return": 93.20000112056732, "episode_return_env": 93.20000112056732, "game_score": 19.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 772, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 772}
+{"episode_id": 9, "episode_return": 15.300000198185444, "episode_return_env": 15.300000198185444, "game_score": 3.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 137, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 137}
+{"episode_id": 11, "episode_return": 13.800000190734863, "episode_return_env": 13.800000190734863, "game_score": 2.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 131, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 131}
+{"episode_id": 13, "episode_return": 17.900000236928463, "episode_return_env": 17.900000236928463, "game_score": 3.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 163, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 163}
+{"episode_id": 15, "episode_return": 9.200000137090683, "episode_return_env": 9.200000137090683, "game_score": 1.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 94, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 94}
+{"episode_id": 17, "episode_return": 10.500000141561031, "episode_return_env": 10.500000141561031, "game_score": 2.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 98, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 98}
+{"episode_id": 19, "episode_return": 101.9000012204051, "episode_return_env": 101.9000012204051, "game_score": 21.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 841, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 841}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..df5353e2ba2ca1bd177f9ba794b7f8e3447c8e98
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 7, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 9, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 11, "episode_return": 67.30000080913305, "episode_return_env": 67.30000080913305, "game_score": 14.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 558, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 558}
+{"episode_id": 13, "episode_return": 440.80000518262386, "episode_return_env": 440.80000518262386, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3573, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3573}
+{"episode_id": 15, "episode_return": 141.80000168085098, "episode_return_env": 141.80000168085098, "game_score": 30.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1159, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1159}
+{"episode_id": 17, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 19, "episode_return": 99.80000118911266, "episode_return_env": 99.80000118911266, "game_score": 21.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 820, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 820}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..6d9084da0ba4636c42390f41cd9b885a5d30d52f
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 2, "episode_return": 214.20000253617764, "episode_return_env": 214.20000253617764, "game_score": 45.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1748, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1748}
+{"episode_id": 4, "episode_return": 55.40000067651272, "episode_return_env": 55.40000067651272, "game_score": 11.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 466, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 466}
+{"episode_id": 6, "episode_return": 90.80000108480453, "episode_return_env": 90.80000108480453, "game_score": 19.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 748, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 748}
+{"episode_id": 8, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 10, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 12, "episode_return": 118.80000141263008, "episode_return_env": 118.80000141263008, "game_score": 25.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 974, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 974}
+{"episode_id": 14, "episode_return": 146.80000174045563, "episode_return_env": 146.80000174045563, "game_score": 31.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1200, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1200}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 348.9000041112304, "episode_return_env": 348.9000041112304, "game_score": 74.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_2750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2834, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2834}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..852d3321a2633adbe423f987a5ed0abd19e80761
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 359.40000423789024, "episode_return_env": 359.40000423789024, "game_score": 76.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2921, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2921}
+{"episode_id": 3, "episode_return": 398.80000469088554, "episode_return_env": 398.80000469088554, "game_score": 85.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3234, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3234}
+{"episode_id": 5, "episode_return": 31.900000400841236, "episode_return_env": 31.900000400841236, "game_score": 6.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 276, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 276}
+{"episode_id": 7, "episode_return": 331.4000039100647, "episode_return_env": 331.4000039100647, "game_score": 70.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2695, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2695}
+{"episode_id": 9, "episode_return": 288.4000034034252, "episode_return_env": 288.4000034034252, "game_score": 61.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2346, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2346}
+{"episode_id": 11, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 81.30000097304583, "episode_return_env": 81.30000097304583, "game_score": 17.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 671, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 671}
+{"episode_id": 19, "episode_return": 20.80000026524067, "episode_return_env": 20.80000026524067, "game_score": 4.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 183, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 183}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..16c4610c8d208a4958861259121e522f1d0d0254
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 7, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 9, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 11, "episode_return": 202.80000239610672, "episode_return_env": 202.80000239610672, "game_score": 43.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1652, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1652}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 361.3000042513013, "episode_return_env": 361.3000042513013, "game_score": 77.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2931, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2931}
+{"episode_id": 17, "episode_return": 328.8000038713217, "episode_return_env": 328.8000038713217, "game_score": 70.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2669, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2669}
+{"episode_id": 19, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..2d151663382b9bd4e56f122cab9617b45081aa38
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 2, "episode_return": 387.600004568696, "episode_return_env": 387.600004568696, "game_score": 83.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3149, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3149}
+{"episode_id": 4, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 6, "episode_return": 258.100003041327, "episode_return_env": 258.100003041327, "game_score": 55.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2097, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2097}
+{"episode_id": 8, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 10, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 12, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 14, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..df2b77d71c01d72578168e43a3a696e53961dfe0
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 393.8000046312809, "episode_return_env": 393.8000046312809, "game_score": 84.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3193, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3193}
+{"episode_id": 7, "episode_return": 328.8000038713217, "episode_return_env": 328.8000038713217, "game_score": 70.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2669, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2669}
+{"episode_id": 9, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 11, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 440.80000518262386, "episode_return_env": 440.80000518262386, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3573, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3573}
+{"episode_id": 17, "episode_return": 328.8000038713217, "episode_return_env": 328.8000038713217, "game_score": 70.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2669, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2669}
+{"episode_id": 19, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_3750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..d4c4da19b9f66f85c879af467dcaf906fdee687e
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 2, "episode_return": 431.30000507086515, "episode_return_env": 431.30000507086515, "game_score": 92.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3496, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3496}
+{"episode_id": 4, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 6, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 8, "episode_return": 79.40000095963478, "episode_return_env": 79.40000095963478, "game_score": 16.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 661, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 661}
+{"episode_id": 10, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 12, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 14, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 289.30000341683626, "episode_return_env": 289.30000341683626, "game_score": 61.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2355, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2355}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..e01ef92ac7bab9ebf403e2a03025d87f35c9e69a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 6.800000101327896, "episode_return_env": 6.800000101327896, "game_score": 1.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 70, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 70}
+{"episode_id": 2, "episode_return": 41.70000051707029, "episode_return_env": 41.70000051707029, "game_score": 8.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 356, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 356}
+{"episode_id": 4, "episode_return": 53.30000064522028, "episode_return_env": 53.30000064522028, "game_score": 11.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 445, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 445}
+{"episode_id": 6, "episode_return": 90.80000108480453, "episode_return_env": 90.80000108480453, "game_score": 19.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 748, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 748}
+{"episode_id": 8, "episode_return": 43.80000053346157, "episode_return_env": 43.80000053346157, "game_score": 9.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 368, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 368}
+{"episode_id": 10, "episode_return": 34.80000042915344, "episode_return_env": 34.80000042915344, "game_score": 7.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 296, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 296}
+{"episode_id": 12, "episode_return": 95.3000011369586, "episode_return_env": 95.3000011369586, "game_score": 20.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 784, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 784}
+{"episode_id": 14, "episode_return": 29.800000369548798, "episode_return_env": 29.800000369548798, "game_score": 6.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 255, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 255}
+{"episode_id": 16, "episode_return": 76.80000092089176, "episode_return_env": 76.80000092089176, "game_score": 16.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 635, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 635}
+{"episode_id": 18, "episode_return": 144.30000171810389, "episode_return_env": 144.30000171810389, "game_score": 30.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1184, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1184}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..154bc2af4b934fd495e968a0992996f8da9d50aa
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 410.10000482946634, "episode_return_env": 410.10000482946634, "game_score": 87.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3329, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3329}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 7, "episode_return": 110.90000132471323, "episode_return_env": 110.90000132471323, "game_score": 23.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 913, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 913}
+{"episode_id": 9, "episode_return": 11.30000015348196, "episode_return_env": 11.30000015348196, "game_score": 2.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 106, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 106}
+{"episode_id": 11, "episode_return": 272.8000032156706, "episode_return_env": 272.8000032156706, "game_score": 58.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2217, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2217}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 328.8000038713217, "episode_return_env": 328.8000038713217, "game_score": 70.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2669, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2669}
+{"episode_id": 17, "episode_return": 66.70000080019236, "episode_return_env": 66.70000080019236, "game_score": 14.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 552, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 552}
+{"episode_id": 19, "episode_return": 178.80000211298466, "episode_return_env": 178.80000211298466, "game_score": 38.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/mid_train/step_750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1457, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1457}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..a6156a995f9d74231e89fd92ceb2fe247c6e7d16
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 25}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..a6156a995f9d74231e89fd92ceb2fe247c6e7d16
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 25}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..8a0278b3f827c2a307cad1f71737fed9d9f53d47
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,25 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 7, "episode_return": 328.8000038713217, "episode_return_env": 328.8000038713217, "game_score": 70.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2669, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2669}
+{"episode_id": 9, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 11, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 13, "episode_return": 83.20000100135803, "episode_return_env": 83.20000100135803, "game_score": 17.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 690, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 690}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 118.80000141263008, "episode_return_env": 118.80000141263008, "game_score": 25.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 974, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 974}
+{"episode_id": 19, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 21, "episode_return": 412.8000048547983, "episode_return_env": 412.8000048547983, "game_score": 88.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 63, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3347, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3347}
+{"episode_id": 23, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 65, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 25, "episode_return": 373.400004401803, "episode_return_env": 373.400004401803, "game_score": 79.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 67, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3034, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3034}
+{"episode_id": 27, "episode_return": 185.90000220388174, "episode_return_env": 185.90000220388174, "game_score": 39.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 69, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1519, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1519}
+{"episode_id": 29, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 71, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 31, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 73, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 33, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 75, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 35, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 77, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 37, "episode_return": 118.80000141263008, "episode_return_env": 118.80000141263008, "game_score": 25.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 79, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 974, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 974}
+{"episode_id": 39, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 81, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 41, "episode_return": 188.80000223219395, "episode_return_env": 188.80000223219395, "game_score": 40.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 83, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1539, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1539}
+{"episode_id": 43, "episode_return": 272.8000032156706, "episode_return_env": 272.8000032156706, "game_score": 58.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 85, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2217, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2217}
+{"episode_id": 45, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 87, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 47, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 89, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 49, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 91, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_plain_multi/latency_bench_eval/post_train/step_4000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/summary.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/summary.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..ce7dad56d7abf845dd57d6f02d76df86c6434c49
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_plain_multi/summary.jsonl
@@ -0,0 +1,8 @@
+{"steps": 500}
+{"steps": 1000}
+{"steps": 1500}
+{"steps": 2000}
+{"steps": 2500}
+{"steps": 3000}
+{"steps": 3500}
+{"steps": 4000}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1000.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1000.json
new file mode 100644
index 0000000000000000000000000000000000000000..9cbcc912a8c34dc5b1da8a98885d9ff2d7fec274
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1000.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 197.2250023353845,
+      "mean_length": 1609.65,
+      "std_reward": 147.19520162716503,
+      "std_length": 1187.3336209760087,
+      "episode_rewards": [
+        79.00000095367432,
+        57.10000068694353,
+        431.30000507086515,
+        421.10000494867563,
+        41.900000520050526,
+        65.60000079870224,
+        6.800000101327896,
+        46.90000057965517,
+        345.5000040754676,
+        234.70000276714563,
+        257.0000030398369,
+        368.6000043451786,
+        210.00000248849392,
+        261.50000309199095,
+        224.0000026524067,
+        289.5000034198165,
+        444.6000052243471,
+        36.60000045597553,
+        80.70000096410513,
+        42.10000052303076
+      ],
+      "episode_lengths": [
+        657,
+        474,
+        3496,
+        3412,
+        358,
+        550,
+        70,
+        399,
+        2809,
+        1908,
+        2095,
+        2995,
+        1715,
+        2131,
+        1828,
+        2357,
+        3600,
+        314,
+        665,
+        360
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 197.2250023353845,
+    "mean_length": 1609.65,
+    "std_reward": 147.19520162716503,
+    "std_length": 1187.3336209760087,
+    "task_count": 1,
+    "macro_mean_reward": 197.2250023353845,
+    "macro_mean_length": 1609.65,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1250.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1250.json
new file mode 100644
index 0000000000000000000000000000000000000000..ab9b4a73645ee510f1cfa920be1a8510f29e6bff
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1250.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 388.7600045718253,
+      "mean_length": 3150.55,
+      "std_reward": 108.89891494959997,
+      "std_length": 876.8407195722608,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        335.6000039577484,
+        249.3000029399991,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        137.30000162869692,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        74.8000009059906,
+        309.20000363886356,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        2728,
+        2027,
+        3600,
+        3600,
+        3600,
+        1123,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        624,
+        2509,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 388.7600045718253,
+    "mean_length": 3150.55,
+    "std_reward": 108.89891494959997,
+    "std_length": 876.8407195722608,
+    "task_count": 1,
+    "macro_mean_reward": 388.7600045718253,
+    "macro_mean_length": 3150.55,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1500.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1500.json
new file mode 100644
index 0000000000000000000000000000000000000000..f825e4f0de82a65e8854633964e30285f0fd26cb
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1500.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 198.610002348572,
+      "mean_length": 1618.8,
+      "std_reward": 156.37761499774624,
+      "std_length": 1261.2708115230448,
+      "episode_rewards": [
+        18.700000248849392,
+        230.80000272393227,
+        169.50000200420618,
+        440.70000518113375,
+        256.20000302791595,
+        200.00000236928463,
+        444.6000052243471,
+        419.2000049352646,
+        71.50000085681677,
+        11.30000015348196,
+        90.80000108480453,
+        69.40000084042549,
+        444.6000052243471,
+        444.6000052243471,
+        125.10000149160624,
+        132.80000157654285,
+        20.700000263750553,
+        90.80000108480453,
+        34.80000042915344,
+        256.10000302642584
+      ],
+      "episode_lengths": [
+        171,
+        1878,
+        1382,
+        3572,
+        2087,
+        1633,
+        3600,
+        3402,
+        591,
+        106,
+        748,
+        579,
+        3600,
+        3600,
+        1028,
+        1087,
+        182,
+        748,
+        296,
+        2086
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 198.610002348572,
+    "mean_length": 1618.8,
+    "std_reward": 156.37761499774624,
+    "std_length": 1261.2708115230448,
+    "task_count": 1,
+    "macro_mean_reward": 198.610002348572,
+    "macro_mean_length": 1618.8,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1750.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1750.json
new file mode 100644
index 0000000000000000000000000000000000000000..04a5ad9e25c71392a895c0318a2b314e53fc22c5
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_1750.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 200.33000236973166,
+      "mean_length": 1633.2,
+      "std_reward": 167.7155352307171,
+      "std_length": 1351.4404389391343,
+      "episode_rewards": [
+        3.90000007301569,
+        139.0000016540289,
+        249.3000029399991,
+        444.6000052243471,
+        426.8000050187111,
+        3.90000007301569,
+        349.1000041142106,
+        3.90000007301569,
+        65.10000079125166,
+        181.40000215172768,
+        64.50000078231096,
+        207.30000244826078,
+        444.6000052243471,
+        444.6000052243471,
+        232.60000275075436,
+        3.90000007301569,
+        216.8000025600195,
+        444.6000052243471,
+        27.400000348687172,
+        53.30000064522028
+      ],
+      "episode_lengths": [
+        50,
+        1140,
+        2027,
+        3600,
+        3460,
+        50,
+        2836,
+        50,
+        545,
+        1483,
+        539,
+        1688,
+        3600,
+        3600,
+        1896,
+        50,
+        1765,
+        3600,
+        240,
+        445
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 1750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 200.33000236973166,
+    "mean_length": 1633.2,
+    "std_reward": 167.7155352307171,
+    "std_length": 1351.4404389391343,
+    "task_count": 1,
+    "macro_mean_reward": 200.33000236973166,
+    "macro_mean_length": 1633.2,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2000.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2000.json
new file mode 100644
index 0000000000000000000000000000000000000000..0d7dca71c573ffc94a5d836b293a1a8db20437ff
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2000.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 360.93000424727796,
+      "mean_length": 2926.9,
+      "std_reward": 141.95610790358714,
+      "std_length": 1143.4599643188212,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        11.30000015348196,
+        140.10000167042017,
+        444.6000052243471,
+        444.6000052243471,
+        277.30000326782465,
+        424.90000500530005,
+        444.6000052243471,
+        263.3000031039119,
+        444.6000052243471,
+        444.6000052243471,
+        18.800000250339508,
+        444.6000052243471,
+        303.1000035777688
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        106,
+        1151,
+        3600,
+        3600,
+        2253,
+        3450,
+        3600,
+        2140,
+        3600,
+        3600,
+        172,
+        3600,
+        2466
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 360.93000424727796,
+    "mean_length": 2926.9,
+    "std_reward": 141.95610790358714,
+    "std_length": 1143.4599643188212,
+    "task_count": 1,
+    "macro_mean_reward": 360.93000424727796,
+    "macro_mean_length": 2926.9,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2250.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2250.json
new file mode 100644
index 0000000000000000000000000000000000000000..02d5d67ee453243e3443ba05b6b22abd80f23794
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2250.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 214.26500253416597,
+      "mean_length": 1746.45,
+      "std_reward": 146.78073708385077,
+      "std_length": 1182.8461639198902,
+      "episode_rewards": [
+        92.90000111609697,
+        92.90000111609697,
+        444.6000052243471,
+        367.9000043347478,
+        97.30000116676092,
+        31.900000400841236,
+        307.40000362694263,
+        181.40000215172768,
+        78.70000094920397,
+        214.90000254660845,
+        90.70000108331442,
+        134.9000016078353,
+        444.6000052243471,
+        444.6000052243471,
+        25.300000317394733,
+        139.100001655519,
+        349.50000412017107,
+        444.6000052243471,
+        176.9000020995736,
+        125.20000149309635
+      ],
+      "episode_lengths": [
+        769,
+        769,
+        3600,
+        2988,
+        804,
+        276,
+        2500,
+        1483,
+        654,
+        1755,
+        747,
+        1108,
+        3600,
+        3600,
+        219,
+        1141,
+        2840,
+        3600,
+        1447,
+        1029
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 214.26500253416597,
+    "mean_length": 1746.45,
+    "std_reward": 146.78073708385077,
+    "std_length": 1182.8461639198902,
+    "task_count": 1,
+    "macro_mean_reward": 214.26500253416597,
+    "macro_mean_length": 1746.45,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_250.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_250.json
new file mode 100644
index 0000000000000000000000000000000000000000..046a637ffda34e44b97b6c0ffa3af296fdc87da4
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_250.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 43.395000534877184,
+      "mean_length": 368.45,
+      "std_reward": 34.15739308424012,
+      "std_length": 275.62700792919406,
+      "episode_rewards": [
+        32.10000040382147,
+        27.400000348687172,
+        9.00000013411045,
+        116.40000139176846,
+        13.600000187754631,
+        132.80000157654285,
+        32.200000405311584,
+        26.90000034123659,
+        43.80000053346157,
+        13.200000181794167,
+        9.500000141561031,
+        55.70000068098307,
+        27.400000348687172,
+        10.600000143051147,
+        25.300000317394733,
+        74.70000090450048,
+        83.20000100135803,
+        65.10000079125166,
+        27.600000351667404,
+        41.400000512599945
+      ],
+      "episode_lengths": [
+        278,
+        240,
+        92,
+        959,
+        129,
+        1087,
+        279,
+        235,
+        368,
+        125,
+        97,
+        469,
+        240,
+        99,
+        219,
+        623,
+        690,
+        545,
+        242,
+        353
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 43.395000534877184,
+    "mean_length": 368.45,
+    "std_reward": 34.15739308424012,
+    "std_length": 275.62700792919406,
+    "task_count": 1,
+    "macro_mean_reward": 43.395000534877184,
+    "macro_mean_length": 368.45,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2500.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2500.json
new file mode 100644
index 0000000000000000000000000000000000000000..c5b7dcb9a7aee4a6ae9e9c624a3009c5cd1cb3f8
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2500.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 275.22000324577095,
+      "mean_length": 2236.8,
+      "std_reward": 170.87475320509728,
+      "std_length": 1376.9918155167081,
+      "episode_rewards": [
+        228.90000271052122,
+        444.6000052243471,
+        444.6000052243471,
+        429.2000050544739,
+        169.40000200271606,
+        444.6000052243471,
+        444.6000052243471,
+        23.500000305473804,
+        79.5000009611249,
+        117.00000140070915,
+        37.50000046938658,
+        444.6000052243471,
+        444.6000052243471,
+        265.60000313818455,
+        43.30000052601099,
+        368.80000434815884,
+        444.6000052243471,
+        444.6000052243471,
+        43.80000053346157,
+        141.10000167042017
+      ],
+      "episode_lengths": [
+        1868,
+        3600,
+        3600,
+        3484,
+        1381,
+        3600,
+        3600,
+        210,
+        662,
+        965,
+        323,
+        3600,
+        3600,
+        2163,
+        363,
+        2997,
+        3600,
+        3600,
+        368,
+        1152
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 275.22000324577095,
+    "mean_length": 2236.8,
+    "std_reward": 170.87475320509728,
+    "std_length": 1376.9918155167081,
+    "task_count": 1,
+    "macro_mean_reward": 275.22000324577095,
+    "macro_mean_length": 2236.8,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2750.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2750.json
new file mode 100644
index 0000000000000000000000000000000000000000..9f4ab24c10f9c6fc5f8c58f72478e5c36bc8a128
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_2750.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 391.8000046081841,
+      "mean_length": 3175.55,
+      "std_reward": 112.64322126288126,
+      "std_length": 906.7831314597776,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        386.40000455081463,
+        444.6000052243471,
+        444.6000052243471,
+        190.4000022560358,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        50.40000061690807,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        381.90000449866056,
+        444.6000052243471,
+        157.9000018760562
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3600,
+        3600,
+        3137,
+        3600,
+        3600,
+        1555,
+        3600,
+        3600,
+        3600,
+        425,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3101,
+        3600,
+        1293
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 2750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 391.8000046081841,
+    "mean_length": 3175.55,
+    "std_reward": 112.64322126288126,
+    "std_length": 906.7831314597776,
+    "task_count": 1,
+    "macro_mean_reward": 391.8000046081841,
+    "macro_mean_length": 3175.55,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3000.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3000.json
new file mode 100644
index 0000000000000000000000000000000000000000..832610d353b00b345a3468af07384226361c2544
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3000.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 358.67500422038137,
+      "mean_length": 2908.4,
+      "std_reward": 133.2495864541459,
+      "std_length": 1073.4027855376564,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        200.00000236928463,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        251.30000296980143,
+        116.00000138580799,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        6.800000101327896,
+        216.8000025600195,
+        382.1000045016408,
+        444.6000052243471,
+        444.6000052243471,
+        220.70000260323286
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        1633,
+        3600,
+        3600,
+        3600,
+        2047,
+        955,
+        3600,
+        3600,
+        3600,
+        70,
+        1765,
+        3103,
+        3600,
+        3600,
+        1795
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 358.67500422038137,
+    "mean_length": 2908.4,
+    "std_reward": 133.2495864541459,
+    "std_length": 1073.4027855376564,
+    "task_count": 1,
+    "macro_mean_reward": 358.67500422038137,
+    "macro_mean_length": 2908.4,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3250.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3250.json
new file mode 100644
index 0000000000000000000000000000000000000000..df16e18f23cc9e96b702a35a8e3ce2896a44a254
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3250.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 398.7300046890974,
+      "mean_length": 3231.45,
+      "std_reward": 83.87884276974052,
+      "std_length": 674.9740346857797,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        246.500002913177,
+        132.80000157654285,
+        293.6000034660101,
+        384.80000452697277,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        363.7000042870641,
+        444.6000052243471,
+        444.6000052243471,
+        328.8000038713217,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3600,
+        3600,
+        2008,
+        1087,
+        2389,
+        3121,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        2955,
+        3600,
+        3600,
+        2669,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3250,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 398.7300046890974,
+    "mean_length": 3231.45,
+    "std_reward": 83.87884276974052,
+    "std_length": 674.9740346857797,
+    "task_count": 1,
+    "macro_mean_reward": 398.7300046890974,
+    "macro_mean_length": 3231.45,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3500.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3500.json
new file mode 100644
index 0000000000000000000000000000000000000000..f9bd903fb4500204e607c7eab0329008ab9a60d3
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3500.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 335.5500039495528,
+      "mean_length": 2721.9,
+      "std_reward": 164.0522634087507,
+      "std_length": 1322.178425931992,
+      "episode_rewards": [
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        384.80000452697277,
+        18.700000248849392,
+        261.60000309348106,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        351.8000041395426,
+        6.800000101327896,
+        444.6000052243471,
+        6.800000101327896,
+        421.80000495910645,
+        62.80000075697899,
+        305.3000035956502,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3121,
+        171,
+        2132,
+        3600,
+        3600,
+        3600,
+        2854,
+        70,
+        3600,
+        70,
+        3419,
+        522,
+        2479,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 335.5500039495528,
+    "mean_length": 2721.9,
+    "std_reward": 164.0522634087507,
+    "std_length": 1322.178425931992,
+    "task_count": 1,
+    "macro_mean_reward": 335.5500039495528,
+    "macro_mean_length": 2721.9,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3750.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3750.json
new file mode 100644
index 0000000000000000000000000000000000000000..6a6e33ceb35a53d5b0a7762f7132ff333dc99277
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_3750.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 277.18000326827166,
+      "mean_length": 2252.45,
+      "std_reward": 168.96234570997657,
+      "std_length": 1361.875525699761,
+      "episode_rewards": [
+        32.40000040829182,
+        358.6000042259693,
+        403.3000047430396,
+        186.40000221133232,
+        76.80000092089176,
+        444.6000052243471,
+        90.80000108480453,
+        444.6000052243471,
+        444.6000052243471,
+        116.30000139027834,
+        444.6000052243471,
+        67.30000080913305,
+        46.300000570714474,
+        444.6000052243471,
+        321.4000037908554,
+        300.80000354349613,
+        444.6000052243471,
+        410.20000483095646,
+        444.6000052243471,
+        20.80000026524067
+      ],
+      "episode_lengths": [
+        281,
+        2913,
+        3270,
+        1524,
+        635,
+        3600,
+        748,
+        3600,
+        3600,
+        958,
+        3600,
+        558,
+        393,
+        3600,
+        2613,
+        2443,
+        3600,
+        3330,
+        3600,
+        183
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 3750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 277.18000326827166,
+    "mean_length": 2252.45,
+    "std_reward": 168.96234570997657,
+    "std_length": 1361.875525699761,
+    "task_count": 1,
+    "macro_mean_reward": 277.18000326827166,
+    "macro_mean_length": 2252.45,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_4000.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_4000.json
new file mode 100644
index 0000000000000000000000000000000000000000..6697aac948a374cec149a2fb68fb0d75678b016f
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_4000.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 323.67500381283463,
+      "mean_length": 2627.55,
+      "std_reward": 145.51053697177636,
+      "std_length": 1171.6111759026542,
+      "episode_rewards": [
+        444.6000052243471,
+        107.2000012844801,
+        444.6000052243471,
+        428.40000504255295,
+        444.6000052243471,
+        185.90000220388174,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        256.80000303685665,
+        265.2000031322241,
+        444.6000052243471,
+        83.70000100880861,
+        444.6000052243471,
+        331.20000390708447,
+        90.80000108480453,
+        59.90000072866678,
+        218.40000258386135
+      ],
+      "episode_lengths": [
+        3600,
+        885,
+        3600,
+        3476,
+        3600,
+        1519,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        2093,
+        2159,
+        3600,
+        695,
+        3600,
+        2693,
+        748,
+        502,
+        1781
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 4000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 323.67500381283463,
+    "mean_length": 2627.55,
+    "std_reward": 145.51053697177636,
+    "std_length": 1171.6111759026542,
+    "task_count": 1,
+    "macro_mean_reward": 323.67500381283463,
+    "macro_mean_length": 2627.55,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_500.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_500.json
new file mode 100644
index 0000000000000000000000000000000000000000..25e98c6db4cc7bb72366cd0e3adcab90ac9954ef
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_500.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 64.18500077836215,
+      "mean_length": 536.3,
+      "std_reward": 70.62001410319165,
+      "std_length": 570.6692649862966,
+      "episode_rewards": [
+        176.40000209212303,
+        23.40000030398369,
+        60.80000074207783,
+        74.8000009059906,
+        37.300000466406345,
+        83.80000101029873,
+        27.500000350177288,
+        18.400000244379044,
+        307.8000036329031,
+        15.800000205636024,
+        18.800000250339508,
+        121.30000144988298,
+        6.800000101327896,
+        111.90000133961439,
+        6.800000101327896,
+        29.800000369548798,
+        41.80000051856041,
+        27.900000356137753,
+        71.80000086128712,
+        20.80000026524067
+      ],
+      "episode_lengths": [
+        1442,
+        209,
+        511,
+        624,
+        321,
+        696,
+        241,
+        168,
+        2504,
+        142,
+        172,
+        999,
+        70,
+        923,
+        70,
+        255,
+        357,
+        245,
+        594,
+        183
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 500,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 64.18500077836215,
+    "mean_length": 536.3,
+    "std_reward": 70.62001410319165,
+    "std_length": 570.6692649862966,
+    "task_count": 1,
+    "macro_mean_reward": 64.18500077836215,
+    "macro_mean_length": 536.3,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_750.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_750.json
new file mode 100644
index 0000000000000000000000000000000000000000..aced4587b724f660972fedaf6a644cb142848a33
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/mid_train/step_750.json
@@ -0,0 +1,120 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 20,
+      "mean_reward": 199.45500236190856,
+      "mean_length": 1627.7,
+      "std_reward": 161.18966493830942,
+      "std_length": 1299.6861197997,
+      "episode_rewards": [
+        444.6000052243471,
+        120.40000143647194,
+        163.50000194460154,
+        186.40000221133232,
+        438.60000516474247,
+        96.90000116080046,
+        60.600000739097595,
+        444.6000052243471,
+        396.50000467151403,
+        69.20000083744526,
+        22.400000289082527,
+        45.90000056475401,
+        124.900001488626,
+        444.6000052243471,
+        134.4000016003847,
+        424.6000050008297,
+        73.90000089257956,
+        45.90000056475401,
+        17.900000236928463,
+        233.30000276118517
+      ],
+      "episode_lengths": [
+        3600,
+        990,
+        1340,
+        1524,
+        3560,
+        800,
+        509,
+        3600,
+        3220,
+        577,
+        199,
+        389,
+        1026,
+        3600,
+        1103,
+        3447,
+        615,
+        389,
+        163,
+        1903
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "mid_train",
+    "step": 750,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 20,
+    "mean_reward": 199.45500236190856,
+    "mean_length": 1627.7,
+    "std_reward": 161.18966493830942,
+    "std_length": 1299.6861197997,
+    "task_count": 1,
+    "macro_mean_reward": 199.45500236190856,
+    "macro_mean_length": 1627.7,
+    "distributed_eval": true
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/post_train/step_4000.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/post_train/step_4000.json
new file mode 100644
index 0000000000000000000000000000000000000000..86f38f3dff26d2ae3e4f73ae85d2330e897ff0d8
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/eval/post_train/step_4000.json
@@ -0,0 +1,440 @@
+{
+  "per_latency": {
+    "flappy/latency_3": {
+      "latency": 3,
+      "num_episodes": 100,
+      "mean_reward": 368.88000434055925,
+      "mean_length": 2991.19,
+      "std_reward": 129.9747321399612,
+      "std_length": 1046.6452951692852,
+      "episode_rewards": [
+        444.6000052243471,
+        107.2000012844801,
+        444.6000052243471,
+        428.40000504255295,
+        444.6000052243471,
+        185.90000220388174,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        256.80000303685665,
+        265.2000031322241,
+        84.00000101327896,
+        444.6000052243471,
+        444.6000052243471,
+        331.20000390708447,
+        90.80000108480453,
+        59.90000072866678,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        111.00000132620335,
+        444.6000052243471,
+        97.70000117272139,
+        321.00000378489494,
+        83.0000009983778,
+        444.6000052243471,
+        65.20000079274178,
+        444.6000052243471,
+        367.9000043347478,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        272.8000032156706,
+        444.6000052243471,
+        442.70000521093607,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        15.800000205636024,
+        444.6000052243471,
+        424.80000500380993,
+        372.500004388392,
+        306.90000361949205,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        400.5000047162175,
+        444.6000052243471,
+        246.500002913177,
+        37.300000466406345,
+        249.3000029399991,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        153.70000182837248,
+        115.90000138431787,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        404.90000476688147,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        442.50000520795584,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        90.80000108480453,
+        444.6000052243471,
+        444.6000052243471,
+        148.40000176429749,
+        444.6000052243471,
+        289.4000034183264,
+        255.9000030234456,
+        444.6000052243471,
+        109.30000130087137,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        444.6000052243471,
+        353.900004170835,
+        444.6000052243471,
+        444.6000052243471
+      ],
+      "episode_lengths": [
+        3600,
+        885,
+        3600,
+        3476,
+        3600,
+        1519,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        2093,
+        2159,
+        698,
+        3600,
+        3600,
+        2693,
+        748,
+        502,
+        3600,
+        3600,
+        3600,
+        3600,
+        914,
+        3600,
+        808,
+        2609,
+        688,
+        3600,
+        546,
+        3600,
+        2988,
+        3600,
+        3600,
+        3600,
+        3600,
+        2217,
+        3600,
+        3592,
+        3600,
+        3600,
+        3600,
+        142,
+        3600,
+        3449,
+        3025,
+        2495,
+        3600,
+        3600,
+        3600,
+        3600,
+        3251,
+        3600,
+        2008,
+        321,
+        2027,
+        3600,
+        3600,
+        3600,
+        3600,
+        1260,
+        954,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3286,
+        3600,
+        3600,
+        3600,
+        3590,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        3600,
+        748,
+        3600,
+        3600,
+        1216,
+        3600,
+        2356,
+        2084,
+        3600,
+        897,
+        3600,
+        3600,
+        3600,
+        3600,
+        2875,
+        3600,
+        3600
+      ],
+      "decoded_action_hist": {},
+      "fixed_episode_seeds": true,
+      "eval_seed": 42,
+      "episode_seeds": [
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null,
+        null
+      ],
+      "episode_indices": [
+        0,
+        1,
+        2,
+        3,
+        4,
+        5,
+        6,
+        7,
+        8,
+        9,
+        10,
+        11,
+        12,
+        13,
+        14,
+        15,
+        16,
+        17,
+        18,
+        19,
+        20,
+        21,
+        22,
+        23,
+        24,
+        25,
+        26,
+        27,
+        28,
+        29,
+        30,
+        31,
+        32,
+        33,
+        34,
+        35,
+        36,
+        37,
+        38,
+        39,
+        40,
+        41,
+        42,
+        43,
+        44,
+        45,
+        46,
+        47,
+        48,
+        49,
+        50,
+        51,
+        52,
+        53,
+        54,
+        55,
+        56,
+        57,
+        58,
+        59,
+        60,
+        61,
+        62,
+        63,
+        64,
+        65,
+        66,
+        67,
+        68,
+        69,
+        70,
+        71,
+        72,
+        73,
+        74,
+        75,
+        76,
+        77,
+        78,
+        79,
+        80,
+        81,
+        82,
+        83,
+        84,
+        85,
+        86,
+        87,
+        88,
+        89,
+        90,
+        91,
+        92,
+        93,
+        94,
+        95,
+        96,
+        97,
+        98,
+        99
+      ]
+    }
+  },
+  "aggregate": {
+    "stage": "post_train",
+    "step": 4000,
+    "task": "flappy",
+    "model_alias": "openvla",
+    "fixed_episode_seeds": true,
+    "eval_seed": 42,
+    "total_episodes": 100,
+    "mean_reward": 368.88000434055925,
+    "mean_length": 2991.19,
+    "std_reward": 129.9747321399612,
+    "std_length": 1046.6452951692852,
+    "task_count": 1,
+    "macro_mean_reward": 368.88000434055925,
+    "macro_mean_length": 2991.19,
+    "distributed_eval": false
+  }
+}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/hydra/.hydra/hydra.yaml b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/hydra/.hydra/hydra.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..8826c220dd4966faf697330116cade7ddfbfbe62
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/hydra/.hydra/hydra.yaml
@@ -0,0 +1,369 @@
+hydra:
+  run:
+    dir: ${run_root_dir}/${run_id}/hydra
+  sweep:
+    dir: multirun/${now:%Y-%m-%d}/${now:%H-%M-%S}
+    subdir: ${hydra.job.num}
+  launcher:
+    _target_: hydra._internal.core_plugins.basic_launcher.BasicLauncher
+  sweeper:
+    _target_: hydra._internal.core_plugins.basic_sweeper.BasicSweeper
+    max_batch_size: null
+    params: null
+  help:
+    app_name: ${hydra.job.name}
+    header: '${hydra.help.app_name} is powered by Hydra.
+
+      '
+    footer: 'Powered by Hydra (https://hydra.cc)
+
+      Use --hydra-help to view Hydra specific help
+
+      '
+    template: '${hydra.help.header}
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (group=option)
+
+
+      $APP_CONFIG_GROUPS
+
+
+      == Config ==
+
+      Override anything in the config (foo.bar=value)
+
+
+      $CONFIG
+
+
+      ${hydra.help.footer}
+
+      '
+  hydra_help:
+    template: 'Hydra (${hydra.runtime.version})
+
+      See https://hydra.cc for more info.
+
+
+      == Flags ==
+
+      $FLAGS_HELP
+
+
+      == Configuration groups ==
+
+      Compose your configuration from those groups (For example, append hydra/job_logging=disabled
+      to command line)
+
+
+      $HYDRA_CONFIG_GROUPS
+
+
+      Use ''--cfg hydra'' to Show the Hydra config.
+
+      '
+    hydra_help: ???
+  hydra_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][HYDRA] %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+    root:
+      level: INFO
+      handlers:
+      - console
+    loggers:
+      logging_example:
+        level: DEBUG
+    disable_existing_loggers: false
+  job_logging:
+    version: 1
+    formatters:
+      simple:
+        format: '[%(asctime)s][%(name)s][%(levelname)s] - %(message)s'
+    handlers:
+      console:
+        class: logging.StreamHandler
+        formatter: simple
+        stream: ext://sys.stdout
+      file:
+        class: logging.FileHandler
+        formatter: simple
+        filename: ${hydra.runtime.output_dir}/${hydra.job.name}.log
+    root:
+      level: INFO
+      handlers:
+      - console
+      - file
+    disable_existing_loggers: false
+  env: {}
+  mode: RUN
+  searchpath: []
+  callbacks: {}
+  output_subdir: .hydra
+  overrides:
+    hydra:
+    - hydra.mode=RUN
+    task:
+    - model=openvla
+    - env=flappy
+    - init=bridge
+    - mode=single
+    - ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++framework.qwenvl.attn_implementation=flash_attention_2
+    - ++framework.qwenvl.enable_gradient_checkpointing=true
+    - ++framework.action_model.state_dim=7
+    - ++framework.action_model.loss_type=discrete_ce
+    - ++framework.action_model.action_horizon=1
+    - ++framework.action_model.future_action_window_size=0
+    - ++framework.action_model.past_action_window_size=0
+    - ++framework.action_model.action_dim=7
+    - ++framework.action_model.action_env_dim=2
+    - ++framework.kv_memory.enabled=false
+    - ++framework.kv_memory.window=4
+    - ++framework.kv_memory.rollout_len=8
+    - ++framework.kv_memory.packed_train=false
+    - ++framework.kv_memory.rebased_sink=true
+    - ++framework.name=QwenOFT
+    - ++datasets.vla_data.dataset_py=lerobot_datasets
+    - ++datasets.vla_data.include_state=true
+    - ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games
+    - ++datasets.vla_data.data_mix=flappy_train
+    - ++datasets.vla_data.eval_data_mix=null
+    - ++datasets.vla_data.custom_mixtures_path=null
+    - ++datasets.vla_data.action_type=discrete
+    - ++datasets.vla_data.sequential_step_sampling=false
+    - ++datasets.vla_data.eval_sequential_step_sampling=null
+    - ++datasets.vla_data.num_workers=8
+    - ++datasets.vla_data.eval_num_workers=8
+    - ++datasets.vla_data.prefetch_factor=4
+    - ++datasets.vla_data.persistent_workers=true
+    - ++datasets.vla_data.pin_memory=true
+    - ++datasets.vla_data.shuffle=true
+    - ++datasets.vla_data.action_balance.enabled=false
+    - ++datasets.vla_data.action_balance.strategy=balanced_epoch
+    - ++datasets.vla_data.action_balance.action_key=action_id
+    - ++datasets.vla_data.action_balance.target_flap_fraction=0.3
+    - ++datasets.vla_data.action_balance.noop_id=0
+    - ++datasets.vla_data.action_balance.flap_id=1
+    - ++datasets.vla_data.latency_curriculum.enabled=false
+    - ++datasets.vla_data.latency_curriculum.strategy=exclusive
+    - ++datasets.vla_data.latency_curriculum.latencies=null
+    - ++datasets.vla_data.latency_curriculum.phase_steps=null
+    - ++datasets.vla_data.latency_curriculum.phase_distributions=null
+    - ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0
+    - ++datasets.vla_data.latency_curriculum.replay_passes=0.25
+    - ++datasets.vla_data.latency_curriculum.target_total_passes=2.0
+    - ++datasets.vla_data.latency_curriculum.final_equalization=true
+    - ++datasets.vla_data.latency_curriculum.step_budget_mode=auto
+    - ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false
+    - ++datasets.vla_data.latency_curriculum.save_at_phase_end=false
+    - ++datasets.vla_data.latency_curriculum.computed_plan=null
+    - ++datasets.vla_data.per_device_batch_size=8
+    - ++datasets.vla_data.load_all_data_for_training=true
+    - ++datasets.vla_data.num_obs_frames=1
+    - ++datasets.vla_data.image_mode=single
+    - ++datasets.vla_data.prompt_mode=raw
+    - ++datasets.vla_data.stitch_grid=[2,2]
+    - ++datasets.vla_data.obs_image_size=null
+    - ++datasets.vla_data.video_backend=torchvision_av
+    - ++dataset.source_hf=
+    - ++dataset.config_name=null
+    - ++dataset.source_subdir=null
+    - ++dataset.converted_name=flappy_train
+    - ++dataset.single_source_hf=
+    - ++dataset.mixed_source_hf=
+    - ++dataset.single_converted_name=flappy_train
+    - ++dataset.mixed_converted_name=flappy_mixed_latency_train
+    - ++dataset.single_latency_filter=null
+    - ++dataset.mixed_latency_filter=null
+    - ++dataset.force_download=false
+    - ++dataset.setup_force=false
+    - ++dataset.skip_verification=false
+    - ++dataset.target_latency_unit=raw_frames
+    - ++dataset.verify_rows=200
+    - ++dataset.max_episodes=null
+    - ++dataset.episodes_per_latency=null
+    - ++dataset.latency_filter=null
+    - ++dataset.debug_subset.enabled=false
+    - ++dataset.debug_subset.max_episodes=5
+    - ++dataset.debug_subset.suffix=debug
+    - ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct
+    - ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1
+    - ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1
+    - ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt
+    - ++trainer.max_train_steps=4000
+    - ++trainer.num_warmup_steps=100
+    - ++trainer.save_interval=500
+    - ++trainer.eval_interval=250
+    - ++trainer.eval_num_batches=50
+    - ++trainer.per_latency_eval_num_batches=null
+    - ++trainer.eval_action_classification=false
+    - ++trainer.eval_action_classification_interval=null
+    - ++trainer.cc_f1_tolerance=1
+    - ++trainer.learning_rate.base=2e-05
+    - ++trainer.learning_rate.qwen_vl_interface=1e-05
+    - ++trainer.learning_rate.action_model=0.0001
+    - ++trainer.lr_scheduler_type=cosine_with_min_lr
+    - ++trainer.scheduler_specific_kwargs.min_lr=1e-06
+    - ++trainer.freeze_modules=
+    - ++trainer.freeze_vit=false
+    - ++trainer.freeze_tied_embedding=false
+    - ++trainer.freeze_llm_layers=[]
+    - ++trainer.loss_scale.vla=1.0
+    - ++trainer.loss_scale.vlm=0.1
+    - ++trainer.max_grad_norm=1.0
+    - ++trainer.weight_decay=0.0
+    - ++trainer.logging_frequency=1
+    - ++trainer.profile_timing.enabled=false
+    - ++trainer.profile_timing.log_interval=10
+    - ++trainer.gradient_clipping=1.0
+    - ++trainer.gradient_accumulation_steps=16
+    - ++trainer.distributed_backend=deepspeed
+    - ++trainer.is_resume=false
+    - ++trainer.pretrained_checkpoint=null
+    - ++trainer.resume_step=0
+    - ++trainer.reload_modules=null
+    - ++trainer.optimizer.name=AdamW
+    - ++trainer.optimizer.betas=[0.9,0.95]
+    - ++trainer.optimizer.eps=1e-08
+    - ++trainer.optimizer.weight_decay=1e-08
+    - ++trainer.optimizer.fused=true
+    - ++trainer.save_format=pt
+    - ++workspace_dir=WORKSPACE_DIR
+    - ++run_root_dir=results/Checkpoints
+    - ++seed=42
+    - ++auth.env_file=null
+    - ++auth.hf_token_env=HF_TOKEN
+    - ++auth.wandb_api_key_env=WANDB_API_KEY
+    - ++paths.run_root_dir=results/Checkpoints
+    - ++paths.dataset_local_dir=data/memory/flappy_fix_latency_3_200ep_7k2steps
+    - ++paths.dataset_cache_dir=null
+    - ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml
+    - ++rl_games.model_alias=openvla
+    - ++rl_games.env_eval.image_size=224
+    - ++rl_games.env_eval.frameskip=1
+    - ++rl_games.env_eval.image_transform=raw_rgb
+    - ++rl_games.env_eval.prompt_mode=raw
+    - ++rl_games.env_eval.ghost_trail.history_frames=5
+    - ++rl_games.env_eval.ghost_trail.gamma=1.3
+    - ++rl_games.env_eval.ghost_trail.min_alpha=35
+    - ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0
+    - ++rl_games.env_eval.ghost_trail.ground_fraction=0.22
+    - ++rl_games.env_eval.seed=42
+    - ++rl_games.env_eval.fixed_episode_seeds=true
+    - ++rl_games.env_eval.latency_seed_stride=0
+    - ++rl_games.env_eval.task_seed_stride=0
+    - ++rl_games.env_eval.task_description=
+    - ++rl_games.env_eval.eval_parallel_envs=5
+    - ++rl_games.env_eval.action_chunk_execution.enabled=false
+    - ++rl_games.env_eval.action_chunk_execution.chunk_size=null
+    - ++rl_games.env_eval.enabled=true
+    - ++rl_games.env_eval.eval_backend=latency_bench
+    - ++rl_games.env_eval.distributed_mode=rank_sharded
+    - ++rl_games.env_eval.vectorized.enabled=false
+    - ++rl_games.env_eval.vectorized.batch_size=1
+    - ++rl_games.env_eval.latency.prompt_map_path=null
+    - ++rl_games.env_eval.latency.mode=single
+    - ++rl_games.env_eval.latency.values=[0]
+    - ++rl_games.env_eval.mid_train.enabled=true
+    - ++rl_games.env_eval.mid_train.interval_steps=250
+    - ++rl_games.env_eval.mid_train.latencies=[3]
+    - ++rl_games.env_eval.mid_train.num_episodes=20
+    - ++rl_games.env_eval.mid_train.max_steps_per_episode=3600
+    - ++rl_games.env_eval.post_train.enabled=true
+    - ++rl_games.env_eval.post_train.latencies=[3]
+    - ++rl_games.env_eval.post_train.num_episodes=50
+    - ++rl_games.env_eval.post_train.max_steps_per_episode=3600
+    - ++rl_games.task=flappy
+    - ++rl_games.initialization_mode=bridge
+    - ++rl_games.action_carrier=bridge
+    - ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action
+    - ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action
+    - ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++checkpoint.load=auto
+    - ++checkpoint.hf_repo_id=null
+    - ++checkpoint.save_best_model=false
+    - ++checkpoint.save_final_model=true
+    - ++checkpoint.save_pt_file=false
+    - ++checkpoint.save_training_state=true
+    - ++checkpoint.save_safetensors_file=true
+    - ++checkpoint.local.keep_last_n=1
+    - ++checkpoint.sync.enabled=false
+    - ++checkpoint.sync.repo_id=null
+    - ++checkpoint.sync.keep_last_n=0
+    - ++checkpoint.sync.sync_every_n_checkpoints=1
+    - ++checkpoint.sync.resume_policy=local_latest
+    - ++run_id=flappy_fix_latency_3_200ep_7k2steps_single_baseline
+    - ++output_dir=null
+    - ++config_yaml=null
+    - ++is_debug=false
+    - ++version_id=0.21
+    - ++run_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints
+    - ++trainer.is_resume=true
+    - ++trainer.pretrained_checkpoint=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state
+    - ++trainer.resume_step=3000
+    - ++datasets.vla_data.data_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps
+    - ++datasets.vla_data.data_mix=flappy_train__bridge
+    - ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val
+    - ++framework.qwenvl.base_vlm=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    - ++rl_games.env_eval.latency.prompt_map_path=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json
+  job:
+    name: train_starvla_hydra
+    chdir: false
+    override_dirname: ++auth.env_file=null,++auth.hf_token_env=HF_TOKEN,++auth.wandb_api_key_env=WANDB_API_KEY,++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action,++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct,++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action,++checkpoint.hf_repo_id=null,++checkpoint.load=auto,++checkpoint.local.keep_last_n=1,++checkpoint.save_best_model=false,++checkpoint.save_final_model=true,++checkpoint.save_pt_file=false,++checkpoint.save_safetensors_file=true,++checkpoint.save_training_state=true,++checkpoint.sync.enabled=false,++checkpoint.sync.keep_last_n=0,++checkpoint.sync.repo_id=null,++checkpoint.sync.resume_policy=local_latest,++checkpoint.sync.sync_every_n_checkpoints=1,++config_yaml=null,++dataset.config_name=null,++dataset.converted_name=flappy_train,++dataset.debug_subset.enabled=false,++dataset.debug_subset.max_episodes=5,++dataset.debug_subset.suffix=debug,++dataset.episodes_per_latency=null,++dataset.force_download=false,++dataset.latency_filter=null,++dataset.max_episodes=null,++dataset.mixed_converted_name=flappy_mixed_latency_train,++dataset.mixed_latency_filter=null,++dataset.mixed_source_hf=,++dataset.setup_force=false,++dataset.single_converted_name=flappy_train,++dataset.single_latency_filter=null,++dataset.single_source_hf=,++dataset.skip_verification=false,++dataset.source_hf=,++dataset.source_subdir=null,++dataset.target_latency_unit=raw_frames,++dataset.verify_rows=200,++datasets.vla_data.action_balance.action_key=action_id,++datasets.vla_data.action_balance.enabled=false,++datasets.vla_data.action_balance.flap_id=1,++datasets.vla_data.action_balance.noop_id=0,++datasets.vla_data.action_balance.strategy=balanced_epoch,++datasets.vla_data.action_balance.target_flap_fraction=0.3,++datasets.vla_data.action_type=discrete,++datasets.vla_data.custom_mixtures_path=null,++datasets.vla_data.data_mix=flappy_train,++datasets.vla_data.data_mix=flappy_train__bridge,++datasets.vla_data.data_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps,++datasets.vla_data.data_root_dir=playground/Datasets/rl_games,++datasets.vla_data.dataset_py=lerobot_datasets,++datasets.vla_data.eval_data_mix=flappy_train__bridge__val,++datasets.vla_data.eval_data_mix=null,++datasets.vla_data.eval_num_workers=8,++datasets.vla_data.eval_sequential_step_sampling=null,++datasets.vla_data.image_mode=single,++datasets.vla_data.include_state=true,++datasets.vla_data.latency_curriculum.computed_plan=null,++datasets.vla_data.latency_curriculum.enabled=false,++datasets.vla_data.latency_curriculum.eval_at_phase_end=false,++datasets.vla_data.latency_curriculum.final_equalization=true,++datasets.vla_data.latency_curriculum.latencies=null,++datasets.vla_data.latency_curriculum.new_latency_passes=1.0,++datasets.vla_data.latency_curriculum.phase_distributions=null,++datasets.vla_data.latency_curriculum.phase_steps=null,++datasets.vla_data.latency_curriculum.replay_passes=0.25,++datasets.vla_data.latency_curriculum.save_at_phase_end=false,++datasets.vla_data.latency_curriculum.step_budget_mode=auto,++datasets.vla_data.latency_curriculum.strategy=exclusive,++datasets.vla_data.latency_curriculum.target_total_passes=2.0,++datasets.vla_data.load_all_data_for_training=true,++datasets.vla_data.num_obs_frames=1,++datasets.vla_data.num_workers=8,++datasets.vla_data.obs_image_size=null,++datasets.vla_data.per_device_batch_size=8,++datasets.vla_data.persistent_workers=true,++datasets.vla_data.pin_memory=true,++datasets.vla_data.prefetch_factor=4,++datasets.vla_data.prompt_mode=raw,++datasets.vla_data.sequential_step_sampling=false,++datasets.vla_data.shuffle=true,++datasets.vla_data.stitch_grid=[2,2],++datasets.vla_data.video_backend=torchvision_av,++framework.action_model.action_dim=7,++framework.action_model.action_env_dim=2,++framework.action_model.action_horizon=1,++framework.action_model.future_action_window_size=0,++framework.action_model.loss_type=discrete_ce,++framework.action_model.past_action_window_size=0,++framework.action_model.state_dim=7,++framework.kv_memory.enabled=false,++framework.kv_memory.packed_train=false,++framework.kv_memory.rebased_sink=true,++framework.kv_memory.rollout_len=8,++framework.kv_memory.window=4,++framework.name=QwenOFT,++framework.qwenvl.attn_implementation=flash_attention_2,++framework.qwenvl.base_vlm=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++framework.qwenvl.enable_gradient_checkpointing=true,++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt,++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1,++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1,++is_debug=false,++output_dir=null,++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml,++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct,++paths.dataset_cache_dir=null,++paths.dataset_local_dir=data/memory/flappy_fix_latency_3_200ep_7k2steps,++paths.run_root_dir=results/Checkpoints,++rl_games.action_carrier=bridge,++rl_games.env_eval.action_chunk_execution.chunk_size=null,++rl_games.env_eval.action_chunk_execution.enabled=false,++rl_games.env_eval.distributed_mode=rank_sharded,++rl_games.env_eval.enabled=true,++rl_games.env_eval.eval_backend=latency_bench,++rl_games.env_eval.eval_parallel_envs=5,++rl_games.env_eval.fixed_episode_seeds=true,++rl_games.env_eval.frameskip=1,++rl_games.env_eval.ghost_trail.gamma=1.3,++rl_games.env_eval.ghost_trail.ground_fraction=0.22,++rl_games.env_eval.ghost_trail.history_frames=5,++rl_games.env_eval.ghost_trail.min_alpha=35,++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0,++rl_games.env_eval.image_size=224,++rl_games.env_eval.image_transform=raw_rgb,++rl_games.env_eval.latency.mode=single,++rl_games.env_eval.latency.prompt_map_path=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json,++rl_games.env_eval.latency.prompt_map_path=null,++rl_games.env_eval.latency.values=[0],++rl_games.env_eval.latency_seed_stride=0,++rl_games.env_eval.mid_train.enabled=true,++rl_games.env_eval.mid_train.interval_steps=250,++rl_games.env_eval.mid_train.latencies=[3],++rl_games.env_eval.mid_train.max_steps_per_episode=3600,++rl_games.env_eval.mid_train.num_episodes=20,++rl_games.env_eval.post_train.enabled=true,++rl_games.env_eval.post_train.latencies=[3],++rl_games.env_eval.post_train.max_steps_per_episode=3600,++rl_games.env_eval.post_train.num_episodes=50,++rl_games.env_eval.prompt_mode=raw,++rl_games.env_eval.seed=42,++rl_games.env_eval.task_description=,++rl_games.env_eval.task_seed_stride=0,++rl_games.env_eval.vectorized.batch_size=1,++rl_games.env_eval.vectorized.enabled=false,++rl_games.initialization_mode=bridge,++rl_games.model_alias=openvla,++rl_games.task=flappy,++run_id=flappy_fix_latency_3_200ep_7k2steps_single_baseline,++run_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints,++run_root_dir=results/Checkpoints,++seed=42,++trainer.cc_f1_tolerance=1,++trainer.distributed_backend=deepspeed,++trainer.eval_action_classification=false,++trainer.eval_action_classification_interval=null,++trainer.eval_interval=250,++trainer.eval_num_batches=50,++trainer.freeze_llm_layers=[],++trainer.freeze_modules=,++trainer.freeze_tied_embedding=false,++trainer.freeze_vit=false,++trainer.gradient_accumulation_steps=16,++trainer.gradient_clipping=1.0,++trainer.is_resume=false,++trainer.is_resume=true,++trainer.learning_rate.action_model=0.0001,++trainer.learning_rate.base=2e-05,++trainer.learning_rate.qwen_vl_interface=1e-05,++trainer.logging_frequency=1,++trainer.loss_scale.vla=1.0,++trainer.loss_scale.vlm=0.1,++trainer.lr_scheduler_type=cosine_with_min_lr,++trainer.max_grad_norm=1.0,++trainer.max_train_steps=4000,++trainer.num_warmup_steps=100,++trainer.optimizer.betas=[0.9,0.95],++trainer.optimizer.eps=1e-08,++trainer.optimizer.fused=true,++trainer.optimizer.name=AdamW,++trainer.optimizer.weight_decay=1e-08,++trainer.per_latency_eval_num_batches=null,++trainer.pretrained_checkpoint=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state,++trainer.pretrained_checkpoint=null,++trainer.profile_timing.enabled=false,++trainer.profile_timing.log_interval=10,++trainer.reload_modules=null,++trainer.resume_step=0,++trainer.resume_step=3000,++trainer.save_format=pt,++trainer.save_interval=500,++trainer.scheduler_specific_kwargs.min_lr=1e-06,++trainer.weight_decay=0.0,++version_id=0.21,++workspace_dir=WORKSPACE_DIR,env=flappy,init=bridge,mode=single,model=openvla
+    id: ???
+    num: ???
+    config_name: train
+    env_set: {}
+    env_copy: []
+    config:
+      override_dirname:
+        kv_sep: '='
+        item_sep: ','
+        exclude_keys: []
+  runtime:
+    version: 1.3.4
+    version_base: '1.1'
+    cwd: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA
+    config_sources:
+    - path: hydra.conf
+      schema: pkg
+      provider: hydra
+    - path: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/examples/rl_games/config
+      schema: file
+      provider: main
+    - path: ''
+      schema: structured
+      provider: schema
+    output_dir: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/hydra
+    choices:
+      cross_task_setup: null
+      checkpoint: default
+      mode: single
+      init: bridge
+      env: flappy
+      model: openvla
+      hydra/env: default
+      hydra/callbacks: null
+      hydra/job_logging: default
+      hydra/hydra_logging: default
+      hydra/hydra_help: default
+      hydra/help: default
+      hydra/sweeper: basic
+      hydra/launcher: basic
+      hydra/output: default
+  verbose: false
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/hydra/.hydra/overrides.yaml b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/hydra/.hydra/overrides.yaml
new file mode 100644
index 0000000000000000000000000000000000000000..9b15d894fe271e811eeaf40d3f021dbf66f6644e
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/hydra/.hydra/overrides.yaml
@@ -0,0 +1,209 @@
+- model=openvla
+- env=flappy
+- init=bridge
+- mode=single
+- ++framework.qwenvl.base_vlm=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++framework.qwenvl.attn_implementation=flash_attention_2
+- ++framework.qwenvl.enable_gradient_checkpointing=true
+- ++framework.action_model.state_dim=7
+- ++framework.action_model.loss_type=discrete_ce
+- ++framework.action_model.action_horizon=1
+- ++framework.action_model.future_action_window_size=0
+- ++framework.action_model.past_action_window_size=0
+- ++framework.action_model.action_dim=7
+- ++framework.action_model.action_env_dim=2
+- ++framework.kv_memory.enabled=false
+- ++framework.kv_memory.window=4
+- ++framework.kv_memory.rollout_len=8
+- ++framework.kv_memory.packed_train=false
+- ++framework.kv_memory.rebased_sink=true
+- ++framework.name=QwenOFT
+- ++datasets.vla_data.dataset_py=lerobot_datasets
+- ++datasets.vla_data.include_state=true
+- ++datasets.vla_data.data_root_dir=playground/Datasets/rl_games
+- ++datasets.vla_data.data_mix=flappy_train
+- ++datasets.vla_data.eval_data_mix=null
+- ++datasets.vla_data.custom_mixtures_path=null
+- ++datasets.vla_data.action_type=discrete
+- ++datasets.vla_data.sequential_step_sampling=false
+- ++datasets.vla_data.eval_sequential_step_sampling=null
+- ++datasets.vla_data.num_workers=8
+- ++datasets.vla_data.eval_num_workers=8
+- ++datasets.vla_data.prefetch_factor=4
+- ++datasets.vla_data.persistent_workers=true
+- ++datasets.vla_data.pin_memory=true
+- ++datasets.vla_data.shuffle=true
+- ++datasets.vla_data.action_balance.enabled=false
+- ++datasets.vla_data.action_balance.strategy=balanced_epoch
+- ++datasets.vla_data.action_balance.action_key=action_id
+- ++datasets.vla_data.action_balance.target_flap_fraction=0.3
+- ++datasets.vla_data.action_balance.noop_id=0
+- ++datasets.vla_data.action_balance.flap_id=1
+- ++datasets.vla_data.latency_curriculum.enabled=false
+- ++datasets.vla_data.latency_curriculum.strategy=exclusive
+- ++datasets.vla_data.latency_curriculum.latencies=null
+- ++datasets.vla_data.latency_curriculum.phase_steps=null
+- ++datasets.vla_data.latency_curriculum.phase_distributions=null
+- ++datasets.vla_data.latency_curriculum.new_latency_passes=1.0
+- ++datasets.vla_data.latency_curriculum.replay_passes=0.25
+- ++datasets.vla_data.latency_curriculum.target_total_passes=2.0
+- ++datasets.vla_data.latency_curriculum.final_equalization=true
+- ++datasets.vla_data.latency_curriculum.step_budget_mode=auto
+- ++datasets.vla_data.latency_curriculum.eval_at_phase_end=false
+- ++datasets.vla_data.latency_curriculum.save_at_phase_end=false
+- ++datasets.vla_data.latency_curriculum.computed_plan=null
+- ++datasets.vla_data.per_device_batch_size=8
+- ++datasets.vla_data.load_all_data_for_training=true
+- ++datasets.vla_data.num_obs_frames=1
+- ++datasets.vla_data.image_mode=single
+- ++datasets.vla_data.prompt_mode=raw
+- ++datasets.vla_data.stitch_grid=[2,2]
+- ++datasets.vla_data.obs_image_size=null
+- ++datasets.vla_data.video_backend=torchvision_av
+- ++dataset.source_hf=
+- ++dataset.config_name=null
+- ++dataset.source_subdir=null
+- ++dataset.converted_name=flappy_train
+- ++dataset.single_source_hf=
+- ++dataset.mixed_source_hf=
+- ++dataset.single_converted_name=flappy_train
+- ++dataset.mixed_converted_name=flappy_mixed_latency_train
+- ++dataset.single_latency_filter=null
+- ++dataset.mixed_latency_filter=null
+- ++dataset.force_download=false
+- ++dataset.setup_force=false
+- ++dataset.skip_verification=false
+- ++dataset.target_latency_unit=raw_frames
+- ++dataset.verify_rows=200
+- ++dataset.max_episodes=null
+- ++dataset.episodes_per_latency=null
+- ++dataset.latency_filter=null
+- ++dataset.debug_subset.enabled=false
+- ++dataset.debug_subset.max_episodes=5
+- ++dataset.debug_subset.suffix=debug
+- ++base_model.repo_id=Qwen/Qwen3-VL-4B-Instruct
+- ++initialization.checkpoint_local_dir=playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1
+- ++initialization.checkpoint_hf_repo_id=StarVLA/Qwen3VL-OFT-Bridge-RT-1
+- ++initialization.checkpoint_filename=checkpoints/steps_5000_pytorch_model.pt
+- ++trainer.max_train_steps=4000
+- ++trainer.num_warmup_steps=100
+- ++trainer.save_interval=500
+- ++trainer.eval_interval=250
+- ++trainer.eval_num_batches=50
+- ++trainer.per_latency_eval_num_batches=null
+- ++trainer.eval_action_classification=false
+- ++trainer.eval_action_classification_interval=null
+- ++trainer.cc_f1_tolerance=1
+- ++trainer.learning_rate.base=2e-05
+- ++trainer.learning_rate.qwen_vl_interface=1e-05
+- ++trainer.learning_rate.action_model=0.0001
+- ++trainer.lr_scheduler_type=cosine_with_min_lr
+- ++trainer.scheduler_specific_kwargs.min_lr=1e-06
+- ++trainer.freeze_modules=
+- ++trainer.freeze_vit=false
+- ++trainer.freeze_tied_embedding=false
+- ++trainer.freeze_llm_layers=[]
+- ++trainer.loss_scale.vla=1.0
+- ++trainer.loss_scale.vlm=0.1
+- ++trainer.max_grad_norm=1.0
+- ++trainer.weight_decay=0.0
+- ++trainer.logging_frequency=1
+- ++trainer.profile_timing.enabled=false
+- ++trainer.profile_timing.log_interval=10
+- ++trainer.gradient_clipping=1.0
+- ++trainer.gradient_accumulation_steps=16
+- ++trainer.distributed_backend=deepspeed
+- ++trainer.is_resume=false
+- ++trainer.pretrained_checkpoint=null
+- ++trainer.resume_step=0
+- ++trainer.reload_modules=null
+- ++trainer.optimizer.name=AdamW
+- ++trainer.optimizer.betas=[0.9,0.95]
+- ++trainer.optimizer.eps=1e-08
+- ++trainer.optimizer.weight_decay=1e-08
+- ++trainer.optimizer.fused=true
+- ++trainer.save_format=pt
+- ++workspace_dir=WORKSPACE_DIR
+- ++run_root_dir=results/Checkpoints
+- ++seed=42
+- ++auth.env_file=null
+- ++auth.hf_token_env=HF_TOKEN
+- ++auth.wandb_api_key_env=WANDB_API_KEY
+- ++paths.run_root_dir=results/Checkpoints
+- ++paths.dataset_local_dir=data/memory/flappy_fix_latency_3_200ep_7k2steps
+- ++paths.dataset_cache_dir=null
+- ++paths.base_model_dir=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++paths.accelerate_config=starVLA/config/deepseeds/deepspeed_zero2.yaml
+- ++rl_games.model_alias=openvla
+- ++rl_games.env_eval.image_size=224
+- ++rl_games.env_eval.frameskip=1
+- ++rl_games.env_eval.image_transform=raw_rgb
+- ++rl_games.env_eval.prompt_mode=raw
+- ++rl_games.env_eval.ghost_trail.history_frames=5
+- ++rl_games.env_eval.ghost_trail.gamma=1.3
+- ++rl_games.env_eval.ghost_trail.min_alpha=35
+- ++rl_games.env_eval.ghost_trail.scroll_px_per_step=4.0
+- ++rl_games.env_eval.ghost_trail.ground_fraction=0.22
+- ++rl_games.env_eval.seed=42
+- ++rl_games.env_eval.fixed_episode_seeds=true
+- ++rl_games.env_eval.latency_seed_stride=0
+- ++rl_games.env_eval.task_seed_stride=0
+- ++rl_games.env_eval.task_description=
+- ++rl_games.env_eval.eval_parallel_envs=5
+- ++rl_games.env_eval.action_chunk_execution.enabled=false
+- ++rl_games.env_eval.action_chunk_execution.chunk_size=null
+- ++rl_games.env_eval.enabled=true
+- ++rl_games.env_eval.eval_backend=latency_bench
+- ++rl_games.env_eval.distributed_mode=rank_sharded
+- ++rl_games.env_eval.vectorized.enabled=false
+- ++rl_games.env_eval.vectorized.batch_size=1
+- ++rl_games.env_eval.latency.prompt_map_path=null
+- ++rl_games.env_eval.latency.mode=single
+- ++rl_games.env_eval.latency.values=[0]
+- ++rl_games.env_eval.mid_train.enabled=true
+- ++rl_games.env_eval.mid_train.interval_steps=250
+- ++rl_games.env_eval.mid_train.latencies=[3]
+- ++rl_games.env_eval.mid_train.num_episodes=20
+- ++rl_games.env_eval.mid_train.max_steps_per_episode=3600
+- ++rl_games.env_eval.post_train.enabled=true
+- ++rl_games.env_eval.post_train.latencies=[3]
+- ++rl_games.env_eval.post_train.num_episodes=50
+- ++rl_games.env_eval.post_train.max_steps_per_episode=3600
+- ++rl_games.task=flappy
+- ++rl_games.initialization_mode=bridge
+- ++rl_games.action_carrier=bridge
+- ++bridge_base_model.repo_id.openvla=Qwen/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.repo_id.pi0=StarVLA/Qwen2.5-VL-3B-Instruct-Action
+- ++bridge_base_model.repo_id.pi05=Qwen/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.repo_id.gr00t=Qwen/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.local_dir.openvla=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.local_dir.pi0=playground/Pretrained_models/Qwen2.5-VL-3B-Instruct-Action
+- ++bridge_base_model.local_dir.pi05=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++bridge_base_model.local_dir.gr00t=playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++checkpoint.load=auto
+- ++checkpoint.hf_repo_id=null
+- ++checkpoint.save_best_model=false
+- ++checkpoint.save_final_model=true
+- ++checkpoint.save_pt_file=false
+- ++checkpoint.save_training_state=true
+- ++checkpoint.save_safetensors_file=true
+- ++checkpoint.local.keep_last_n=1
+- ++checkpoint.sync.enabled=false
+- ++checkpoint.sync.repo_id=null
+- ++checkpoint.sync.keep_last_n=0
+- ++checkpoint.sync.sync_every_n_checkpoints=1
+- ++checkpoint.sync.resume_policy=local_latest
+- ++run_id=flappy_fix_latency_3_200ep_7k2steps_single_baseline
+- ++output_dir=null
+- ++config_yaml=null
+- ++is_debug=false
+- ++version_id=0.21
+- ++run_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints
+- ++trainer.is_resume=true
+- ++trainer.pretrained_checkpoint=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state
+- ++trainer.resume_step=3000
+- ++datasets.vla_data.data_root_dir=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps
+- ++datasets.vla_data.data_mix=flappy_train__bridge
+- ++datasets.vla_data.eval_data_mix=flappy_train__bridge__val
+- ++framework.qwenvl.base_vlm=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/playground/Pretrained_models/Qwen3-VL-4B-Instruct
+- ++rl_games.env_eval.latency.prompt_map_path=/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/data/memory/flappy_fix_latency_3_200ep_7k2steps/flappy_train__bridge/latency_prompt_map.json
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/hydra/train_starvla_hydra.log b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/hydra/train_starvla_hydra.log
new file mode 100644
index 0000000000000000000000000000000000000000..c5b6fe4c326a06ab59f70ce2ab1d0a784bf805a4
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/hydra/train_starvla_hydra.log
@@ -0,0 +1,5271 @@
+[2026-07-28 01:53:02,069][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21'
+[2026-07-28 01:53:02,070][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 01:53:02,071][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 01:53:02,291][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 01:53:02,291][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 01:53:02,291][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 01:53:02,291][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 01:53:03,278][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 01:53:03,293][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 01:53:03,346][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 01:53:03,361][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 01:53:35,652][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt
+[2026-07-28 01:53:35,655][starVLA.training.train_starvla][INFO] - Preloaded model checkpoint before Accelerator/DeepSpeed initialization: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/playground/Pretrained_models/Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt
+[2026-07-28 01:53:35,665][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 01:53:35,666][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 01:53:36,623][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 01:53:36,624][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 01:53:36,624][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 01:53:36,624][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 01:53:36,624][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 01:53:36,624][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 01:53:36,625][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 01:53:36,625][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 01:53:36,625][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 01:53:36,625][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 01:53:36,626][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 01:53:36,626][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 01:53:36,626][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 01:53:36,627][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 01:53:36,627][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 01:53:36,627][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 01:53:36,628][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 01:53:36,628][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 01:53:36,628][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 01:53:36,628][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 01:53:36,628][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 01:53:36,629][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 01:53:36,629][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 01:53:36,629][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 01:53:36,629][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 01:53:36,629][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 01:53:36,629][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 01:53:36,629][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 01:53:36,630][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 01:53:36,630][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 01:53:36,630][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 01:53:36,630][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 01:53:36,630][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 01:53:36,630][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 01:53:36,630][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 01:53:36,631][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 01:53:36,631][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 01:53:36,631][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 01:53:36,631][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 01:53:36,631][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 01:53:36,631][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 01:53:36,632][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 01:53:36,632][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 01:53:36,632][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 01:53:36,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 01:53:36,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 01:53:36,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 01:53:36,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 01:53:36,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 01:53:36,633][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 01:53:36,634][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 01:53:36,634][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 01:53:36,634][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 01:53:36,635][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 01:53:36,635][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 01:53:36,635][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 01:53:36,635][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 01:53:36,635][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 01:53:36,636][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 01:53:36,636][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 01:53:36,636][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 01:53:36,636][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 01:53:37,991][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 01:53:37,996][starVLA.dataloader][INFO] - Created train VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 01:53:37,998][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 01:53:38,014][starVLA.dataloader][INFO] - Created eval VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 01:53:38,019][starVLA.training.train_starvla][INFO] - LR Group qwen_vl_interface: lr=1e-05, num_params=713
+[2026-07-28 01:53:38,019][starVLA.training.train_starvla][INFO] - LR Group action_model: lr=0.0001, num_params=16
+[2026-07-28 01:53:39,282][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch.
+[2026-07-28 01:53:39,291][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/config.full.yaml
+[2026-07-28 01:53:39,302][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/config.yaml
+[2026-07-28 01:53:39,303][starVLA.training.train_starvla][INFO] - No pretrained checkpoint provided. Starting training from scratch.
+[2026-07-28 01:53:39,307][accelerate.accelerator][INFO] - Since you passed both train and evaluation dataloader, `is_train_batch_min` (here True will decide the `train_batch_size` (8).
+[2026-07-28 01:53:50,656][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-07-28 01:53:50,656][starVLA.training.train_starvla][INFO] -   Total optimization steps = 4000
+[2026-07-28 01:53:50,656][starVLA.training.train_starvla][INFO] -   Per device batch size = 8
+[2026-07-28 01:53:50,657][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-07-28 01:53:50,657][starVLA.training.train_starvla][INFO] -   Total batch size = 256
+[2026-07-28 01:55:05,746][starVLA.training.train_starvla][INFO] - Step 1, Loss: {'train/loss': 0.6968364119529724, 'train/grad_norm_pre_clip': 33.22444534301758, 'timing/data': 0.00038653286173939705, 'timing/model': 1.150137902237475, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.0})
+[2026-07-28 01:55:10,091][starVLA.training.train_starvla][INFO] - Step 2, Loss: {'train/loss': 0.6962391138076782, 'train/grad_norm_pre_clip': 30.7915096282959, 'timing/data': 0.0003964710049331188, 'timing/model': 0.47678671637549996, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.0})
+[2026-07-28 01:55:14,245][starVLA.training.train_starvla][INFO] - Step 3, Loss: {'train/loss': 0.6959848403930664, 'train/grad_norm_pre_clip': 31.578458786010742, 'timing/data': 0.0002616927959024906, 'timing/model': 0.33957912493497133, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.0})
+[2026-07-28 01:55:18,407][starVLA.training.train_starvla][INFO] - Step 4, Loss: {'train/loss': 0.693832278251648, 'train/grad_norm_pre_clip': 33.420135498046875, 'timing/data': 0.0005320850759744644, 'timing/model': 0.3325315802358091, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.0})
+[2026-07-28 01:55:22,535][starVLA.training.train_starvla][INFO] - Step 5, Loss: {'train/loss': 0.68753981590271, 'train/grad_norm_pre_clip': 31.693761825561523, 'timing/data': 0.0002987799234688282, 'timing/model': 0.33302693301811814, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.0})
+[2026-07-28 01:55:26,630][starVLA.training.train_starvla][INFO] - Step 6, Loss: {'train/loss': 0.6660332679748535, 'train/grad_norm_pre_clip': 30.703161239624023, 'timing/data': 0.00028752023354172707, 'timing/model': 0.33481680508702993, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.0})
+[2026-07-28 01:55:30,757][starVLA.training.train_starvla][INFO] - Step 7, Loss: {'train/loss': 0.6541649699211121, 'train/grad_norm_pre_clip': 34.08931350708008, 'timing/data': 0.00028402265161275864, 'timing/model': 0.33234589407220483, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.0})
+[2026-07-28 01:55:34,848][starVLA.training.train_starvla][INFO] - Step 8, Loss: {'train/loss': 0.5689445734024048, 'train/grad_norm_pre_clip': 28.968576431274414, 'timing/data': 0.00038146693259477615, 'timing/model': 0.3314359150826931, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.0})
+[2026-07-28 01:55:38,941][starVLA.training.train_starvla][INFO] - Step 9, Loss: {'train/loss': 0.5555559396743774, 'train/grad_norm_pre_clip': 24.901180267333984, 'timing/data': 0.00031264591962099075, 'timing/model': 0.33231036365032196, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.0})
+[2026-07-28 01:55:43,041][starVLA.training.train_starvla][INFO] - Step 10, Loss: {'train/loss': 0.4515530467033386, 'train/grad_norm_pre_clip': 7.87453031539917, 'timing/data': 0.000414176844060421, 'timing/model': 0.33405636670067906, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.0})
+[2026-07-28 01:55:47,120][starVLA.training.train_starvla][INFO] - Step 11, Loss: {'train/loss': 0.452229768037796, 'train/grad_norm_pre_clip': 4.553825855255127, 'timing/data': 0.00027421489357948303, 'timing/model': 0.3314596572890878, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.0})
+[2026-07-28 01:55:51,218][starVLA.training.train_starvla][INFO] - Step 12, Loss: {'train/loss': 0.4588742256164551, 'train/grad_norm_pre_clip': 0.9649564027786255, 'timing/data': 0.0002859528176486492, 'timing/model': 0.33128544595092535, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.0})
+[2026-07-28 01:55:55,281][starVLA.training.train_starvla][INFO] - Step 13, Loss: {'train/loss': 0.45436298847198486, 'train/grad_norm_pre_clip': 2.8477466106414795, 'timing/data': 0.0002787439152598381, 'timing/model': 0.3322036061435938, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.0})
+[2026-07-28 01:55:59,350][starVLA.training.train_starvla][INFO] - Step 14, Loss: {'train/loss': 0.4995079040527344, 'train/grad_norm_pre_clip': 6.205190181732178, 'timing/data': 0.00030780304223299026, 'timing/model': 0.3331617210060358, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.0})
+[2026-07-28 01:56:03,415][starVLA.training.train_starvla][INFO] - Step 15, Loss: {'train/loss': 0.4510555863380432, 'train/grad_norm_pre_clip': 3.221482276916504, 'timing/data': 0.0002870718017220497, 'timing/model': 0.3312325980514288, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.0})
+[2026-07-28 01:56:07,476][starVLA.training.train_starvla][INFO] - Step 16, Loss: {'train/loss': 0.4539657533168793, 'train/grad_norm_pre_clip': 2.4494760036468506, 'timing/data': 0.00030132709071040154, 'timing/model': 0.33366177789866924, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.0})
+[2026-07-28 01:56:11,533][starVLA.training.train_starvla][INFO] - Step 17, Loss: {'train/loss': 0.4342711567878723, 'train/grad_norm_pre_clip': 1.9113987684249878, 'timing/data': 0.0002812170423567295, 'timing/model': 0.3313219710253179, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.0})
+[2026-07-28 01:56:15,582][starVLA.training.train_starvla][INFO] - Step 18, Loss: {'train/loss': 0.517727792263031, 'train/grad_norm_pre_clip': 6.186933517456055, 'timing/data': 0.00038868095725774765, 'timing/model': 0.32976411702111363, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.0})
+[2026-07-28 01:56:19,650][starVLA.training.train_starvla][INFO] - Step 19, Loss: {'train/loss': 0.42259541153907776, 'train/grad_norm_pre_clip': 3.9559543132781982, 'timing/data': 0.0003261198289692402, 'timing/model': 0.3312577558681369, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.0})
+[2026-07-28 01:56:23,725][starVLA.training.train_starvla][INFO] - Step 20, Loss: {'train/loss': 0.5361746549606323, 'train/grad_norm_pre_clip': 11.656529426574707, 'timing/data': 0.0002808910794556141, 'timing/model': 0.3292960487306118, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.0})
+[2026-07-28 01:56:27,824][starVLA.training.train_starvla][INFO] - Step 21, Loss: {'train/loss': 0.4069287180900574, 'train/grad_norm_pre_clip': 4.09623908996582, 'timing/data': 0.00039678905159235, 'timing/model': 0.33320950297638774, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.0})
+[2026-07-28 01:56:31,884][starVLA.training.train_starvla][INFO] - Step 22, Loss: {'train/loss': 0.4597986042499542, 'train/grad_norm_pre_clip': 6.975360870361328, 'timing/data': 0.0002868478186428547, 'timing/model': 0.33200092473998666, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.0})
+[2026-07-28 01:56:35,944][starVLA.training.train_starvla][INFO] - Step 23, Loss: {'train/loss': 0.45976030826568604, 'train/grad_norm_pre_clip': 8.137288093566895, 'timing/data': 0.00027823587879538536, 'timing/model': 0.3318849499337375, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.0})
+[2026-07-28 01:56:39,994][starVLA.training.train_starvla][INFO] - Step 24, Loss: {'train/loss': 0.39828425645828247, 'train/grad_norm_pre_clip': 4.191092491149902, 'timing/data': 0.000299651175737381, 'timing/model': 0.33013294404372573, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.0})
+[2026-07-28 01:56:44,058][starVLA.training.train_starvla][INFO] - Step 25, Loss: {'train/loss': 0.3799653947353363, 'train/grad_norm_pre_clip': 10.090250015258789, 'timing/data': 0.00029709283262491226, 'timing/model': 0.33336862828582525, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.0})
+[2026-07-28 01:56:48,110][starVLA.training.train_starvla][INFO] - Step 26, Loss: {'train/loss': 0.34653639793395996, 'train/grad_norm_pre_clip': 4.332661151885986, 'timing/data': 0.00027694134041666985, 'timing/model': 0.3305051480419934, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.0})
+[2026-07-28 01:56:52,169][starVLA.training.train_starvla][INFO] - Step 27, Loss: {'train/loss': 0.38377368450164795, 'train/grad_norm_pre_clip': 9.21958065032959, 'timing/data': 0.0002790880389511585, 'timing/model': 0.3320365399122238, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.0})
+[2026-07-28 01:56:56,234][starVLA.training.train_starvla][INFO] - Step 28, Loss: {'train/loss': 0.39477261900901794, 'train/grad_norm_pre_clip': 19.293027877807617, 'timing/data': 0.0002923347055912018, 'timing/model': 0.33534796023741364, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.01})
+[2026-07-28 01:57:00,330][starVLA.training.train_starvla][INFO] - Step 29, Loss: {'train/loss': 0.28097355365753174, 'train/grad_norm_pre_clip': 11.83454418182373, 'timing/data': 0.000295375008136034, 'timing/model': 0.3315671840682626, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.01})
+[2026-07-28 01:57:04,459][starVLA.training.train_starvla][INFO] - Step 30, Loss: {'train/loss': 0.42442119121551514, 'train/grad_norm_pre_clip': 32.491119384765625, 'timing/data': 0.0003045089542865753, 'timing/model': 0.3292319653555751, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.01})
+[2026-07-28 01:57:08,596][starVLA.training.train_starvla][INFO] - Step 31, Loss: {'train/loss': 0.4448468089103699, 'train/grad_norm_pre_clip': 29.845684051513672, 'timing/data': 0.0007161782123148441, 'timing/model': 0.3323943023569882, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.01})
+[2026-07-28 01:57:12,657][starVLA.training.train_starvla][INFO] - Step 32, Loss: {'train/loss': 0.38288015127182007, 'train/grad_norm_pre_clip': 29.561077117919922, 'timing/data': 0.0002775280736386776, 'timing/model': 0.33087014220654964, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.01})
+[2026-07-28 01:57:16,704][starVLA.training.train_starvla][INFO] - Step 33, Loss: {'train/loss': 0.3588103652000427, 'train/grad_norm_pre_clip': 11.951409339904785, 'timing/data': 0.00028167711570858955, 'timing/model': 0.3310966622084379, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.01})
+[2026-07-28 01:57:20,755][starVLA.training.train_starvla][INFO] - Step 34, Loss: {'train/loss': 0.4214519262313843, 'train/grad_norm_pre_clip': 55.808589935302734, 'timing/data': 0.00028524594381451607, 'timing/model': 0.3339735511690378, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.01})
+[2026-07-28 01:57:24,820][starVLA.training.train_starvla][INFO] - Step 35, Loss: {'train/loss': 0.4067703187465668, 'train/grad_norm_pre_clip': 24.04734230041504, 'timing/data': 0.0004039178602397442, 'timing/model': 0.3329843902029097, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.01})
+[2026-07-28 01:57:29,152][starVLA.training.train_starvla][INFO] - Step 36, Loss: {'train/loss': 0.3221471905708313, 'train/grad_norm_pre_clip': 14.581399917602539, 'timing/data': 0.0002844058908522129, 'timing/model': 0.33189270831644535, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.01})
+[2026-07-28 01:57:33,498][starVLA.training.train_starvla][INFO] - Step 37, Loss: {'train/loss': 0.33704373240470886, 'train/grad_norm_pre_clip': 10.90261459350586, 'timing/data': 0.00025970395654439926, 'timing/model': 0.3293785508722067, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.01})
+[2026-07-28 01:57:37,548][starVLA.training.train_starvla][INFO] - Step 38, Loss: {'train/loss': 0.2548055052757263, 'train/grad_norm_pre_clip': 17.65854263305664, 'timing/data': 0.0002996032126247883, 'timing/model': 0.33067181007936597, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.01})
+[2026-07-28 01:57:41,588][starVLA.training.train_starvla][INFO] - Step 39, Loss: {'train/loss': 0.34655189514160156, 'train/grad_norm_pre_clip': 9.834115982055664, 'timing/data': 0.0003600437194108963, 'timing/model': 0.3293494158424437, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.01})
+[2026-07-28 01:57:45,640][starVLA.training.train_starvla][INFO] - Step 40, Loss: {'train/loss': 0.2151864767074585, 'train/grad_norm_pre_clip': 10.17291259765625, 'timing/data': 0.00038474705070257187, 'timing/model': 0.3321018679998815, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.01})
+[2026-07-28 01:57:49,706][starVLA.training.train_starvla][INFO] - Step 41, Loss: {'train/loss': 0.2815706729888916, 'train/grad_norm_pre_clip': 7.068741798400879, 'timing/data': 0.0003083692863583565, 'timing/model': 0.3314509876072407, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.01})
+[2026-07-28 01:57:53,751][starVLA.training.train_starvla][INFO] - Step 42, Loss: {'train/loss': 0.2884249687194824, 'train/grad_norm_pre_clip': 6.555479526519775, 'timing/data': 0.0002859393134713173, 'timing/model': 0.3316491488367319, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.01})
+[2026-07-28 01:57:57,803][starVLA.training.train_starvla][INFO] - Step 43, Loss: {'train/loss': 0.21924808621406555, 'train/grad_norm_pre_clip': 11.519065856933594, 'timing/data': 0.00028402870520949364, 'timing/model': 0.33067686622962356, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.01})
+[2026-07-28 01:58:01,874][starVLA.training.train_starvla][INFO] - Step 44, Loss: {'train/loss': 0.2376474142074585, 'train/grad_norm_pre_clip': 20.988384246826172, 'timing/data': 0.00029374193400144577, 'timing/model': 0.3311215811409056, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.01})
+[2026-07-28 01:58:05,917][starVLA.training.train_starvla][INFO] - Step 45, Loss: {'train/loss': 0.2507634460926056, 'train/grad_norm_pre_clip': 14.130392074584961, 'timing/data': 0.000285138376057148, 'timing/model': 0.33076360169798136, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.01})
+[2026-07-28 01:58:09,959][starVLA.training.train_starvla][INFO] - Step 46, Loss: {'train/loss': 0.30044931173324585, 'train/grad_norm_pre_clip': 20.220069885253906, 'timing/data': 0.00028092600405216217, 'timing/model': 0.3287718119099736, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.01})
+[2026-07-28 01:58:14,013][starVLA.training.train_starvla][INFO] - Step 47, Loss: {'train/loss': 0.28648701310157776, 'train/grad_norm_pre_clip': 10.422154426574707, 'timing/data': 0.0003107301890850067, 'timing/model': 0.3315402651205659, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.01})
+[2026-07-28 01:58:18,073][starVLA.training.train_starvla][INFO] - Step 48, Loss: {'train/loss': 0.2779616713523865, 'train/grad_norm_pre_clip': 14.305591583251953, 'timing/data': 0.00027709873393177986, 'timing/model': 0.32984435372054577, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.01})
+[2026-07-28 01:58:22,165][starVLA.training.train_starvla][INFO] - Step 49, Loss: {'train/loss': 0.32890036702156067, 'train/grad_norm_pre_clip': 61.02674102783203, 'timing/data': 0.0002841297537088394, 'timing/model': 0.3323629768565297, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.01})
+[2026-07-28 01:58:26,222][starVLA.training.train_starvla][INFO] - Step 50, Loss: {'train/loss': 0.2300860583782196, 'train/grad_norm_pre_clip': 13.77906608581543, 'timing/data': 0.0003052116371691227, 'timing/model': 0.3304663128219545, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.01})
+[2026-07-28 01:58:30,320][starVLA.training.train_starvla][INFO] - Step 51, Loss: {'train/loss': 0.5528312921524048, 'train/grad_norm_pre_clip': 46.421669006347656, 'timing/data': 0.00028769485652446747, 'timing/model': 0.3325670100748539, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.01})
+[2026-07-28 01:58:34,402][starVLA.training.train_starvla][INFO] - Step 52, Loss: {'train/loss': 0.44394341111183167, 'train/grad_norm_pre_clip': 44.49131393432617, 'timing/data': 0.00029072724282741547, 'timing/model': 0.3305796873755753, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.01})
+[2026-07-28 01:58:38,486][starVLA.training.train_starvla][INFO] - Step 53, Loss: {'train/loss': 0.21964848041534424, 'train/grad_norm_pre_clip': 21.217361450195312, 'timing/data': 0.0002887556329369545, 'timing/model': 0.3455605520866811, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.01})
+[2026-07-28 01:58:42,543][starVLA.training.train_starvla][INFO] - Step 54, Loss: {'train/loss': 0.6497321128845215, 'train/grad_norm_pre_clip': 96.0418930053711, 'timing/data': 0.0002879425883293152, 'timing/model': 0.3303689998574555, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.01})
+[2026-07-28 01:58:46,601][starVLA.training.train_starvla][INFO] - Step 55, Loss: {'train/loss': 0.7703109979629517, 'train/grad_norm_pre_clip': 107.41295623779297, 'timing/data': 0.0002694888971745968, 'timing/model': 0.3305485490709543, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.01})
+[2026-07-28 01:58:50,683][starVLA.training.train_starvla][INFO] - Step 56, Loss: {'train/loss': 0.2997353672981262, 'train/grad_norm_pre_clip': 30.183361053466797, 'timing/data': 0.00029579922556877136, 'timing/model': 0.338062081951648, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.01})
+[2026-07-28 01:58:54,737][starVLA.training.train_starvla][INFO] - Step 57, Loss: {'train/loss': 0.17399698495864868, 'train/grad_norm_pre_clip': 4.39428186416626, 'timing/data': 0.00029454100877046585, 'timing/model': 0.3310128250159323, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.01})
+[2026-07-28 01:58:58,824][starVLA.training.train_starvla][INFO] - Step 58, Loss: {'train/loss': 0.3781200051307678, 'train/grad_norm_pre_clip': 25.412670135498047, 'timing/data': 0.0002860007807612419, 'timing/model': 0.3322929418645799, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.01})
+[2026-07-28 01:59:02,931][starVLA.training.train_starvla][INFO] - Step 59, Loss: {'train/loss': 0.22348225116729736, 'train/grad_norm_pre_clip': 9.862617492675781, 'timing/data': 0.0002871369943022728, 'timing/model': 0.33375933207571507, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.01})
+[2026-07-28 01:59:06,987][starVLA.training.train_starvla][INFO] - Step 60, Loss: {'train/loss': 0.2514772117137909, 'train/grad_norm_pre_clip': 19.05960464477539, 'timing/data': 0.00027682119980454445, 'timing/model': 0.331696602050215, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.01})
+[2026-07-28 01:59:11,077][starVLA.training.train_starvla][INFO] - Step 61, Loss: {'train/loss': 0.20888099074363708, 'train/grad_norm_pre_clip': 13.740832328796387, 'timing/data': 0.00027921469882130623, 'timing/model': 0.3322199690155685, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.01})
+[2026-07-28 01:59:15,167][starVLA.training.train_starvla][INFO] - Step 62, Loss: {'train/loss': 0.2225627303123474, 'train/grad_norm_pre_clip': 10.9115629196167, 'timing/data': 0.00027345819398760796, 'timing/model': 0.3438924178481102, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.01})
+[2026-07-28 01:59:19,277][starVLA.training.train_starvla][INFO] - Step 63, Loss: {'train/loss': 0.20371802151203156, 'train/grad_norm_pre_clip': 12.025599479675293, 'timing/data': 0.00034550437703728676, 'timing/model': 0.33058809908106923, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.01})
+[2026-07-28 01:59:23,389][starVLA.training.train_starvla][INFO] - Step 64, Loss: {'train/loss': 0.14824846386909485, 'train/grad_norm_pre_clip': 13.551434516906738, 'timing/data': 0.00027927104383707047, 'timing/model': 0.34664247184991837, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.01})
+[2026-07-28 01:59:27,466][starVLA.training.train_starvla][INFO] - Step 65, Loss: {'train/loss': 0.16950571537017822, 'train/grad_norm_pre_clip': 4.752241134643555, 'timing/data': 0.00026146695017814636, 'timing/model': 0.32959411572664976, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.01})
+[2026-07-28 01:59:31,507][starVLA.training.train_starvla][INFO] - Step 66, Loss: {'train/loss': 0.18239566683769226, 'train/grad_norm_pre_clip': 7.5288286209106445, 'timing/data': 0.00039242813363671303, 'timing/model': 0.3297410439699888, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.01})
+[2026-07-28 01:59:35,562][starVLA.training.train_starvla][INFO] - Step 67, Loss: {'train/loss': 0.10047340393066406, 'train/grad_norm_pre_clip': 4.454858303070068, 'timing/data': 0.00026527605950832367, 'timing/model': 0.33118804497644305, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.01})
+[2026-07-28 01:59:39,608][starVLA.training.train_starvla][INFO] - Step 68, Loss: {'train/loss': 0.22769466042518616, 'train/grad_norm_pre_clip': 18.37525749206543, 'timing/data': 0.00027818605303764343, 'timing/model': 0.3297023018822074, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.01})
+[2026-07-28 01:59:43,653][starVLA.training.train_starvla][INFO] - Step 69, Loss: {'train/loss': 0.25941741466522217, 'train/grad_norm_pre_clip': 14.239717483520508, 'timing/data': 0.00027426425367593765, 'timing/model': 0.33156808093190193, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.01})
+[2026-07-28 01:59:47,695][starVLA.training.train_starvla][INFO] - Step 70, Loss: {'train/loss': 0.1686270534992218, 'train/grad_norm_pre_clip': 9.136213302612305, 'timing/data': 0.00029049720615148544, 'timing/model': 0.33172257896512747, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.01})
+[2026-07-28 01:59:51,763][starVLA.training.train_starvla][INFO] - Step 71, Loss: {'train/loss': 0.20342496037483215, 'train/grad_norm_pre_clip': 9.206730842590332, 'timing/data': 0.0002772510051727295, 'timing/model': 0.3303095460869372, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.01})
+[2026-07-28 01:59:55,805][starVLA.training.train_starvla][INFO] - Step 72, Loss: {'train/loss': 0.20016184449195862, 'train/grad_norm_pre_clip': 9.122693061828613, 'timing/data': 0.00027772877365350723, 'timing/model': 0.3311839629895985, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.01})
+[2026-07-28 01:59:59,851][starVLA.training.train_starvla][INFO] - Step 73, Loss: {'train/loss': 0.2022220939397812, 'train/grad_norm_pre_clip': 6.98841667175293, 'timing/data': 0.00027110008522868156, 'timing/model': 0.330471508204937, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.01})
+[2026-07-28 02:00:03,917][starVLA.training.train_starvla][INFO] - Step 74, Loss: {'train/loss': 0.19061484932899475, 'train/grad_norm_pre_clip': 6.096477031707764, 'timing/data': 0.00034841708838939667, 'timing/model': 0.32874819869175553, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.01})
+[2026-07-28 02:00:07,980][starVLA.training.train_starvla][INFO] - Step 75, Loss: {'train/loss': 0.1444563865661621, 'train/grad_norm_pre_clip': 10.8096284866333, 'timing/data': 0.00025569507852196693, 'timing/model': 0.33162327017635107, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.01})
+[2026-07-28 02:00:12,026][starVLA.training.train_starvla][INFO] - Step 76, Loss: {'train/loss': 0.16190284490585327, 'train/grad_norm_pre_clip': 9.164570808410645, 'timing/data': 0.00027569709345698357, 'timing/model': 0.33126545092090964, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.01})
+[2026-07-28 02:00:16,074][starVLA.training.train_starvla][INFO] - Step 77, Loss: {'train/loss': 0.1688651740550995, 'train/grad_norm_pre_clip': 7.738804817199707, 'timing/data': 0.00027404818683862686, 'timing/model': 0.33101570000872016, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.01})
+[2026-07-28 02:00:20,127][starVLA.training.train_starvla][INFO] - Step 78, Loss: {'train/loss': 0.1898217499256134, 'train/grad_norm_pre_clip': 9.816823959350586, 'timing/data': 0.00036837998777627945, 'timing/model': 0.33146159816533327, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.01})
+[2026-07-28 02:00:24,203][starVLA.training.train_starvla][INFO] - Step 79, Loss: {'train/loss': 0.21455851197242737, 'train/grad_norm_pre_clip': 8.581974029541016, 'timing/data': 0.0002696169540286064, 'timing/model': 0.33264687983319163, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.01})
+[2026-07-28 02:00:28,573][starVLA.training.train_starvla][INFO] - Step 80, Loss: {'train/loss': 0.2757815718650818, 'train/grad_norm_pre_clip': 16.83482551574707, 'timing/data': 0.0002719340845942497, 'timing/model': 0.3293891088105738, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.01})
+[2026-07-28 02:00:32,683][starVLA.training.train_starvla][INFO] - Step 81, Loss: {'train/loss': 0.185243159532547, 'train/grad_norm_pre_clip': 8.388617515563965, 'timing/data': 0.0002443287521600723, 'timing/model': 0.3447104380466044, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.01})
+[2026-07-28 02:00:36,751][starVLA.training.train_starvla][INFO] - Step 82, Loss: {'train/loss': 0.185081347823143, 'train/grad_norm_pre_clip': 8.80974292755127, 'timing/data': 0.0003502010367810726, 'timing/model': 0.3293944913893938, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.02})
+[2026-07-28 02:00:41,245][starVLA.training.train_starvla][INFO] - Step 83, Loss: {'train/loss': 0.22498460114002228, 'train/grad_norm_pre_clip': 9.899009704589844, 'timing/data': 0.0002524876035749912, 'timing/model': 0.32883167266845703, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.02})
+[2026-07-28 02:00:45,538][starVLA.training.train_starvla][INFO] - Step 84, Loss: {'train/loss': 0.19458827376365662, 'train/grad_norm_pre_clip': 11.68953800201416, 'timing/data': 0.0002571321092545986, 'timing/model': 0.3294439469464123, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.02})
+[2026-07-28 02:00:49,588][starVLA.training.train_starvla][INFO] - Step 85, Loss: {'train/loss': 0.2350764274597168, 'train/grad_norm_pre_clip': 22.019390106201172, 'timing/data': 0.0007322868332266808, 'timing/model': 0.32961568888276815, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.02})
+[2026-07-28 02:00:53,637][starVLA.training.train_starvla][INFO] - Step 86, Loss: {'train/loss': 0.20418964326381683, 'train/grad_norm_pre_clip': 13.081796646118164, 'timing/data': 0.00027184607461094856, 'timing/model': 0.331083161290735, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.02})
+[2026-07-28 02:00:57,691][starVLA.training.train_starvla][INFO] - Step 87, Loss: {'train/loss': 0.16731607913970947, 'train/grad_norm_pre_clip': 7.2599382400512695, 'timing/data': 0.00027185119688510895, 'timing/model': 0.33135194005444646, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.02})
+[2026-07-28 02:01:01,748][starVLA.training.train_starvla][INFO] - Step 88, Loss: {'train/loss': 0.20115265250205994, 'train/grad_norm_pre_clip': 10.96242904663086, 'timing/data': 0.00026524532586336136, 'timing/model': 0.33744279108941555, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.02})
+[2026-07-28 02:01:05,796][starVLA.training.train_starvla][INFO] - Step 89, Loss: {'train/loss': 0.18096037209033966, 'train/grad_norm_pre_clip': 12.963250160217285, 'timing/data': 0.00028474628925323486, 'timing/model': 0.33068365696817636, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.02})
+[2026-07-28 02:01:09,877][starVLA.training.train_starvla][INFO] - Step 90, Loss: {'train/loss': 0.1650686264038086, 'train/grad_norm_pre_clip': 5.572188854217529, 'timing/data': 0.0002673477865755558, 'timing/model': 0.33011416997760534, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.02})
+[2026-07-28 02:01:13,963][starVLA.training.train_starvla][INFO] - Step 91, Loss: {'train/loss': 0.200815349817276, 'train/grad_norm_pre_clip': 11.534757614135742, 'timing/data': 0.00028935493901371956, 'timing/model': 0.33855529595166445, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.02})
+[2026-07-28 02:01:18,039][starVLA.training.train_starvla][INFO] - Step 92, Loss: {'train/loss': 0.1613377183675766, 'train/grad_norm_pre_clip': 6.2356977462768555, 'timing/data': 0.0002800598740577698, 'timing/model': 0.3296574801206589, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.02})
+[2026-07-28 02:01:22,117][starVLA.training.train_starvla][INFO] - Step 93, Loss: {'train/loss': 0.2693541646003723, 'train/grad_norm_pre_clip': 17.853485107421875, 'timing/data': 0.0002708602696657181, 'timing/model': 0.3360647913068533, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.02})
+[2026-07-28 02:01:26,187][starVLA.training.train_starvla][INFO] - Step 94, Loss: {'train/loss': 0.17076535522937775, 'train/grad_norm_pre_clip': 7.438172817230225, 'timing/data': 0.00029194075614213943, 'timing/model': 0.3401314248330891, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.02})
+[2026-07-28 02:01:30,263][starVLA.training.train_starvla][INFO] - Step 95, Loss: {'train/loss': 0.2354375720024109, 'train/grad_norm_pre_clip': 15.112689971923828, 'timing/data': 0.00035541411489248276, 'timing/model': 0.33633071603253484, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.02})
+[2026-07-28 02:01:34,321][starVLA.training.train_starvla][INFO] - Step 96, Loss: {'train/loss': 0.24073389172554016, 'train/grad_norm_pre_clip': 16.951595306396484, 'timing/data': 0.00026186229661107063, 'timing/model': 0.3293254850432277, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.02})
+[2026-07-28 02:01:38,364][starVLA.training.train_starvla][INFO] - Step 97, Loss: {'train/loss': 0.1823498010635376, 'train/grad_norm_pre_clip': 5.915168762207031, 'timing/data': 0.0002719671465456486, 'timing/model': 0.3305677091702819, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.02})
+[2026-07-28 02:01:42,409][starVLA.training.train_starvla][INFO] - Step 98, Loss: {'train/loss': 0.33782875537872314, 'train/grad_norm_pre_clip': 38.13237380981445, 'timing/data': 0.000278230756521225, 'timing/model': 0.33074617292732, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.02})
+[2026-07-28 02:01:46,466][starVLA.training.train_starvla][INFO] - Step 99, Loss: {'train/loss': 0.5493913292884827, 'train/grad_norm_pre_clip': 29.12552261352539, 'timing/data': 0.00027184002101421356, 'timing/model': 0.33079968160018325, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.02})
+[2026-07-28 02:01:50,523][starVLA.training.train_starvla][INFO] - Step 100, Loss: {'train/loss': 0.38316094875335693, 'train/grad_norm_pre_clip': 23.60953712463379, 'timing/data': 0.000278988853096962, 'timing/model': 0.3308973768725991, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.02})
+[2026-07-28 02:01:54,584][starVLA.training.train_starvla][INFO] - Step 101, Loss: {'train/loss': 0.1382225602865219, 'train/grad_norm_pre_clip': 5.105812072753906, 'timing/data': 0.0002902550622820854, 'timing/model': 0.3314773580059409, 'learning_rate/qwen_vl_interface': 9.999998458888285e-06, 'learning_rate/action_model': 9.999998458888285e-05, 'epoch': 0.02})
+[2026-07-28 02:01:58,647][starVLA.training.train_starvla][INFO] - Step 102, Loss: {'train/loss': 0.1993173062801361, 'train/grad_norm_pre_clip': 8.408282279968262, 'timing/data': 0.00028764596208930016, 'timing/model': 0.33152961265295744, 'learning_rate/qwen_vl_interface': 9.999993835554141e-06, 'learning_rate/action_model': 9.999993835554141e-05, 'epoch': 0.02})
+[2026-07-28 02:02:02,702][starVLA.training.train_starvla][INFO] - Step 103, Loss: {'train/loss': 0.22732135653495789, 'train/grad_norm_pre_clip': 14.743783950805664, 'timing/data': 0.00028544897213578224, 'timing/model': 0.32898847106844187, 'learning_rate/qwen_vl_interface': 9.999986130000566e-06, 'learning_rate/action_model': 9.999986130000566e-05, 'epoch': 0.02})
+[2026-07-28 02:02:06,761][starVLA.training.train_starvla][INFO] - Step 104, Loss: {'train/loss': 0.16074657440185547, 'train/grad_norm_pre_clip': 9.551880836486816, 'timing/data': 0.00028473883867263794, 'timing/model': 0.33056537294760346, 'learning_rate/qwen_vl_interface': 9.99997534223256e-06, 'learning_rate/action_model': 9.99997534223256e-05, 'epoch': 0.02})
+[2026-07-28 02:02:10,830][starVLA.training.train_starvla][INFO] - Step 105, Loss: {'train/loss': 0.14834028482437134, 'train/grad_norm_pre_clip': 11.623702049255371, 'timing/data': 0.000359369907528162, 'timing/model': 0.3307826407253742, 'learning_rate/qwen_vl_interface': 9.999961472257127e-06, 'learning_rate/action_model': 9.999961472257127e-05, 'epoch': 0.02})
+[2026-07-28 02:02:14,876][starVLA.training.train_starvla][INFO] - Step 106, Loss: {'train/loss': 0.18477612733840942, 'train/grad_norm_pre_clip': 15.694981575012207, 'timing/data': 0.00028114719316363335, 'timing/model': 0.3308285949751735, 'learning_rate/qwen_vl_interface': 9.999944520083262e-06, 'learning_rate/action_model': 9.999944520083261e-05, 'epoch': 0.02})
+[2026-07-28 02:02:18,949][starVLA.training.train_starvla][INFO] - Step 107, Loss: {'train/loss': 0.2538977265357971, 'train/grad_norm_pre_clip': 11.179375648498535, 'timing/data': 0.0002767401747405529, 'timing/model': 0.3449313333258033, 'learning_rate/qwen_vl_interface': 9.999924485721968e-06, 'learning_rate/action_model': 9.999924485721968e-05, 'epoch': 0.02})
+[2026-07-28 02:02:23,007][starVLA.training.train_starvla][INFO] - Step 108, Loss: {'train/loss': 0.20544204115867615, 'train/grad_norm_pre_clip': 7.339386463165283, 'timing/data': 0.00035370420664548874, 'timing/model': 0.33206934621557593, 'learning_rate/qwen_vl_interface': 9.999901369186244e-06, 'learning_rate/action_model': 9.999901369186244e-05, 'epoch': 0.02})
+[2026-07-28 02:02:27,061][starVLA.training.train_starvla][INFO] - Step 109, Loss: {'train/loss': 0.16933894157409668, 'train/grad_norm_pre_clip': 6.13411283493042, 'timing/data': 0.00027045002207159996, 'timing/model': 0.3306000679731369, 'learning_rate/qwen_vl_interface': 9.999875170491092e-06, 'learning_rate/action_model': 9.999875170491091e-05, 'epoch': 0.02})
+[2026-07-28 02:02:31,139][starVLA.training.train_starvla][INFO] - Step 110, Loss: {'train/loss': 0.17493095993995667, 'train/grad_norm_pre_clip': 7.1768693923950195, 'timing/data': 0.0003515887074172497, 'timing/model': 0.3302070409990847, 'learning_rate/qwen_vl_interface': 9.999845889653509e-06, 'learning_rate/action_model': 9.999845889653508e-05, 'epoch': 0.02})
+[2026-07-28 02:02:35,224][starVLA.training.train_starvla][INFO] - Step 111, Loss: {'train/loss': 0.11736992746591568, 'train/grad_norm_pre_clip': 1.4842469692230225, 'timing/data': 0.00027427123859524727, 'timing/model': 0.331306170206517, 'learning_rate/qwen_vl_interface': 9.999813526692496e-06, 'learning_rate/action_model': 9.999813526692495e-05, 'epoch': 0.02})
+[2026-07-28 02:02:39,287][starVLA.training.train_starvla][INFO] - Step 112, Loss: {'train/loss': 0.18025442957878113, 'train/grad_norm_pre_clip': 5.910248756408691, 'timing/data': 0.0006064139306545258, 'timing/model': 0.3307740590535104, 'learning_rate/qwen_vl_interface': 9.999778081629055e-06, 'learning_rate/action_model': 9.999778081629054e-05, 'epoch': 0.02})
+[2026-07-28 02:02:43,352][starVLA.training.train_starvla][INFO] - Step 113, Loss: {'train/loss': 0.1469208300113678, 'train/grad_norm_pre_clip': 10.032269477844238, 'timing/data': 0.0002726488746702671, 'timing/model': 0.33276426140218973, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.02})
+[2026-07-28 02:02:47,402][starVLA.training.train_starvla][INFO] - Step 114, Loss: {'train/loss': 0.19640393555164337, 'train/grad_norm_pre_clip': 8.401324272155762, 'timing/data': 0.0003779907710850239, 'timing/model': 0.3305110349319875, 'learning_rate/qwen_vl_interface': 9.999697945288881e-06, 'learning_rate/action_model': 9.999697945288881e-05, 'epoch': 0.02})
+[2026-07-28 02:02:51,461][starVLA.training.train_starvla][INFO] - Step 115, Loss: {'train/loss': 0.26636818051338196, 'train/grad_norm_pre_clip': 17.018783569335938, 'timing/data': 0.0002864398993551731, 'timing/model': 0.33265105867758393, 'learning_rate/qwen_vl_interface': 9.999653254064153e-06, 'learning_rate/action_model': 9.999653254064152e-05, 'epoch': 0.02})
+[2026-07-28 02:02:55,531][starVLA.training.train_starvla][INFO] - Step 116, Loss: {'train/loss': 0.29756051301956177, 'train/grad_norm_pre_clip': 19.491445541381836, 'timing/data': 0.00028455909341573715, 'timing/model': 0.3328696247190237, 'learning_rate/qwen_vl_interface': 9.99960548084099e-06, 'learning_rate/action_model': 9.99960548084099e-05, 'epoch': 0.02})
+[2026-07-28 02:02:59,645][starVLA.training.train_starvla][INFO] - Step 117, Loss: {'train/loss': 0.20190948247909546, 'train/grad_norm_pre_clip': 11.256608963012695, 'timing/data': 0.0002898252569139004, 'timing/model': 0.3312899679876864, 'learning_rate/qwen_vl_interface': 9.999554625650398e-06, 'learning_rate/action_model': 9.999554625650399e-05, 'epoch': 0.02})
+[2026-07-28 02:03:03,745][starVLA.training.train_starvla][INFO] - Step 118, Loss: {'train/loss': 0.14143604040145874, 'train/grad_norm_pre_clip': 3.4177627563476562, 'timing/data': 0.00035625696182250977, 'timing/model': 0.34180666878819466, 'learning_rate/qwen_vl_interface': 9.999500688525376e-06, 'learning_rate/action_model': 9.999500688525375e-05, 'epoch': 0.02})
+[2026-07-28 02:03:07,843][starVLA.training.train_starvla][INFO] - Step 119, Loss: {'train/loss': 0.19851791858673096, 'train/grad_norm_pre_clip': 12.092617988586426, 'timing/data': 0.0003869328647851944, 'timing/model': 0.33688311791047454, 'learning_rate/qwen_vl_interface': 9.99944366950092e-06, 'learning_rate/action_model': 9.999443669500919e-05, 'epoch': 0.02})
+[2026-07-28 02:03:11,964][starVLA.training.train_starvla][INFO] - Step 120, Loss: {'train/loss': 0.1484312117099762, 'train/grad_norm_pre_clip': 3.9966981410980225, 'timing/data': 0.0002847299911081791, 'timing/model': 0.3372010630555451, 'learning_rate/qwen_vl_interface': 9.99938356861403e-06, 'learning_rate/action_model': 9.99938356861403e-05, 'epoch': 0.02})
+[2026-07-28 02:03:16,096][starVLA.training.train_starvla][INFO] - Step 121, Loss: {'train/loss': 0.17916959524154663, 'train/grad_norm_pre_clip': 5.495428562164307, 'timing/data': 0.00029570283368229866, 'timing/model': 0.3322339067235589, 'learning_rate/qwen_vl_interface': 9.999320385903709e-06, 'learning_rate/action_model': 9.999320385903708e-05, 'epoch': 0.02})
+[2026-07-28 02:03:20,163][starVLA.training.train_starvla][INFO] - Step 122, Loss: {'train/loss': 0.1839975267648697, 'train/grad_norm_pre_clip': 7.809604644775391, 'timing/data': 0.000396686140447855, 'timing/model': 0.33248963207006454, 'learning_rate/qwen_vl_interface': 9.999254121410948e-06, 'learning_rate/action_model': 9.999254121410949e-05, 'epoch': 0.02})
+[2026-07-28 02:03:24,222][starVLA.training.train_starvla][INFO] - Step 123, Loss: {'train/loss': 0.1890394389629364, 'train/grad_norm_pre_clip': 5.9609575271606445, 'timing/data': 0.00026928214356303215, 'timing/model': 0.3351592402905226, 'learning_rate/qwen_vl_interface': 9.999184775178751e-06, 'learning_rate/action_model': 9.999184775178751e-05, 'epoch': 0.02})
+[2026-07-28 02:03:28,292][starVLA.training.train_starvla][INFO] - Step 124, Loss: {'train/loss': 0.15464046597480774, 'train/grad_norm_pre_clip': 5.618047714233398, 'timing/data': 0.00028813816606998444, 'timing/model': 0.33186135115101933, 'learning_rate/qwen_vl_interface': 9.999112347252117e-06, 'learning_rate/action_model': 9.999112347252116e-05, 'epoch': 0.02})
+[2026-07-28 02:03:32,642][starVLA.training.train_starvla][INFO] - Step 125, Loss: {'train/loss': 0.1544049084186554, 'train/grad_norm_pre_clip': 7.9981770515441895, 'timing/data': 0.000266076996922493, 'timing/model': 0.3313222290016711, 'learning_rate/qwen_vl_interface': 9.99903683767804e-06, 'learning_rate/action_model': 9.99903683767804e-05, 'epoch': 0.02})
+[2026-07-28 02:03:36,701][starVLA.training.train_starvla][INFO] - Step 126, Loss: {'train/loss': 0.12221561372280121, 'train/grad_norm_pre_clip': 2.666649580001831, 'timing/data': 0.0002928967587649822, 'timing/model': 0.3335785251110792, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.02})
+[2026-07-28 02:03:40,790][starVLA.training.train_starvla][INFO] - Step 127, Loss: {'train/loss': 0.14827585220336914, 'train/grad_norm_pre_clip': 7.078985691070557, 'timing/data': 0.0002871127799153328, 'timing/model': 0.3326559467241168, 'learning_rate/qwen_vl_interface': 9.998876573785547e-06, 'learning_rate/action_model': 9.998876573785547e-05, 'epoch': 0.02})
+[2026-07-28 02:03:44,854][starVLA.training.train_starvla][INFO] - Step 128, Loss: {'train/loss': 0.2080959975719452, 'train/grad_norm_pre_clip': 7.738907337188721, 'timing/data': 0.0002767592668533325, 'timing/model': 0.3301362879574299, 'learning_rate/qwen_vl_interface': 9.998791819571126e-06, 'learning_rate/action_model': 9.998791819571127e-05, 'epoch': 0.02})
+[2026-07-28 02:03:49,200][starVLA.training.train_starvla][INFO] - Step 129, Loss: {'train/loss': 0.1659223735332489, 'train/grad_norm_pre_clip': 10.856832504272461, 'timing/data': 0.00029896991327404976, 'timing/model': 0.33197005186229944, 'learning_rate/qwen_vl_interface': 9.998703983917252e-06, 'learning_rate/action_model': 9.998703983917252e-05, 'epoch': 0.02})
+[2026-07-28 02:03:53,250][starVLA.training.train_starvla][INFO] - Step 130, Loss: {'train/loss': 0.21828675270080566, 'train/grad_norm_pre_clip': 5.543349266052246, 'timing/data': 0.0002831472083926201, 'timing/model': 0.3316448559053242, 'learning_rate/qwen_vl_interface': 9.998613066880919e-06, 'learning_rate/action_model': 9.998613066880918e-05, 'epoch': 0.02})
+[2026-07-28 02:03:57,305][starVLA.training.train_starvla][INFO] - Step 131, Loss: {'train/loss': 0.14237630367279053, 'train/grad_norm_pre_clip': 8.82574462890625, 'timing/data': 0.0002782149240374565, 'timing/model': 0.33114844327792525, 'learning_rate/qwen_vl_interface': 9.99851906852112e-06, 'learning_rate/action_model': 9.998519068521119e-05, 'epoch': 0.02})
+[2026-07-28 02:04:01,365][starVLA.training.train_starvla][INFO] - Step 132, Loss: {'train/loss': 0.176812082529068, 'train/grad_norm_pre_clip': 3.572985887527466, 'timing/data': 0.0002926546148955822, 'timing/model': 0.33151046000421047, 'learning_rate/qwen_vl_interface': 9.998421988898851e-06, 'learning_rate/action_model': 9.998421988898851e-05, 'epoch': 0.02})
+[2026-07-28 02:04:05,451][starVLA.training.train_starvla][INFO] - Step 133, Loss: {'train/loss': 0.12116007506847382, 'train/grad_norm_pre_clip': 2.5805320739746094, 'timing/data': 0.00027273595333099365, 'timing/model': 0.33400047989562154, 'learning_rate/qwen_vl_interface': 9.998321828077107e-06, 'learning_rate/action_model': 9.998321828077106e-05, 'epoch': 0.02})
+[2026-07-28 02:04:09,522][starVLA.training.train_starvla][INFO] - Step 134, Loss: {'train/loss': 0.10346338152885437, 'train/grad_norm_pre_clip': 2.6814024448394775, 'timing/data': 0.0002784677781164646, 'timing/model': 0.33345820009708405, 'learning_rate/qwen_vl_interface': 9.99821858612088e-06, 'learning_rate/action_model': 9.99821858612088e-05, 'epoch': 0.02})
+[2026-07-28 02:04:13,602][starVLA.training.train_starvla][INFO] - Step 135, Loss: {'train/loss': 0.1437252163887024, 'train/grad_norm_pre_clip': 4.568337440490723, 'timing/data': 0.00046253763139247894, 'timing/model': 0.33183825900778174, 'learning_rate/qwen_vl_interface': 9.998112263097164e-06, 'learning_rate/action_model': 9.998112263097165e-05, 'epoch': 0.02})
+[2026-07-28 02:04:17,664][starVLA.training.train_starvla][INFO] - Step 136, Loss: {'train/loss': 0.1749119609594345, 'train/grad_norm_pre_clip': 6.446915149688721, 'timing/data': 0.0002879383973777294, 'timing/model': 0.3325125789269805, 'learning_rate/qwen_vl_interface': 9.99800285907495e-06, 'learning_rate/action_model': 9.99800285907495e-05, 'epoch': 0.02})
+[2026-07-28 02:04:21,725][starVLA.training.train_starvla][INFO] - Step 137, Loss: {'train/loss': 0.14490783214569092, 'train/grad_norm_pre_clip': 4.022778034210205, 'timing/data': 0.00028959475457668304, 'timing/model': 0.3315145540982485, 'learning_rate/qwen_vl_interface': 9.997890374125228e-06, 'learning_rate/action_model': 9.997890374125227e-05, 'epoch': 0.03})
+[2026-07-28 02:04:25,777][starVLA.training.train_starvla][INFO] - Step 138, Loss: {'train/loss': 0.2149273157119751, 'train/grad_norm_pre_clip': 6.038002014160156, 'timing/data': 0.0002907118760049343, 'timing/model': 0.3322652578353882, 'learning_rate/qwen_vl_interface': 9.997774808320992e-06, 'learning_rate/action_model': 9.997774808320992e-05, 'epoch': 0.03})
+[2026-07-28 02:04:29,870][starVLA.training.train_starvla][INFO] - Step 139, Loss: {'train/loss': 0.17583496868610382, 'train/grad_norm_pre_clip': 8.000056266784668, 'timing/data': 0.000608762726187706, 'timing/model': 0.33222005516290665, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.03})
+[2026-07-28 02:04:33,936][starVLA.training.train_starvla][INFO] - Step 140, Loss: {'train/loss': 0.1366795301437378, 'train/grad_norm_pre_clip': 3.872737407684326, 'timing/data': 0.000384393148124218, 'timing/model': 0.33130100183188915, 'learning_rate/qwen_vl_interface': 9.997534434450922e-06, 'learning_rate/action_model': 9.997534434450921e-05, 'epoch': 0.03})
+[2026-07-28 02:04:38,011][starVLA.training.train_starvla][INFO] - Step 141, Loss: {'train/loss': 0.14114537835121155, 'train/grad_norm_pre_clip': 4.390803337097168, 'timing/data': 0.0002719201147556305, 'timing/model': 0.3337217951193452, 'learning_rate/qwen_vl_interface': 9.997409626541065e-06, 'learning_rate/action_model': 9.997409626541066e-05, 'epoch': 0.03})
+[2026-07-28 02:04:42,070][starVLA.training.train_starvla][INFO] - Step 142, Loss: {'train/loss': 0.16666746139526367, 'train/grad_norm_pre_clip': 5.721374034881592, 'timing/data': 0.0002769553102552891, 'timing/model': 0.3330263220705092, 'learning_rate/qwen_vl_interface': 9.997281738088647e-06, 'learning_rate/action_model': 9.997281738088646e-05, 'epoch': 0.03})
+[2026-07-28 02:04:46,179][starVLA.training.train_starvla][INFO] - Step 143, Loss: {'train/loss': 0.09925548732280731, 'train/grad_norm_pre_clip': 4.9288129806518555, 'timing/data': 0.00029016798362135887, 'timing/model': 0.3326222742907703, 'learning_rate/qwen_vl_interface': 9.997150769176648e-06, 'learning_rate/action_model': 9.997150769176647e-05, 'epoch': 0.03})
+[2026-07-28 02:04:50,300][starVLA.training.train_starvla][INFO] - Step 144, Loss: {'train/loss': 0.11746200919151306, 'train/grad_norm_pre_clip': 2.8538384437561035, 'timing/data': 0.00027362629771232605, 'timing/model': 0.33787526935338974, 'learning_rate/qwen_vl_interface': 9.997016719890052e-06, 'learning_rate/action_model': 9.997016719890052e-05, 'epoch': 0.03})
+[2026-07-28 02:04:54,409][starVLA.training.train_starvla][INFO] - Step 145, Loss: {'train/loss': 0.1353064924478531, 'train/grad_norm_pre_clip': 3.5216431617736816, 'timing/data': 0.0002906699664890766, 'timing/model': 0.33727570716291666, 'learning_rate/qwen_vl_interface': 9.996879590315847e-06, 'learning_rate/action_model': 9.996879590315846e-05, 'epoch': 0.03})
+[2026-07-28 02:04:58,564][starVLA.training.train_starvla][INFO] - Step 146, Loss: {'train/loss': 0.11535748094320297, 'train/grad_norm_pre_clip': 2.6440157890319824, 'timing/data': 0.00028338003903627396, 'timing/model': 0.348545104265213, 'learning_rate/qwen_vl_interface': 9.99673938054301e-06, 'learning_rate/action_model': 9.996739380543009e-05, 'epoch': 0.03})
+[2026-07-28 02:05:02,638][starVLA.training.train_starvla][INFO] - Step 147, Loss: {'train/loss': 0.158670574426651, 'train/grad_norm_pre_clip': 6.348477363586426, 'timing/data': 0.00028744805604219437, 'timing/model': 0.33215389493852854, 'learning_rate/qwen_vl_interface': 9.996596090662525e-06, 'learning_rate/action_model': 9.996596090662525e-05, 'epoch': 0.03})
+[2026-07-28 02:05:06,696][starVLA.training.train_starvla][INFO] - Step 148, Loss: {'train/loss': 0.1517539769411087, 'train/grad_norm_pre_clip': 7.293130397796631, 'timing/data': 0.0002907710149884224, 'timing/model': 0.33285203902050853, 'learning_rate/qwen_vl_interface': 9.996449720767369e-06, 'learning_rate/action_model': 9.996449720767368e-05, 'epoch': 0.03})
+[2026-07-28 02:05:10,775][starVLA.training.train_starvla][INFO] - Step 149, Loss: {'train/loss': 0.14945538341999054, 'train/grad_norm_pre_clip': 9.206847190856934, 'timing/data': 0.00028630904853343964, 'timing/model': 0.33323579281568527, 'learning_rate/qwen_vl_interface': 9.99630027095252e-06, 'learning_rate/action_model': 9.99630027095252e-05, 'epoch': 0.03})
+[2026-07-28 02:05:14,849][starVLA.training.train_starvla][INFO] - Step 150, Loss: {'train/loss': 0.13568809628486633, 'train/grad_norm_pre_clip': 6.423350811004639, 'timing/data': 0.000299657229334116, 'timing/model': 0.3327207202091813, 'learning_rate/qwen_vl_interface': 9.996147741314956e-06, 'learning_rate/action_model': 9.996147741314957e-05, 'epoch': 0.03})
+[2026-07-28 02:05:18,917][starVLA.training.train_starvla][INFO] - Step 151, Loss: {'train/loss': 0.08873675763607025, 'train/grad_norm_pre_clip': 6.665724754333496, 'timing/data': 0.00039689289405941963, 'timing/model': 0.3319060178473592, 'learning_rate/qwen_vl_interface': 9.99599213195365e-06, 'learning_rate/action_model': 9.99599213195365e-05, 'epoch': 0.03})
+[2026-07-28 02:05:22,987][starVLA.training.train_starvla][INFO] - Step 152, Loss: {'train/loss': 0.15587420761585236, 'train/grad_norm_pre_clip': 4.24544620513916, 'timing/data': 0.00046391692012548447, 'timing/model': 0.33121410431340337, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.03})
+[2026-07-28 02:05:27,067][starVLA.training.train_starvla][INFO] - Step 153, Loss: {'train/loss': 0.18807581067085266, 'train/grad_norm_pre_clip': 11.816627502441406, 'timing/data': 0.00027035409584641457, 'timing/model': 0.3317185961641371, 'learning_rate/qwen_vl_interface': 9.995671674465709e-06, 'learning_rate/action_model': 9.995671674465709e-05, 'epoch': 0.03})
+[2026-07-28 02:05:31,136][starVLA.training.train_starvla][INFO] - Step 154, Loss: {'train/loss': 0.17885959148406982, 'train/grad_norm_pre_clip': 7.5199503898620605, 'timing/data': 0.000282221008092165, 'timing/model': 0.33004593290388584, 'learning_rate/qwen_vl_interface': 9.99550682654701e-06, 'learning_rate/action_model': 9.995506826547011e-05, 'epoch': 0.03})
+[2026-07-28 02:05:35,188][starVLA.training.train_starvla][INFO] - Step 155, Loss: {'train/loss': 0.17241346836090088, 'train/grad_norm_pre_clip': 4.750372409820557, 'timing/data': 0.0002743038348853588, 'timing/model': 0.3315399377606809, 'learning_rate/qwen_vl_interface': 9.995338899320457e-06, 'learning_rate/action_model': 9.995338899320456e-05, 'epoch': 0.03})
+[2026-07-28 02:05:39,239][starVLA.training.train_starvla][INFO] - Step 156, Loss: {'train/loss': 0.1705664098262787, 'train/grad_norm_pre_clip': 9.71023941040039, 'timing/data': 0.0003033997491002083, 'timing/model': 0.3315895120613277, 'learning_rate/qwen_vl_interface': 9.995167892895011e-06, 'learning_rate/action_model': 9.99516789289501e-05, 'epoch': 0.03})
+[2026-07-28 02:05:43,298][starVLA.training.train_starvla][INFO] - Step 157, Loss: {'train/loss': 0.1357884705066681, 'train/grad_norm_pre_clip': 3.55293345451355, 'timing/data': 0.00038027483969926834, 'timing/model': 0.336484638042748, 'learning_rate/qwen_vl_interface': 9.994993807381635e-06, 'learning_rate/action_model': 9.994993807381635e-05, 'epoch': 0.03})
+[2026-07-28 02:05:47,386][starVLA.training.train_starvla][INFO] - Step 158, Loss: {'train/loss': 0.1890290081501007, 'train/grad_norm_pre_clip': 11.464838027954102, 'timing/data': 0.0002743019722402096, 'timing/model': 0.33163241809234023, 'learning_rate/qwen_vl_interface': 9.994816642893294e-06, 'learning_rate/action_model': 9.994816642893293e-05, 'epoch': 0.03})
+[2026-07-28 02:05:51,490][starVLA.training.train_starvla][INFO] - Step 159, Loss: {'train/loss': 0.13115881383419037, 'train/grad_norm_pre_clip': 5.099859714508057, 'timing/data': 0.00027407705783843994, 'timing/model': 0.33132361713796854, 'learning_rate/qwen_vl_interface': 9.994636399544948e-06, 'learning_rate/action_model': 9.994636399544947e-05, 'epoch': 0.03})
+[2026-07-28 02:05:55,572][starVLA.training.train_starvla][INFO] - Step 160, Loss: {'train/loss': 0.15779590606689453, 'train/grad_norm_pre_clip': 3.6633291244506836, 'timing/data': 0.000288297887891531, 'timing/model': 0.3373846272006631, 'learning_rate/qwen_vl_interface': 9.994453077453552e-06, 'learning_rate/action_model': 9.994453077453552e-05, 'epoch': 0.03})
+[2026-07-28 02:05:59,660][starVLA.training.train_starvla][INFO] - Step 161, Loss: {'train/loss': 0.11476209759712219, 'train/grad_norm_pre_clip': 4.919507026672363, 'timing/data': 0.0003041899763047695, 'timing/model': 0.3332353918813169, 'learning_rate/qwen_vl_interface': 9.994266676738065e-06, 'learning_rate/action_model': 9.994266676738065e-05, 'epoch': 0.03})
+[2026-07-28 02:06:03,764][starVLA.training.train_starvla][INFO] - Step 162, Loss: {'train/loss': 0.2624296247959137, 'train/grad_norm_pre_clip': 10.902992248535156, 'timing/data': 0.0003021480515599251, 'timing/model': 0.3373257229104638, 'learning_rate/qwen_vl_interface': 9.99407719751944e-06, 'learning_rate/action_model': 9.994077197519439e-05, 'epoch': 0.03})
+[2026-07-28 02:06:07,851][starVLA.training.train_starvla][INFO] - Step 163, Loss: {'train/loss': 0.1801128387451172, 'train/grad_norm_pre_clip': 7.881343841552734, 'timing/data': 0.00027410220354795456, 'timing/model': 0.33888819394633174, 'learning_rate/qwen_vl_interface': 9.993884639920626e-06, 'learning_rate/action_model': 9.993884639920626e-05, 'epoch': 0.03})
+[2026-07-28 02:06:11,946][starVLA.training.train_starvla][INFO] - Step 164, Loss: {'train/loss': 0.10291536152362823, 'train/grad_norm_pre_clip': 1.9446591138839722, 'timing/data': 0.00029999110847711563, 'timing/model': 0.3345709340646863, 'learning_rate/qwen_vl_interface': 9.993689004066573e-06, 'learning_rate/action_model': 9.993689004066573e-05, 'epoch': 0.03})
+[2026-07-28 02:06:16,016][starVLA.training.train_starvla][INFO] - Step 165, Loss: {'train/loss': 0.15156839787960052, 'train/grad_norm_pre_clip': 4.7402825355529785, 'timing/data': 0.0002782018855214119, 'timing/model': 0.3314880640245974, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.03})
+[2026-07-28 02:06:20,096][starVLA.training.train_starvla][INFO] - Step 166, Loss: {'train/loss': 0.19132092595100403, 'train/grad_norm_pre_clip': 9.291820526123047, 'timing/data': 0.0006757150404155254, 'timing/model': 0.33313130075111985, 'learning_rate/qwen_vl_interface': 9.993288498102531e-06, 'learning_rate/action_model': 9.993288498102532e-05, 'epoch': 0.03})
+[2026-07-28 02:06:24,151][starVLA.training.train_starvla][INFO] - Step 167, Loss: {'train/loss': 0.20854313671588898, 'train/grad_norm_pre_clip': 10.047141075134277, 'timing/data': 0.0002868650481104851, 'timing/model': 0.3362890090793371, 'learning_rate/qwen_vl_interface': 9.993083628252424e-06, 'learning_rate/action_model': 9.993083628252424e-05, 'epoch': 0.03})
+[2026-07-28 02:06:28,215][starVLA.training.train_starvla][INFO] - Step 168, Loss: {'train/loss': 0.1370389461517334, 'train/grad_norm_pre_clip': 4.146509647369385, 'timing/data': 0.0003778459504246712, 'timing/model': 0.3345526340417564, 'learning_rate/qwen_vl_interface': 9.992875680666849e-06, 'learning_rate/action_model': 9.992875680666849e-05, 'epoch': 0.03})
+[2026-07-28 02:06:32,274][starVLA.training.train_starvla][INFO] - Step 169, Loss: {'train/loss': 0.1372397541999817, 'train/grad_norm_pre_clip': 4.426989555358887, 'timing/data': 0.0002686181105673313, 'timing/model': 0.33015470393002033, 'learning_rate/qwen_vl_interface': 9.992664655480736e-06, 'learning_rate/action_model': 9.992664655480735e-05, 'epoch': 0.03})
+[2026-07-28 02:06:36,634][starVLA.training.train_starvla][INFO] - Step 170, Loss: {'train/loss': 0.1670166552066803, 'train/grad_norm_pre_clip': 7.48771333694458, 'timing/data': 0.00028706295415759087, 'timing/model': 0.33084495225921273, 'learning_rate/qwen_vl_interface': 9.992450552831018e-06, 'learning_rate/action_model': 9.992450552831017e-05, 'epoch': 0.03})
+[2026-07-28 02:06:40,683][starVLA.training.train_starvla][INFO] - Step 171, Loss: {'train/loss': 0.12343056499958038, 'train/grad_norm_pre_clip': 4.474286079406738, 'timing/data': 0.00029567116871476173, 'timing/model': 0.331815374083817, 'learning_rate/qwen_vl_interface': 9.992233372856623e-06, 'learning_rate/action_model': 9.992233372856624e-05, 'epoch': 0.03})
+[2026-07-28 02:06:44,741][starVLA.training.train_starvla][INFO] - Step 172, Loss: {'train/loss': 0.132629856467247, 'train/grad_norm_pre_clip': 3.776144027709961, 'timing/data': 0.0002953680232167244, 'timing/model': 0.3324224161915481, 'learning_rate/qwen_vl_interface': 9.99201311569848e-06, 'learning_rate/action_model': 9.99201311569848e-05, 'epoch': 0.03})
+[2026-07-28 02:06:48,811][starVLA.training.train_starvla][INFO] - Step 173, Loss: {'train/loss': 0.12804661691188812, 'train/grad_norm_pre_clip': 2.5571563243865967, 'timing/data': 0.00028378376737236977, 'timing/model': 0.3315269579179585, 'learning_rate/qwen_vl_interface': 9.99178978149951e-06, 'learning_rate/action_model': 9.99178978149951e-05, 'epoch': 0.03})
+[2026-07-28 02:06:52,881][starVLA.training.train_starvla][INFO] - Step 174, Loss: {'train/loss': 0.09415093809366226, 'train/grad_norm_pre_clip': 3.5989041328430176, 'timing/data': 0.00029032398015260696, 'timing/model': 0.3322920990176499, 'learning_rate/qwen_vl_interface': 9.99156337040463e-06, 'learning_rate/action_model': 9.991563370404629e-05, 'epoch': 0.03})
+[2026-07-28 02:06:56,950][starVLA.training.train_starvla][INFO] - Step 175, Loss: {'train/loss': 0.10570117831230164, 'train/grad_norm_pre_clip': 2.5781466960906982, 'timing/data': 0.000365271233022213, 'timing/model': 0.3336380710825324, 'learning_rate/qwen_vl_interface': 9.991333882560759e-06, 'learning_rate/action_model': 9.991333882560759e-05, 'epoch': 0.03})
+[2026-07-28 02:07:01,302][starVLA.training.train_starvla][INFO] - Step 176, Loss: {'train/loss': 0.1808924674987793, 'train/grad_norm_pre_clip': 2.7282910346984863, 'timing/data': 0.00029272306710481644, 'timing/model': 0.33081417018547654, 'learning_rate/qwen_vl_interface': 9.991101318116806e-06, 'learning_rate/action_model': 9.991101318116806e-05, 'epoch': 0.03})
+[2026-07-28 02:07:05,384][starVLA.training.train_starvla][INFO] - Step 177, Loss: {'train/loss': 0.2019379436969757, 'train/grad_norm_pre_clip': 5.071471214294434, 'timing/data': 0.0002788756974041462, 'timing/model': 0.33982588490471244, 'learning_rate/qwen_vl_interface': 9.990865677223684e-06, 'learning_rate/action_model': 9.990865677223684e-05, 'epoch': 0.03})
+[2026-07-28 02:07:09,445][starVLA.training.train_starvla][INFO] - Step 178, Loss: {'train/loss': 0.14579932391643524, 'train/grad_norm_pre_clip': 4.5404205322265625, 'timing/data': 0.00034590112045407295, 'timing/model': 0.33186038490384817, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.03})
+[2026-07-28 02:07:13,509][starVLA.training.train_starvla][INFO] - Step 179, Loss: {'train/loss': 0.11293978989124298, 'train/grad_norm_pre_clip': 2.2114386558532715, 'timing/data': 0.0002960227429866791, 'timing/model': 0.33251668279990554, 'learning_rate/qwen_vl_interface': 9.990385166703534e-06, 'learning_rate/action_model': 9.990385166703533e-05, 'epoch': 0.03})
+[2026-07-28 02:07:17,573][starVLA.training.train_starvla][INFO] - Step 180, Loss: {'train/loss': 0.16633760929107666, 'train/grad_norm_pre_clip': 6.630430698394775, 'timing/data': 0.00028117699548602104, 'timing/model': 0.3326772218570113, 'learning_rate/qwen_vl_interface': 9.990140297388306e-06, 'learning_rate/action_model': 9.990140297388307e-05, 'epoch': 0.03})
+[2026-07-28 02:07:21,631][starVLA.training.train_starvla][INFO] - Step 181, Loss: {'train/loss': 0.1452430784702301, 'train/grad_norm_pre_clip': 14.945809364318848, 'timing/data': 0.0002959580160677433, 'timing/model': 0.3316687638871372, 'learning_rate/qwen_vl_interface': 9.989892352247505e-06, 'learning_rate/action_model': 9.989892352247504e-05, 'epoch': 0.03})
+[2026-07-28 02:07:25,689][starVLA.training.train_starvla][INFO] - Step 182, Loss: {'train/loss': 0.13350968062877655, 'train/grad_norm_pre_clip': 5.169404029846191, 'timing/data': 0.0003536129370331764, 'timing/model': 0.3306772857904434, 'learning_rate/qwen_vl_interface': 9.989641331442014e-06, 'learning_rate/action_model': 9.989641331442013e-05, 'epoch': 0.03})
+[2026-07-28 02:07:29,756][starVLA.training.train_starvla][INFO] - Step 183, Loss: {'train/loss': 0.15555882453918457, 'train/grad_norm_pre_clip': 2.8254621028900146, 'timing/data': 0.00037065334618091583, 'timing/model': 0.33346664579585195, 'learning_rate/qwen_vl_interface': 9.989387235134722e-06, 'learning_rate/action_model': 9.989387235134722e-05, 'epoch': 0.03})
+[2026-07-28 02:07:33,832][starVLA.training.train_starvla][INFO] - Step 184, Loss: {'train/loss': 0.08887245506048203, 'train/grad_norm_pre_clip': 2.4346301555633545, 'timing/data': 0.0003386051394045353, 'timing/model': 0.33212291123345494, 'learning_rate/qwen_vl_interface': 9.98913006349051e-06, 'learning_rate/action_model': 9.989130063490508e-05, 'epoch': 0.03})
+[2026-07-28 02:07:37,898][starVLA.training.train_starvla][INFO] - Step 185, Loss: {'train/loss': 0.1448580026626587, 'train/grad_norm_pre_clip': 5.498573303222656, 'timing/data': 0.0002758018672466278, 'timing/model': 0.3305294797755778, 'learning_rate/qwen_vl_interface': 9.988869816676247e-06, 'learning_rate/action_model': 9.988869816676246e-05, 'epoch': 0.03})
+[2026-07-28 02:07:41,969][starVLA.training.train_starvla][INFO] - Step 186, Loss: {'train/loss': 0.14890319108963013, 'train/grad_norm_pre_clip': 5.353523254394531, 'timing/data': 0.00037154508754611015, 'timing/model': 0.33190191304311156, 'learning_rate/qwen_vl_interface': 9.988606494860811e-06, 'learning_rate/action_model': 9.988606494860812e-05, 'epoch': 0.03})
+[2026-07-28 02:07:46,087][starVLA.training.train_starvla][INFO] - Step 187, Loss: {'train/loss': 0.21040698885917664, 'train/grad_norm_pre_clip': 4.490992546081543, 'timing/data': 0.0003504902124404907, 'timing/model': 0.331503814086318, 'learning_rate/qwen_vl_interface': 9.98834009821507e-06, 'learning_rate/action_model': 9.988340098215069e-05, 'epoch': 0.03})
+[2026-07-28 02:07:50,164][starVLA.training.train_starvla][INFO] - Step 188, Loss: {'train/loss': 0.14094969630241394, 'train/grad_norm_pre_clip': 3.8380448818206787, 'timing/data': 0.0004810201935470104, 'timing/model': 0.33650247985497117, 'learning_rate/qwen_vl_interface': 9.98807062691188e-06, 'learning_rate/action_model': 9.988070626911879e-05, 'epoch': 0.03})
+[2026-07-28 02:07:54,270][starVLA.training.train_starvla][INFO] - Step 189, Loss: {'train/loss': 0.1221824437379837, 'train/grad_norm_pre_clip': 3.7366697788238525, 'timing/data': 0.00026911264285445213, 'timing/model': 0.34753035428002477, 'learning_rate/qwen_vl_interface': 9.987798081126099e-06, 'learning_rate/action_model': 9.987798081126099e-05, 'epoch': 0.03})
+[2026-07-28 02:07:58,339][starVLA.training.train_starvla][INFO] - Step 190, Loss: {'train/loss': 0.14890429377555847, 'train/grad_norm_pre_clip': 3.4526591300964355, 'timing/data': 0.0002645389176905155, 'timing/model': 0.32962742960080504, 'learning_rate/qwen_vl_interface': 9.987522461034584e-06, 'learning_rate/action_model': 9.987522461034584e-05, 'epoch': 0.03})
+[2026-07-28 02:08:02,415][starVLA.training.train_starvla][INFO] - Step 191, Loss: {'train/loss': 0.11116774380207062, 'train/grad_norm_pre_clip': 4.541999816894531, 'timing/data': 0.0002847672440111637, 'timing/model': 0.33238503616303205, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.04})
+[2026-07-28 02:08:06,477][starVLA.training.train_starvla][INFO] - Step 192, Loss: {'train/loss': 0.10500591993331909, 'train/grad_norm_pre_clip': 3.282912015914917, 'timing/data': 0.00029230816289782524, 'timing/model': 0.3333399901166558, 'learning_rate/qwen_vl_interface': 9.98696199865172e-06, 'learning_rate/action_model': 9.98696199865172e-05, 'epoch': 0.04})
+[2026-07-28 02:08:10,546][starVLA.training.train_starvla][INFO] - Step 193, Loss: {'train/loss': 0.11297309398651123, 'train/grad_norm_pre_clip': 2.566087245941162, 'timing/data': 0.000863680150359869, 'timing/model': 0.3320859419181943, 'learning_rate/qwen_vl_interface': 9.986677156724054e-06, 'learning_rate/action_model': 9.986677156724054e-05, 'epoch': 0.04})
+[2026-07-28 02:08:14,598][starVLA.training.train_starvla][INFO] - Step 194, Loss: {'train/loss': 0.09248248487710953, 'train/grad_norm_pre_clip': 5.503769874572754, 'timing/data': 0.00028108805418014526, 'timing/model': 0.3320723297074437, 'learning_rate/qwen_vl_interface': 9.986389241218003e-06, 'learning_rate/action_model': 9.986389241218003e-05, 'epoch': 0.04})
+[2026-07-28 02:08:18,688][starVLA.training.train_starvla][INFO] - Step 195, Loss: {'train/loss': 0.21937701106071472, 'train/grad_norm_pre_clip': 11.431180953979492, 'timing/data': 0.00028728414326906204, 'timing/model': 0.33672789484262466, 'learning_rate/qwen_vl_interface': 9.986098252320397e-06, 'learning_rate/action_model': 9.986098252320397e-05, 'epoch': 0.04})
+[2026-07-28 02:08:22,773][starVLA.training.train_starvla][INFO] - Step 196, Loss: {'train/loss': 0.19018495082855225, 'train/grad_norm_pre_clip': 4.901306629180908, 'timing/data': 0.00027650734409689903, 'timing/model': 0.332672453019768, 'learning_rate/qwen_vl_interface': 9.985804190220054e-06, 'learning_rate/action_model': 9.985804190220053e-05, 'epoch': 0.04})
+[2026-07-28 02:08:26,847][starVLA.training.train_starvla][INFO] - Step 197, Loss: {'train/loss': 0.14838774502277374, 'train/grad_norm_pre_clip': 7.836228847503662, 'timing/data': 0.00028003891929984093, 'timing/model': 0.3331657010130584, 'learning_rate/qwen_vl_interface': 9.985507055107788e-06, 'learning_rate/action_model': 9.985507055107788e-05, 'epoch': 0.04})
+[2026-07-28 02:08:30,918][starVLA.training.train_starvla][INFO] - Step 198, Loss: {'train/loss': 0.15907695889472961, 'train/grad_norm_pre_clip': 3.1099209785461426, 'timing/data': 0.00029344717040657997, 'timing/model': 0.3351476793177426, 'learning_rate/qwen_vl_interface': 9.985206847176406e-06, 'learning_rate/action_model': 9.985206847176406e-05, 'epoch': 0.04})
+[2026-07-28 02:08:34,977][starVLA.training.train_starvla][INFO] - Step 199, Loss: {'train/loss': 0.13294798135757446, 'train/grad_norm_pre_clip': 3.5677623748779297, 'timing/data': 0.0002692011184990406, 'timing/model': 0.33218067372217774, 'learning_rate/qwen_vl_interface': 9.984903566620711e-06, 'learning_rate/action_model': 9.984903566620712e-05, 'epoch': 0.04})
+[2026-07-28 02:08:39,061][starVLA.training.train_starvla][INFO] - Step 200, Loss: {'train/loss': 0.17531991004943848, 'train/grad_norm_pre_clip': 3.7525298595428467, 'timing/data': 0.000295502133667469, 'timing/model': 0.33312711399048567, 'learning_rate/qwen_vl_interface': 9.984597213637497e-06, 'learning_rate/action_model': 9.984597213637497e-05, 'epoch': 0.04})
+[2026-07-28 02:08:43,138][starVLA.training.train_starvla][INFO] - Step 201, Loss: {'train/loss': 0.15067949891090393, 'train/grad_norm_pre_clip': 2.1877408027648926, 'timing/data': 0.0002808570861816406, 'timing/model': 0.3331309179775417, 'learning_rate/qwen_vl_interface': 9.984287788425555e-06, 'learning_rate/action_model': 9.984287788425555e-05, 'epoch': 0.04})
+[2026-07-28 02:08:47,209][starVLA.training.train_starvla][INFO] - Step 202, Loss: {'train/loss': 0.11652646213769913, 'train/grad_norm_pre_clip': 3.8940632343292236, 'timing/data': 0.000287666916847229, 'timing/model': 0.3311810023151338, 'learning_rate/qwen_vl_interface': 9.983975291185666e-06, 'learning_rate/action_model': 9.983975291185667e-05, 'epoch': 0.04})
+[2026-07-28 02:08:51,300][starVLA.training.train_starvla][INFO] - Step 203, Loss: {'train/loss': 0.16943703591823578, 'train/grad_norm_pre_clip': 5.187643527984619, 'timing/data': 0.00028189271688461304, 'timing/model': 0.33166950894519687, 'learning_rate/qwen_vl_interface': 9.983659722120607e-06, 'learning_rate/action_model': 9.983659722120608e-05, 'epoch': 0.04})
+[2026-07-28 02:08:55,357][starVLA.training.train_starvla][INFO] - Step 204, Loss: {'train/loss': 0.13137246668338776, 'train/grad_norm_pre_clip': 2.3811817169189453, 'timing/data': 0.0003011869266629219, 'timing/model': 0.3331874259747565, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.04})
+[2026-07-28 02:08:59,416][starVLA.training.train_starvla][INFO] - Step 205, Loss: {'train/loss': 0.13469040393829346, 'train/grad_norm_pre_clip': 1.2097305059432983, 'timing/data': 0.0002853972837328911, 'timing/model': 0.33298606611788273, 'learning_rate/qwen_vl_interface': 9.983019369336048e-06, 'learning_rate/action_model': 9.983019369336048e-05, 'epoch': 0.04})
+[2026-07-28 02:09:03,476][starVLA.training.train_starvla][INFO] - Step 206, Loss: {'train/loss': 0.1435643434524536, 'train/grad_norm_pre_clip': 4.140404224395752, 'timing/data': 0.0003033881075680256, 'timing/model': 0.33252884168177843, 'learning_rate/qwen_vl_interface': 9.982694586032066e-06, 'learning_rate/action_model': 9.982694586032065e-05, 'epoch': 0.04})
+[2026-07-28 02:09:07,537][starVLA.training.train_starvla][INFO] - Step 207, Loss: {'train/loss': 0.12903347611427307, 'train/grad_norm_pre_clip': 3.3325998783111572, 'timing/data': 0.00030324794352054596, 'timing/model': 0.3315176470205188, 'learning_rate/qwen_vl_interface': 9.982366731733948e-06, 'learning_rate/action_model': 9.982366731733947e-05, 'epoch': 0.04})
+[2026-07-28 02:09:11,586][starVLA.training.train_starvla][INFO] - Step 208, Loss: {'train/loss': 0.14407990872859955, 'train/grad_norm_pre_clip': 5.013622760772705, 'timing/data': 0.00028571533039212227, 'timing/model': 0.33077283576130867, 'learning_rate/qwen_vl_interface': 9.982035806654437e-06, 'learning_rate/action_model': 9.982035806654437e-05, 'epoch': 0.04})
+[2026-07-28 02:09:15,648][starVLA.training.train_starvla][INFO] - Step 209, Loss: {'train/loss': 0.15112707018852234, 'train/grad_norm_pre_clip': 14.601350784301758, 'timing/data': 0.0002954639494419098, 'timing/model': 0.33192831091582775, 'learning_rate/qwen_vl_interface': 9.981701811008267e-06, 'learning_rate/action_model': 9.981701811008267e-05, 'epoch': 0.04})
+[2026-07-28 02:09:19,707][starVLA.training.train_starvla][INFO] - Step 210, Loss: {'train/loss': 0.14784105122089386, 'train/grad_norm_pre_clip': 5.759326934814453, 'timing/data': 0.0002906522713601589, 'timing/model': 0.33310242695733905, 'learning_rate/qwen_vl_interface': 9.981364745012163e-06, 'learning_rate/action_model': 9.981364745012163e-05, 'epoch': 0.04})
+[2026-07-28 02:09:23,784][starVLA.training.train_starvla][INFO] - Step 211, Loss: {'train/loss': 0.13940422236919403, 'train/grad_norm_pre_clip': 2.6767454147338867, 'timing/data': 0.00028339633718132973, 'timing/model': 0.3315032352693379, 'learning_rate/qwen_vl_interface': 9.981024608884842e-06, 'learning_rate/action_model': 9.981024608884843e-05, 'epoch': 0.04})
+[2026-07-28 02:09:27,861][starVLA.training.train_starvla][INFO] - Step 212, Loss: {'train/loss': 0.09457121789455414, 'train/grad_norm_pre_clip': 3.371119737625122, 'timing/data': 0.00028839195147156715, 'timing/model': 0.3482200987637043, 'learning_rate/qwen_vl_interface': 9.980681402847017e-06, 'learning_rate/action_model': 9.980681402847016e-05, 'epoch': 0.04})
+[2026-07-28 02:09:32,191][starVLA.training.train_starvla][INFO] - Step 213, Loss: {'train/loss': 0.20036983489990234, 'train/grad_norm_pre_clip': 4.6169586181640625, 'timing/data': 0.0002861260436475277, 'timing/model': 0.5868504950776696, 'learning_rate/qwen_vl_interface': 9.980335127121391e-06, 'learning_rate/action_model': 9.98033512712139e-05, 'epoch': 0.04})
+[2026-07-28 02:09:36,266][starVLA.training.train_starvla][INFO] - Step 214, Loss: {'train/loss': 0.11907173693180084, 'train/grad_norm_pre_clip': 3.1372952461242676, 'timing/data': 0.00028599798679351807, 'timing/model': 0.33190571004524827, 'learning_rate/qwen_vl_interface': 9.979985781932658e-06, 'learning_rate/action_model': 9.979985781932657e-05, 'epoch': 0.04})
+[2026-07-28 02:09:40,338][starVLA.training.train_starvla][INFO] - Step 215, Loss: {'train/loss': 0.1735689640045166, 'train/grad_norm_pre_clip': 6.420784950256348, 'timing/data': 0.0002759872004389763, 'timing/model': 0.3320766310207546, 'learning_rate/qwen_vl_interface': 9.979633367507501e-06, 'learning_rate/action_model': 9.979633367507502e-05, 'epoch': 0.04})
+[2026-07-28 02:09:44,402][starVLA.training.train_starvla][INFO] - Step 216, Loss: {'train/loss': 0.19840475916862488, 'train/grad_norm_pre_clip': 6.920050144195557, 'timing/data': 0.0002741958014667034, 'timing/model': 0.32991263503208756, 'learning_rate/qwen_vl_interface': 9.979277884074604e-06, 'learning_rate/action_model': 9.979277884074604e-05, 'epoch': 0.04})
+[2026-07-28 02:09:48,457][starVLA.training.train_starvla][INFO] - Step 217, Loss: {'train/loss': 0.14174509048461914, 'train/grad_norm_pre_clip': 2.8937907218933105, 'timing/data': 0.0002911188639700413, 'timing/model': 0.332035340834409, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.04})
+[2026-07-28 02:09:52,513][starVLA.training.train_starvla][INFO] - Step 218, Loss: {'train/loss': 0.13714341819286346, 'train/grad_norm_pre_clip': 5.02678918838501, 'timing/data': 0.00028803199529647827, 'timing/model': 0.332376510836184, 'learning_rate/qwen_vl_interface': 9.978557711110245e-06, 'learning_rate/action_model': 9.978557711110244e-05, 'epoch': 0.04})
+[2026-07-28 02:09:56,569][starVLA.training.train_starvla][INFO] - Step 219, Loss: {'train/loss': 0.15669116377830505, 'train/grad_norm_pre_clip': 6.731563091278076, 'timing/data': 0.00030335597693920135, 'timing/model': 0.33218602277338505, 'learning_rate/qwen_vl_interface': 9.978193022046097e-06, 'learning_rate/action_model': 9.978193022046097e-05, 'epoch': 0.04})
+[2026-07-28 02:10:00,641][starVLA.training.train_starvla][INFO] - Step 220, Loss: {'train/loss': 0.1570652425289154, 'train/grad_norm_pre_clip': 3.7537708282470703, 'timing/data': 0.0007130121812224388, 'timing/model': 0.3299201070331037, 'learning_rate/qwen_vl_interface': 9.977825264908832e-06, 'learning_rate/action_model': 9.977825264908832e-05, 'epoch': 0.04})
+[2026-07-28 02:10:04,694][starVLA.training.train_starvla][INFO] - Step 221, Loss: {'train/loss': 0.12957200407981873, 'train/grad_norm_pre_clip': 4.146987438201904, 'timing/data': 0.00030045071616768837, 'timing/model': 0.3309115869924426, 'learning_rate/qwen_vl_interface': 9.97745443993708e-06, 'learning_rate/action_model': 9.97745443993708e-05, 'epoch': 0.04})
+[2026-07-28 02:10:08,747][starVLA.training.train_starvla][INFO] - Step 222, Loss: {'train/loss': 0.08211345970630646, 'train/grad_norm_pre_clip': 4.40971565246582, 'timing/data': 0.0003834557719528675, 'timing/model': 0.3314081188291311, 'learning_rate/qwen_vl_interface': 9.977080547371468e-06, 'learning_rate/action_model': 9.977080547371468e-05, 'epoch': 0.04})
+[2026-07-28 02:10:13,102][starVLA.training.train_starvla][INFO] - Step 223, Loss: {'train/loss': 0.15965917706489563, 'train/grad_norm_pre_clip': 7.598133563995361, 'timing/data': 0.00028762128204107285, 'timing/model': 0.33023190731182694, 'learning_rate/qwen_vl_interface': 9.97670358745461e-06, 'learning_rate/action_model': 9.97670358745461e-05, 'epoch': 0.04})
+[2026-07-28 02:10:17,150][starVLA.training.train_starvla][INFO] - Step 224, Loss: {'train/loss': 0.19096174836158752, 'train/grad_norm_pre_clip': 7.932681560516357, 'timing/data': 0.000294520054012537, 'timing/model': 0.33193041384220123, 'learning_rate/qwen_vl_interface': 9.976323560431112e-06, 'learning_rate/action_model': 9.976323560431111e-05, 'epoch': 0.04})
+[2026-07-28 02:10:21,237][starVLA.training.train_starvla][INFO] - Step 225, Loss: {'train/loss': 0.0755089819431305, 'train/grad_norm_pre_clip': 3.002504825592041, 'timing/data': 0.00029233191162347794, 'timing/model': 0.3316377936862409, 'learning_rate/qwen_vl_interface': 9.975940466547567e-06, 'learning_rate/action_model': 9.975940466547567e-05, 'epoch': 0.04})
+[2026-07-28 02:10:25,285][starVLA.training.train_starvla][INFO] - Step 226, Loss: {'train/loss': 0.13226290047168732, 'train/grad_norm_pre_clip': 3.1124765872955322, 'timing/data': 0.00029029231518507004, 'timing/model': 0.3302354118786752, 'learning_rate/qwen_vl_interface': 9.975554306052563e-06, 'learning_rate/action_model': 9.975554306052562e-05, 'epoch': 0.04})
+[2026-07-28 02:10:29,341][starVLA.training.train_starvla][INFO] - Step 227, Loss: {'train/loss': 0.11966779828071594, 'train/grad_norm_pre_clip': 5.694964408874512, 'timing/data': 0.00026967283338308334, 'timing/model': 0.33074934128671885, 'learning_rate/qwen_vl_interface': 9.975165079196674e-06, 'learning_rate/action_model': 9.975165079196673e-05, 'epoch': 0.04})
+[2026-07-28 02:10:33,407][starVLA.training.train_starvla][INFO] - Step 228, Loss: {'train/loss': 0.12443782389163971, 'train/grad_norm_pre_clip': 5.704428672790527, 'timing/data': 0.0003009759820997715, 'timing/model': 0.3321216949261725, 'learning_rate/qwen_vl_interface': 9.974772786232466e-06, 'learning_rate/action_model': 9.974772786232466e-05, 'epoch': 0.04})
+[2026-07-28 02:10:37,465][starVLA.training.train_starvla][INFO] - Step 229, Loss: {'train/loss': 0.13416007161140442, 'train/grad_norm_pre_clip': 2.258579969406128, 'timing/data': 0.00028595514595508575, 'timing/model': 0.33293267292901874, 'learning_rate/qwen_vl_interface': 9.974377427414493e-06, 'learning_rate/action_model': 9.974377427414493e-05, 'epoch': 0.04})
+[2026-07-28 02:10:41,522][starVLA.training.train_starvla][INFO] - Step 230, Loss: {'train/loss': 0.08559950441122055, 'train/grad_norm_pre_clip': 5.588371753692627, 'timing/data': 0.00030273105949163437, 'timing/model': 0.33229560405015945, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.04})
+[2026-07-28 02:10:45,584][starVLA.training.train_starvla][INFO] - Step 231, Loss: {'train/loss': 0.16795296967029572, 'train/grad_norm_pre_clip': 7.035068035125732, 'timing/data': 0.00044101011008024216, 'timing/model': 0.33266196912154555, 'learning_rate/qwen_vl_interface': 9.973577513245417e-06, 'learning_rate/action_model': 9.973577513245417e-05, 'epoch': 0.04})
+[2026-07-28 02:10:49,640][starVLA.training.train_starvla][INFO] - Step 232, Loss: {'train/loss': 0.16911491751670837, 'train/grad_norm_pre_clip': 6.979156017303467, 'timing/data': 0.00030068960040807724, 'timing/model': 0.33690665336325765, 'learning_rate/qwen_vl_interface': 9.973172958413372e-06, 'learning_rate/action_model': 9.97317295841337e-05, 'epoch': 0.04})
+[2026-07-28 02:10:53,717][starVLA.training.train_starvla][INFO] - Step 233, Loss: {'train/loss': 0.09129677712917328, 'train/grad_norm_pre_clip': 3.5525472164154053, 'timing/data': 0.0002909158356487751, 'timing/model': 0.33229197980836034, 'learning_rate/qwen_vl_interface': 9.97276533876567e-06, 'learning_rate/action_model': 9.97276533876567e-05, 'epoch': 0.04})
+[2026-07-28 02:10:57,789][starVLA.training.train_starvla][INFO] - Step 234, Loss: {'train/loss': 0.13819411396980286, 'train/grad_norm_pre_clip': 3.548116445541382, 'timing/data': 0.00028688693419098854, 'timing/model': 0.3333024582825601, 'learning_rate/qwen_vl_interface': 9.972354654566816e-06, 'learning_rate/action_model': 9.972354654566816e-05, 'epoch': 0.04})
+[2026-07-28 02:11:02,028][starVLA.training.train_starvla][INFO] - Step 235, Loss: {'train/loss': 0.07918021082878113, 'train/grad_norm_pre_clip': 1.6930407285690308, 'timing/data': 0.00041285203769803047, 'timing/model': 0.3341004103422165, 'learning_rate/qwen_vl_interface': 9.971940906083295e-06, 'learning_rate/action_model': 9.971940906083295e-05, 'epoch': 0.04})
+[2026-07-28 02:11:06,183][starVLA.training.train_starvla][INFO] - Step 236, Loss: {'train/loss': 0.1461586356163025, 'train/grad_norm_pre_clip': 4.444005012512207, 'timing/data': 0.0002939789555966854, 'timing/model': 0.3304494530893862, 'learning_rate/qwen_vl_interface': 9.971524093583586e-06, 'learning_rate/action_model': 9.971524093583586e-05, 'epoch': 0.04})
+[2026-07-28 02:11:10,272][starVLA.training.train_starvla][INFO] - Step 237, Loss: {'train/loss': 0.18695992231369019, 'train/grad_norm_pre_clip': 10.013982772827148, 'timing/data': 0.000286902766674757, 'timing/model': 0.3314102990552783, 'learning_rate/qwen_vl_interface': 9.971104217338154e-06, 'learning_rate/action_model': 9.971104217338153e-05, 'epoch': 0.04})
+[2026-07-28 02:11:14,331][starVLA.training.train_starvla][INFO] - Step 238, Loss: {'train/loss': 0.15022751688957214, 'train/grad_norm_pre_clip': 6.121479511260986, 'timing/data': 0.0002739978954195976, 'timing/model': 0.3324960032477975, 'learning_rate/qwen_vl_interface': 9.970681277619452e-06, 'learning_rate/action_model': 9.970681277619452e-05, 'epoch': 0.04})
+[2026-07-28 02:11:18,388][starVLA.training.train_starvla][INFO] - Step 239, Loss: {'train/loss': 0.10711467266082764, 'train/grad_norm_pre_clip': 3.2442514896392822, 'timing/data': 0.00028368039056658745, 'timing/model': 0.33084478694945574, 'learning_rate/qwen_vl_interface': 9.97025527470192e-06, 'learning_rate/action_model': 9.970255274701919e-05, 'epoch': 0.04})
+[2026-07-28 02:11:22,466][starVLA.training.train_starvla][INFO] - Step 240, Loss: {'train/loss': 0.09952862560749054, 'train/grad_norm_pre_clip': 2.3966755867004395, 'timing/data': 0.00026257894933223724, 'timing/model': 0.33165688905864954, 'learning_rate/qwen_vl_interface': 9.969826208861987e-06, 'learning_rate/action_model': 9.969826208861987e-05, 'epoch': 0.04})
+[2026-07-28 02:11:26,526][starVLA.training.train_starvla][INFO] - Step 241, Loss: {'train/loss': 0.12768866121768951, 'train/grad_norm_pre_clip': 4.668665409088135, 'timing/data': 0.0003048270009458065, 'timing/model': 0.3305072784423828, 'learning_rate/qwen_vl_interface': 9.96939408037807e-06, 'learning_rate/action_model': 9.96939408037807e-05, 'epoch': 0.04})
+[2026-07-28 02:11:30,604][starVLA.training.train_starvla][INFO] - Step 242, Loss: {'train/loss': 0.11732025444507599, 'train/grad_norm_pre_clip': 6.025201797485352, 'timing/data': 0.00030889688059687614, 'timing/model': 0.33329080417752266, 'learning_rate/qwen_vl_interface': 9.968958889530573e-06, 'learning_rate/action_model': 9.968958889530573e-05, 'epoch': 0.04})
+[2026-07-28 02:11:34,683][starVLA.training.train_starvla][INFO] - Step 243, Loss: {'train/loss': 0.1336061805486679, 'train/grad_norm_pre_clip': 2.8995347023010254, 'timing/data': 0.00027632666751742363, 'timing/model': 0.33329584263265133, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.04})
+[2026-07-28 02:11:38,751][starVLA.training.train_starvla][INFO] - Step 244, Loss: {'train/loss': 0.12714126706123352, 'train/grad_norm_pre_clip': 3.561408519744873, 'timing/data': 0.00029749562963843346, 'timing/model': 0.3303518882021308, 'learning_rate/qwen_vl_interface': 9.968079321876382e-06, 'learning_rate/action_model': 9.968079321876382e-05, 'epoch': 0.04})
+[2026-07-28 02:11:42,805][starVLA.training.train_starvla][INFO] - Step 245, Loss: {'train/loss': 0.1611814796924591, 'train/grad_norm_pre_clip': 4.416878700256348, 'timing/data': 0.000288376584649086, 'timing/model': 0.33340215822681785, 'learning_rate/qwen_vl_interface': 9.967634945640432e-06, 'learning_rate/action_model': 9.967634945640431e-05, 'epoch': 0.05})
+[2026-07-28 02:11:46,860][starVLA.training.train_starvla][INFO] - Step 246, Loss: {'train/loss': 0.12647327780723572, 'train/grad_norm_pre_clip': 3.2893218994140625, 'timing/data': 0.00028673699125647545, 'timing/model': 0.3311906922608614, 'learning_rate/qwen_vl_interface': 9.967187508182385e-06, 'learning_rate/action_model': 9.967187508182385e-05, 'epoch': 0.05})
+[2026-07-28 02:11:50,918][starVLA.training.train_starvla][INFO] - Step 247, Loss: {'train/loss': 0.10975527763366699, 'train/grad_norm_pre_clip': 2.421558141708374, 'timing/data': 0.0007621701806783676, 'timing/model': 0.3356122858822346, 'learning_rate/qwen_vl_interface': 9.966737009792577e-06, 'learning_rate/action_model': 9.966737009792577e-05, 'epoch': 0.05})
+[2026-07-28 02:11:55,022][starVLA.training.train_starvla][INFO] - Step 248, Loss: {'train/loss': 0.11548672616481781, 'train/grad_norm_pre_clip': 1.9186837673187256, 'timing/data': 0.0002720346674323082, 'timing/model': 0.32914251321926713, 'learning_rate/qwen_vl_interface': 9.966283450763333e-06, 'learning_rate/action_model': 9.966283450763333e-05, 'epoch': 0.05})
+[2026-07-28 02:11:59,134][starVLA.training.train_starvla][INFO] - Step 249, Loss: {'train/loss': 0.19723419845104218, 'train/grad_norm_pre_clip': 5.73639440536499, 'timing/data': 0.0002942439168691635, 'timing/model': 0.33170580910518765, 'learning_rate/qwen_vl_interface': 9.965826831388962e-06, 'learning_rate/action_model': 9.965826831388961e-05, 'epoch': 0.05})
+[2026-07-28 02:12:03,184][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 250
+[2026-07-28 02:13:14,994][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 250
+[2026-07-28 02:14:29,399][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 250
+[2026-07-28 02:14:29,400][starVLA.training.train_starvla][INFO] - Step 250, Loss: {'train/loss': 0.11521629989147186, 'train/grad_norm_pre_clip': 2.345209836959839, 'eval/action_loss/seconds': 71.80595580302179, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.12103698402643204, 'eval/latency_3/loss': 0.12103698402643204, 'eval/flappy/loss': 0.12103698402643204, 'eval/flappy/latency_3/loss': 0.12103698402643204, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 43.395000534877184, 'rl_games_eval/mid_train/mean_length': 368.45, 'rl_games_eval/mid_train/std_reward': 34.15739308424012, 'rl_games_eval/mid_train/std_length': 275.62700792919406, 'rl_games_eval/mid_train/macro_mean_reward': 43.395000534877184, 'rl_games_eval/mid_train/macro_mean_length': 368.45, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 43.395000534877184, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 368.45, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 34.15739308424012, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 275.62700792919406, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 43.395000534877184, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00035878783091902733, 'timing/model': 0.3307340289466083, 'learning_rate/qwen_vl_interface': 9.965367151965757e-06, 'learning_rate/action_model': 9.965367151965756e-05, 'epoch': 0.05})
+[2026-07-28 02:14:35,849][starVLA.training.train_starvla][INFO] - Step 251, Loss: {'train/loss': 0.11636252701282501, 'train/grad_norm_pre_clip': 3.535496234893799, 'timing/data': 0.00028623221442103386, 'timing/model': 0.6442784471437335, 'learning_rate/qwen_vl_interface': 9.964904412792002e-06, 'learning_rate/action_model': 9.964904412792002e-05, 'epoch': 0.05})
+[2026-07-28 02:14:41,729][starVLA.training.train_starvla][INFO] - Step 252, Loss: {'train/loss': 0.09113067388534546, 'train/grad_norm_pre_clip': 2.1826820373535156, 'timing/data': 0.00027101999148726463, 'timing/model': 0.4126845118589699, 'learning_rate/qwen_vl_interface': 9.964438614167964e-06, 'learning_rate/action_model': 9.964438614167964e-05, 'epoch': 0.05})
+[2026-07-28 02:14:47,606][starVLA.training.train_starvla][INFO] - Step 253, Loss: {'train/loss': 0.09930369257926941, 'train/grad_norm_pre_clip': 3.095297336578369, 'timing/data': 0.0002648881636559963, 'timing/model': 0.41382026113569736, 'learning_rate/qwen_vl_interface': 9.963969756395892e-06, 'learning_rate/action_model': 9.96396975639589e-05, 'epoch': 0.05})
+[2026-07-28 02:14:53,502][starVLA.training.train_starvla][INFO] - Step 254, Loss: {'train/loss': 0.1860489547252655, 'train/grad_norm_pre_clip': 5.7377824783325195, 'timing/data': 0.0002941782586276531, 'timing/model': 0.4222773318178952, 'learning_rate/qwen_vl_interface': 9.963497839780025e-06, 'learning_rate/action_model': 9.963497839780024e-05, 'epoch': 0.05})
+[2026-07-28 02:14:59,200][starVLA.training.train_starvla][INFO] - Step 255, Loss: {'train/loss': 0.06438697129487991, 'train/grad_norm_pre_clip': 1.5264068841934204, 'timing/data': 0.00029243389144539833, 'timing/model': 0.4209192506968975, 'learning_rate/qwen_vl_interface': 9.963022864626582e-06, 'learning_rate/action_model': 9.963022864626581e-05, 'epoch': 0.05})
+[2026-07-28 02:15:05,140][starVLA.training.train_starvla][INFO] - Step 256, Loss: {'train/loss': 0.1547044813632965, 'train/grad_norm_pre_clip': 5.828305721282959, 'timing/data': 0.00027817394584417343, 'timing/model': 0.41630358900874853, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.05})
+[2026-07-28 02:15:11,350][starVLA.training.train_starvla][INFO] - Step 257, Loss: {'train/loss': 0.14653700590133667, 'train/grad_norm_pre_clip': 4.057582378387451, 'timing/data': 0.0002540731802582741, 'timing/model': 0.46702622110024095, 'learning_rate/qwen_vl_interface': 9.962063739941783e-06, 'learning_rate/action_model': 9.962063739941783e-05, 'epoch': 0.05})
+[2026-07-28 02:15:17,259][starVLA.training.train_starvla][INFO] - Step 258, Loss: {'train/loss': 0.10900861024856567, 'train/grad_norm_pre_clip': 4.117964267730713, 'timing/data': 0.00025929510593414307, 'timing/model': 0.417811315972358, 'learning_rate/qwen_vl_interface': 9.961579591032791e-06, 'learning_rate/action_model': 9.96157959103279e-05, 'epoch': 0.05})
+[2026-07-28 02:15:22,936][starVLA.training.train_starvla][INFO] - Step 259, Loss: {'train/loss': 0.1295580267906189, 'train/grad_norm_pre_clip': 4.7617387771606445, 'timing/data': 0.0002652718685567379, 'timing/model': 0.41427113907411695, 'learning_rate/qwen_vl_interface': 9.961092384830956e-06, 'learning_rate/action_model': 9.961092384830956e-05, 'epoch': 0.05})
+[2026-07-28 02:15:28,845][starVLA.training.train_starvla][INFO] - Step 260, Loss: {'train/loss': 0.17343804240226746, 'train/grad_norm_pre_clip': 5.194177150726318, 'timing/data': 0.00028920313343405724, 'timing/model': 0.4166573849506676, 'learning_rate/qwen_vl_interface': 9.96060212165242e-06, 'learning_rate/action_model': 9.96060212165242e-05, 'epoch': 0.05})
+[2026-07-28 02:15:35,178][starVLA.training.train_starvla][INFO] - Step 261, Loss: {'train/loss': 0.13480165600776672, 'train/grad_norm_pre_clip': 4.0251665115356445, 'timing/data': 0.00026055285707116127, 'timing/model': 0.6483256951905787, 'learning_rate/qwen_vl_interface': 9.96010880181531e-06, 'learning_rate/action_model': 9.960108801815309e-05, 'epoch': 0.05})
+[2026-07-28 02:15:41,045][starVLA.training.train_starvla][INFO] - Step 262, Loss: {'train/loss': 0.13195843994617462, 'train/grad_norm_pre_clip': 3.3355844020843506, 'timing/data': 0.00037722522392868996, 'timing/model': 0.4155211318284273, 'learning_rate/qwen_vl_interface': 9.959612425639736e-06, 'learning_rate/action_model': 9.959612425639735e-05, 'epoch': 0.05})
+[2026-07-28 02:15:46,969][starVLA.training.train_starvla][INFO] - Step 263, Loss: {'train/loss': 0.1194547489285469, 'train/grad_norm_pre_clip': 3.753742218017578, 'timing/data': 0.0003742780536413193, 'timing/model': 0.41826124070212245, 'learning_rate/qwen_vl_interface': 9.95911299344779e-06, 'learning_rate/action_model': 9.959112993447789e-05, 'epoch': 0.05})
+[2026-07-28 02:15:53,136][starVLA.training.train_starvla][INFO] - Step 264, Loss: {'train/loss': 0.133554145693779, 'train/grad_norm_pre_clip': 3.925992250442505, 'timing/data': 0.0003629331476986408, 'timing/model': 0.6561281969770789, 'learning_rate/qwen_vl_interface': 9.958610505563546e-06, 'learning_rate/action_model': 9.958610505563546e-05, 'epoch': 0.05})
+[2026-07-28 02:15:59,326][starVLA.training.train_starvla][INFO] - Step 265, Loss: {'train/loss': 0.12170127779245377, 'train/grad_norm_pre_clip': 2.5822014808654785, 'timing/data': 0.00025538494810462, 'timing/model': 0.414100157096982, 'learning_rate/qwen_vl_interface': 9.95810496231307e-06, 'learning_rate/action_model': 9.95810496231307e-05, 'epoch': 0.05})
+[2026-07-28 02:16:05,233][starVLA.training.train_starvla][INFO] - Step 266, Loss: {'train/loss': 0.08480489999055862, 'train/grad_norm_pre_clip': 1.7003566026687622, 'timing/data': 0.00036339694634079933, 'timing/model': 0.41579227801412344, 'learning_rate/qwen_vl_interface': 9.957596364024396e-06, 'learning_rate/action_model': 9.957596364024396e-05, 'epoch': 0.05})
+[2026-07-28 02:16:11,106][starVLA.training.train_starvla][INFO] - Step 267, Loss: {'train/loss': 0.12303043901920319, 'train/grad_norm_pre_clip': 3.268399953842163, 'timing/data': 0.0003441115841269493, 'timing/model': 0.41509357327595353, 'learning_rate/qwen_vl_interface': 9.957084711027553e-06, 'learning_rate/action_model': 9.957084711027552e-05, 'epoch': 0.05})
+[2026-07-28 02:16:17,216][starVLA.training.train_starvla][INFO] - Step 268, Loss: {'train/loss': 0.07783885300159454, 'train/grad_norm_pre_clip': 3.301135540008545, 'timing/data': 0.00026268837973475456, 'timing/model': 0.41418080497533083, 'learning_rate/qwen_vl_interface': 9.956570003654544e-06, 'learning_rate/action_model': 9.956570003654544e-05, 'epoch': 0.05})
+[2026-07-28 02:16:23,336][starVLA.training.train_starvla][INFO] - Step 269, Loss: {'train/loss': 0.12657158076763153, 'train/grad_norm_pre_clip': 3.523197650909424, 'timing/data': 0.0003424868918955326, 'timing/model': 0.4120238162577152, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.05})
+[2026-07-28 02:16:29,206][starVLA.training.train_starvla][INFO] - Step 270, Loss: {'train/loss': 0.12854629755020142, 'train/grad_norm_pre_clip': 6.191515922546387, 'timing/data': 0.00033440813422203064, 'timing/model': 0.4151768400333822, 'learning_rate/qwen_vl_interface': 9.955531427117964e-06, 'learning_rate/action_model': 9.955531427117964e-05, 'epoch': 0.05})
+[2026-07-28 02:16:35,300][starVLA.training.train_starvla][INFO] - Step 271, Loss: {'train/loss': 0.09310311079025269, 'train/grad_norm_pre_clip': 2.586273431777954, 'timing/data': 0.00028580985963344574, 'timing/model': 0.4151760577224195, 'learning_rate/qwen_vl_interface': 9.955007558628316e-06, 'learning_rate/action_model': 9.955007558628316e-05, 'epoch': 0.05})
+[2026-07-28 02:16:41,384][starVLA.training.train_starvla][INFO] - Step 272, Loss: {'train/loss': 0.09360289573669434, 'train/grad_norm_pre_clip': 4.178228855133057, 'timing/data': 0.000347641296684742, 'timing/model': 0.6523900008760393, 'learning_rate/qwen_vl_interface': 9.954480637110346e-06, 'learning_rate/action_model': 9.954480637110344e-05, 'epoch': 0.05})
+[2026-07-28 02:16:47,325][starVLA.training.train_starvla][INFO] - Step 273, Loss: {'train/loss': 0.24001674354076385, 'train/grad_norm_pre_clip': 11.806707382202148, 'timing/data': 0.0002907668240368366, 'timing/model': 0.4168231748044491, 'learning_rate/qwen_vl_interface': 9.953950662905963e-06, 'learning_rate/action_model': 9.953950662905964e-05, 'epoch': 0.05})
+[2026-07-28 02:16:53,460][starVLA.training.train_starvla][INFO] - Step 274, Loss: {'train/loss': 0.15035206079483032, 'train/grad_norm_pre_clip': 7.770575046539307, 'timing/data': 0.0003913366235792637, 'timing/model': 0.5958353360183537, 'learning_rate/qwen_vl_interface': 9.953417636359067e-06, 'learning_rate/action_model': 9.953417636359067e-05, 'epoch': 0.05})
+[2026-07-28 02:16:59,345][starVLA.training.train_starvla][INFO] - Step 275, Loss: {'train/loss': 0.08675924688577652, 'train/grad_norm_pre_clip': 3.502577304840088, 'timing/data': 0.0003622532822191715, 'timing/model': 0.4174640872515738, 'learning_rate/qwen_vl_interface': 9.95288155781553e-06, 'learning_rate/action_model': 9.95288155781553e-05, 'epoch': 0.05})
+[2026-07-28 02:17:05,263][starVLA.training.train_starvla][INFO] - Step 276, Loss: {'train/loss': 0.12423748522996902, 'train/grad_norm_pre_clip': 5.846474647521973, 'timing/data': 0.00034736795350909233, 'timing/model': 0.4156125490553677, 'learning_rate/qwen_vl_interface': 9.952342427623209e-06, 'learning_rate/action_model': 9.952342427623208e-05, 'epoch': 0.05})
+[2026-07-28 02:17:11,536][starVLA.training.train_starvla][INFO] - Step 277, Loss: {'train/loss': 0.10963626205921173, 'train/grad_norm_pre_clip': 3.53210186958313, 'timing/data': 0.0002781120128929615, 'timing/model': 0.65404262393713, 'learning_rate/qwen_vl_interface': 9.95180024613194e-06, 'learning_rate/action_model': 9.951800246131939e-05, 'epoch': 0.05})
+[2026-07-28 02:17:17,598][starVLA.training.train_starvla][INFO] - Step 278, Loss: {'train/loss': 0.11195814609527588, 'train/grad_norm_pre_clip': 3.8504045009613037, 'timing/data': 0.0002589798532426357, 'timing/model': 0.41442670300602913, 'learning_rate/qwen_vl_interface': 9.951255013693537e-06, 'learning_rate/action_model': 9.951255013693536e-05, 'epoch': 0.05})
+[2026-07-28 02:17:23,464][starVLA.training.train_starvla][INFO] - Step 279, Loss: {'train/loss': 0.10980868339538574, 'train/grad_norm_pre_clip': 3.6450586318969727, 'timing/data': 0.00027931900694966316, 'timing/model': 0.415787850972265, 'learning_rate/qwen_vl_interface': 9.950706730661797e-06, 'learning_rate/action_model': 9.950706730661797e-05, 'epoch': 0.05})
+[2026-07-28 02:17:29,349][starVLA.training.train_starvla][INFO] - Step 280, Loss: {'train/loss': 0.0840722918510437, 'train/grad_norm_pre_clip': 3.3732335567474365, 'timing/data': 0.0002689110115170479, 'timing/model': 0.41716993087902665, 'learning_rate/qwen_vl_interface': 9.950155397392494e-06, 'learning_rate/action_model': 9.950155397392493e-05, 'epoch': 0.05})
+[2026-07-28 02:17:35,202][starVLA.training.train_starvla][INFO] - Step 281, Loss: {'train/loss': 0.11123407632112503, 'train/grad_norm_pre_clip': 2.4586398601531982, 'timing/data': 0.00028059864416718483, 'timing/model': 0.4167681219987571, 'learning_rate/qwen_vl_interface': 9.949601014243382e-06, 'learning_rate/action_model': 9.949601014243382e-05, 'epoch': 0.05})
+[2026-07-28 02:17:41,378][starVLA.training.train_starvla][INFO] - Step 282, Loss: {'train/loss': 0.08309891819953918, 'train/grad_norm_pre_clip': 3.880082368850708, 'timing/data': 0.0002933642826974392, 'timing/model': 0.6475274008698761, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.05})
+[2026-07-28 02:17:47,589][starVLA.training.train_starvla][INFO] - Step 283, Loss: {'train/loss': 0.13139808177947998, 'train/grad_norm_pre_clip': 3.010976791381836, 'timing/data': 0.0002693808637559414, 'timing/model': 0.6617120732553303, 'learning_rate/qwen_vl_interface': 9.948483099746645e-06, 'learning_rate/action_model': 9.948483099746645e-05, 'epoch': 0.05})
+[2026-07-28 02:17:53,518][starVLA.training.train_starvla][INFO] - Step 284, Loss: {'train/loss': 0.18973389267921448, 'train/grad_norm_pre_clip': 6.292774677276611, 'timing/data': 0.0002668900415301323, 'timing/model': 0.41330302413553, 'learning_rate/qwen_vl_interface': 9.947919569124423e-06, 'learning_rate/action_model': 9.947919569124422e-05, 'epoch': 0.05})
+[2026-07-28 02:17:59,491][starVLA.training.train_starvla][INFO] - Step 285, Loss: {'train/loss': 0.12847119569778442, 'train/grad_norm_pre_clip': 2.847395658493042, 'timing/data': 0.00026890961453318596, 'timing/model': 0.41666919086128473, 'learning_rate/qwen_vl_interface': 9.947352990073196e-06, 'learning_rate/action_model': 9.947352990073195e-05, 'epoch': 0.05})
+[2026-07-28 02:18:05,396][starVLA.training.train_starvla][INFO] - Step 286, Loss: {'train/loss': 0.12913942337036133, 'train/grad_norm_pre_clip': 2.8446621894836426, 'timing/data': 0.00036400509998202324, 'timing/model': 0.41597283026203513, 'learning_rate/qwen_vl_interface': 9.946783362960612e-06, 'learning_rate/action_model': 9.946783362960612e-05, 'epoch': 0.05})
+[2026-07-28 02:18:11,285][starVLA.training.train_starvla][INFO] - Step 287, Loss: {'train/loss': 0.10544461756944656, 'train/grad_norm_pre_clip': 2.880591630935669, 'timing/data': 0.0002799299545586109, 'timing/model': 0.4149922551587224, 'learning_rate/qwen_vl_interface': 9.946210688156297e-06, 'learning_rate/action_model': 9.946210688156297e-05, 'epoch': 0.05})
+[2026-07-28 02:18:17,425][starVLA.training.train_starvla][INFO] - Step 288, Loss: {'train/loss': 0.14493660628795624, 'train/grad_norm_pre_clip': 3.7137796878814697, 'timing/data': 0.0002606939524412155, 'timing/model': 0.41450073989108205, 'learning_rate/qwen_vl_interface': 9.945634966031852e-06, 'learning_rate/action_model': 9.945634966031851e-05, 'epoch': 0.05})
+[2026-07-28 02:18:23,083][starVLA.training.train_starvla][INFO] - Step 289, Loss: {'train/loss': 0.12083644419908524, 'train/grad_norm_pre_clip': 4.973198890686035, 'timing/data': 0.00028028665110468864, 'timing/model': 0.41633614897727966, 'learning_rate/qwen_vl_interface': 9.945056196960858e-06, 'learning_rate/action_model': 9.945056196960858e-05, 'epoch': 0.05})
+[2026-07-28 02:18:28,788][starVLA.training.train_starvla][INFO] - Step 290, Loss: {'train/loss': 0.12317678332328796, 'train/grad_norm_pre_clip': 2.1073172092437744, 'timing/data': 0.00029369164258241653, 'timing/model': 0.41779075004160404, 'learning_rate/qwen_vl_interface': 9.944474381318871e-06, 'learning_rate/action_model': 9.94447438131887e-05, 'epoch': 0.05})
+[2026-07-28 02:18:34,684][starVLA.training.train_starvla][INFO] - Step 291, Loss: {'train/loss': 0.15161675214767456, 'train/grad_norm_pre_clip': 4.516184329986572, 'timing/data': 0.000398057047277689, 'timing/model': 0.41602744767442346, 'learning_rate/qwen_vl_interface': 9.943889519483424e-06, 'learning_rate/action_model': 9.943889519483424e-05, 'epoch': 0.05})
+[2026-07-28 02:18:40,861][starVLA.training.train_starvla][INFO] - Step 292, Loss: {'train/loss': 0.18700474500656128, 'train/grad_norm_pre_clip': 5.573974609375, 'timing/data': 0.00025294581428170204, 'timing/model': 0.4130070782266557, 'learning_rate/qwen_vl_interface': 9.94330161183403e-06, 'learning_rate/action_model': 9.94330161183403e-05, 'epoch': 0.05})
+[2026-07-28 02:18:46,542][starVLA.training.train_starvla][INFO] - Step 293, Loss: {'train/loss': 0.12832412123680115, 'train/grad_norm_pre_clip': 4.82210111618042, 'timing/data': 0.00029073096811771393, 'timing/model': 0.4157566740177572, 'learning_rate/qwen_vl_interface': 9.942710658752175e-06, 'learning_rate/action_model': 9.942710658752174e-05, 'epoch': 0.05})
+[2026-07-28 02:18:52,437][starVLA.training.train_starvla][INFO] - Step 294, Loss: {'train/loss': 0.0817892998456955, 'train/grad_norm_pre_clip': 1.3692035675048828, 'timing/data': 0.0002877330407500267, 'timing/model': 0.41720230504870415, 'learning_rate/qwen_vl_interface': 9.942116660621319e-06, 'learning_rate/action_model': 9.94211666062132e-05, 'epoch': 0.05})
+[2026-07-28 02:18:58,570][starVLA.training.train_starvla][INFO] - Step 295, Loss: {'train/loss': 0.09378834068775177, 'train/grad_norm_pre_clip': 3.022937297821045, 'timing/data': 0.00028340425342321396, 'timing/model': 0.6695738448761404, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.05})
+[2026-07-28 02:19:04,406][starVLA.training.train_starvla][INFO] - Step 296, Loss: {'train/loss': 0.11962775141000748, 'train/grad_norm_pre_clip': 4.797567367553711, 'timing/data': 0.00029006972908973694, 'timing/model': 0.41625814978033304, 'learning_rate/qwen_vl_interface': 9.940919530756347e-06, 'learning_rate/action_model': 9.940919530756347e-05, 'epoch': 0.05})
+[2026-07-28 02:19:10,518][starVLA.training.train_starvla][INFO] - Step 297, Loss: {'train/loss': 0.12340535968542099, 'train/grad_norm_pre_clip': 2.9435198307037354, 'timing/data': 0.00026804208755493164, 'timing/model': 0.6525586689822376, 'learning_rate/qwen_vl_interface': 9.940316399799032e-06, 'learning_rate/action_model': 9.940316399799033e-05, 'epoch': 0.05})
+[2026-07-28 02:19:16,399][starVLA.training.train_starvla][INFO] - Step 298, Loss: {'train/loss': 0.09473970532417297, 'train/grad_norm_pre_clip': 1.5665693283081055, 'timing/data': 0.0002662278711795807, 'timing/model': 0.4167392533272505, 'learning_rate/qwen_vl_interface': 9.939710225346329e-06, 'learning_rate/action_model': 9.939710225346329e-05, 'epoch': 0.05})
+[2026-07-28 02:19:22,516][starVLA.training.train_starvla][INFO] - Step 299, Loss: {'train/loss': 0.1534596085548401, 'train/grad_norm_pre_clip': 2.8233680725097656, 'timing/data': 0.000341558363288641, 'timing/model': 0.41404922725632787, 'learning_rate/qwen_vl_interface': 9.939101007791575e-06, 'learning_rate/action_model': 9.939101007791575e-05, 'epoch': 0.05})
+[2026-07-28 02:19:28,173][starVLA.training.train_starvla][INFO] - Step 300, Loss: {'train/loss': 0.16000986099243164, 'train/grad_norm_pre_clip': 1.657281517982483, 'timing/data': 0.00035588908940553665, 'timing/model': 0.419338540174067, 'learning_rate/qwen_vl_interface': 9.938488747530087e-06, 'learning_rate/action_model': 9.938488747530086e-05, 'epoch': 0.06})
+[2026-07-28 02:19:34,120][starVLA.training.train_starvla][INFO] - Step 301, Loss: {'train/loss': 0.14830204844474792, 'train/grad_norm_pre_clip': 6.159046649932861, 'timing/data': 0.00043596699833869934, 'timing/model': 0.41893757740035653, 'learning_rate/qwen_vl_interface': 9.937873444959151e-06, 'learning_rate/action_model': 9.937873444959151e-05, 'epoch': 0.06})
+[2026-07-28 02:19:40,022][starVLA.training.train_starvla][INFO] - Step 302, Loss: {'train/loss': 0.1322929859161377, 'train/grad_norm_pre_clip': 2.827134609222412, 'timing/data': 0.0003022383898496628, 'timing/model': 0.416078879032284, 'learning_rate/qwen_vl_interface': 9.937255100478036e-06, 'learning_rate/action_model': 9.937255100478036e-05, 'epoch': 0.06})
+[2026-07-28 02:19:46,205][starVLA.training.train_starvla][INFO] - Step 303, Loss: {'train/loss': 0.17279404401779175, 'train/grad_norm_pre_clip': 3.4617583751678467, 'timing/data': 0.0002791578881442547, 'timing/model': 0.7052791179157794, 'learning_rate/qwen_vl_interface': 9.93663371448797e-06, 'learning_rate/action_model': 9.93663371448797e-05, 'epoch': 0.06})
+[2026-07-28 02:19:52,294][starVLA.training.train_starvla][INFO] - Step 304, Loss: {'train/loss': 0.11684131622314453, 'train/grad_norm_pre_clip': 2.135678291320801, 'timing/data': 0.0002633240073919296, 'timing/model': 0.4130290490575135, 'learning_rate/qwen_vl_interface': 9.936009287392173e-06, 'learning_rate/action_model': 9.936009287392171e-05, 'epoch': 0.06})
+[2026-07-28 02:19:57,934][starVLA.training.train_starvla][INFO] - Step 305, Loss: {'train/loss': 0.12188870459794998, 'train/grad_norm_pre_clip': 2.0892202854156494, 'timing/data': 0.0002866378054022789, 'timing/model': 0.4173921109177172, 'learning_rate/qwen_vl_interface': 9.935381819595822e-06, 'learning_rate/action_model': 9.935381819595822e-05, 'epoch': 0.06})
+[2026-07-28 02:20:03,628][starVLA.training.train_starvla][INFO] - Step 306, Loss: {'train/loss': 0.12387260794639587, 'train/grad_norm_pre_clip': 2.1755383014678955, 'timing/data': 0.00028158584609627724, 'timing/model': 0.41463130759075284, 'learning_rate/qwen_vl_interface': 9.934751311506078e-06, 'learning_rate/action_model': 9.934751311506077e-05, 'epoch': 0.06})
+[2026-07-28 02:20:09,522][starVLA.training.train_starvla][INFO] - Step 307, Loss: {'train/loss': 0.12223957479000092, 'train/grad_norm_pre_clip': 2.025648355484009, 'timing/data': 0.0003421660512685776, 'timing/model': 0.4171441369690001, 'learning_rate/qwen_vl_interface': 9.93411776353207e-06, 'learning_rate/action_model': 9.934117763532069e-05, 'epoch': 0.06})
+[2026-07-28 02:20:15,605][starVLA.training.train_starvla][INFO] - Step 308, Loss: {'train/loss': 0.1482061743736267, 'train/grad_norm_pre_clip': 6.039999008178711, 'timing/data': 0.000348665751516819, 'timing/model': 0.65279675507918, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.06})
+[2026-07-28 02:20:21,677][starVLA.training.train_starvla][INFO] - Step 309, Loss: {'train/loss': 0.12637333571910858, 'train/grad_norm_pre_clip': 3.121053695678711, 'timing/data': 0.00034032296389341354, 'timing/model': 0.4151919623836875, 'learning_rate/qwen_vl_interface': 9.932841549577642e-06, 'learning_rate/action_model': 9.932841549577641e-05, 'epoch': 0.06})
+[2026-07-28 02:20:27,570][starVLA.training.train_starvla][INFO] - Step 310, Loss: {'train/loss': 0.18864291906356812, 'train/grad_norm_pre_clip': 2.6711461544036865, 'timing/data': 0.0002709627151489258, 'timing/model': 0.41706973873078823, 'learning_rate/qwen_vl_interface': 9.932198884425343e-06, 'learning_rate/action_model': 9.932198884425342e-05, 'epoch': 0.06})
+[2026-07-28 02:20:33,562][starVLA.training.train_starvla][INFO] - Step 311, Loss: {'train/loss': 0.11501103639602661, 'train/grad_norm_pre_clip': 2.689885377883911, 'timing/data': 0.0002943756990134716, 'timing/model': 0.41729270201176405, 'learning_rate/qwen_vl_interface': 9.931553181045022e-06, 'learning_rate/action_model': 9.931553181045022e-05, 'epoch': 0.06})
+[2026-07-28 02:20:39,675][starVLA.training.train_starvla][INFO] - Step 312, Loss: {'train/loss': 0.0813106894493103, 'train/grad_norm_pre_clip': 2.354517936706543, 'timing/data': 0.0002920846454799175, 'timing/model': 0.4140151129104197, 'learning_rate/qwen_vl_interface': 9.930904439855667e-06, 'learning_rate/action_model': 9.930904439855667e-05, 'epoch': 0.06})
+[2026-07-28 02:20:45,807][starVLA.training.train_starvla][INFO] - Step 313, Loss: {'train/loss': 0.09881198406219482, 'train/grad_norm_pre_clip': 2.895260810852051, 'timing/data': 0.00032712798565626144, 'timing/model': 0.41824736911803484, 'learning_rate/qwen_vl_interface': 9.930252661278245e-06, 'learning_rate/action_model': 9.930252661278244e-05, 'epoch': 0.06})
+[2026-07-28 02:20:51,714][starVLA.training.train_starvla][INFO] - Step 314, Loss: {'train/loss': 0.14543768763542175, 'train/grad_norm_pre_clip': 3.235436201095581, 'timing/data': 0.0003008781932294369, 'timing/model': 0.41665770998224616, 'learning_rate/qwen_vl_interface': 9.92959784573568e-06, 'learning_rate/action_model': 9.92959784573568e-05, 'epoch': 0.06})
+[2026-07-28 02:20:57,633][starVLA.training.train_starvla][INFO] - Step 315, Loss: {'train/loss': 0.11919090151786804, 'train/grad_norm_pre_clip': 2.5182323455810547, 'timing/data': 0.0002721427008509636, 'timing/model': 0.4177947612479329, 'learning_rate/qwen_vl_interface': 9.92893999365288e-06, 'learning_rate/action_model': 9.92893999365288e-05, 'epoch': 0.06})
+[2026-07-28 02:21:03,972][starVLA.training.train_starvla][INFO] - Step 316, Loss: {'train/loss': 0.11221499741077423, 'train/grad_norm_pre_clip': 2.5181596279144287, 'timing/data': 0.0003328039310872555, 'timing/model': 0.6459967629052699, 'learning_rate/qwen_vl_interface': 9.928279105456717e-06, 'learning_rate/action_model': 9.928279105456718e-05, 'epoch': 0.06})
+[2026-07-28 02:21:09,857][starVLA.training.train_starvla][INFO] - Step 317, Loss: {'train/loss': 0.07641541957855225, 'train/grad_norm_pre_clip': 2.691904306411743, 'timing/data': 0.0002754512242972851, 'timing/model': 0.41585903242230415, 'learning_rate/qwen_vl_interface': 9.927615181576034e-06, 'learning_rate/action_model': 9.927615181576034e-05, 'epoch': 0.06})
+[2026-07-28 02:21:15,543][starVLA.training.train_starvla][INFO] - Step 318, Loss: {'train/loss': 0.10047127306461334, 'train/grad_norm_pre_clip': 4.003772735595703, 'timing/data': 0.00034451624378561974, 'timing/model': 0.4214129177853465, 'learning_rate/qwen_vl_interface': 9.926948222441644e-06, 'learning_rate/action_model': 9.926948222441643e-05, 'epoch': 0.06})
+[2026-07-28 02:21:21,709][starVLA.training.train_starvla][INFO] - Step 319, Loss: {'train/loss': 0.06864319741725922, 'train/grad_norm_pre_clip': 1.5731593370437622, 'timing/data': 0.0002767411060631275, 'timing/model': 0.6677322858013213, 'learning_rate/qwen_vl_interface': 9.926278228486328e-06, 'learning_rate/action_model': 9.926278228486327e-05, 'epoch': 0.06})
+[2026-07-28 02:21:27,614][starVLA.training.train_starvla][INFO] - Step 320, Loss: {'train/loss': 0.17064596712589264, 'train/grad_norm_pre_clip': 4.2225213050842285, 'timing/data': 0.0002816002815961838, 'timing/model': 0.42120909597724676, 'learning_rate/qwen_vl_interface': 9.92560520014484e-06, 'learning_rate/action_model': 9.92560520014484e-05, 'epoch': 0.06})
+[2026-07-28 02:21:33,723][starVLA.training.train_starvla][INFO] - Step 321, Loss: {'train/loss': 0.08303218334913254, 'train/grad_norm_pre_clip': 3.11091947555542, 'timing/data': 0.0002638474106788635, 'timing/model': 0.4149150871671736, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.06})
+[2026-07-28 02:21:39,591][starVLA.training.train_starvla][INFO] - Step 322, Loss: {'train/loss': 0.11125823855400085, 'train/grad_norm_pre_clip': 5.372093200683594, 'timing/data': 0.00028972234576940536, 'timing/model': 0.4165235529653728, 'learning_rate/qwen_vl_interface': 9.924250042052197e-06, 'learning_rate/action_model': 9.924250042052197e-05, 'epoch': 0.06})
+[2026-07-28 02:21:45,477][starVLA.training.train_starvla][INFO] - Step 323, Loss: {'train/loss': 0.1460079699754715, 'train/grad_norm_pre_clip': 4.6807475090026855, 'timing/data': 0.0002842419780790806, 'timing/model': 0.4166027931496501, 'learning_rate/qwen_vl_interface': 9.92356791318039e-06, 'learning_rate/action_model': 9.92356791318039e-05, 'epoch': 0.06})
+[2026-07-28 02:21:51,605][starVLA.training.train_starvla][INFO] - Step 324, Loss: {'train/loss': 0.14158934354782104, 'train/grad_norm_pre_clip': 4.328695774078369, 'timing/data': 0.00028688181191682816, 'timing/model': 0.6653018919751048, 'learning_rate/qwen_vl_interface': 9.922882751681104e-06, 'learning_rate/action_model': 9.922882751681104e-05, 'epoch': 0.06})
+[2026-07-28 02:21:57,489][starVLA.training.train_starvla][INFO] - Step 325, Loss: {'train/loss': 0.06665287911891937, 'train/grad_norm_pre_clip': 1.4851086139678955, 'timing/data': 0.00031978217884898186, 'timing/model': 0.4136945349164307, 'learning_rate/qwen_vl_interface': 9.922194557998932e-06, 'learning_rate/action_model': 9.922194557998933e-05, 'epoch': 0.06})
+[2026-07-28 02:22:03,554][starVLA.training.train_starvla][INFO] - Step 326, Loss: {'train/loss': 0.0888548418879509, 'train/grad_norm_pre_clip': 3.4812707901000977, 'timing/data': 0.00043457793071866035, 'timing/model': 0.6489635347388685, 'learning_rate/qwen_vl_interface': 9.921503332580439e-06, 'learning_rate/action_model': 9.921503332580439e-05, 'epoch': 0.06})
+[2026-07-28 02:22:09,460][starVLA.training.train_starvla][INFO] - Step 327, Loss: {'train/loss': 0.13571825623512268, 'train/grad_norm_pre_clip': 4.62252950668335, 'timing/data': 0.00027894042432308197, 'timing/model': 0.4161505959928036, 'learning_rate/qwen_vl_interface': 9.92080907587415e-06, 'learning_rate/action_model': 9.92080907587415e-05, 'epoch': 0.06})
+[2026-07-28 02:22:15,391][starVLA.training.train_starvla][INFO] - Step 328, Loss: {'train/loss': 0.13678285479545593, 'train/grad_norm_pre_clip': 3.921591281890869, 'timing/data': 0.00033759791404008865, 'timing/model': 0.4235256230458617, 'learning_rate/qwen_vl_interface': 9.92011178833056e-06, 'learning_rate/action_model': 9.92011178833056e-05, 'epoch': 0.06})
+[2026-07-28 02:22:21,276][starVLA.training.train_starvla][INFO] - Step 329, Loss: {'train/loss': 0.10161177814006805, 'train/grad_norm_pre_clip': 2.0233871936798096, 'timing/data': 0.0003672027960419655, 'timing/model': 0.42395347775891423, 'learning_rate/qwen_vl_interface': 9.919411470402136e-06, 'learning_rate/action_model': 9.919411470402136e-05, 'epoch': 0.06})
+[2026-07-28 02:22:27,369][starVLA.training.train_starvla][INFO] - Step 330, Loss: {'train/loss': 0.06363064050674438, 'train/grad_norm_pre_clip': 2.12282657623291, 'timing/data': 0.0003622230142354965, 'timing/model': 0.6665995828807354, 'learning_rate/qwen_vl_interface': 9.918708122543305e-06, 'learning_rate/action_model': 9.918708122543305e-05, 'epoch': 0.06})
+[2026-07-28 02:22:33,456][starVLA.training.train_starvla][INFO] - Step 331, Loss: {'train/loss': 0.10215619951486588, 'train/grad_norm_pre_clip': 1.7249433994293213, 'timing/data': 0.00026196613907814026, 'timing/model': 0.41517871897667646, 'learning_rate/qwen_vl_interface': 9.918001745210458e-06, 'learning_rate/action_model': 9.918001745210456e-05, 'epoch': 0.06})
+[2026-07-28 02:22:39,325][starVLA.training.train_starvla][INFO] - Step 332, Loss: {'train/loss': 0.21237894892692566, 'train/grad_norm_pre_clip': 8.834199905395508, 'timing/data': 0.00033892691135406494, 'timing/model': 0.42117701936513186, 'learning_rate/qwen_vl_interface': 9.917292338861958e-06, 'learning_rate/action_model': 9.917292338861957e-05, 'epoch': 0.06})
+[2026-07-28 02:22:45,211][starVLA.training.train_starvla][INFO] - Step 333, Loss: {'train/loss': 0.19676001369953156, 'train/grad_norm_pre_clip': 7.643396854400635, 'timing/data': 0.0003540809266269207, 'timing/model': 0.42168582417070866, 'learning_rate/qwen_vl_interface': 9.916579903958134e-06, 'learning_rate/action_model': 9.916579903958133e-05, 'epoch': 0.06})
+[2026-07-28 02:22:51,365][starVLA.training.train_starvla][INFO] - Step 334, Loss: {'train/loss': 0.07840066403150558, 'train/grad_norm_pre_clip': 4.5979766845703125, 'timing/data': 0.00025629205629229546, 'timing/model': 0.41358260391280055, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.06})
+[2026-07-28 02:22:57,493][starVLA.training.train_starvla][INFO] - Step 335, Loss: {'train/loss': 0.09134254604578018, 'train/grad_norm_pre_clip': 2.238579511642456, 'timing/data': 0.00026516569778323174, 'timing/model': 0.415196412242949, 'learning_rate/qwen_vl_interface': 9.915145950335631e-06, 'learning_rate/action_model': 9.915145950335632e-05, 'epoch': 0.06})
+[2026-07-28 02:23:03,329][starVLA.training.train_starvla][INFO] - Step 336, Loss: {'train/loss': 0.07548215240240097, 'train/grad_norm_pre_clip': 2.9785940647125244, 'timing/data': 0.0003670700825750828, 'timing/model': 0.41679254826158285, 'learning_rate/qwen_vl_interface': 9.914424432547434e-06, 'learning_rate/action_model': 9.914424432547433e-05, 'epoch': 0.06})
+[2026-07-28 02:23:09,333][starVLA.training.train_starvla][INFO] - Step 337, Loss: {'train/loss': 0.18115738034248352, 'train/grad_norm_pre_clip': 6.380116939544678, 'timing/data': 0.0003624670207500458, 'timing/model': 0.5865814676508307, 'learning_rate/qwen_vl_interface': 9.913699888064862e-06, 'learning_rate/action_model': 9.913699888064861e-05, 'epoch': 0.06})
+[2026-07-28 02:23:15,186][starVLA.training.train_starvla][INFO] - Step 338, Loss: {'train/loss': 0.1366083323955536, 'train/grad_norm_pre_clip': 4.641314506530762, 'timing/data': 0.00037820497527718544, 'timing/model': 0.415472649037838, 'learning_rate/qwen_vl_interface': 9.912972317358066e-06, 'learning_rate/action_model': 9.912972317358066e-05, 'epoch': 0.06})
+[2026-07-28 02:23:21,297][starVLA.training.train_starvla][INFO] - Step 339, Loss: {'train/loss': 0.12643536925315857, 'train/grad_norm_pre_clip': 2.938779592514038, 'timing/data': 0.0003239437937736511, 'timing/model': 0.42077779304236174, 'learning_rate/qwen_vl_interface': 9.912241720899157e-06, 'learning_rate/action_model': 9.912241720899157e-05, 'epoch': 0.06})
+[2026-07-28 02:23:27,220][starVLA.training.train_starvla][INFO] - Step 340, Loss: {'train/loss': 0.07806390523910522, 'train/grad_norm_pre_clip': 3.677175998687744, 'timing/data': 0.0002839118242263794, 'timing/model': 0.4152716719545424, 'learning_rate/qwen_vl_interface': 9.911508099162215e-06, 'learning_rate/action_model': 9.911508099162214e-05, 'epoch': 0.06})
+[2026-07-28 02:23:33,333][starVLA.training.train_starvla][INFO] - Step 341, Loss: {'train/loss': 0.07433858513832092, 'train/grad_norm_pre_clip': 2.146094799041748, 'timing/data': 0.00028357328847050667, 'timing/model': 0.4155981224030256, 'learning_rate/qwen_vl_interface': 9.910771452623276e-06, 'learning_rate/action_model': 9.910771452623275e-05, 'epoch': 0.06})
+[2026-07-28 02:23:39,204][starVLA.training.train_starvla][INFO] - Step 342, Loss: {'train/loss': 0.14113739132881165, 'train/grad_norm_pre_clip': 3.2894647121429443, 'timing/data': 0.0002834298647940159, 'timing/model': 0.4306998918764293, 'learning_rate/qwen_vl_interface': 9.910031781760342e-06, 'learning_rate/action_model': 9.910031781760342e-05, 'epoch': 0.06})
+[2026-07-28 02:23:45,051][starVLA.training.train_starvla][INFO] - Step 343, Loss: {'train/loss': 0.07770706713199615, 'train/grad_norm_pre_clip': 4.244843006134033, 'timing/data': 0.0002879314124584198, 'timing/model': 0.4155230140313506, 'learning_rate/qwen_vl_interface': 9.909289087053378e-06, 'learning_rate/action_model': 9.909289087053378e-05, 'epoch': 0.06})
+[2026-07-28 02:23:51,115][starVLA.training.train_starvla][INFO] - Step 344, Loss: {'train/loss': 0.13566724956035614, 'train/grad_norm_pre_clip': 4.47156286239624, 'timing/data': 0.0002896040678024292, 'timing/model': 0.6525650941766798, 'learning_rate/qwen_vl_interface': 9.90854336898431e-06, 'learning_rate/action_model': 9.90854336898431e-05, 'epoch': 0.06})
+[2026-07-28 02:23:57,180][starVLA.training.train_starvla][INFO] - Step 345, Loss: {'train/loss': 0.08970675617456436, 'train/grad_norm_pre_clip': 2.9719276428222656, 'timing/data': 0.00026546092703938484, 'timing/model': 0.41407022485509515, 'learning_rate/qwen_vl_interface': 9.907794628037028e-06, 'learning_rate/action_model': 9.907794628037028e-05, 'epoch': 0.06})
+[2026-07-28 02:24:03,020][starVLA.training.train_starvla][INFO] - Step 346, Loss: {'train/loss': 0.08281879872083664, 'train/grad_norm_pre_clip': 1.960221529006958, 'timing/data': 0.0002697957679629326, 'timing/model': 0.41564772091805935, 'learning_rate/qwen_vl_interface': 9.90704286469738e-06, 'learning_rate/action_model': 9.90704286469738e-05, 'epoch': 0.06})
+[2026-07-28 02:24:08,924][starVLA.training.train_starvla][INFO] - Step 347, Loss: {'train/loss': 0.14539381861686707, 'train/grad_norm_pre_clip': 4.372816562652588, 'timing/data': 0.00028213905170559883, 'timing/model': 0.4154847180470824, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.06})
+[2026-07-28 02:24:14,580][starVLA.training.train_starvla][INFO] - Step 348, Loss: {'train/loss': 0.1164436787366867, 'train/grad_norm_pre_clip': 4.246457099914551, 'timing/data': 0.00027175480499863625, 'timing/model': 0.41576950112357736, 'learning_rate/qwen_vl_interface': 9.905530272794195e-06, 'learning_rate/action_model': 9.905530272794195e-05, 'epoch': 0.06})
+[2026-07-28 02:24:20,224][starVLA.training.train_starvla][INFO] - Step 349, Loss: {'train/loss': 0.09297265112400055, 'train/grad_norm_pre_clip': 1.5188970565795898, 'timing/data': 0.00027890317142009735, 'timing/model': 0.41645685490220785, 'learning_rate/qwen_vl_interface': 9.904769445212163e-06, 'learning_rate/action_model': 9.904769445212162e-05, 'epoch': 0.06})
+[2026-07-28 02:24:26,119][starVLA.training.train_starvla][INFO] - Step 350, Loss: {'train/loss': 0.073763906955719, 'train/grad_norm_pre_clip': 2.280393123626709, 'timing/data': 0.0002756640315055847, 'timing/model': 0.6528640231117606, 'learning_rate/qwen_vl_interface': 9.904005597200772e-06, 'learning_rate/action_model': 9.904005597200772e-05, 'epoch': 0.06})
+[2026-07-28 02:24:31,960][starVLA.training.train_starvla][INFO] - Step 351, Loss: {'train/loss': 0.06803728640079498, 'train/grad_norm_pre_clip': 1.7283594608306885, 'timing/data': 0.00026176590472459793, 'timing/model': 0.414793130941689, 'learning_rate/qwen_vl_interface': 9.903238729255678e-06, 'learning_rate/action_model': 9.903238729255677e-05, 'epoch': 0.06})
+[2026-07-28 02:24:37,634][starVLA.training.train_starvla][INFO] - Step 352, Loss: {'train/loss': 0.08353748917579651, 'train/grad_norm_pre_clip': 1.72127366065979, 'timing/data': 0.0002970821224153042, 'timing/model': 0.41686005098745227, 'learning_rate/qwen_vl_interface': 9.902468841874492e-06, 'learning_rate/action_model': 9.902468841874492e-05, 'epoch': 0.06})
+[2026-07-28 02:24:43,545][starVLA.training.train_starvla][INFO] - Step 353, Loss: {'train/loss': 0.13205468654632568, 'train/grad_norm_pre_clip': 2.958587169647217, 'timing/data': 0.000294737983494997, 'timing/model': 0.4191768793389201, 'learning_rate/qwen_vl_interface': 9.901695935556784e-06, 'learning_rate/action_model': 9.901695935556783e-05, 'epoch': 0.06})
+[2026-07-28 02:24:49,436][starVLA.training.train_starvla][INFO] - Step 354, Loss: {'train/loss': 0.08909253776073456, 'train/grad_norm_pre_clip': 2.9412357807159424, 'timing/data': 0.00029031187295913696, 'timing/model': 0.4227181817404926, 'learning_rate/qwen_vl_interface': 9.900920010804088e-06, 'learning_rate/action_model': 9.900920010804088e-05, 'epoch': 0.07})
+[2026-07-28 02:24:55,837][starVLA.training.train_starvla][INFO] - Step 355, Loss: {'train/loss': 0.15106871724128723, 'train/grad_norm_pre_clip': 3.9901645183563232, 'timing/data': 0.00035482319071888924, 'timing/model': 0.6456767339259386, 'learning_rate/qwen_vl_interface': 9.900141068119891e-06, 'learning_rate/action_model': 9.900141068119891e-05, 'epoch': 0.07})
+[2026-07-28 02:25:01,729][starVLA.training.train_starvla][INFO] - Step 356, Loss: {'train/loss': 0.07385490834712982, 'train/grad_norm_pre_clip': 1.1807724237442017, 'timing/data': 0.00028326408937573433, 'timing/model': 0.42397563392296433, 'learning_rate/qwen_vl_interface': 9.89935910800964e-06, 'learning_rate/action_model': 9.89935910800964e-05, 'epoch': 0.07})
+[2026-07-28 02:25:07,615][starVLA.training.train_starvla][INFO] - Step 357, Loss: {'train/loss': 0.12986084818840027, 'train/grad_norm_pre_clip': 3.970337152481079, 'timing/data': 0.0002858629450201988, 'timing/model': 0.42303569708019495, 'learning_rate/qwen_vl_interface': 9.898574130980743e-06, 'learning_rate/action_model': 9.898574130980742e-05, 'epoch': 0.07})
+[2026-07-28 02:25:13,715][starVLA.training.train_starvla][INFO] - Step 358, Loss: {'train/loss': 0.12247580289840698, 'train/grad_norm_pre_clip': 4.380904197692871, 'timing/data': 0.0002573039382696152, 'timing/model': 0.4137539337389171, 'learning_rate/qwen_vl_interface': 9.89778613754256e-06, 'learning_rate/action_model': 9.89778613754256e-05, 'epoch': 0.07})
+[2026-07-28 02:25:19,816][starVLA.training.train_starvla][INFO] - Step 359, Loss: {'train/loss': 0.05871782824397087, 'train/grad_norm_pre_clip': 1.7197481393814087, 'timing/data': 0.0002542068250477314, 'timing/model': 0.4134702510200441, 'learning_rate/qwen_vl_interface': 9.896995128206412e-06, 'learning_rate/action_model': 9.896995128206411e-05, 'epoch': 0.07})
+[2026-07-28 02:25:25,689][starVLA.training.train_starvla][INFO] - Step 360, Loss: {'train/loss': 0.12962649762630463, 'train/grad_norm_pre_clip': 3.093048095703125, 'timing/data': 0.000276234932243824, 'timing/model': 0.4161217659711838, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.07})
+[2026-07-28 02:25:31,584][starVLA.training.train_starvla][INFO] - Step 361, Loss: {'train/loss': 0.13258494436740875, 'train/grad_norm_pre_clip': 6.547626972198486, 'timing/data': 0.00025481218472123146, 'timing/model': 0.4159396942704916, 'learning_rate/qwen_vl_interface': 9.89540406389529e-06, 'learning_rate/action_model': 9.89540406389529e-05, 'epoch': 0.07})
+[2026-07-28 02:25:37,714][starVLA.training.train_starvla][INFO] - Step 362, Loss: {'train/loss': 0.14184686541557312, 'train/grad_norm_pre_clip': 2.313386917114258, 'timing/data': 0.000276896171271801, 'timing/model': 0.6749966200441122, 'learning_rate/qwen_vl_interface': 9.894604009952742e-06, 'learning_rate/action_model': 9.89460400995274e-05, 'epoch': 0.07})
+[2026-07-28 02:25:43,559][starVLA.training.train_starvla][INFO] - Step 363, Loss: {'train/loss': 0.0677042156457901, 'train/grad_norm_pre_clip': 1.32203209400177, 'timing/data': 0.00025432975962758064, 'timing/model': 0.4152980959042907, 'learning_rate/qwen_vl_interface': 9.893800942177073e-06, 'learning_rate/action_model': 9.893800942177073e-05, 'epoch': 0.07})
+[2026-07-28 02:25:49,431][starVLA.training.train_starvla][INFO] - Step 364, Loss: {'train/loss': 0.1085776835680008, 'train/grad_norm_pre_clip': 3.0294272899627686, 'timing/data': 0.000271419994533062, 'timing/model': 0.4150360831990838, 'learning_rate/qwen_vl_interface': 9.892994861089393e-06, 'learning_rate/action_model': 9.892994861089393e-05, 'epoch': 0.07})
+[2026-07-28 02:25:55,294][starVLA.training.train_starvla][INFO] - Step 365, Loss: {'train/loss': 0.08425413072109222, 'train/grad_norm_pre_clip': 2.764347553253174, 'timing/data': 0.0002897670492529869, 'timing/model': 0.4153617527335882, 'learning_rate/qwen_vl_interface': 9.892185767212756e-06, 'learning_rate/action_model': 9.892185767212756e-05, 'epoch': 0.07})
+[2026-07-28 02:26:01,379][starVLA.training.train_starvla][INFO] - Step 366, Loss: {'train/loss': 0.08861222863197327, 'train/grad_norm_pre_clip': 2.6560909748077393, 'timing/data': 0.0002772812731564045, 'timing/model': 0.6509287660010159, 'learning_rate/qwen_vl_interface': 9.891373661072172e-06, 'learning_rate/action_model': 9.891373661072173e-05, 'epoch': 0.07})
+[2026-07-28 02:26:07,254][starVLA.training.train_starvla][INFO] - Step 367, Loss: {'train/loss': 0.08164499700069427, 'train/grad_norm_pre_clip': 2.842223644256592, 'timing/data': 0.00026422785595059395, 'timing/model': 0.4154948089271784, 'learning_rate/qwen_vl_interface': 9.89055854319461e-06, 'learning_rate/action_model': 9.890558543194611e-05, 'epoch': 0.07})
+[2026-07-28 02:26:13,353][starVLA.training.train_starvla][INFO] - Step 368, Loss: {'train/loss': 0.10609469562768936, 'train/grad_norm_pre_clip': 4.707466125488281, 'timing/data': 0.0002485858276486397, 'timing/model': 0.41387283988296986, 'learning_rate/qwen_vl_interface': 9.889740414108994e-06, 'learning_rate/action_model': 9.889740414108995e-05, 'epoch': 0.07})
+[2026-07-28 02:26:19,247][starVLA.training.train_starvla][INFO] - Step 369, Loss: {'train/loss': 0.07962200045585632, 'train/grad_norm_pre_clip': 2.29290771484375, 'timing/data': 0.00028221821412444115, 'timing/model': 0.4164527510292828, 'learning_rate/qwen_vl_interface': 9.888919274346196e-06, 'learning_rate/action_model': 9.888919274346195e-05, 'epoch': 0.07})
+[2026-07-28 02:26:25,085][starVLA.training.train_starvla][INFO] - Step 370, Loss: {'train/loss': 0.09574413299560547, 'train/grad_norm_pre_clip': 2.189114570617676, 'timing/data': 0.00027174316346645355, 'timing/model': 0.4163708509877324, 'learning_rate/qwen_vl_interface': 9.888095124439043e-06, 'learning_rate/action_model': 9.888095124439043e-05, 'epoch': 0.07})
+[2026-07-28 02:26:31,170][starVLA.training.train_starvla][INFO] - Step 371, Loss: {'train/loss': 0.11030591279268265, 'train/grad_norm_pre_clip': 4.131537914276123, 'timing/data': 0.0002752970904111862, 'timing/model': 0.661241078749299, 'learning_rate/qwen_vl_interface': 9.88726796492232e-06, 'learning_rate/action_model': 9.88726796492232e-05, 'epoch': 0.07})
+[2026-07-28 02:26:37,288][starVLA.training.train_starvla][INFO] - Step 372, Loss: {'train/loss': 0.07196931540966034, 'train/grad_norm_pre_clip': 1.1048924922943115, 'timing/data': 0.00025454023852944374, 'timing/model': 0.4140884126536548, 'learning_rate/qwen_vl_interface': 9.886437796332759e-06, 'learning_rate/action_model': 9.88643779633276e-05, 'epoch': 0.07})
+[2026-07-28 02:26:43,136][starVLA.training.train_starvla][INFO] - Step 373, Loss: {'train/loss': 0.13247138261795044, 'train/grad_norm_pre_clip': 5.691636085510254, 'timing/data': 0.0002625659108161926, 'timing/model': 0.4169042268767953, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.07})
+[2026-07-28 02:26:48,999][starVLA.training.train_starvla][INFO] - Step 374, Loss: {'train/loss': 0.12155283242464066, 'train/grad_norm_pre_clip': 3.850526809692383, 'timing/data': 0.0002796747721731663, 'timing/model': 0.41587178176268935, 'learning_rate/qwen_vl_interface': 9.884768434091832e-06, 'learning_rate/action_model': 9.884768434091832e-05, 'epoch': 0.07})
+[2026-07-28 02:26:55,088][starVLA.training.train_starvla][INFO] - Step 375, Loss: {'train/loss': 0.07432909309864044, 'train/grad_norm_pre_clip': 3.31660532951355, 'timing/data': 0.00023450003936886787, 'timing/model': 0.41451421985402703, 'learning_rate/qwen_vl_interface': 9.883929241523698e-06, 'learning_rate/action_model': 9.883929241523697e-05, 'epoch': 0.07})
+[2026-07-28 02:27:00,947][starVLA.training.train_starvla][INFO] - Step 376, Loss: {'train/loss': 0.13500355184078217, 'train/grad_norm_pre_clip': 3.6771814823150635, 'timing/data': 0.00028221495449543, 'timing/model': 0.41561434185132384, 'learning_rate/qwen_vl_interface': 9.883087042049185e-06, 'learning_rate/action_model': 9.883087042049186e-05, 'epoch': 0.07})
+[2026-07-28 02:27:07,156][starVLA.training.train_starvla][INFO] - Step 377, Loss: {'train/loss': 0.11142134666442871, 'train/grad_norm_pre_clip': 2.432856798171997, 'timing/data': 0.0002721003256738186, 'timing/model': 0.6623070510104299, 'learning_rate/qwen_vl_interface': 9.882241836214794e-06, 'learning_rate/action_model': 9.882241836214794e-05, 'epoch': 0.07})
+[2026-07-28 02:27:13,154][starVLA.training.train_starvla][INFO] - Step 378, Loss: {'train/loss': 0.16200436651706696, 'train/grad_norm_pre_clip': 5.268353462219238, 'timing/data': 0.00024727871641516685, 'timing/model': 0.5868072509765625, 'learning_rate/qwen_vl_interface': 9.881393624568965e-06, 'learning_rate/action_model': 9.881393624568965e-05, 'epoch': 0.07})
+[2026-07-28 02:27:19,027][starVLA.training.train_starvla][INFO] - Step 379, Loss: {'train/loss': 0.12319888174533844, 'train/grad_norm_pre_clip': 2.7943227291107178, 'timing/data': 0.0002722828648984432, 'timing/model': 0.41788176679983735, 'learning_rate/qwen_vl_interface': 9.880542407662098e-06, 'learning_rate/action_model': 9.880542407662097e-05, 'epoch': 0.07})
+[2026-07-28 02:27:24,946][starVLA.training.train_starvla][INFO] - Step 380, Loss: {'train/loss': 0.13470877707004547, 'train/grad_norm_pre_clip': 3.157202959060669, 'timing/data': 0.00029559899121522903, 'timing/model': 0.42419614316895604, 'learning_rate/qwen_vl_interface': 9.879688186046535e-06, 'learning_rate/action_model': 9.879688186046534e-05, 'epoch': 0.07})
+[2026-07-28 02:27:30,849][starVLA.training.train_starvla][INFO] - Step 381, Loss: {'train/loss': 0.1566530466079712, 'train/grad_norm_pre_clip': 4.500113487243652, 'timing/data': 0.00039024418219923973, 'timing/model': 0.41593687096610665, 'learning_rate/qwen_vl_interface': 9.87883096027657e-06, 'learning_rate/action_model': 9.87883096027657e-05, 'epoch': 0.07})
+[2026-07-28 02:27:36,991][starVLA.training.train_starvla][INFO] - Step 382, Loss: {'train/loss': 0.14922404289245605, 'train/grad_norm_pre_clip': 2.7446048259735107, 'timing/data': 0.0003057699650526047, 'timing/model': 0.41500577004626393, 'learning_rate/qwen_vl_interface': 9.87797073090845e-06, 'learning_rate/action_model': 9.87797073090845e-05, 'epoch': 0.07})
+[2026-07-28 02:27:43,092][starVLA.training.train_starvla][INFO] - Step 383, Loss: {'train/loss': 0.11711739003658295, 'train/grad_norm_pre_clip': 2.522289514541626, 'timing/data': 0.0002657161094248295, 'timing/model': 0.4156334437429905, 'learning_rate/qwen_vl_interface': 9.877107498500368e-06, 'learning_rate/action_model': 9.877107498500367e-05, 'epoch': 0.07})
+[2026-07-28 02:27:48,950][starVLA.training.train_starvla][INFO] - Step 384, Loss: {'train/loss': 0.11671945452690125, 'train/grad_norm_pre_clip': 4.584192276000977, 'timing/data': 0.00027159135788679123, 'timing/model': 0.4160216189920902, 'learning_rate/qwen_vl_interface': 9.876241263612467e-06, 'learning_rate/action_model': 9.876241263612465e-05, 'epoch': 0.07})
+[2026-07-28 02:27:54,819][starVLA.training.train_starvla][INFO] - Step 385, Loss: {'train/loss': 0.11854543536901474, 'train/grad_norm_pre_clip': 2.9946892261505127, 'timing/data': 0.0003618369810283184, 'timing/model': 0.4161159610375762, 'learning_rate/qwen_vl_interface': 9.875372026806835e-06, 'learning_rate/action_model': 9.875372026806835e-05, 'epoch': 0.07})
+[2026-07-28 02:28:00,914][starVLA.training.train_starvla][INFO] - Step 386, Loss: {'train/loss': 0.11387831717729568, 'train/grad_norm_pre_clip': 1.8248467445373535, 'timing/data': 0.00025658123195171356, 'timing/model': 0.4148340751416981, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.07})
+[2026-07-28 02:28:06,800][starVLA.training.train_starvla][INFO] - Step 387, Loss: {'train/loss': 0.0808405727148056, 'train/grad_norm_pre_clip': 1.6954642534255981, 'timing/data': 0.0003501041792333126, 'timing/model': 0.4161764611490071, 'learning_rate/qwen_vl_interface': 9.873624549700485e-06, 'learning_rate/action_model': 9.873624549700484e-05, 'epoch': 0.07})
+[2026-07-28 02:28:12,929][starVLA.training.train_starvla][INFO] - Step 388, Loss: {'train/loss': 0.12841197848320007, 'train/grad_norm_pre_clip': 3.7242186069488525, 'timing/data': 0.00026642484590411186, 'timing/model': 0.6497075706720352, 'learning_rate/qwen_vl_interface': 9.872746310533683e-06, 'learning_rate/action_model': 9.872746310533684e-05, 'epoch': 0.07})
+[2026-07-28 02:28:18,802][starVLA.training.train_starvla][INFO] - Step 389, Loss: {'train/loss': 0.14738979935646057, 'train/grad_norm_pre_clip': 4.5105814933776855, 'timing/data': 0.0002597067505121231, 'timing/model': 0.41432652389630675, 'learning_rate/qwen_vl_interface': 9.871865071716991e-06, 'learning_rate/action_model': 9.87186507171699e-05, 'epoch': 0.07})
+[2026-07-28 02:28:24,692][starVLA.training.train_starvla][INFO] - Step 390, Loss: {'train/loss': 0.06848844140768051, 'train/grad_norm_pre_clip': 2.9305148124694824, 'timing/data': 0.000273779034614563, 'timing/model': 0.416660760063678, 'learning_rate/qwen_vl_interface': 9.870980833822231e-06, 'learning_rate/action_model': 9.87098083382223e-05, 'epoch': 0.07})
+[2026-07-28 02:28:30,791][starVLA.training.train_starvla][INFO] - Step 391, Loss: {'train/loss': 0.1420554220676422, 'train/grad_norm_pre_clip': 2.6851110458374023, 'timing/data': 0.0002772710286080837, 'timing/model': 0.6514446446672082, 'learning_rate/qwen_vl_interface': 9.87009359742318e-06, 'learning_rate/action_model': 9.87009359742318e-05, 'epoch': 0.07})
+[2026-07-28 02:28:36,682][starVLA.training.train_starvla][INFO] - Step 392, Loss: {'train/loss': 0.10325756669044495, 'train/grad_norm_pre_clip': 5.116493225097656, 'timing/data': 0.00027342885732650757, 'timing/model': 0.4136909060180187, 'learning_rate/qwen_vl_interface': 9.86920336309555e-06, 'learning_rate/action_model': 9.86920336309555e-05, 'epoch': 0.07})
+[2026-07-28 02:28:42,768][starVLA.training.train_starvla][INFO] - Step 393, Loss: {'train/loss': 0.12361954152584076, 'train/grad_norm_pre_clip': 5.178313255310059, 'timing/data': 0.00034277327358722687, 'timing/model': 0.4146798369474709, 'learning_rate/qwen_vl_interface': 9.86831013141701e-06, 'learning_rate/action_model': 9.86831013141701e-05, 'epoch': 0.07})
+[2026-07-28 02:28:48,428][starVLA.training.train_starvla][INFO] - Step 394, Loss: {'train/loss': 0.10026321560144424, 'train/grad_norm_pre_clip': 2.207216501235962, 'timing/data': 0.00028181029483675957, 'timing/model': 0.4166729236021638, 'learning_rate/qwen_vl_interface': 9.867413902967166e-06, 'learning_rate/action_model': 9.867413902967166e-05, 'epoch': 0.07})
+[2026-07-28 02:28:54,313][starVLA.training.train_starvla][INFO] - Step 395, Loss: {'train/loss': 0.12193378061056137, 'train/grad_norm_pre_clip': 3.899773120880127, 'timing/data': 0.0003322651609778404, 'timing/model': 0.41564213018864393, 'learning_rate/qwen_vl_interface': 9.86651467832757e-06, 'learning_rate/action_model': 9.86651467832757e-05, 'epoch': 0.07})
+[2026-07-28 02:29:00,439][starVLA.training.train_starvla][INFO] - Step 396, Loss: {'train/loss': 0.08298724889755249, 'train/grad_norm_pre_clip': 2.2041354179382324, 'timing/data': 0.0002680569887161255, 'timing/model': 0.41784948110580444, 'learning_rate/qwen_vl_interface': 9.865612458081718e-06, 'learning_rate/action_model': 9.865612458081719e-05, 'epoch': 0.07})
+[2026-07-28 02:29:06,306][starVLA.training.train_starvla][INFO] - Step 397, Loss: {'train/loss': 0.10819893330335617, 'train/grad_norm_pre_clip': 1.6474374532699585, 'timing/data': 0.00034217117354273796, 'timing/model': 0.41678147204220295, 'learning_rate/qwen_vl_interface': 9.864707242815057e-06, 'learning_rate/action_model': 9.864707242815056e-05, 'epoch': 0.07})
+[2026-07-28 02:29:12,417][starVLA.training.train_starvla][INFO] - Step 398, Loss: {'train/loss': 0.07830765843391418, 'train/grad_norm_pre_clip': 1.8946698904037476, 'timing/data': 0.00025170668959617615, 'timing/model': 0.4137763720937073, 'learning_rate/qwen_vl_interface': 9.863799033114966e-06, 'learning_rate/action_model': 9.863799033114966e-05, 'epoch': 0.07})
+[2026-07-28 02:29:18,293][starVLA.training.train_starvla][INFO] - Step 399, Loss: {'train/loss': 0.06361927837133408, 'train/grad_norm_pre_clip': 1.8811558485031128, 'timing/data': 0.0002827690914273262, 'timing/model': 0.4172863126732409, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.07})
+[2026-07-28 02:29:24,446][starVLA.training.train_starvla][INFO] - Step 400, Loss: {'train/loss': 0.13929234445095062, 'train/grad_norm_pre_clip': 2.1804020404815674, 'timing/data': 0.00027952389791607857, 'timing/model': 0.6466406602412462, 'learning_rate/qwen_vl_interface': 9.861973632773748e-06, 'learning_rate/action_model': 9.861973632773748e-05, 'epoch': 0.07})
+[2026-07-28 02:29:30,299][starVLA.training.train_starvla][INFO] - Step 401, Loss: {'train/loss': 0.07291126251220703, 'train/grad_norm_pre_clip': 1.9546574354171753, 'timing/data': 0.00026674196124076843, 'timing/model': 0.41365530993789434, 'learning_rate/qwen_vl_interface': 9.861056443317105e-06, 'learning_rate/action_model': 9.861056443317105e-05, 'epoch': 0.07})
+[2026-07-28 02:29:36,362][starVLA.training.train_starvla][INFO] - Step 402, Loss: {'train/loss': 0.10445726662874222, 'train/grad_norm_pre_clip': 2.6412193775177, 'timing/data': 0.0002856752835214138, 'timing/model': 0.6530121732503176, 'learning_rate/qwen_vl_interface': 9.860136261796e-06, 'learning_rate/action_model': 9.860136261795998e-05, 'epoch': 0.07})
+[2026-07-28 02:29:42,214][starVLA.training.train_starvla][INFO] - Step 403, Loss: {'train/loss': 0.11013805866241455, 'train/grad_norm_pre_clip': 2.3835957050323486, 'timing/data': 0.00028567900881171227, 'timing/model': 0.4165616100654006, 'learning_rate/qwen_vl_interface': 9.859213088807521e-06, 'learning_rate/action_model': 9.85921308880752e-05, 'epoch': 0.07})
+[2026-07-28 02:29:48,080][starVLA.training.train_starvla][INFO] - Step 404, Loss: {'train/loss': 0.13796831667423248, 'train/grad_norm_pre_clip': 4.180176734924316, 'timing/data': 0.0002789651043713093, 'timing/model': 0.41573171596974134, 'learning_rate/qwen_vl_interface': 9.858286924950714e-06, 'learning_rate/action_model': 9.858286924950713e-05, 'epoch': 0.07})
+[2026-07-28 02:29:53,975][starVLA.training.train_starvla][INFO] - Step 405, Loss: {'train/loss': 0.0776362419128418, 'train/grad_norm_pre_clip': 3.1444969177246094, 'timing/data': 0.00027801981195807457, 'timing/model': 0.4148292248137295, 'learning_rate/qwen_vl_interface': 9.85735777082655e-06, 'learning_rate/action_model': 9.857357770826551e-05, 'epoch': 0.07})
+[2026-07-28 02:30:00,127][starVLA.training.train_starvla][INFO] - Step 406, Loss: {'train/loss': 0.07317804545164108, 'train/grad_norm_pre_clip': 1.240861415863037, 'timing/data': 0.0002568592317402363, 'timing/model': 0.4156691422685981, 'learning_rate/qwen_vl_interface': 9.856425627037951e-06, 'learning_rate/action_model': 9.856425627037952e-05, 'epoch': 0.07})
+[2026-07-28 02:30:05,792][starVLA.training.train_starvla][INFO] - Step 407, Loss: {'train/loss': 0.0783543586730957, 'train/grad_norm_pre_clip': 3.2323572635650635, 'timing/data': 0.0002779979258775711, 'timing/model': 0.41962618893012404, 'learning_rate/qwen_vl_interface': 9.855490494189777e-06, 'learning_rate/action_model': 9.855490494189776e-05, 'epoch': 0.07})
+[2026-07-28 02:30:11,693][starVLA.training.train_starvla][INFO] - Step 408, Loss: {'train/loss': 0.11859577894210815, 'train/grad_norm_pre_clip': 5.213680744171143, 'timing/data': 0.0002819378860294819, 'timing/model': 0.4315059897489846, 'learning_rate/qwen_vl_interface': 9.854552372888819e-06, 'learning_rate/action_model': 9.854552372888819e-05, 'epoch': 0.07})
+[2026-07-28 02:30:17,549][starVLA.training.train_starvla][INFO] - Step 409, Loss: {'train/loss': 0.1169334277510643, 'train/grad_norm_pre_clip': 3.071854829788208, 'timing/data': 0.0006800456903874874, 'timing/model': 0.4165094937197864, 'learning_rate/qwen_vl_interface': 9.853611263743817e-06, 'learning_rate/action_model': 9.853611263743817e-05, 'epoch': 0.08})
+[2026-07-28 02:30:23,702][starVLA.training.train_starvla][INFO] - Step 410, Loss: {'train/loss': 0.10138507187366486, 'train/grad_norm_pre_clip': 2.8536534309387207, 'timing/data': 0.0002556322142481804, 'timing/model': 0.4143951595760882, 'learning_rate/qwen_vl_interface': 9.85266716736545e-06, 'learning_rate/action_model': 9.852667167365449e-05, 'epoch': 0.08})
+[2026-07-28 02:30:29,390][starVLA.training.train_starvla][INFO] - Step 411, Loss: {'train/loss': 0.09791998565196991, 'train/grad_norm_pre_clip': 3.7690343856811523, 'timing/data': 0.0002783290110528469, 'timing/model': 0.416817604098469, 'learning_rate/qwen_vl_interface': 9.851720084366324e-06, 'learning_rate/action_model': 9.851720084366324e-05, 'epoch': 0.08})
+[2026-07-28 02:30:35,305][starVLA.training.train_starvla][INFO] - Step 412, Loss: {'train/loss': 0.15675072371959686, 'train/grad_norm_pre_clip': 5.472712516784668, 'timing/data': 0.0002932688221335411, 'timing/model': 0.4326665010303259, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.08})
+[2026-07-28 02:30:41,418][starVLA.training.train_starvla][INFO] - Step 413, Loss: {'train/loss': 0.16016337275505066, 'train/grad_norm_pre_clip': 7.280455112457275, 'timing/data': 0.00026297662407159805, 'timing/model': 0.4146633897908032, 'learning_rate/qwen_vl_interface': 9.849816960965959e-06, 'learning_rate/action_model': 9.849816960965959e-05, 'epoch': 0.08})
+[2026-07-28 02:30:47,531][starVLA.training.train_starvla][INFO] - Step 414, Loss: {'train/loss': 0.061043791472911835, 'train/grad_norm_pre_clip': 3.567365884780884, 'timing/data': 0.0002560848370194435, 'timing/model': 0.4157614829018712, 'learning_rate/qwen_vl_interface': 9.848860921799633e-06, 'learning_rate/action_model': 9.848860921799632e-05, 'epoch': 0.08})
+[2026-07-28 02:30:53,402][starVLA.training.train_starvla][INFO] - Step 415, Loss: {'train/loss': 0.09645920991897583, 'train/grad_norm_pre_clip': 2.779162883758545, 'timing/data': 0.00031711068004369736, 'timing/model': 0.420252104755491, 'learning_rate/qwen_vl_interface': 9.847901898482385e-06, 'learning_rate/action_model': 9.847901898482385e-05, 'epoch': 0.08})
+[2026-07-28 02:30:59,253][starVLA.training.train_starvla][INFO] - Step 416, Loss: {'train/loss': 0.15136800706386566, 'train/grad_norm_pre_clip': 9.197437286376953, 'timing/data': 0.000363851897418499, 'timing/model': 0.419393063057214, 'learning_rate/qwen_vl_interface': 9.846939891636512e-06, 'learning_rate/action_model': 9.846939891636512e-05, 'epoch': 0.08})
+[2026-07-28 02:31:05,637][starVLA.training.train_starvla][INFO] - Step 417, Loss: {'train/loss': 0.09822021424770355, 'train/grad_norm_pre_clip': 2.136260509490967, 'timing/data': 0.0002853567712008953, 'timing/model': 0.6598257599398494, 'learning_rate/qwen_vl_interface': 9.845974901886253e-06, 'learning_rate/action_model': 9.845974901886252e-05, 'epoch': 0.08})
+[2026-07-28 02:31:11,528][starVLA.training.train_starvla][INFO] - Step 418, Loss: {'train/loss': 0.10984089970588684, 'train/grad_norm_pre_clip': 1.9348582029342651, 'timing/data': 0.00027279090136289597, 'timing/model': 0.4185986458323896, 'learning_rate/qwen_vl_interface': 9.845006929857777e-06, 'learning_rate/action_model': 9.845006929857776e-05, 'epoch': 0.08})
+[2026-07-28 02:31:17,414][starVLA.training.train_starvla][INFO] - Step 419, Loss: {'train/loss': 0.09129171073436737, 'train/grad_norm_pre_clip': 2.5057616233825684, 'timing/data': 0.0003041103482246399, 'timing/model': 0.41640874184668064, 'learning_rate/qwen_vl_interface': 9.844035976179193e-06, 'learning_rate/action_model': 9.844035976179192e-05, 'epoch': 0.08})
+[2026-07-28 02:31:23,278][starVLA.training.train_starvla][INFO] - Step 420, Loss: {'train/loss': 0.07642745971679688, 'train/grad_norm_pre_clip': 1.524397611618042, 'timing/data': 0.0003453390672802925, 'timing/model': 0.41539462422952056, 'learning_rate/qwen_vl_interface': 9.843062041480538e-06, 'learning_rate/action_model': 9.843062041480539e-05, 'epoch': 0.08})
+[2026-07-28 02:31:29,162][starVLA.training.train_starvla][INFO] - Step 421, Loss: {'train/loss': 0.10442022979259491, 'train/grad_norm_pre_clip': 2.490509033203125, 'timing/data': 0.0002625100314617157, 'timing/model': 0.415283627808094, 'learning_rate/qwen_vl_interface': 9.842085126393793e-06, 'learning_rate/action_model': 9.842085126393792e-05, 'epoch': 0.08})
+[2026-07-28 02:31:35,025][starVLA.training.train_starvla][INFO] - Step 422, Loss: {'train/loss': 0.0677928775548935, 'train/grad_norm_pre_clip': 2.0100150108337402, 'timing/data': 0.0002845423296093941, 'timing/model': 0.416047889739275, 'learning_rate/qwen_vl_interface': 9.841105231552863e-06, 'learning_rate/action_model': 9.841105231552864e-05, 'epoch': 0.08})
+[2026-07-28 02:31:40,988][starVLA.training.train_starvla][INFO] - Step 423, Loss: {'train/loss': 0.10738372802734375, 'train/grad_norm_pre_clip': 4.003664016723633, 'timing/data': 0.00027691107243299484, 'timing/model': 0.43297390080988407, 'learning_rate/qwen_vl_interface': 9.840122357593594e-06, 'learning_rate/action_model': 9.840122357593594e-05, 'epoch': 0.08})
+[2026-07-28 02:31:46,970][starVLA.training.train_starvla][INFO] - Step 424, Loss: {'train/loss': 0.09087786823511124, 'train/grad_norm_pre_clip': 4.03011417388916, 'timing/data': 0.00027839839458465576, 'timing/model': 0.4320601341314614, 'learning_rate/qwen_vl_interface': 9.839136505153762e-06, 'learning_rate/action_model': 9.839136505153762e-05, 'epoch': 0.08})
+[2026-07-28 02:31:53,128][starVLA.training.train_starvla][INFO] - Step 425, Loss: {'train/loss': 0.05816110596060753, 'train/grad_norm_pre_clip': 2.328645944595337, 'timing/data': 0.00025980593636631966, 'timing/model': 0.4144335291348398, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.08})
+[2026-07-28 02:31:59,247][starVLA.training.train_starvla][INFO] - Step 426, Loss: {'train/loss': 0.11535966396331787, 'train/grad_norm_pre_clip': 2.796797037124634, 'timing/data': 0.00029416196048259735, 'timing/model': 0.6541577549651265, 'learning_rate/qwen_vl_interface': 9.837155867393176e-06, 'learning_rate/action_model': 9.837155867393175e-05, 'epoch': 0.08})
+[2026-07-28 02:32:05,154][starVLA.training.train_starvla][INFO] - Step 427, Loss: {'train/loss': 0.1355208307504654, 'train/grad_norm_pre_clip': 5.099108695983887, 'timing/data': 0.00027777766808867455, 'timing/model': 0.4150056717917323, 'learning_rate/qwen_vl_interface': 9.836161083357635e-06, 'learning_rate/action_model': 9.836161083357635e-05, 'epoch': 0.08})
+[2026-07-28 02:32:11,272][starVLA.training.train_starvla][INFO] - Step 428, Loss: {'train/loss': 0.20486357808113098, 'train/grad_norm_pre_clip': 6.667947769165039, 'timing/data': 0.00028866203501820564, 'timing/model': 0.6561528658494353, 'learning_rate/qwen_vl_interface': 9.835163323411957e-06, 'learning_rate/action_model': 9.835163323411958e-05, 'epoch': 0.08})
+[2026-07-28 02:32:17,142][starVLA.training.train_starvla][INFO] - Step 429, Loss: {'train/loss': 0.1460641622543335, 'train/grad_norm_pre_clip': 3.2646093368530273, 'timing/data': 0.0002732211723923683, 'timing/model': 0.417929548304528, 'learning_rate/qwen_vl_interface': 9.834162588203581e-06, 'learning_rate/action_model': 9.83416258820358e-05, 'epoch': 0.08})
+[2026-07-28 02:32:23,113][starVLA.training.train_starvla][INFO] - Step 430, Loss: {'train/loss': 0.07932990044355392, 'train/grad_norm_pre_clip': 2.07706880569458, 'timing/data': 0.00029246602207422256, 'timing/model': 0.4164813836105168, 'learning_rate/qwen_vl_interface': 9.833158878381872e-06, 'learning_rate/action_model': 9.83315887838187e-05, 'epoch': 0.08})
+[2026-07-28 02:32:29,245][starVLA.training.train_starvla][INFO] - Step 431, Loss: {'train/loss': 0.08756282925605774, 'train/grad_norm_pre_clip': 2.5631415843963623, 'timing/data': 0.00025143614038825035, 'timing/model': 0.41510668816044927, 'learning_rate/qwen_vl_interface': 9.832152194598123e-06, 'learning_rate/action_model': 9.832152194598122e-05, 'epoch': 0.08})
+[2026-07-28 02:32:35,132][starVLA.training.train_starvla][INFO] - Step 432, Loss: {'train/loss': 0.0915147215127945, 'train/grad_norm_pre_clip': 1.9736366271972656, 'timing/data': 0.00029315799474716187, 'timing/model': 0.41784193413332105, 'learning_rate/qwen_vl_interface': 9.831142537505563e-06, 'learning_rate/action_model': 9.831142537505562e-05, 'epoch': 0.08})
+[2026-07-28 02:32:41,269][starVLA.training.train_starvla][INFO] - Step 433, Loss: {'train/loss': 0.1002020314335823, 'train/grad_norm_pre_clip': 3.1888427734375, 'timing/data': 0.0002898010425269604, 'timing/model': 0.6626810440793633, 'learning_rate/qwen_vl_interface': 9.830129907759347e-06, 'learning_rate/action_model': 9.830129907759347e-05, 'epoch': 0.08})
+[2026-07-28 02:32:47,420][starVLA.training.train_starvla][INFO] - Step 434, Loss: {'train/loss': 0.08608794212341309, 'train/grad_norm_pre_clip': 2.135129690170288, 'timing/data': 0.00026415614411234856, 'timing/model': 0.4212404047138989, 'learning_rate/qwen_vl_interface': 9.82911430601656e-06, 'learning_rate/action_model': 9.82911430601656e-05, 'epoch': 0.08})
+[2026-07-28 02:32:53,371][starVLA.training.train_starvla][INFO] - Step 435, Loss: {'train/loss': 0.09761818498373032, 'train/grad_norm_pre_clip': 2.8847031593322754, 'timing/data': 0.0002819132059812546, 'timing/model': 0.43298396095633507, 'learning_rate/qwen_vl_interface': 9.828095732936216e-06, 'learning_rate/action_model': 9.828095732936216e-05, 'epoch': 0.08})
+[2026-07-28 02:32:59,268][starVLA.training.train_starvla][INFO] - Step 436, Loss: {'train/loss': 0.0921643078327179, 'train/grad_norm_pre_clip': 1.4785079956054688, 'timing/data': 0.0003009648062288761, 'timing/model': 0.41512224776670337, 'learning_rate/qwen_vl_interface': 9.827074189179254e-06, 'learning_rate/action_model': 9.827074189179253e-05, 'epoch': 0.08})
+[2026-07-28 02:33:05,191][starVLA.training.train_starvla][INFO] - Step 437, Loss: {'train/loss': 0.1351746916770935, 'train/grad_norm_pre_clip': 4.919901371002197, 'timing/data': 0.00028569111600518227, 'timing/model': 0.4167167139239609, 'learning_rate/qwen_vl_interface': 9.82604967540854e-06, 'learning_rate/action_model': 9.82604967540854e-05, 'epoch': 0.08})
+[2026-07-28 02:33:11,288][starVLA.training.train_starvla][INFO] - Step 438, Loss: {'train/loss': 0.09587658941745758, 'train/grad_norm_pre_clip': 3.2949764728546143, 'timing/data': 0.0005588461644947529, 'timing/model': 0.6451739631593227, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.08})
+[2026-07-28 02:33:17,382][starVLA.training.train_starvla][INFO] - Step 439, Loss: {'train/loss': 0.07638216018676758, 'train/grad_norm_pre_clip': 2.2463862895965576, 'timing/data': 0.00026030000299215317, 'timing/model': 0.6507106176577508, 'learning_rate/qwen_vl_interface': 9.823991740486982e-06, 'learning_rate/action_model': 9.823991740486982e-05, 'epoch': 0.08})
+[2026-07-28 02:33:23,242][starVLA.training.train_starvla][INFO] - Step 440, Loss: {'train/loss': 0.05923917889595032, 'train/grad_norm_pre_clip': 1.789125919342041, 'timing/data': 0.0002671601250767708, 'timing/model': 0.41377435764297843, 'learning_rate/qwen_vl_interface': 9.822958320671506e-06, 'learning_rate/action_model': 9.822958320671506e-05, 'epoch': 0.08})
+[2026-07-28 02:33:29,257][starVLA.training.train_starvla][INFO] - Step 441, Loss: {'train/loss': 0.07639772444963455, 'train/grad_norm_pre_clip': 7.915600299835205, 'timing/data': 0.00026558805257081985, 'timing/model': 0.4166981289163232, 'learning_rate/qwen_vl_interface': 9.821921933513023e-06, 'learning_rate/action_model': 9.821921933513023e-05, 'epoch': 0.08})
+[2026-07-28 02:33:35,134][starVLA.training.train_starvla][INFO] - Step 442, Loss: {'train/loss': 0.1478150188922882, 'train/grad_norm_pre_clip': 7.535977363586426, 'timing/data': 0.0002802358940243721, 'timing/model': 0.4173330566845834, 'learning_rate/qwen_vl_interface': 9.820882579684037e-06, 'learning_rate/action_model': 9.820882579684037e-05, 'epoch': 0.08})
+[2026-07-28 02:33:41,034][starVLA.training.train_starvla][INFO] - Step 443, Loss: {'train/loss': 0.297002375125885, 'train/grad_norm_pre_clip': 13.454888343811035, 'timing/data': 0.0002798340283334255, 'timing/model': 0.4172955248504877, 'learning_rate/qwen_vl_interface': 9.819840259858968e-06, 'learning_rate/action_model': 9.819840259858968e-05, 'epoch': 0.08})
+[2026-07-28 02:33:47,397][starVLA.training.train_starvla][INFO] - Step 444, Loss: {'train/loss': 0.21327200531959534, 'train/grad_norm_pre_clip': 9.474011421203613, 'timing/data': 0.0002767927944660187, 'timing/model': 0.4175980337895453, 'learning_rate/qwen_vl_interface': 9.818794974714167e-06, 'learning_rate/action_model': 9.818794974714168e-05, 'epoch': 0.08})
+[2026-07-28 02:33:53,285][starVLA.training.train_starvla][INFO] - Step 445, Loss: {'train/loss': 0.09773409366607666, 'train/grad_norm_pre_clip': 5.794893264770508, 'timing/data': 0.0002793017774820328, 'timing/model': 0.41774915158748627, 'learning_rate/qwen_vl_interface': 9.817746724927912e-06, 'learning_rate/action_model': 9.817746724927912e-05, 'epoch': 0.08})
+[2026-07-28 02:33:59,205][starVLA.training.train_starvla][INFO] - Step 446, Loss: {'train/loss': 0.10691751539707184, 'train/grad_norm_pre_clip': 3.5653796195983887, 'timing/data': 0.00028851674869656563, 'timing/model': 0.417365456931293, 'learning_rate/qwen_vl_interface': 9.8166955111804e-06, 'learning_rate/action_model': 9.816695511180399e-05, 'epoch': 0.08})
+[2026-07-28 02:34:05,129][starVLA.training.train_starvla][INFO] - Step 447, Loss: {'train/loss': 0.12028098106384277, 'train/grad_norm_pre_clip': 2.9798319339752197, 'timing/data': 0.0002675699070096016, 'timing/model': 0.4176361998543143, 'learning_rate/qwen_vl_interface': 9.815641334153746e-06, 'learning_rate/action_model': 9.815641334153746e-05, 'epoch': 0.08})
+[2026-07-28 02:34:11,528][starVLA.training.train_starvla][INFO] - Step 448, Loss: {'train/loss': 0.08966900408267975, 'train/grad_norm_pre_clip': 3.458501100540161, 'timing/data': 0.000271861907094717, 'timing/model': 0.4150183843448758, 'learning_rate/qwen_vl_interface': 9.814584194532001e-06, 'learning_rate/action_model': 9.814584194532001e-05, 'epoch': 0.08})
+[2026-07-28 02:34:17,417][starVLA.training.train_starvla][INFO] - Step 449, Loss: {'train/loss': 0.08707084506750107, 'train/grad_norm_pre_clip': 2.9605863094329834, 'timing/data': 0.00027900421991944313, 'timing/model': 0.4171303943730891, 'learning_rate/qwen_vl_interface': 9.813524093001128e-06, 'learning_rate/action_model': 9.813524093001129e-05, 'epoch': 0.08})
+[2026-07-28 02:34:23,329][starVLA.training.train_starvla][INFO] - Step 450, Loss: {'train/loss': 0.10482259094715118, 'train/grad_norm_pre_clip': 2.951594352722168, 'timing/data': 0.00035354308784008026, 'timing/model': 0.4166115033440292, 'learning_rate/qwen_vl_interface': 9.81246103024902e-06, 'learning_rate/action_model': 9.81246103024902e-05, 'epoch': 0.08})
+[2026-07-28 02:34:29,479][starVLA.training.train_starvla][INFO] - Step 451, Loss: {'train/loss': 0.08727109432220459, 'train/grad_norm_pre_clip': 3.0779941082000732, 'timing/data': 0.0002682791091501713, 'timing/model': 0.4158947980031371, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.08})
+[2026-07-28 02:34:35,369][starVLA.training.train_starvla][INFO] - Step 452, Loss: {'train/loss': 0.1256900131702423, 'train/grad_norm_pre_clip': 4.5406904220581055, 'timing/data': 0.0002829679287970066, 'timing/model': 0.41677771881222725, 'learning_rate/qwen_vl_interface': 9.810326023842245e-06, 'learning_rate/action_model': 9.810326023842245e-05, 'epoch': 0.08})
+[2026-07-28 02:34:41,513][starVLA.training.train_starvla][INFO] - Step 453, Loss: {'train/loss': 0.07472244650125504, 'train/grad_norm_pre_clip': 2.9073240756988525, 'timing/data': 0.0002881172113120556, 'timing/model': 0.6846937388181686, 'learning_rate/qwen_vl_interface': 9.809254081572963e-06, 'learning_rate/action_model': 9.809254081572961e-05, 'epoch': 0.08})
+[2026-07-28 02:34:47,548][starVLA.training.train_starvla][INFO] - Step 454, Loss: {'train/loss': 0.11457589268684387, 'train/grad_norm_pre_clip': 3.3923113346099854, 'timing/data': 0.0002507097087800503, 'timing/model': 0.5832407488487661, 'learning_rate/qwen_vl_interface': 9.808179180853204e-06, 'learning_rate/action_model': 9.808179180853204e-05, 'epoch': 0.08})
+[2026-07-28 02:34:53,408][starVLA.training.train_starvla][INFO] - Step 455, Loss: {'train/loss': 0.14765194058418274, 'train/grad_norm_pre_clip': 4.369499206542969, 'timing/data': 0.0002833460457623005, 'timing/model': 0.4166735438629985, 'learning_rate/qwen_vl_interface': 9.807101322380464e-06, 'learning_rate/action_model': 9.807101322380464e-05, 'epoch': 0.08})
+[2026-07-28 02:34:59,292][starVLA.training.train_starvla][INFO] - Step 456, Loss: {'train/loss': 0.10643348097801208, 'train/grad_norm_pre_clip': 2.1231582164764404, 'timing/data': 0.000277920626103878, 'timing/model': 0.4153995788656175, 'learning_rate/qwen_vl_interface': 9.806020506854152e-06, 'learning_rate/action_model': 9.806020506854153e-05, 'epoch': 0.08})
+[2026-07-28 02:35:05,194][starVLA.training.train_starvla][INFO] - Step 457, Loss: {'train/loss': 0.10897362232208252, 'train/grad_norm_pre_clip': 3.019303798675537, 'timing/data': 0.0002771681174635887, 'timing/model': 0.41729891020804644, 'learning_rate/qwen_vl_interface': 9.804936734975596e-06, 'learning_rate/action_model': 9.804936734975596e-05, 'epoch': 0.08})
+[2026-07-28 02:35:11,323][starVLA.training.train_starvla][INFO] - Step 458, Loss: {'train/loss': 0.1230137050151825, 'train/grad_norm_pre_clip': 3.0432989597320557, 'timing/data': 0.00034964317455887794, 'timing/model': 0.4167763232253492, 'learning_rate/qwen_vl_interface': 9.803850007448046e-06, 'learning_rate/action_model': 9.803850007448045e-05, 'epoch': 0.08})
+[2026-07-28 02:35:17,197][starVLA.training.train_starvla][INFO] - Step 459, Loss: {'train/loss': 0.09012670814990997, 'train/grad_norm_pre_clip': 1.5261300802230835, 'timing/data': 0.0002767927944660187, 'timing/model': 0.41700800135731697, 'learning_rate/qwen_vl_interface': 9.802760324976666e-06, 'learning_rate/action_model': 9.802760324976667e-05, 'epoch': 0.08})
+[2026-07-28 02:35:23,339][starVLA.training.train_starvla][INFO] - Step 460, Loss: {'train/loss': 0.06092973053455353, 'train/grad_norm_pre_clip': 1.6239334344863892, 'timing/data': 0.0002454528585076332, 'timing/model': 0.41458993637934327, 'learning_rate/qwen_vl_interface': 9.80166768826854e-06, 'learning_rate/action_model': 9.80166768826854e-05, 'epoch': 0.08})
+[2026-07-28 02:35:29,221][starVLA.training.train_starvla][INFO] - Step 461, Loss: {'train/loss': 0.0772443562746048, 'train/grad_norm_pre_clip': 1.577549934387207, 'timing/data': 0.0002860482782125473, 'timing/model': 0.41708183381706476, 'learning_rate/qwen_vl_interface': 9.800572098032668e-06, 'learning_rate/action_model': 9.800572098032668e-05, 'epoch': 0.08})
+[2026-07-28 02:35:35,352][starVLA.training.train_starvla][INFO] - Step 462, Loss: {'train/loss': 0.09139920771121979, 'train/grad_norm_pre_clip': 2.736022472381592, 'timing/data': 0.00024641770869493484, 'timing/model': 0.41566351242363453, 'learning_rate/qwen_vl_interface': 9.799473554979966e-06, 'learning_rate/action_model': 9.799473554979965e-05, 'epoch': 0.08})
+[2026-07-28 02:35:41,242][starVLA.training.train_starvla][INFO] - Step 463, Loss: {'train/loss': 0.08433537185192108, 'train/grad_norm_pre_clip': 1.5335246324539185, 'timing/data': 0.0003024502657353878, 'timing/model': 0.41752093425020576, 'learning_rate/qwen_vl_interface': 9.798372059823268e-06, 'learning_rate/action_model': 9.798372059823268e-05, 'epoch': 0.09})
+[2026-07-28 02:35:47,334][starVLA.training.train_starvla][INFO] - Step 464, Loss: {'train/loss': 0.09995844215154648, 'train/grad_norm_pre_clip': 2.9814796447753906, 'timing/data': 0.0002728579565882683, 'timing/model': 0.6507757580839097, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.09})
+[2026-07-28 02:35:53,189][starVLA.training.train_starvla][INFO] - Step 465, Loss: {'train/loss': 0.23828405141830444, 'train/grad_norm_pre_clip': 6.8114166259765625, 'timing/data': 0.0002801166847348213, 'timing/model': 0.41717860428616405, 'learning_rate/qwen_vl_interface': 9.79616021605879e-06, 'learning_rate/action_model': 9.796160216058788e-05, 'epoch': 0.09})
+[2026-07-28 02:35:59,067][starVLA.training.train_starvla][INFO] - Step 466, Loss: {'train/loss': 0.14189328253269196, 'train/grad_norm_pre_clip': 6.98835563659668, 'timing/data': 0.0002768179401755333, 'timing/model': 0.4164204029366374, 'learning_rate/qwen_vl_interface': 9.795049868886247e-06, 'learning_rate/action_model': 9.795049868886247e-05, 'epoch': 0.09})
+[2026-07-28 02:36:05,191][starVLA.training.train_starvla][INFO] - Step 467, Loss: {'train/loss': 0.08707588911056519, 'train/grad_norm_pre_clip': 2.6595041751861572, 'timing/data': 0.00027605192735791206, 'timing/model': 0.6515157292596996, 'learning_rate/qwen_vl_interface': 9.793936572480192e-06, 'learning_rate/action_model': 9.793936572480192e-05, 'epoch': 0.09})
+[2026-07-28 02:36:11,073][starVLA.training.train_starvla][INFO] - Step 468, Loss: {'train/loss': 0.07208945602178574, 'train/grad_norm_pre_clip': 1.5653105974197388, 'timing/data': 0.0002540429122745991, 'timing/model': 0.4138363399542868, 'learning_rate/qwen_vl_interface': 9.792820327563028e-06, 'learning_rate/action_model': 9.792820327563028e-05, 'epoch': 0.09})
+[2026-07-28 02:36:16,740][starVLA.training.train_starvla][INFO] - Step 469, Loss: {'train/loss': 0.11525861918926239, 'train/grad_norm_pre_clip': 3.377593755722046, 'timing/data': 0.00029481295496225357, 'timing/model': 0.41710539488121867, 'learning_rate/qwen_vl_interface': 9.791701134859076e-06, 'learning_rate/action_model': 9.791701134859075e-05, 'epoch': 0.09})
+[2026-07-28 02:36:22,630][starVLA.training.train_starvla][INFO] - Step 470, Loss: {'train/loss': 0.09302090108394623, 'train/grad_norm_pre_clip': 1.6786043643951416, 'timing/data': 0.0002916748635470867, 'timing/model': 0.4172167028300464, 'learning_rate/qwen_vl_interface': 9.790578995094565e-06, 'learning_rate/action_model': 9.790578995094563e-05, 'epoch': 0.09})
+[2026-07-28 02:36:28,546][starVLA.training.train_starvla][INFO] - Step 471, Loss: {'train/loss': 0.10357896983623505, 'train/grad_norm_pre_clip': 2.2677860260009766, 'timing/data': 0.00029536522924900055, 'timing/model': 0.4175294269807637, 'learning_rate/qwen_vl_interface': 9.789453908997639e-06, 'learning_rate/action_model': 9.789453908997638e-05, 'epoch': 0.09})
+[2026-07-28 02:36:34,452][starVLA.training.train_starvla][INFO] - Step 472, Loss: {'train/loss': 0.11544051766395569, 'train/grad_norm_pre_clip': 5.044946193695068, 'timing/data': 0.00026262830942869186, 'timing/model': 0.4154480779543519, 'learning_rate/qwen_vl_interface': 9.788325877298355e-06, 'learning_rate/action_model': 9.788325877298355e-05, 'epoch': 0.09})
+[2026-07-28 02:36:40,143][starVLA.training.train_starvla][INFO] - Step 473, Loss: {'train/loss': 0.1057477593421936, 'train/grad_norm_pre_clip': 2.0897445678710938, 'timing/data': 0.0002816561609506607, 'timing/model': 0.41761932522058487, 'learning_rate/qwen_vl_interface': 9.787194900728682e-06, 'learning_rate/action_model': 9.787194900728683e-05, 'epoch': 0.09})
+[2026-07-28 02:36:46,025][starVLA.training.train_starvla][INFO] - Step 474, Loss: {'train/loss': 0.11283142864704132, 'train/grad_norm_pre_clip': 2.0363481044769287, 'timing/data': 0.00028590811416506767, 'timing/model': 0.4171669529750943, 'learning_rate/qwen_vl_interface': 9.786060980022499e-06, 'learning_rate/action_model': 9.786060980022498e-05, 'epoch': 0.09})
+[2026-07-28 02:36:52,340][starVLA.training.train_starvla][INFO] - Step 475, Loss: {'train/loss': 0.09595368057489395, 'train/grad_norm_pre_clip': 2.7649149894714355, 'timing/data': 0.0002654660493135452, 'timing/model': 0.6536494912579656, 'learning_rate/qwen_vl_interface': 9.784924115915592e-06, 'learning_rate/action_model': 9.784924115915592e-05, 'epoch': 0.09})
+[2026-07-28 02:36:58,244][starVLA.training.train_starvla][INFO] - Step 476, Loss: {'train/loss': 0.10174460709095001, 'train/grad_norm_pre_clip': 1.9470336437225342, 'timing/data': 0.000294641125947237, 'timing/model': 0.41739504411816597, 'learning_rate/qwen_vl_interface': 9.783784309145663e-06, 'learning_rate/action_model': 9.783784309145661e-05, 'epoch': 0.09})
+[2026-07-28 02:37:04,131][starVLA.training.train_starvla][INFO] - Step 477, Loss: {'train/loss': 0.0891772136092186, 'train/grad_norm_pre_clip': 3.8010709285736084, 'timing/data': 0.00027366960421204567, 'timing/model': 0.41625099908560514, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.09})
+[2026-07-28 02:37:10,020][starVLA.training.train_starvla][INFO] - Step 478, Loss: {'train/loss': 0.0549226850271225, 'train/grad_norm_pre_clip': 2.0368998050689697, 'timing/data': 0.00027949176728725433, 'timing/model': 0.42009413288906217, 'learning_rate/qwen_vl_interface': 9.78149587057707e-06, 'learning_rate/action_model': 9.78149587057707e-05, 'epoch': 0.09})
+[2026-07-28 02:37:16,142][starVLA.training.train_starvla][INFO] - Step 479, Loss: {'train/loss': 0.14786316454410553, 'train/grad_norm_pre_clip': 2.969370126724243, 'timing/data': 0.0002835686318576336, 'timing/model': 0.6537993713282049, 'learning_rate/qwen_vl_interface': 9.780347240263354e-06, 'learning_rate/action_model': 9.780347240263353e-05, 'epoch': 0.09})
+[2026-07-28 02:37:22,288][starVLA.training.train_starvla][INFO] - Step 480, Loss: {'train/loss': 0.15874475240707397, 'train/grad_norm_pre_clip': 4.790773868560791, 'timing/data': 0.00024852389469742775, 'timing/model': 0.6593480939045548, 'learning_rate/qwen_vl_interface': 9.779195670256498e-06, 'learning_rate/action_model': 9.779195670256497e-05, 'epoch': 0.09})
+[2026-07-28 02:37:28,189][starVLA.training.train_starvla][INFO] - Step 481, Loss: {'train/loss': 0.057665470987558365, 'train/grad_norm_pre_clip': 1.4935635328292847, 'timing/data': 0.0002512871287763119, 'timing/model': 0.4168736869469285, 'learning_rate/qwen_vl_interface': 9.778041161303746e-06, 'learning_rate/action_model': 9.778041161303745e-05, 'epoch': 0.09})
+[2026-07-28 02:37:34,055][starVLA.training.train_starvla][INFO] - Step 482, Loss: {'train/loss': 0.08708318322896957, 'train/grad_norm_pre_clip': 3.8556811809539795, 'timing/data': 0.0002864468842744827, 'timing/model': 0.4201223901472986, 'learning_rate/qwen_vl_interface': 9.776883714154242e-06, 'learning_rate/action_model': 9.776883714154241e-05, 'epoch': 0.09})
+[2026-07-28 02:37:39,991][starVLA.training.train_starvla][INFO] - Step 483, Loss: {'train/loss': 0.06567472964525223, 'train/grad_norm_pre_clip': 1.645392656326294, 'timing/data': 0.0002754051238298416, 'timing/model': 0.41871743090450764, 'learning_rate/qwen_vl_interface': 9.775723329559044e-06, 'learning_rate/action_model': 9.775723329559044e-05, 'epoch': 0.09})
+[2026-07-28 02:37:45,898][starVLA.training.train_starvla][INFO] - Step 484, Loss: {'train/loss': 0.06873995065689087, 'train/grad_norm_pre_clip': 1.764016032218933, 'timing/data': 0.00028319377452135086, 'timing/model': 0.41859625885263085, 'learning_rate/qwen_vl_interface': 9.774560008271114e-06, 'learning_rate/action_model': 9.774560008271114e-05, 'epoch': 0.09})
+[2026-07-28 02:37:52,037][starVLA.training.train_starvla][INFO] - Step 485, Loss: {'train/loss': 0.06267344951629639, 'train/grad_norm_pre_clip': 2.479412078857422, 'timing/data': 0.000250890851020813, 'timing/model': 0.4143270617350936, 'learning_rate/qwen_vl_interface': 9.773393751045315e-06, 'learning_rate/action_model': 9.773393751045315e-05, 'epoch': 0.09})
+[2026-07-28 02:37:58,121][starVLA.training.train_starvla][INFO] - Step 486, Loss: {'train/loss': 0.08400917798280716, 'train/grad_norm_pre_clip': 1.3530184030532837, 'timing/data': 0.00029595382511615753, 'timing/model': 0.6518769571557641, 'learning_rate/qwen_vl_interface': 9.772224558638422e-06, 'learning_rate/action_model': 9.772224558638422e-05, 'epoch': 0.09})
+[2026-07-28 02:38:04,010][starVLA.training.train_starvla][INFO] - Step 487, Loss: {'train/loss': 0.10568322241306305, 'train/grad_norm_pre_clip': 1.6918143033981323, 'timing/data': 0.0002468191087245941, 'timing/model': 0.4146352792158723, 'learning_rate/qwen_vl_interface': 9.77105243180911e-06, 'learning_rate/action_model': 9.771052431809109e-05, 'epoch': 0.09})
+[2026-07-28 02:38:09,855][starVLA.training.train_starvla][INFO] - Step 488, Loss: {'train/loss': 0.08108280599117279, 'train/grad_norm_pre_clip': 2.027409076690674, 'timing/data': 0.00028865598142147064, 'timing/model': 0.4166723908856511, 'learning_rate/qwen_vl_interface': 9.76987737131796e-06, 'learning_rate/action_model': 9.76987737131796e-05, 'epoch': 0.09})
+[2026-07-28 02:38:15,957][starVLA.training.train_starvla][INFO] - Step 489, Loss: {'train/loss': 0.07857252657413483, 'train/grad_norm_pre_clip': 1.465133547782898, 'timing/data': 0.0002877446822822094, 'timing/model': 0.41705123288556933, 'learning_rate/qwen_vl_interface': 9.768699377927454e-06, 'learning_rate/action_model': 9.768699377927453e-05, 'epoch': 0.09})
+[2026-07-28 02:38:21,827][starVLA.training.train_starvla][INFO] - Step 490, Loss: {'train/loss': 0.0763971135020256, 'train/grad_norm_pre_clip': 2.2187106609344482, 'timing/data': 0.00027573807165026665, 'timing/model': 0.41738647082820535, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.09})
+[2026-07-28 02:38:27,968][starVLA.training.train_starvla][INFO] - Step 491, Loss: {'train/loss': 0.07948313653469086, 'train/grad_norm_pre_clip': 2.869596004486084, 'timing/data': 0.0002527981996536255, 'timing/model': 0.4152069306001067, 'learning_rate/qwen_vl_interface': 9.76633459550783e-06, 'learning_rate/action_model': 9.76633459550783e-05, 'epoch': 0.09})
+[2026-07-28 02:38:34,103][starVLA.training.train_starvla][INFO] - Step 492, Loss: {'train/loss': 0.08192503452301025, 'train/grad_norm_pre_clip': 1.7420183420181274, 'timing/data': 0.00024871760979294777, 'timing/model': 0.4142103218473494, 'learning_rate/qwen_vl_interface': 9.765147808013193e-06, 'learning_rate/action_model': 9.765147808013192e-05, 'epoch': 0.09})
+[2026-07-28 02:38:39,969][starVLA.training.train_starvla][INFO] - Step 493, Loss: {'train/loss': 0.12881088256835938, 'train/grad_norm_pre_clip': 3.748445749282837, 'timing/data': 0.00028133299201726913, 'timing/model': 0.4165038038045168, 'learning_rate/qwen_vl_interface': 9.763958090688165e-06, 'learning_rate/action_model': 9.763958090688164e-05, 'epoch': 0.09})
+[2026-07-28 02:38:45,840][starVLA.training.train_starvla][INFO] - Step 494, Loss: {'train/loss': 0.11155300587415695, 'train/grad_norm_pre_clip': 3.354975700378418, 'timing/data': 0.0002876478247344494, 'timing/model': 0.41618803702294827, 'learning_rate/qwen_vl_interface': 9.762765444304737e-06, 'learning_rate/action_model': 9.762765444304737e-05, 'epoch': 0.09})
+[2026-07-28 02:38:52,109][starVLA.training.train_starvla][INFO] - Step 495, Loss: {'train/loss': 0.09371364861726761, 'train/grad_norm_pre_clip': 2.2401881217956543, 'timing/data': 0.00025988323614001274, 'timing/model': 0.6013783710077405, 'learning_rate/qwen_vl_interface': 9.761569869636808e-06, 'learning_rate/action_model': 9.761569869636808e-05, 'epoch': 0.09})
+[2026-07-28 02:38:57,996][starVLA.training.train_starvla][INFO] - Step 496, Loss: {'train/loss': 0.12093549221754074, 'train/grad_norm_pre_clip': 2.983205556869507, 'timing/data': 0.0002849227748811245, 'timing/model': 0.41545080579817295, 'learning_rate/qwen_vl_interface': 9.760371367460172e-06, 'learning_rate/action_model': 9.760371367460171e-05, 'epoch': 0.09})
+[2026-07-28 02:39:03,650][starVLA.training.train_starvla][INFO] - Step 497, Loss: {'train/loss': 0.08336656540632248, 'train/grad_norm_pre_clip': 1.3967574834823608, 'timing/data': 0.00033724820241332054, 'timing/model': 0.41779086692258716, 'learning_rate/qwen_vl_interface': 9.759169938552524e-06, 'learning_rate/action_model': 9.759169938552524e-05, 'epoch': 0.09})
+[2026-07-28 02:39:09,313][starVLA.training.train_starvla][INFO] - Step 498, Loss: {'train/loss': 0.13045096397399902, 'train/grad_norm_pre_clip': 3.8692095279693604, 'timing/data': 0.0002868599258363247, 'timing/model': 0.4187904172576964, 'learning_rate/qwen_vl_interface': 9.757965583693458e-06, 'learning_rate/action_model': 9.757965583693457e-05, 'epoch': 0.09})
+[2026-07-28 02:39:15,409][starVLA.training.train_starvla][INFO] - Step 499, Loss: {'train/loss': 0.08815070986747742, 'train/grad_norm_pre_clip': 2.818453550338745, 'timing/data': 0.00025420915335416794, 'timing/model': 0.4147537499666214, 'learning_rate/qwen_vl_interface': 9.756758303664467e-06, 'learning_rate/action_model': 9.756758303664468e-05, 'epoch': 0.09})
+[2026-07-28 02:39:21,532][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 500
+[2026-07-28 02:39:26,407][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 500
+[2026-07-28 02:41:31,017][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 500
+[2026-07-28 02:41:31,019][starVLA.training.train_starvla][INFO] - Step 500, Loss: {'train/loss': 0.08630054444074631, 'train/grad_norm_pre_clip': 1.5897154808044434, 'eval/action_loss/seconds': 4.872626513708383, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.08101459592580795, 'eval/latency_3/loss': 0.08101459592580795, 'eval/flappy/loss': 0.08101459592580795, 'eval/flappy/latency_3/loss': 0.08101459592580795, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 64.18500077836215, 'rl_games_eval/mid_train/mean_length': 536.3, 'rl_games_eval/mid_train/std_reward': 70.62001410319165, 'rl_games_eval/mid_train/std_length': 570.6692649862966, 'rl_games_eval/mid_train/macro_mean_reward': 64.18500077836215, 'rl_games_eval/mid_train/macro_mean_length': 536.3, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 64.18500077836215, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 536.3, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 70.62001410319165, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 570.6692649862966, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 64.18500077836215, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000267575029283762, 'timing/model': 0.6558567374013364, 'learning_rate/qwen_vl_interface': 9.755548099248942e-06, 'learning_rate/action_model': 9.75554809924894e-05, 'epoch': 0.09})
+[2026-07-28 02:41:31,021][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_500_state
+[2026-07-28 02:41:31,021][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 02:42:08,165][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_500_state/pytorch_model
+[2026-07-28 02:42:08,166][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_500_state/sampler.bin
+[2026-07-28 02:42:08,166][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_500_state/sampler_1.bin
+[2026-07-28 02:42:08,169][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_500_state/random_states_0.pkl
+[2026-07-28 02:42:16,145][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 500
+[2026-07-28 02:42:16,145][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 02:42:16,167][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 02:42:22,250][starVLA.training.train_starvla][INFO] - Step 501, Loss: {'train/loss': 0.08374081552028656, 'train/grad_norm_pre_clip': 1.1848093271255493, 'timing/data': 0.0002970462664961815, 'timing/model': 0.4237213069573045, 'learning_rate/qwen_vl_interface': 9.75433497123217e-06, 'learning_rate/action_model': 9.754334971232171e-05, 'epoch': 0.09})
+[2026-07-28 02:42:28,378][starVLA.training.train_starvla][INFO] - Step 502, Loss: {'train/loss': 0.09518608450889587, 'train/grad_norm_pre_clip': 1.4941622018814087, 'timing/data': 0.00025707390159368515, 'timing/model': 0.41428656270727515, 'learning_rate/qwen_vl_interface': 9.753118920401341e-06, 'learning_rate/action_model': 9.75311892040134e-05, 'epoch': 0.09})
+[2026-07-28 02:42:34,284][starVLA.training.train_starvla][INFO] - Step 503, Loss: {'train/loss': 0.06224970147013664, 'train/grad_norm_pre_clip': 2.4897384643554688, 'timing/data': 0.0002602129243314266, 'timing/model': 0.4160833419300616, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.09})
+[2026-07-28 02:42:40,407][starVLA.training.train_starvla][INFO] - Step 504, Loss: {'train/loss': 0.033691368997097015, 'train/grad_norm_pre_clip': 1.7002675533294678, 'timing/data': 0.00025610998272895813, 'timing/model': 0.41879451274871826, 'learning_rate/qwen_vl_interface': 9.750678053455725e-06, 'learning_rate/action_model': 9.750678053455725e-05, 'epoch': 0.09})
+[2026-07-28 02:42:46,565][starVLA.training.train_starvla][INFO] - Step 505, Loss: {'train/loss': 0.10324487835168839, 'train/grad_norm_pre_clip': 2.7132468223571777, 'timing/data': 0.0002696285955607891, 'timing/model': 0.6818249528296292, 'learning_rate/qwen_vl_interface': 9.749453238924793e-06, 'learning_rate/action_model': 9.749453238924792e-05, 'epoch': 0.09})
+[2026-07-28 02:42:52,417][starVLA.training.train_starvla][INFO] - Step 506, Loss: {'train/loss': 0.1077815592288971, 'train/grad_norm_pre_clip': 3.6513795852661133, 'timing/data': 0.000271407887339592, 'timing/model': 0.4155078297480941, 'learning_rate/qwen_vl_interface': 9.7482255047475e-06, 'learning_rate/action_model': 9.7482255047475e-05, 'epoch': 0.09})
+[2026-07-28 02:42:58,092][starVLA.training.train_starvla][INFO] - Step 507, Loss: {'train/loss': 0.06705398857593536, 'train/grad_norm_pre_clip': 2.5927793979644775, 'timing/data': 0.00035114120692014694, 'timing/model': 0.4169643372297287, 'learning_rate/qwen_vl_interface': 9.746994851720518e-06, 'learning_rate/action_model': 9.746994851720517e-05, 'epoch': 0.09})
+[2026-07-28 02:43:03,963][starVLA.training.train_starvla][INFO] - Step 508, Loss: {'train/loss': 0.0923258364200592, 'train/grad_norm_pre_clip': 4.93021297454834, 'timing/data': 0.0002734898589551449, 'timing/model': 0.41688676504418254, 'learning_rate/qwen_vl_interface': 9.745761280642397e-06, 'learning_rate/action_model': 9.745761280642397e-05, 'epoch': 0.09})
+[2026-07-28 02:43:10,050][starVLA.training.train_starvla][INFO] - Step 509, Loss: {'train/loss': 0.0674939975142479, 'train/grad_norm_pre_clip': 4.296432018280029, 'timing/data': 0.0002473238855600357, 'timing/model': 0.41413332521915436, 'learning_rate/qwen_vl_interface': 9.744524792313591e-06, 'learning_rate/action_model': 9.74452479231359e-05, 'epoch': 0.09})
+[2026-07-28 02:43:16,140][starVLA.training.train_starvla][INFO] - Step 510, Loss: {'train/loss': 0.17198817431926727, 'train/grad_norm_pre_clip': 3.9198782444000244, 'timing/data': 0.00024931831285357475, 'timing/model': 0.4165638736449182, 'learning_rate/qwen_vl_interface': 9.743285387536445e-06, 'learning_rate/action_model': 9.743285387536444e-05, 'epoch': 0.09})
+[2026-07-28 02:43:22,007][starVLA.training.train_starvla][INFO] - Step 511, Loss: {'train/loss': 0.13862526416778564, 'train/grad_norm_pre_clip': 8.476778984069824, 'timing/data': 0.00026983581483364105, 'timing/model': 0.4169581909663975, 'learning_rate/qwen_vl_interface': 9.742043067115193e-06, 'learning_rate/action_model': 9.742043067115192e-05, 'epoch': 0.09})
+[2026-07-28 02:43:28,119][starVLA.training.train_starvla][INFO] - Step 512, Loss: {'train/loss': 0.11737759411334991, 'train/grad_norm_pre_clip': 5.23016357421875, 'timing/data': 0.00025129178538918495, 'timing/model': 0.41519397497177124, 'learning_rate/qwen_vl_interface': 9.740797831855965e-06, 'learning_rate/action_model': 9.740797831855965e-05, 'epoch': 0.09})
+[2026-07-28 02:43:33,974][starVLA.training.train_starvla][INFO] - Step 513, Loss: {'train/loss': 0.1555211842060089, 'train/grad_norm_pre_clip': 4.926634311676025, 'timing/data': 0.00028273602947592735, 'timing/model': 0.4142073872499168, 'learning_rate/qwen_vl_interface': 9.73954968256678e-06, 'learning_rate/action_model': 9.73954968256678e-05, 'epoch': 0.09})
+[2026-07-28 02:43:39,841][starVLA.training.train_starvla][INFO] - Step 514, Loss: {'train/loss': 0.09369935095310211, 'train/grad_norm_pre_clip': 6.301610469818115, 'timing/data': 0.0002736188471317291, 'timing/model': 0.4156008381396532, 'learning_rate/qwen_vl_interface': 9.738298620057547e-06, 'learning_rate/action_model': 9.738298620057546e-05, 'epoch': 0.09})
+[2026-07-28 02:43:46,056][starVLA.training.train_starvla][INFO] - Step 515, Loss: {'train/loss': 0.13299024105072021, 'train/grad_norm_pre_clip': 3.354893922805786, 'timing/data': 0.00025748182088136673, 'timing/model': 0.5815941276960075, 'learning_rate/qwen_vl_interface': 9.73704464514007e-06, 'learning_rate/action_model': 9.737044645140068e-05, 'epoch': 0.09})
+[2026-07-28 02:43:51,914][starVLA.training.train_starvla][INFO] - Step 516, Loss: {'train/loss': 0.09086471796035767, 'train/grad_norm_pre_clip': 3.2668046951293945, 'timing/data': 0.000279801432043314, 'timing/model': 0.41659657191485167, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.09})
+[2026-07-28 02:43:57,825][starVLA.training.train_starvla][INFO] - Step 517, Loss: {'train/loss': 0.09892624616622925, 'train/grad_norm_pre_clip': 3.288167715072632, 'timing/data': 0.0003142780624330044, 'timing/model': 0.41761783976107836, 'learning_rate/qwen_vl_interface': 9.734527961337028e-06, 'learning_rate/action_model': 9.734527961337028e-05, 'epoch': 0.09})
+[2026-07-28 02:44:03,917][starVLA.training.train_starvla][INFO] - Step 518, Loss: {'train/loss': 0.07715947180986404, 'train/grad_norm_pre_clip': 3.206174373626709, 'timing/data': 0.00029037799686193466, 'timing/model': 0.6558980820700526, 'learning_rate/qwen_vl_interface': 9.733265254084514e-06, 'learning_rate/action_model': 9.733265254084513e-05, 'epoch': 0.1})
+[2026-07-28 02:44:10,015][starVLA.training.train_starvla][INFO] - Step 519, Loss: {'train/loss': 0.08652108907699585, 'train/grad_norm_pre_clip': 0.9937701225280762, 'timing/data': 0.0002618408761918545, 'timing/model': 0.4139891806989908, 'learning_rate/qwen_vl_interface': 9.73199963768985e-06, 'learning_rate/action_model': 9.731999637689849e-05, 'epoch': 0.1})
+[2026-07-28 02:44:15,875][starVLA.training.train_starvla][INFO] - Step 520, Loss: {'train/loss': 0.13180744647979736, 'train/grad_norm_pre_clip': 2.0810272693634033, 'timing/data': 0.00029393890872597694, 'timing/model': 0.41664565075188875, 'learning_rate/qwen_vl_interface': 9.73073111297428e-06, 'learning_rate/action_model': 9.73073111297428e-05, 'epoch': 0.1})
+[2026-07-28 02:44:21,525][starVLA.training.train_starvla][INFO] - Step 521, Loss: {'train/loss': 0.09361258894205093, 'train/grad_norm_pre_clip': 1.2939800024032593, 'timing/data': 0.00027345679700374603, 'timing/model': 0.4156064800918102, 'learning_rate/qwen_vl_interface': 9.72945968076094e-06, 'learning_rate/action_model': 9.729459680760939e-05, 'epoch': 0.1})
+[2026-07-28 02:44:27,153][starVLA.training.train_starvla][INFO] - Step 522, Loss: {'train/loss': 0.08492386341094971, 'train/grad_norm_pre_clip': 1.3383651971817017, 'timing/data': 0.0002710600383579731, 'timing/model': 0.41722250590100884, 'learning_rate/qwen_vl_interface': 9.728185341874844e-06, 'learning_rate/action_model': 9.728185341874844e-05, 'epoch': 0.1})
+[2026-07-28 02:44:33,020][starVLA.training.train_starvla][INFO] - Step 523, Loss: {'train/loss': 0.08299082517623901, 'train/grad_norm_pre_clip': 2.472186803817749, 'timing/data': 0.00027899211272597313, 'timing/model': 0.655748344026506, 'learning_rate/qwen_vl_interface': 9.7269080971429e-06, 'learning_rate/action_model': 9.726908097142899e-05, 'epoch': 0.1})
+[2026-07-28 02:44:38,683][starVLA.training.train_starvla][INFO] - Step 524, Loss: {'train/loss': 0.14388251304626465, 'train/grad_norm_pre_clip': 2.659360885620117, 'timing/data': 0.00035182293504476547, 'timing/model': 0.414328427053988, 'learning_rate/qwen_vl_interface': 9.725627947393896e-06, 'learning_rate/action_model': 9.725627947393896e-05, 'epoch': 0.1})
+[2026-07-28 02:44:44,513][starVLA.training.train_starvla][INFO] - Step 525, Loss: {'train/loss': 0.0827101320028305, 'train/grad_norm_pre_clip': 2.6307897567749023, 'timing/data': 0.0002958248369395733, 'timing/model': 0.4160696417093277, 'learning_rate/qwen_vl_interface': 9.72434489345851e-06, 'learning_rate/action_model': 9.72434489345851e-05, 'epoch': 0.1})
+[2026-07-28 02:44:50,384][starVLA.training.train_starvla][INFO] - Step 526, Loss: {'train/loss': 0.0823962539434433, 'train/grad_norm_pre_clip': 2.7532386779785156, 'timing/data': 0.00027135713025927544, 'timing/model': 0.41787027614191175, 'learning_rate/qwen_vl_interface': 9.723058936169298e-06, 'learning_rate/action_model': 9.723058936169297e-05, 'epoch': 0.1})
+[2026-07-28 02:44:56,256][starVLA.training.train_starvla][INFO] - Step 527, Loss: {'train/loss': 0.07405302673578262, 'train/grad_norm_pre_clip': 2.5991740226745605, 'timing/data': 0.00030165771022439003, 'timing/model': 0.4167041629552841, 'learning_rate/qwen_vl_interface': 9.721770076360707e-06, 'learning_rate/action_model': 9.721770076360706e-05, 'epoch': 0.1})
+[2026-07-28 02:45:02,107][starVLA.training.train_starvla][INFO] - Step 528, Loss: {'train/loss': 0.0661666989326477, 'train/grad_norm_pre_clip': 1.2119029760360718, 'timing/data': 0.0002895002253353596, 'timing/model': 0.41714289505034685, 'learning_rate/qwen_vl_interface': 9.720478314869061e-06, 'learning_rate/action_model': 9.72047831486906e-05, 'epoch': 0.1})
+[2026-07-28 02:45:08,428][starVLA.training.train_starvla][INFO] - Step 529, Loss: {'train/loss': 0.1633448302745819, 'train/grad_norm_pre_clip': 6.591625213623047, 'timing/data': 0.0002680630423128605, 'timing/model': 0.6472788229584694, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.1})
+[2026-07-28 02:45:14,270][starVLA.training.train_starvla][INFO] - Step 530, Loss: {'train/loss': 0.14778609573841095, 'train/grad_norm_pre_clip': 4.270474433898926, 'timing/data': 0.00025367410853505135, 'timing/model': 0.41596690099686384, 'learning_rate/qwen_vl_interface': 9.717886090191334e-06, 'learning_rate/action_model': 9.717886090191333e-05, 'epoch': 0.1})
+[2026-07-28 02:45:20,172][starVLA.training.train_starvla][INFO] - Step 531, Loss: {'train/loss': 0.08088192343711853, 'train/grad_norm_pre_clip': 4.22953462600708, 'timing/data': 0.00034873513504862785, 'timing/model': 0.4159615682438016, 'learning_rate/qwen_vl_interface': 9.716585628687317e-06, 'learning_rate/action_model': 9.716585628687316e-05, 'epoch': 0.1})
+[2026-07-28 02:45:26,065][starVLA.training.train_starvla][INFO] - Step 532, Loss: {'train/loss': 0.09678751230239868, 'train/grad_norm_pre_clip': 2.7979917526245117, 'timing/data': 0.0002713068388402462, 'timing/model': 0.4155903710052371, 'learning_rate/qwen_vl_interface': 9.715282268864379e-06, 'learning_rate/action_model': 9.715282268864378e-05, 'epoch': 0.1})
+[2026-07-28 02:45:32,287][starVLA.training.train_starvla][INFO] - Step 533, Loss: {'train/loss': 0.08205641806125641, 'train/grad_norm_pre_clip': 1.3441756963729858, 'timing/data': 0.0002543828450143337, 'timing/model': 0.4144121571443975, 'learning_rate/qwen_vl_interface': 9.713976011568256e-06, 'learning_rate/action_model': 9.713976011568256e-05, 'epoch': 0.1})
+[2026-07-28 02:45:38,398][starVLA.training.train_starvla][INFO] - Step 534, Loss: {'train/loss': 0.10174541175365448, 'train/grad_norm_pre_clip': 2.3847317695617676, 'timing/data': 0.0002535260282456875, 'timing/model': 0.41650292091071606, 'learning_rate/qwen_vl_interface': 9.712666857646565e-06, 'learning_rate/action_model': 9.712666857646564e-05, 'epoch': 0.1})
+[2026-07-28 02:45:44,254][starVLA.training.train_starvla][INFO] - Step 535, Loss: {'train/loss': 0.08034518361091614, 'train/grad_norm_pre_clip': 1.1462985277175903, 'timing/data': 0.00029943184927105904, 'timing/model': 0.4169638962484896, 'learning_rate/qwen_vl_interface': 9.711354807948798e-06, 'learning_rate/action_model': 9.711354807948798e-05, 'epoch': 0.1})
+[2026-07-28 02:45:50,333][starVLA.training.train_starvla][INFO] - Step 536, Loss: {'train/loss': 0.06721891462802887, 'train/grad_norm_pre_clip': 2.20528244972229, 'timing/data': 0.00027008121833205223, 'timing/model': 0.649792576674372, 'learning_rate/qwen_vl_interface': 9.710039863326334e-06, 'learning_rate/action_model': 9.710039863326334e-05, 'epoch': 0.1})
+[2026-07-28 02:45:56,175][starVLA.training.train_starvla][INFO] - Step 537, Loss: {'train/loss': 0.06863826513290405, 'train/grad_norm_pre_clip': 0.919116735458374, 'timing/data': 0.0002765129320323467, 'timing/model': 0.41711516212671995, 'learning_rate/qwen_vl_interface': 9.708722024632425e-06, 'learning_rate/action_model': 9.708722024632424e-05, 'epoch': 0.1})
+[2026-07-28 02:46:02,274][starVLA.training.train_starvla][INFO] - Step 538, Loss: {'train/loss': 0.05162244290113449, 'train/grad_norm_pre_clip': 1.6950428485870361, 'timing/data': 0.00025627901777625084, 'timing/model': 0.41617720341309905, 'learning_rate/qwen_vl_interface': 9.7074012927222e-06, 'learning_rate/action_model': 9.7074012927222e-05, 'epoch': 0.1})
+[2026-07-28 02:46:08,138][starVLA.training.train_starvla][INFO] - Step 539, Loss: {'train/loss': 0.10316258668899536, 'train/grad_norm_pre_clip': 4.006539821624756, 'timing/data': 0.0002787481062114239, 'timing/model': 0.41683713300153613, 'learning_rate/qwen_vl_interface': 9.706077668452671e-06, 'learning_rate/action_model': 9.706077668452671e-05, 'epoch': 0.1})
+[2026-07-28 02:46:14,228][starVLA.training.train_starvla][INFO] - Step 540, Loss: {'train/loss': 0.08331907540559769, 'train/grad_norm_pre_clip': 1.7190546989440918, 'timing/data': 0.00025587715208530426, 'timing/model': 0.41437281016260386, 'learning_rate/qwen_vl_interface': 9.704751152682722e-06, 'learning_rate/action_model': 9.704751152682721e-05, 'epoch': 0.1})
+[2026-07-28 02:46:20,090][starVLA.training.train_starvla][INFO] - Step 541, Loss: {'train/loss': 0.08352712541818619, 'train/grad_norm_pre_clip': 11.695413589477539, 'timing/data': 0.0002794060856103897, 'timing/model': 0.41666119964793324, 'learning_rate/qwen_vl_interface': 9.703421746273115e-06, 'learning_rate/action_model': 9.703421746273114e-05, 'epoch': 0.1})
+[2026-07-28 02:46:25,961][starVLA.training.train_starvla][INFO] - Step 542, Loss: {'train/loss': 0.12201599776744843, 'train/grad_norm_pre_clip': 2.9217183589935303, 'timing/data': 0.00027430523186922073, 'timing/model': 0.41576972510665655, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.1})
+[2026-07-28 02:46:32,294][starVLA.training.train_starvla][INFO] - Step 543, Loss: {'train/loss': 0.10862938314676285, 'train/grad_norm_pre_clip': 2.430661916732788, 'timing/data': 0.0002585500478744507, 'timing/model': 0.6471605990082026, 'learning_rate/qwen_vl_interface': 9.700754264987347e-06, 'learning_rate/action_model': 9.700754264987347e-05, 'epoch': 0.1})
+[2026-07-28 02:46:38,179][starVLA.training.train_starvla][INFO] - Step 544, Loss: {'train/loss': 0.09648852795362473, 'train/grad_norm_pre_clip': 4.464982509613037, 'timing/data': 0.0004078410565853119, 'timing/model': 0.41399119794368744, 'learning_rate/qwen_vl_interface': 9.69941619184209e-06, 'learning_rate/action_model': 9.699416191842089e-05, 'epoch': 0.1})
+[2026-07-28 02:46:44,042][starVLA.training.train_starvla][INFO] - Step 545, Loss: {'train/loss': 0.14476129412651062, 'train/grad_norm_pre_clip': 10.51608657836914, 'timing/data': 0.00026750192046165466, 'timing/model': 0.4166540182195604, 'learning_rate/qwen_vl_interface': 9.69807523151897e-06, 'learning_rate/action_model': 9.698075231518969e-05, 'epoch': 0.1})
+[2026-07-28 02:46:50,122][starVLA.training.train_starvla][INFO] - Step 546, Loss: {'train/loss': 0.08442400395870209, 'train/grad_norm_pre_clip': 1.9676889181137085, 'timing/data': 0.0002948208712041378, 'timing/model': 0.6322895209304988, 'learning_rate/qwen_vl_interface': 9.696731384888125e-06, 'learning_rate/action_model': 9.696731384888125e-05, 'epoch': 0.1})
+[2026-07-28 02:46:55,992][starVLA.training.train_starvla][INFO] - Step 547, Loss: {'train/loss': 0.07860037684440613, 'train/grad_norm_pre_clip': 1.9897472858428955, 'timing/data': 0.00027680816128849983, 'timing/model': 0.4170688516460359, 'learning_rate/qwen_vl_interface': 9.695384652821562e-06, 'learning_rate/action_model': 9.695384652821561e-05, 'epoch': 0.1})
+[2026-07-28 02:47:01,879][starVLA.training.train_starvla][INFO] - Step 548, Loss: {'train/loss': 0.10783441364765167, 'train/grad_norm_pre_clip': 2.271843433380127, 'timing/data': 0.00028730323538184166, 'timing/model': 0.41662088176235557, 'learning_rate/qwen_vl_interface': 9.694035036193162e-06, 'learning_rate/action_model': 9.694035036193161e-05, 'epoch': 0.1})
+[2026-07-28 02:47:07,778][starVLA.training.train_starvla][INFO] - Step 549, Loss: {'train/loss': 0.0777149647474289, 'train/grad_norm_pre_clip': 1.8973478078842163, 'timing/data': 0.0002803080715239048, 'timing/model': 0.4169776546768844, 'learning_rate/qwen_vl_interface': 9.692682535878674e-06, 'learning_rate/action_model': 9.692682535878673e-05, 'epoch': 0.1})
+[2026-07-28 02:47:13,909][starVLA.training.train_starvla][INFO] - Step 550, Loss: {'train/loss': 0.08782389760017395, 'train/grad_norm_pre_clip': 2.7655105590820312, 'timing/data': 0.00032242992892861366, 'timing/model': 0.4144021878018975, 'learning_rate/qwen_vl_interface': 9.691327152755721e-06, 'learning_rate/action_model': 9.691327152755722e-05, 'epoch': 0.1})
+[2026-07-28 02:47:19,809][starVLA.training.train_starvla][INFO] - Step 551, Loss: {'train/loss': 0.09804562479257584, 'train/grad_norm_pre_clip': 1.8789215087890625, 'timing/data': 0.00035162828862667084, 'timing/model': 0.42112355679273605, 'learning_rate/qwen_vl_interface': 9.6899688877038e-06, 'learning_rate/action_model': 9.689968887703798e-05, 'epoch': 0.1})
+[2026-07-28 02:47:25,670][starVLA.training.train_starvla][INFO] - Step 552, Loss: {'train/loss': 0.06854349374771118, 'train/grad_norm_pre_clip': 1.4275356531143188, 'timing/data': 0.0002643708139657974, 'timing/model': 0.4184979749843478, 'learning_rate/qwen_vl_interface': 9.688607741604269e-06, 'learning_rate/action_model': 9.688607741604269e-05, 'epoch': 0.1})
+[2026-07-28 02:47:31,512][starVLA.training.train_starvla][INFO] - Step 553, Loss: {'train/loss': 0.0833391547203064, 'train/grad_norm_pre_clip': 1.2056025266647339, 'timing/data': 0.00028017908334732056, 'timing/model': 0.41581684397533536, 'learning_rate/qwen_vl_interface': 9.687243715340365e-06, 'learning_rate/action_model': 9.687243715340364e-05, 'epoch': 0.1})
+[2026-07-28 02:47:37,830][starVLA.training.train_starvla][INFO] - Step 554, Loss: {'train/loss': 0.09792853891849518, 'train/grad_norm_pre_clip': 3.341048002243042, 'timing/data': 0.00025837216526269913, 'timing/model': 0.6465137409977615, 'learning_rate/qwen_vl_interface': 9.685876809797188e-06, 'learning_rate/action_model': 9.685876809797188e-05, 'epoch': 0.1})
+[2026-07-28 02:47:43,467][starVLA.training.train_starvla][INFO] - Step 555, Loss: {'train/loss': 0.08671039342880249, 'train/grad_norm_pre_clip': 5.152888774871826, 'timing/data': 0.00034490227699279785, 'timing/model': 0.4190342277288437, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.1})
+[2026-07-28 02:47:49,631][starVLA.training.train_starvla][INFO] - Step 556, Loss: {'train/loss': 0.07207606732845306, 'train/grad_norm_pre_clip': 2.1397037506103516, 'timing/data': 0.0002773487940430641, 'timing/model': 0.4163538538850844, 'learning_rate/qwen_vl_interface': 9.683134364422766e-06, 'learning_rate/action_model': 9.683134364422765e-05, 'epoch': 0.1})
+[2026-07-28 02:47:55,527][starVLA.training.train_starvla][INFO] - Step 557, Loss: {'train/loss': 0.11373430490493774, 'train/grad_norm_pre_clip': 1.7917041778564453, 'timing/data': 0.00026832008734345436, 'timing/model': 0.41747067496180534, 'learning_rate/qwen_vl_interface': 9.681758826371064e-06, 'learning_rate/action_model': 9.681758826371063e-05, 'epoch': 0.1})
+[2026-07-28 02:48:01,643][starVLA.training.train_starvla][INFO] - Step 558, Loss: {'train/loss': 0.12215468287467957, 'train/grad_norm_pre_clip': 4.753739833831787, 'timing/data': 0.00025966810062527657, 'timing/model': 0.4141756482422352, 'learning_rate/qwen_vl_interface': 9.680380412599174e-06, 'learning_rate/action_model': 9.680380412599173e-05, 'epoch': 0.1})
+[2026-07-28 02:48:07,785][starVLA.training.train_starvla][INFO] - Step 559, Loss: {'train/loss': 0.05359836667776108, 'train/grad_norm_pre_clip': 1.5462168455123901, 'timing/data': 0.0003571747802197933, 'timing/model': 0.6485216319561005, 'learning_rate/qwen_vl_interface': 9.678999124001533e-06, 'learning_rate/action_model': 9.678999124001533e-05, 'epoch': 0.1})
+[2026-07-28 02:48:13,697][starVLA.training.train_starvla][INFO] - Step 560, Loss: {'train/loss': 0.07892004400491714, 'train/grad_norm_pre_clip': 2.1142935752868652, 'timing/data': 0.00028132088482379913, 'timing/model': 0.4154429459013045, 'learning_rate/qwen_vl_interface': 9.67761496147445e-06, 'learning_rate/action_model': 9.677614961474449e-05, 'epoch': 0.1})
+[2026-07-28 02:48:19,590][starVLA.training.train_starvla][INFO] - Step 561, Loss: {'train/loss': 0.1138206273317337, 'train/grad_norm_pre_clip': 2.095578670501709, 'timing/data': 0.00033712899312376976, 'timing/model': 0.41877256892621517, 'learning_rate/qwen_vl_interface': 9.676227925916082e-06, 'learning_rate/action_model': 9.676227925916082e-05, 'epoch': 0.1})
+[2026-07-28 02:48:25,727][starVLA.training.train_starvla][INFO] - Step 562, Loss: {'train/loss': 0.08618734776973724, 'train/grad_norm_pre_clip': 2.0126945972442627, 'timing/data': 0.00026428699493408203, 'timing/model': 0.41550912288948894, 'learning_rate/qwen_vl_interface': 9.674838018226472e-06, 'learning_rate/action_model': 9.674838018226471e-05, 'epoch': 0.1})
+[2026-07-28 02:48:31,598][starVLA.training.train_starvla][INFO] - Step 563, Loss: {'train/loss': 0.11161211133003235, 'train/grad_norm_pre_clip': 4.293213367462158, 'timing/data': 0.00037417467683553696, 'timing/model': 0.41597200697287917, 'learning_rate/qwen_vl_interface': 9.67344523930751e-06, 'learning_rate/action_model': 9.673445239307511e-05, 'epoch': 0.1})
+[2026-07-28 02:48:37,746][starVLA.training.train_starvla][INFO] - Step 564, Loss: {'train/loss': 0.08788225799798965, 'train/grad_norm_pre_clip': 4.372175693511963, 'timing/data': 0.00026018405333161354, 'timing/model': 0.4152990961447358, 'learning_rate/qwen_vl_interface': 9.672049590062959e-06, 'learning_rate/action_model': 9.672049590062958e-05, 'epoch': 0.1})
+[2026-07-28 02:48:43,866][starVLA.training.train_starvla][INFO] - Step 565, Loss: {'train/loss': 0.07677562534809113, 'train/grad_norm_pre_clip': 2.6866495609283447, 'timing/data': 0.0003575221635401249, 'timing/model': 0.6546446867287159, 'learning_rate/qwen_vl_interface': 9.670651071398436e-06, 'learning_rate/action_model': 9.670651071398436e-05, 'epoch': 0.1})
+[2026-07-28 02:48:49,715][starVLA.training.train_starvla][INFO] - Step 566, Loss: {'train/loss': 0.11965645849704742, 'train/grad_norm_pre_clip': 2.5117666721343994, 'timing/data': 0.0003526504151523113, 'timing/model': 0.4159838082268834, 'learning_rate/qwen_vl_interface': 9.669249684221427e-06, 'learning_rate/action_model': 9.669249684221428e-05, 'epoch': 0.1})
+[2026-07-28 02:48:55,810][starVLA.training.train_starvla][INFO] - Step 567, Loss: {'train/loss': 0.03434024378657341, 'train/grad_norm_pre_clip': 0.8075659871101379, 'timing/data': 0.00026662973687052727, 'timing/model': 0.4152870289981365, 'learning_rate/qwen_vl_interface': 9.66784542944128e-06, 'learning_rate/action_model': 9.66784542944128e-05, 'epoch': 0.1})
+[2026-07-28 02:49:01,664][starVLA.training.train_starvla][INFO] - Step 568, Loss: {'train/loss': 0.11772380769252777, 'train/grad_norm_pre_clip': 2.974327325820923, 'timing/data': 0.00026988890022039413, 'timing/model': 0.4150559105910361, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.1})
+[2026-07-28 02:49:07,510][starVLA.training.train_starvla][INFO] - Step 569, Loss: {'train/loss': 0.08098883181810379, 'train/grad_norm_pre_clip': 1.4501687288284302, 'timing/data': 0.0003642980009317398, 'timing/model': 0.4168802029453218, 'learning_rate/qwen_vl_interface': 9.66502832071824e-06, 'learning_rate/action_model': 9.66502832071824e-05, 'epoch': 0.1})
+[2026-07-28 02:49:13,623][starVLA.training.train_starvla][INFO] - Step 570, Loss: {'train/loss': 0.10767132043838501, 'train/grad_norm_pre_clip': 2.4879567623138428, 'timing/data': 0.00026516010984778404, 'timing/model': 0.6608027420006692, 'learning_rate/qwen_vl_interface': 9.663615468603344e-06, 'learning_rate/action_model': 9.663615468603343e-05, 'epoch': 0.1})
+[2026-07-28 02:49:19,717][starVLA.training.train_starvla][INFO] - Step 571, Loss: {'train/loss': 0.0683501735329628, 'train/grad_norm_pre_clip': 1.4925745725631714, 'timing/data': 0.00029936665669083595, 'timing/model': 0.6436626962386072, 'learning_rate/qwen_vl_interface': 9.662199752541287e-06, 'learning_rate/action_model': 9.662199752541287e-05, 'epoch': 0.1})
+[2026-07-28 02:49:25,554][starVLA.training.train_starvla][INFO] - Step 572, Loss: {'train/loss': 0.06620795279741287, 'train/grad_norm_pre_clip': 1.9255638122558594, 'timing/data': 0.0002598571591079235, 'timing/model': 0.41340891318395734, 'learning_rate/qwen_vl_interface': 9.660781173450711e-06, 'learning_rate/action_model': 9.660781173450712e-05, 'epoch': 0.11})
+[2026-07-28 02:49:31,391][starVLA.training.train_starvla][INFO] - Step 573, Loss: {'train/loss': 0.09967166930437088, 'train/grad_norm_pre_clip': 4.286738395690918, 'timing/data': 0.0003514229319989681, 'timing/model': 0.4162815138697624, 'learning_rate/qwen_vl_interface': 9.659359732252121e-06, 'learning_rate/action_model': 9.659359732252121e-05, 'epoch': 0.11})
+[2026-07-28 02:49:37,238][starVLA.training.train_starvla][INFO] - Step 574, Loss: {'train/loss': 0.06331813335418701, 'train/grad_norm_pre_clip': 2.356590509414673, 'timing/data': 0.00028496095910668373, 'timing/model': 0.4169731358997524, 'learning_rate/qwen_vl_interface': 9.657935429867872e-06, 'learning_rate/action_model': 9.657935429867873e-05, 'epoch': 0.11})
+[2026-07-28 02:49:42,872][starVLA.training.train_starvla][INFO] - Step 575, Loss: {'train/loss': 0.10086353123188019, 'train/grad_norm_pre_clip': 2.0362555980682373, 'timing/data': 0.00027276715263724327, 'timing/model': 0.4164272830821574, 'learning_rate/qwen_vl_interface': 9.65650826722218e-06, 'learning_rate/action_model': 9.656508267222179e-05, 'epoch': 0.11})
+[2026-07-28 02:49:48,736][starVLA.training.train_starvla][INFO] - Step 576, Loss: {'train/loss': 0.11578899621963501, 'train/grad_norm_pre_clip': 3.52405047416687, 'timing/data': 0.0002833898179233074, 'timing/model': 0.6538141053169966, 'learning_rate/qwen_vl_interface': 9.655078245241114e-06, 'learning_rate/action_model': 9.655078245241112e-05, 'epoch': 0.11})
+[2026-07-28 02:49:54,588][starVLA.training.train_starvla][INFO] - Step 577, Loss: {'train/loss': 0.0835503339767456, 'train/grad_norm_pre_clip': 2.472938060760498, 'timing/data': 0.0002635810524225235, 'timing/model': 0.413837821688503, 'learning_rate/qwen_vl_interface': 9.653645364852598e-06, 'learning_rate/action_model': 9.653645364852597e-05, 'epoch': 0.11})
+[2026-07-28 02:50:00,217][starVLA.training.train_starvla][INFO] - Step 578, Loss: {'train/loss': 0.10830441862344742, 'train/grad_norm_pre_clip': 2.479173421859741, 'timing/data': 0.00036679208278656006, 'timing/model': 0.4200836247764528, 'learning_rate/qwen_vl_interface': 9.652209626986414e-06, 'learning_rate/action_model': 9.652209626986413e-05, 'epoch': 0.11})
+[2026-07-28 02:50:05,849][starVLA.training.train_starvla][INFO] - Step 579, Loss: {'train/loss': 0.06511010229587555, 'train/grad_norm_pre_clip': 2.2627339363098145, 'timing/data': 0.0002882680855691433, 'timing/model': 0.4167865626513958, 'learning_rate/qwen_vl_interface': 9.6507710325742e-06, 'learning_rate/action_model': 9.6507710325742e-05, 'epoch': 0.11})
+[2026-07-28 02:50:11,679][starVLA.training.train_starvla][INFO] - Step 580, Loss: {'train/loss': 0.058260269463062286, 'train/grad_norm_pre_clip': 2.6345767974853516, 'timing/data': 0.0003628791309893131, 'timing/model': 0.41645543929189444, 'learning_rate/qwen_vl_interface': 9.649329582549439e-06, 'learning_rate/action_model': 9.649329582549439e-05, 'epoch': 0.11})
+[2026-07-28 02:50:18,006][starVLA.training.train_starvla][INFO] - Step 581, Loss: {'train/loss': 0.07450360059738159, 'train/grad_norm_pre_clip': 2.8072850704193115, 'timing/data': 0.0002636718563735485, 'timing/model': 0.41599466698244214, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.11})
+[2026-07-28 02:50:23,845][starVLA.training.train_starvla][INFO] - Step 582, Loss: {'train/loss': 0.08678524196147919, 'train/grad_norm_pre_clip': 1.2322015762329102, 'timing/data': 0.0002899877727031708, 'timing/model': 0.41711689764633775, 'learning_rate/qwen_vl_interface': 9.646438119405503e-06, 'learning_rate/action_model': 9.646438119405503e-05, 'epoch': 0.11})
+[2026-07-28 02:50:29,708][starVLA.training.train_starvla][INFO] - Step 583, Loss: {'train/loss': 0.055899932980537415, 'train/grad_norm_pre_clip': 1.8814560174942017, 'timing/data': 0.00027351127937436104, 'timing/model': 0.41641825437545776, 'learning_rate/qwen_vl_interface': 9.644988108162566e-06, 'learning_rate/action_model': 9.644988108162567e-05, 'epoch': 0.11})
+[2026-07-28 02:50:35,803][starVLA.training.train_starvla][INFO] - Step 584, Loss: {'train/loss': 0.10867367684841156, 'train/grad_norm_pre_clip': 1.3625062704086304, 'timing/data': 0.00026864372193813324, 'timing/model': 0.4158223890699446, 'learning_rate/qwen_vl_interface': 9.643535245059562e-06, 'learning_rate/action_model': 9.643535245059561e-05, 'epoch': 0.11})
+[2026-07-28 02:50:41,908][starVLA.training.train_starvla][INFO] - Step 585, Loss: {'train/loss': 0.07058936357498169, 'train/grad_norm_pre_clip': 1.3642945289611816, 'timing/data': 0.0002645188942551613, 'timing/model': 0.4202322168275714, 'learning_rate/qwen_vl_interface': 9.642079531039235e-06, 'learning_rate/action_model': 9.642079531039235e-05, 'epoch': 0.11})
+[2026-07-28 02:50:47,770][starVLA.training.train_starvla][INFO] - Step 586, Loss: {'train/loss': 0.062248699367046356, 'train/grad_norm_pre_clip': 2.9997880458831787, 'timing/data': 0.00028264569118618965, 'timing/model': 0.41702666506171227, 'learning_rate/qwen_vl_interface': 9.640620967046186e-06, 'learning_rate/action_model': 9.640620967046186e-05, 'epoch': 0.11})
+[2026-07-28 02:50:53,785][starVLA.training.train_starvla][INFO] - Step 587, Loss: {'train/loss': 0.020778581500053406, 'train/grad_norm_pre_clip': 1.001367211341858, 'timing/data': 0.00028565945103764534, 'timing/model': 0.5872258129529655, 'learning_rate/qwen_vl_interface': 9.63915955402686e-06, 'learning_rate/action_model': 9.63915955402686e-05, 'epoch': 0.11})
+[2026-07-28 02:50:59,680][starVLA.training.train_starvla][INFO] - Step 588, Loss: {'train/loss': 0.07308150827884674, 'train/grad_norm_pre_clip': 2.080142021179199, 'timing/data': 0.00028325291350483894, 'timing/model': 0.41633265698328614, 'learning_rate/qwen_vl_interface': 9.637695292929552e-06, 'learning_rate/action_model': 9.637695292929552e-05, 'epoch': 0.11})
+[2026-07-28 02:51:05,570][starVLA.training.train_starvla][INFO] - Step 589, Loss: {'train/loss': 0.11188772320747375, 'train/grad_norm_pre_clip': 4.467380523681641, 'timing/data': 0.0003558560274541378, 'timing/model': 0.4169723242521286, 'learning_rate/qwen_vl_interface': 9.636228184704405e-06, 'learning_rate/action_model': 9.636228184704405e-05, 'epoch': 0.11})
+[2026-07-28 02:51:11,699][starVLA.training.train_starvla][INFO] - Step 590, Loss: {'train/loss': 0.061043255031108856, 'train/grad_norm_pre_clip': 3.0641796588897705, 'timing/data': 0.00035759806632995605, 'timing/model': 0.6599521790631115, 'learning_rate/qwen_vl_interface': 9.634758230303408e-06, 'learning_rate/action_model': 9.634758230303408e-05, 'epoch': 0.11})
+[2026-07-28 02:51:17,809][starVLA.training.train_starvla][INFO] - Step 591, Loss: {'train/loss': 0.06479081511497498, 'train/grad_norm_pre_clip': 4.734361171722412, 'timing/data': 0.00027476483955979347, 'timing/model': 0.4138619899749756, 'learning_rate/qwen_vl_interface': 9.633285430680401e-06, 'learning_rate/action_model': 9.633285430680401e-05, 'epoch': 0.11})
+[2026-07-28 02:51:23,659][starVLA.training.train_starvla][INFO] - Step 592, Loss: {'train/loss': 0.13140830397605896, 'train/grad_norm_pre_clip': 2.8487789630889893, 'timing/data': 0.0003351578488945961, 'timing/model': 0.42015627678483725, 'learning_rate/qwen_vl_interface': 9.631809786791066e-06, 'learning_rate/action_model': 9.631809786791066e-05, 'epoch': 0.11})
+[2026-07-28 02:51:29,530][starVLA.training.train_starvla][INFO] - Step 593, Loss: {'train/loss': 0.0735972672700882, 'train/grad_norm_pre_clip': 2.3851683139801025, 'timing/data': 0.00034120818600058556, 'timing/model': 0.41740273125469685, 'learning_rate/qwen_vl_interface': 9.630331299592932e-06, 'learning_rate/action_model': 9.630331299592933e-05, 'epoch': 0.11})
+[2026-07-28 02:51:35,371][starVLA.training.train_starvla][INFO] - Step 594, Loss: {'train/loss': 0.09033606946468353, 'train/grad_norm_pre_clip': 3.177952289581299, 'timing/data': 0.0003536720760166645, 'timing/model': 0.4176860717125237, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.11})
+[2026-07-28 02:51:41,474][starVLA.training.train_starvla][INFO] - Step 595, Loss: {'train/loss': 0.15211036801338196, 'train/grad_norm_pre_clip': 3.64674973487854, 'timing/data': 0.00031828414648771286, 'timing/model': 0.4151941267773509, 'learning_rate/qwen_vl_interface': 9.627365799109617e-06, 'learning_rate/action_model': 9.627365799109617e-05, 'epoch': 0.11})
+[2026-07-28 02:51:47,335][starVLA.training.train_starvla][INFO] - Step 596, Loss: {'train/loss': 0.17963558435440063, 'train/grad_norm_pre_clip': 4.397218227386475, 'timing/data': 0.00033627916127443314, 'timing/model': 0.41853710636496544, 'learning_rate/qwen_vl_interface': 9.625878787748712e-06, 'learning_rate/action_model': 9.625878787748713e-05, 'epoch': 0.11})
+[2026-07-28 02:51:53,434][starVLA.training.train_starvla][INFO] - Step 597, Loss: {'train/loss': 0.08237801492214203, 'train/grad_norm_pre_clip': 2.3218631744384766, 'timing/data': 0.000310542993247509, 'timing/model': 0.41533585963770747, 'learning_rate/qwen_vl_interface': 9.624388936927572e-06, 'learning_rate/action_model': 9.624388936927571e-05, 'epoch': 0.11})
+[2026-07-28 02:51:59,324][starVLA.training.train_starvla][INFO] - Step 598, Loss: {'train/loss': 0.12256506085395813, 'train/grad_norm_pre_clip': 4.811216354370117, 'timing/data': 0.0004701642319560051, 'timing/model': 0.4170962409116328, 'learning_rate/qwen_vl_interface': 9.622896247612943e-06, 'learning_rate/action_model': 9.622896247612942e-05, 'epoch': 0.11})
+[2026-07-28 02:52:05,203][starVLA.training.train_starvla][INFO] - Step 599, Loss: {'train/loss': 0.12850768864154816, 'train/grad_norm_pre_clip': 7.6109466552734375, 'timing/data': 0.0003496212884783745, 'timing/model': 0.4178552171215415, 'learning_rate/qwen_vl_interface': 9.621400720773414e-06, 'learning_rate/action_model': 9.621400720773414e-05, 'epoch': 0.11})
+[2026-07-28 02:52:11,307][starVLA.training.train_starvla][INFO] - Step 600, Loss: {'train/loss': 0.09615423530340195, 'train/grad_norm_pre_clip': 1.1762170791625977, 'timing/data': 0.0003460790030658245, 'timing/model': 0.6487073870375752, 'learning_rate/qwen_vl_interface': 9.619902357379416e-06, 'learning_rate/action_model': 9.619902357379415e-05, 'epoch': 0.11})
+[2026-07-28 02:52:17,419][starVLA.training.train_starvla][INFO] - Step 601, Loss: {'train/loss': 0.10556481778621674, 'train/grad_norm_pre_clip': 1.336467981338501, 'timing/data': 0.00032444484531879425, 'timing/model': 0.6518542147241533, 'learning_rate/qwen_vl_interface': 9.618401158403222e-06, 'learning_rate/action_model': 9.618401158403222e-05, 'epoch': 0.11})
+[2026-07-28 02:52:23,297][starVLA.training.train_starvla][INFO] - Step 602, Loss: {'train/loss': 0.11488573253154755, 'train/grad_norm_pre_clip': 2.027801990509033, 'timing/data': 0.00034856703132390976, 'timing/model': 0.4175397348590195, 'learning_rate/qwen_vl_interface': 9.616897124818944e-06, 'learning_rate/action_model': 9.616897124818945e-05, 'epoch': 0.11})
+[2026-07-28 02:52:29,163][starVLA.training.train_starvla][INFO] - Step 603, Loss: {'train/loss': 0.09819786250591278, 'train/grad_norm_pre_clip': 2.8745036125183105, 'timing/data': 0.0003487630747258663, 'timing/model': 0.4160768291912973, 'learning_rate/qwen_vl_interface': 9.615390257602532e-06, 'learning_rate/action_model': 9.615390257602531e-05, 'epoch': 0.11})
+[2026-07-28 02:52:35,002][starVLA.training.train_starvla][INFO] - Step 604, Loss: {'train/loss': 0.05483204126358032, 'train/grad_norm_pre_clip': 2.904585123062134, 'timing/data': 0.000301106832921505, 'timing/model': 0.41577882692217827, 'learning_rate/qwen_vl_interface': 9.613880557731776e-06, 'learning_rate/action_model': 9.613880557731776e-05, 'epoch': 0.11})
+[2026-07-28 02:52:41,110][starVLA.training.train_starvla][INFO] - Step 605, Loss: {'train/loss': 0.08318538218736649, 'train/grad_norm_pre_clip': 1.6182998418807983, 'timing/data': 0.00026840390637516975, 'timing/model': 0.4132440038956702, 'learning_rate/qwen_vl_interface': 9.612368026186304e-06, 'learning_rate/action_model': 9.612368026186304e-05, 'epoch': 0.11})
+[2026-07-28 02:52:46,962][starVLA.training.train_starvla][INFO] - Step 606, Loss: {'train/loss': 0.06752631068229675, 'train/grad_norm_pre_clip': 1.4215842485427856, 'timing/data': 0.00038025621324777603, 'timing/model': 0.4168747761286795, 'learning_rate/qwen_vl_interface': 9.610852663947582e-06, 'learning_rate/action_model': 9.610852663947583e-05, 'epoch': 0.11})
+[2026-07-28 02:52:52,815][starVLA.training.train_starvla][INFO] - Step 607, Loss: {'train/loss': 0.10653778910636902, 'train/grad_norm_pre_clip': 3.2121834754943848, 'timing/data': 0.00028473371639847755, 'timing/model': 0.4188866619952023, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.11})
+[2026-07-28 02:52:58,907][starVLA.training.train_starvla][INFO] - Step 608, Loss: {'train/loss': 0.06269694864749908, 'train/grad_norm_pre_clip': 0.884455144405365, 'timing/data': 0.00026490725576877594, 'timing/model': 0.4142129668034613, 'learning_rate/qwen_vl_interface': 9.607813451325431e-06, 'learning_rate/action_model': 9.60781345132543e-05, 'epoch': 0.11})
+[2026-07-28 02:53:04,981][starVLA.training.train_starvla][INFO] - Step 609, Loss: {'train/loss': 0.14314387738704681, 'train/grad_norm_pre_clip': 4.429029941558838, 'timing/data': 0.0003031226806342602, 'timing/model': 0.6521849748678505, 'learning_rate/qwen_vl_interface': 9.606289602914113e-06, 'learning_rate/action_model': 9.606289602914112e-05, 'epoch': 0.11})
+[2026-07-28 02:53:10,809][starVLA.training.train_starvla][INFO] - Step 610, Loss: {'train/loss': 0.10879839956760406, 'train/grad_norm_pre_clip': 3.218047618865967, 'timing/data': 0.00028772465884685516, 'timing/model': 0.41648509493097663, 'learning_rate/qwen_vl_interface': 9.604762927753768e-06, 'learning_rate/action_model': 9.604762927753768e-05, 'epoch': 0.11})
+[2026-07-28 02:53:16,673][starVLA.training.train_starvla][INFO] - Step 611, Loss: {'train/loss': 0.11209817230701447, 'train/grad_norm_pre_clip': 2.9304773807525635, 'timing/data': 0.00028102193027734756, 'timing/model': 0.41713492712005973, 'learning_rate/qwen_vl_interface': 9.603233426835037e-06, 'learning_rate/action_model': 9.603233426835036e-05, 'epoch': 0.11})
+[2026-07-28 02:53:22,772][starVLA.training.train_starvla][INFO] - Step 612, Loss: {'train/loss': 0.0672767162322998, 'train/grad_norm_pre_clip': 1.443595290184021, 'timing/data': 0.0002876198850572109, 'timing/model': 0.6484026438556612, 'learning_rate/qwen_vl_interface': 9.601701101150399e-06, 'learning_rate/action_model': 9.601701101150398e-05, 'epoch': 0.11})
+[2026-07-28 02:53:28,651][starVLA.training.train_starvla][INFO] - Step 613, Loss: {'train/loss': 0.0953642725944519, 'train/grad_norm_pre_clip': 2.3739876747131348, 'timing/data': 0.00026424694806337357, 'timing/model': 0.41437479481101036, 'learning_rate/qwen_vl_interface': 9.600165951694162e-06, 'learning_rate/action_model': 9.60016595169416e-05, 'epoch': 0.11})
+[2026-07-28 02:53:34,741][starVLA.training.train_starvla][INFO] - Step 614, Loss: {'train/loss': 0.0929848849773407, 'train/grad_norm_pre_clip': 4.121966361999512, 'timing/data': 0.0002895081415772438, 'timing/model': 0.6517472271807492, 'learning_rate/qwen_vl_interface': 9.598627979462466e-06, 'learning_rate/action_model': 9.598627979462467e-05, 'epoch': 0.11})
+[2026-07-28 02:53:40,604][starVLA.training.train_starvla][INFO] - Step 615, Loss: {'train/loss': 0.05006755143404007, 'train/grad_norm_pre_clip': 1.0168980360031128, 'timing/data': 0.00033053290098905563, 'timing/model': 0.4139216602779925, 'learning_rate/qwen_vl_interface': 9.59708718545329e-06, 'learning_rate/action_model': 9.597087185453288e-05, 'epoch': 0.11})
+[2026-07-28 02:53:46,440][starVLA.training.train_starvla][INFO] - Step 616, Loss: {'train/loss': 0.07184846699237823, 'train/grad_norm_pre_clip': 1.7806156873703003, 'timing/data': 0.0002840189263224602, 'timing/model': 0.41581870010122657, 'learning_rate/qwen_vl_interface': 9.595543570666433e-06, 'learning_rate/action_model': 9.595543570666432e-05, 'epoch': 0.11})
+[2026-07-28 02:53:52,290][starVLA.training.train_starvla][INFO] - Step 617, Loss: {'train/loss': 0.09407225251197815, 'train/grad_norm_pre_clip': 2.67380952835083, 'timing/data': 0.00027630990371108055, 'timing/model': 0.4178164820186794, 'learning_rate/qwen_vl_interface': 9.593997136103533e-06, 'learning_rate/action_model': 9.593997136103533e-05, 'epoch': 0.11})
+[2026-07-28 02:53:58,396][starVLA.training.train_starvla][INFO] - Step 618, Loss: {'train/loss': 0.11278676986694336, 'train/grad_norm_pre_clip': 5.299453258514404, 'timing/data': 0.00026648491621017456, 'timing/model': 0.41436158074066043, 'learning_rate/qwen_vl_interface': 9.592447882768053e-06, 'learning_rate/action_model': 9.592447882768052e-05, 'epoch': 0.11})
+[2026-07-28 02:54:04,279][starVLA.training.train_starvla][INFO] - Step 619, Loss: {'train/loss': 0.0778205394744873, 'train/grad_norm_pre_clip': 1.3577702045440674, 'timing/data': 0.0002842778339982033, 'timing/model': 0.42446026485413313, 'learning_rate/qwen_vl_interface': 9.590895811665284e-06, 'learning_rate/action_model': 9.590895811665285e-05, 'epoch': 0.11})
+[2026-07-28 02:54:10,375][starVLA.training.train_starvla][INFO] - Step 620, Loss: {'train/loss': 0.06430409848690033, 'train/grad_norm_pre_clip': 2.4270706176757812, 'timing/data': 0.00029238592833280563, 'timing/model': 0.6569296629168093, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.11})
+[2026-07-28 02:54:16,240][starVLA.training.train_starvla][INFO] - Step 621, Loss: {'train/loss': 0.08013634383678436, 'train/grad_norm_pre_clip': 3.1018192768096924, 'timing/data': 0.00026694079861044884, 'timing/model': 0.4174820319749415, 'learning_rate/qwen_vl_interface': 9.587783220188212e-06, 'learning_rate/action_model': 9.587783220188212e-05, 'epoch': 0.11})
+[2026-07-28 02:54:22,343][starVLA.training.train_starvla][INFO] - Step 622, Loss: {'train/loss': 0.1113259568810463, 'train/grad_norm_pre_clip': 1.5901042222976685, 'timing/data': 0.00028525106608867645, 'timing/model': 0.42634680680930614, 'learning_rate/qwen_vl_interface': 9.586222701833631e-06, 'learning_rate/action_model': 9.58622270183363e-05, 'epoch': 0.11})
+[2026-07-28 02:54:28,237][starVLA.training.train_starvla][INFO] - Step 623, Loss: {'train/loss': 0.12463807314634323, 'train/grad_norm_pre_clip': 8.116002082824707, 'timing/data': 0.0002887677401304245, 'timing/model': 0.41673986334353685, 'learning_rate/qwen_vl_interface': 9.584659369751216e-06, 'learning_rate/action_model': 9.584659369751215e-05, 'epoch': 0.11})
+[2026-07-28 02:54:34,133][starVLA.training.train_starvla][INFO] - Step 624, Loss: {'train/loss': 0.07520212233066559, 'train/grad_norm_pre_clip': 1.9132411479949951, 'timing/data': 0.00027283141389489174, 'timing/model': 0.41733011417090893, 'learning_rate/qwen_vl_interface': 9.583093224955397e-06, 'learning_rate/action_model': 9.583093224955397e-05, 'epoch': 0.11})
+[2026-07-28 02:54:40,234][starVLA.training.train_starvla][INFO] - Step 625, Loss: {'train/loss': 0.08428822457790375, 'train/grad_norm_pre_clip': 1.7859392166137695, 'timing/data': 0.00030191102996468544, 'timing/model': 0.41431126510724425, 'learning_rate/qwen_vl_interface': 9.58152426846243e-06, 'learning_rate/action_model': 9.581524268462429e-05, 'epoch': 0.11})
+[2026-07-28 02:54:46,093][starVLA.training.train_starvla][INFO] - Step 626, Loss: {'train/loss': 0.09448032826185226, 'train/grad_norm_pre_clip': 2.822092056274414, 'timing/data': 0.0002826778218150139, 'timing/model': 0.4172548488713801, 'learning_rate/qwen_vl_interface': 9.579952501290388e-06, 'learning_rate/action_model': 9.579952501290387e-05, 'epoch': 0.11})
+[2026-07-28 02:54:51,768][starVLA.training.train_starvla][INFO] - Step 627, Loss: {'train/loss': 0.09977786988019943, 'train/grad_norm_pre_clip': 3.103661298751831, 'timing/data': 0.00027988292276859283, 'timing/model': 0.4176727430894971, 'learning_rate/qwen_vl_interface': 9.57837792445918e-06, 'learning_rate/action_model': 9.578377924459181e-05, 'epoch': 0.12})
+[2026-07-28 02:54:57,799][starVLA.training.train_starvla][INFO] - Step 628, Loss: {'train/loss': 0.06259046494960785, 'train/grad_norm_pre_clip': 1.4744683504104614, 'timing/data': 0.00028229597955942154, 'timing/model': 0.5868830592371523, 'learning_rate/qwen_vl_interface': 9.57680053899053e-06, 'learning_rate/action_model': 9.576800538990529e-05, 'epoch': 0.12})
+[2026-07-28 02:55:03,899][starVLA.training.train_starvla][INFO] - Step 629, Loss: {'train/loss': 0.1391230821609497, 'train/grad_norm_pre_clip': 3.279123306274414, 'timing/data': 0.00025963829830288887, 'timing/model': 0.416521614883095, 'learning_rate/qwen_vl_interface': 9.575220345907984e-06, 'learning_rate/action_model': 9.575220345907984e-05, 'epoch': 0.12})
+[2026-07-28 02:55:09,766][starVLA.training.train_starvla][INFO] - Step 630, Loss: {'train/loss': 0.09871795773506165, 'train/grad_norm_pre_clip': 4.136014461517334, 'timing/data': 0.0002960353158414364, 'timing/model': 0.41595812886953354, 'learning_rate/qwen_vl_interface': 9.573637346236915e-06, 'learning_rate/action_model': 9.573637346236915e-05, 'epoch': 0.12})
+[2026-07-28 02:55:15,720][starVLA.training.train_starvla][INFO] - Step 631, Loss: {'train/loss': 0.1031724214553833, 'train/grad_norm_pre_clip': 5.546494007110596, 'timing/data': 0.0002737920731306076, 'timing/model': 0.4704415709711611, 'learning_rate/qwen_vl_interface': 9.572051541004513e-06, 'learning_rate/action_model': 9.572051541004513e-05, 'epoch': 0.12})
+[2026-07-28 02:55:21,923][starVLA.training.train_starvla][INFO] - Step 632, Loss: {'train/loss': 0.055667437613010406, 'train/grad_norm_pre_clip': 1.7535569667816162, 'timing/data': 0.0003864467144012451, 'timing/model': 0.4179973821155727, 'learning_rate/qwen_vl_interface': 9.57046293123979e-06, 'learning_rate/action_model': 9.570462931239789e-05, 'epoch': 0.12})
+[2026-07-28 02:55:27,825][starVLA.training.train_starvla][INFO] - Step 633, Loss: {'train/loss': 0.05994543060660362, 'train/grad_norm_pre_clip': 1.5858113765716553, 'timing/data': 0.0003059566952288151, 'timing/model': 0.41912057995796204, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.12})
+[2026-07-28 02:55:34,010][starVLA.training.train_starvla][INFO] - Step 634, Loss: {'train/loss': 0.07504114508628845, 'train/grad_norm_pre_clip': 2.0357627868652344, 'timing/data': 0.00033671827986836433, 'timing/model': 0.6591639709658921, 'learning_rate/qwen_vl_interface': 9.567277302238524e-06, 'learning_rate/action_model': 9.567277302238524e-05, 'epoch': 0.12})
+[2026-07-28 02:55:39,940][starVLA.training.train_starvla][INFO] - Step 635, Loss: {'train/loss': 0.10381640493869781, 'train/grad_norm_pre_clip': 4.256503105163574, 'timing/data': 0.0002806819975376129, 'timing/model': 0.41546938195824623, 'learning_rate/qwen_vl_interface': 9.565680285069104e-06, 'learning_rate/action_model': 9.565680285069103e-05, 'epoch': 0.12})
+[2026-07-28 02:55:46,076][starVLA.training.train_starvla][INFO] - Step 636, Loss: {'train/loss': 0.07446198165416718, 'train/grad_norm_pre_clip': 1.4784231185913086, 'timing/data': 0.00030896300449967384, 'timing/model': 0.41753967478871346, 'learning_rate/qwen_vl_interface': 9.564080467501601e-06, 'learning_rate/action_model': 9.5640804675016e-05, 'epoch': 0.12})
+[2026-07-28 02:55:52,003][starVLA.training.train_starvla][INFO] - Step 637, Loss: {'train/loss': 0.06694024801254272, 'train/grad_norm_pre_clip': 2.1456756591796875, 'timing/data': 0.00030030542984604836, 'timing/model': 0.4179740874096751, 'learning_rate/qwen_vl_interface': 9.562477850574118e-06, 'learning_rate/action_model': 9.562477850574119e-05, 'epoch': 0.12})
+[2026-07-28 02:55:57,890][starVLA.training.train_starvla][INFO] - Step 638, Loss: {'train/loss': 0.09854844957590103, 'train/grad_norm_pre_clip': 7.172954559326172, 'timing/data': 0.0003607040271162987, 'timing/model': 0.4171987809240818, 'learning_rate/qwen_vl_interface': 9.56087243532658e-06, 'learning_rate/action_model': 9.56087243532658e-05, 'epoch': 0.12})
+[2026-07-28 02:56:03,820][starVLA.training.train_starvla][INFO] - Step 639, Loss: {'train/loss': 0.11934246867895126, 'train/grad_norm_pre_clip': 7.043262481689453, 'timing/data': 0.00030721817165613174, 'timing/model': 0.4161219811066985, 'learning_rate/qwen_vl_interface': 9.55926422280072e-06, 'learning_rate/action_model': 9.55926422280072e-05, 'epoch': 0.12})
+[2026-07-28 02:56:09,704][starVLA.training.train_starvla][INFO] - Step 640, Loss: {'train/loss': 0.16388490796089172, 'train/grad_norm_pre_clip': 2.4014739990234375, 'timing/data': 0.0003024018369615078, 'timing/model': 0.4212464368902147, 'learning_rate/qwen_vl_interface': 9.55765321404009e-06, 'learning_rate/action_model': 9.557653214040091e-05, 'epoch': 0.12})
+[2026-07-28 02:56:16,135][starVLA.training.train_starvla][INFO] - Step 641, Loss: {'train/loss': 0.048668503761291504, 'train/grad_norm_pre_clip': 1.6171668767929077, 'timing/data': 0.00034403614699840546, 'timing/model': 0.6472799312323332, 'learning_rate/qwen_vl_interface': 9.556039410090061e-06, 'learning_rate/action_model': 9.556039410090061e-05, 'epoch': 0.12})
+[2026-07-28 02:56:22,014][starVLA.training.train_starvla][INFO] - Step 642, Loss: {'train/loss': 0.04717680811882019, 'train/grad_norm_pre_clip': 1.4047893285751343, 'timing/data': 0.00025855889543890953, 'timing/model': 0.4158975798636675, 'learning_rate/qwen_vl_interface': 9.554422811997804e-06, 'learning_rate/action_model': 9.554422811997805e-05, 'epoch': 0.12})
+[2026-07-28 02:56:27,908][starVLA.training.train_starvla][INFO] - Step 643, Loss: {'train/loss': 0.09578479826450348, 'train/grad_norm_pre_clip': 2.806661605834961, 'timing/data': 0.0002984851598739624, 'timing/model': 0.41776661993935704, 'learning_rate/qwen_vl_interface': 9.55280342081232e-06, 'learning_rate/action_model': 9.55280342081232e-05, 'epoch': 0.12})
+[2026-07-28 02:56:33,799][starVLA.training.train_starvla][INFO] - Step 644, Loss: {'train/loss': 0.09862597286701202, 'train/grad_norm_pre_clip': 2.8865609169006348, 'timing/data': 0.00036411872133612633, 'timing/model': 0.4172376091592014, 'learning_rate/qwen_vl_interface': 9.55118123758441e-06, 'learning_rate/action_model': 9.551181237584409e-05, 'epoch': 0.12})
+[2026-07-28 02:56:39,909][starVLA.training.train_starvla][INFO] - Step 645, Loss: {'train/loss': 0.07515592873096466, 'train/grad_norm_pre_clip': 1.4996453523635864, 'timing/data': 0.00032226182520389557, 'timing/model': 0.659997743088752, 'learning_rate/qwen_vl_interface': 9.549556263366691e-06, 'learning_rate/action_model': 9.549556263366692e-05, 'epoch': 0.12})
+[2026-07-28 02:56:46,060][starVLA.training.train_starvla][INFO] - Step 646, Loss: {'train/loss': 0.08088023960590363, 'train/grad_norm_pre_clip': 1.8322163820266724, 'timing/data': 0.0002678362652659416, 'timing/model': 0.4236746309325099, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.12})
+[2026-07-28 02:56:52,040][starVLA.training.train_starvla][INFO] - Step 647, Loss: {'train/loss': 0.07755313068628311, 'train/grad_norm_pre_clip': 2.8276031017303467, 'timing/data': 0.0003414400853216648, 'timing/model': 0.4254806046374142, 'learning_rate/qwen_vl_interface': 9.546297946181354e-06, 'learning_rate/action_model': 9.546297946181354e-05, 'epoch': 0.12})
+[2026-07-28 02:56:57,911][starVLA.training.train_starvla][INFO] - Step 648, Loss: {'train/loss': 0.07191778719425201, 'train/grad_norm_pre_clip': 1.2714614868164062, 'timing/data': 0.0002811201848089695, 'timing/model': 0.41721611795946956, 'learning_rate/qwen_vl_interface': 9.544664605328022e-06, 'learning_rate/action_model': 9.544664605328022e-05, 'epoch': 0.12})
+[2026-07-28 02:57:04,311][starVLA.training.train_starvla][INFO] - Step 649, Loss: {'train/loss': 0.08183518052101135, 'train/grad_norm_pre_clip': 4.359447002410889, 'timing/data': 0.0003060069866478443, 'timing/model': 0.4159206850454211, 'learning_rate/qwen_vl_interface': 9.543028477713455e-06, 'learning_rate/action_model': 9.543028477713453e-05, 'epoch': 0.12})
+[2026-07-28 02:57:10,002][starVLA.training.train_starvla][INFO] - Step 650, Loss: {'train/loss': 0.10819409042596817, 'train/grad_norm_pre_clip': 2.6214938163757324, 'timing/data': 0.0002995324321091175, 'timing/model': 0.41712650703266263, 'learning_rate/qwen_vl_interface': 9.541389564399316e-06, 'learning_rate/action_model': 9.541389564399315e-05, 'epoch': 0.12})
+[2026-07-28 02:57:15,918][starVLA.training.train_starvla][INFO] - Step 651, Loss: {'train/loss': 0.07403774559497833, 'train/grad_norm_pre_clip': 3.0760982036590576, 'timing/data': 0.0003786529414355755, 'timing/model': 0.41673882212489843, 'learning_rate/qwen_vl_interface': 9.53974786644908e-06, 'learning_rate/action_model': 9.53974786644908e-05, 'epoch': 0.12})
+[2026-07-28 02:57:21,565][starVLA.training.train_starvla][INFO] - Step 652, Loss: {'train/loss': 0.08605438470840454, 'train/grad_norm_pre_clip': 2.2465097904205322, 'timing/data': 0.00047216610983014107, 'timing/model': 0.4169257618486881, 'learning_rate/qwen_vl_interface': 9.538103384928029e-06, 'learning_rate/action_model': 9.538103384928027e-05, 'epoch': 0.12})
+[2026-07-28 02:57:27,956][starVLA.training.train_starvla][INFO] - Step 653, Loss: {'train/loss': 0.073635995388031, 'train/grad_norm_pre_clip': 1.8311793804168701, 'timing/data': 0.0002701221965253353, 'timing/model': 0.650886164046824, 'learning_rate/qwen_vl_interface': 9.536456120903244e-06, 'learning_rate/action_model': 9.536456120903244e-05, 'epoch': 0.12})
+[2026-07-28 02:57:33,820][starVLA.training.train_starvla][INFO] - Step 654, Loss: {'train/loss': 0.05253364145755768, 'train/grad_norm_pre_clip': 2.7829740047454834, 'timing/data': 0.0002733343280851841, 'timing/model': 0.41463616816326976, 'learning_rate/qwen_vl_interface': 9.534806075443619e-06, 'learning_rate/action_model': 9.534806075443619e-05, 'epoch': 0.12})
+[2026-07-28 02:57:39,709][starVLA.training.train_starvla][INFO] - Step 655, Loss: {'train/loss': 0.06012313440442085, 'train/grad_norm_pre_clip': 1.7524479627609253, 'timing/data': 0.0002903677523136139, 'timing/model': 0.41602953616529703, 'learning_rate/qwen_vl_interface': 9.533153249619853e-06, 'learning_rate/action_model': 9.533153249619853e-05, 'epoch': 0.12})
+[2026-07-28 02:57:45,821][starVLA.training.train_starvla][INFO] - Step 656, Loss: {'train/loss': 0.07051975280046463, 'train/grad_norm_pre_clip': 2.0926454067230225, 'timing/data': 0.00027678580954670906, 'timing/model': 0.65414112014696, 'learning_rate/qwen_vl_interface': 9.531497644504444e-06, 'learning_rate/action_model': 9.531497644504445e-05, 'epoch': 0.12})
+[2026-07-28 02:57:51,672][starVLA.training.train_starvla][INFO] - Step 657, Loss: {'train/loss': 0.07255305349826813, 'train/grad_norm_pre_clip': 2.388467311859131, 'timing/data': 0.00027730921283364296, 'timing/model': 0.41521778516471386, 'learning_rate/qwen_vl_interface': 9.529839261171698e-06, 'learning_rate/action_model': 9.529839261171696e-05, 'epoch': 0.12})
+[2026-07-28 02:57:57,538][starVLA.training.train_starvla][INFO] - Step 658, Loss: {'train/loss': 0.10989125072956085, 'train/grad_norm_pre_clip': 2.881387710571289, 'timing/data': 0.0002788580022752285, 'timing/model': 0.4152228687889874, 'learning_rate/qwen_vl_interface': 9.52817810069772e-06, 'learning_rate/action_model': 9.52817810069772e-05, 'epoch': 0.12})
+[2026-07-28 02:58:03,681][starVLA.training.train_starvla][INFO] - Step 659, Loss: {'train/loss': 0.110084667801857, 'train/grad_norm_pre_clip': 2.699573278427124, 'timing/data': 0.0002653207629919052, 'timing/model': 0.41513060685247183, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.12})
+[2026-07-28 02:58:09,585][starVLA.training.train_starvla][INFO] - Step 660, Loss: {'train/loss': 0.13691388070583344, 'train/grad_norm_pre_clip': 3.977572202682495, 'timing/data': 0.0002785632386803627, 'timing/model': 0.41585886385291815, 'learning_rate/qwen_vl_interface': 9.524847452639509e-06, 'learning_rate/action_model': 9.524847452639509e-05, 'epoch': 0.12})
+[2026-07-28 02:58:15,703][starVLA.training.train_starvla][INFO] - Step 661, Loss: {'train/loss': 0.09839267283678055, 'train/grad_norm_pre_clip': 2.6261861324310303, 'timing/data': 0.00029729772359132767, 'timing/model': 0.66520243184641, 'learning_rate/qwen_vl_interface': 9.523177967216496e-06, 'learning_rate/action_model': 9.523177967216496e-05, 'epoch': 0.12})
+[2026-07-28 02:58:21,621][starVLA.training.train_starvla][INFO] - Step 662, Loss: {'train/loss': 0.11078799515962601, 'train/grad_norm_pre_clip': 3.0837059020996094, 'timing/data': 0.00031442008912563324, 'timing/model': 0.41443406511098146, 'learning_rate/qwen_vl_interface': 9.521505708974694e-06, 'learning_rate/action_model': 9.521505708974694e-05, 'epoch': 0.12})
+[2026-07-28 02:58:27,675][starVLA.training.train_starvla][INFO] - Step 663, Loss: {'train/loss': 0.07418406009674072, 'train/grad_norm_pre_clip': 0.8090174794197083, 'timing/data': 0.00027858791872859, 'timing/model': 0.4168683891184628, 'learning_rate/qwen_vl_interface': 9.519830678999214e-06, 'learning_rate/action_model': 9.519830678999213e-05, 'epoch': 0.12})
+[2026-07-28 02:58:33,298][starVLA.training.train_starvla][INFO] - Step 664, Loss: {'train/loss': 0.08321110904216766, 'train/grad_norm_pre_clip': 1.5889427661895752, 'timing/data': 0.0002698376774787903, 'timing/model': 0.4154870267957449, 'learning_rate/qwen_vl_interface': 9.518152878376961e-06, 'learning_rate/action_model': 9.51815287837696e-05, 'epoch': 0.12})
+[2026-07-28 02:58:39,180][starVLA.training.train_starvla][INFO] - Step 665, Loss: {'train/loss': 0.10085843503475189, 'train/grad_norm_pre_clip': 1.5442862510681152, 'timing/data': 0.0002897498197853565, 'timing/model': 0.4171510962769389, 'learning_rate/qwen_vl_interface': 9.516472308196646e-06, 'learning_rate/action_model': 9.516472308196645e-05, 'epoch': 0.12})
+[2026-07-28 02:58:45,396][starVLA.training.train_starvla][INFO] - Step 666, Loss: {'train/loss': 0.0704353079199791, 'train/grad_norm_pre_clip': 1.3058664798736572, 'timing/data': 0.00031396327540278435, 'timing/model': 0.4147337879985571, 'learning_rate/qwen_vl_interface': 9.51478896954877e-06, 'learning_rate/action_model': 9.51478896954877e-05, 'epoch': 0.12})
+[2026-07-28 02:58:51,547][starVLA.training.train_starvla][INFO] - Step 667, Loss: {'train/loss': 0.0709429681301117, 'train/grad_norm_pre_clip': 1.2753185033798218, 'timing/data': 0.000283082015812397, 'timing/model': 0.6717516295611858, 'learning_rate/qwen_vl_interface': 9.51310286352563e-06, 'learning_rate/action_model': 9.51310286352563e-05, 'epoch': 0.12})
+[2026-07-28 02:58:57,416][starVLA.training.train_starvla][INFO] - Step 668, Loss: {'train/loss': 0.061629414558410645, 'train/grad_norm_pre_clip': 1.1518456935882568, 'timing/data': 0.0002673463895916939, 'timing/model': 0.4140886119566858, 'learning_rate/qwen_vl_interface': 9.51141399122133e-06, 'learning_rate/action_model': 9.511413991221331e-05, 'epoch': 0.12})
+[2026-07-28 02:59:03,419][starVLA.training.train_starvla][INFO] - Step 669, Loss: {'train/loss': 0.09583266079425812, 'train/grad_norm_pre_clip': 2.6825475692749023, 'timing/data': 0.00028199702501296997, 'timing/model': 0.5877400869503617, 'learning_rate/qwen_vl_interface': 9.509722353731757e-06, 'learning_rate/action_model': 9.509722353731756e-05, 'epoch': 0.12})
+[2026-07-28 02:59:09,287][starVLA.training.train_starvla][INFO] - Step 670, Loss: {'train/loss': 0.08556897193193436, 'train/grad_norm_pre_clip': 1.914435863494873, 'timing/data': 0.0002902070991694927, 'timing/model': 0.41631289198994637, 'learning_rate/qwen_vl_interface': 9.508027952154595e-06, 'learning_rate/action_model': 9.508027952154593e-05, 'epoch': 0.12})
+[2026-07-28 02:59:15,150][starVLA.training.train_starvla][INFO] - Step 671, Loss: {'train/loss': 0.09070701897144318, 'train/grad_norm_pre_clip': 4.038106918334961, 'timing/data': 0.00028479425236582756, 'timing/model': 0.41640934022143483, 'learning_rate/qwen_vl_interface': 9.506330787589321e-06, 'learning_rate/action_model': 9.506330787589322e-05, 'epoch': 0.12})
+[2026-07-28 02:59:21,256][starVLA.training.train_starvla][INFO] - Step 672, Loss: {'train/loss': 0.1104479506611824, 'train/grad_norm_pre_clip': 2.276911973953247, 'timing/data': 0.0002694223076105118, 'timing/model': 0.41596663231030107, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.12})
+[2026-07-28 02:59:27,320][starVLA.training.train_starvla][INFO] - Step 673, Loss: {'train/loss': 0.11528186500072479, 'train/grad_norm_pre_clip': 2.8451457023620605, 'timing/data': 0.00029543228447437286, 'timing/model': 0.41632131626829505, 'learning_rate/qwen_vl_interface': 9.50292817390133e-06, 'learning_rate/action_model': 9.502928173901329e-05, 'epoch': 0.12})
+[2026-07-28 02:59:33,257][starVLA.training.train_starvla][INFO] - Step 674, Loss: {'train/loss': 0.06666238605976105, 'train/grad_norm_pre_clip': 2.996853828430176, 'timing/data': 0.00032715778797864914, 'timing/model': 0.4181154100224376, 'learning_rate/qwen_vl_interface': 9.501222726986525e-06, 'learning_rate/action_model': 9.501222726986524e-05, 'epoch': 0.12})
+[2026-07-28 02:59:39,143][starVLA.training.train_starvla][INFO] - Step 675, Loss: {'train/loss': 0.07116681337356567, 'train/grad_norm_pre_clip': 1.537193775177002, 'timing/data': 0.0002837567590177059, 'timing/model': 0.41670131077989936, 'learning_rate/qwen_vl_interface': 9.49951452149945e-06, 'learning_rate/action_model': 9.499514521499451e-05, 'epoch': 0.12})
+[2026-07-28 02:59:45,209][starVLA.training.train_starvla][INFO] - Step 676, Loss: {'train/loss': 0.10457885265350342, 'train/grad_norm_pre_clip': 2.060123920440674, 'timing/data': 0.0002584499306976795, 'timing/model': 0.41632759291678667, 'learning_rate/qwen_vl_interface': 9.497803558548539e-06, 'learning_rate/action_model': 9.497803558548539e-05, 'epoch': 0.12})
+[2026-07-28 02:59:51,296][starVLA.training.train_starvla][INFO] - Step 677, Loss: {'train/loss': 0.09846952557563782, 'train/grad_norm_pre_clip': 2.5915815830230713, 'timing/data': 0.000266107264906168, 'timing/model': 0.4151679868809879, 'learning_rate/qwen_vl_interface': 9.496089839244015e-06, 'learning_rate/action_model': 9.496089839244015e-05, 'epoch': 0.12})
+[2026-07-28 02:59:57,171][starVLA.training.train_starvla][INFO] - Step 678, Loss: {'train/loss': 0.0887795016169548, 'train/grad_norm_pre_clip': 1.9163819551467896, 'timing/data': 0.0003116619773209095, 'timing/model': 0.41989372996613383, 'learning_rate/qwen_vl_interface': 9.494373364697895e-06, 'learning_rate/action_model': 9.494373364697894e-05, 'epoch': 0.12})
+[2026-07-28 03:00:03,075][starVLA.training.train_starvla][INFO] - Step 679, Loss: {'train/loss': 0.08236411213874817, 'train/grad_norm_pre_clip': 2.081850528717041, 'timing/data': 0.0002850252203643322, 'timing/model': 0.41746398620307446, 'learning_rate/qwen_vl_interface': 9.492654136023978e-06, 'learning_rate/action_model': 9.492654136023979e-05, 'epoch': 0.12})
+[2026-07-28 03:00:08,984][starVLA.training.train_starvla][INFO] - Step 680, Loss: {'train/loss': 0.0738978236913681, 'train/grad_norm_pre_clip': 2.0023982524871826, 'timing/data': 0.00029024481773376465, 'timing/model': 0.4162443126551807, 'learning_rate/qwen_vl_interface': 9.490932154337855e-06, 'learning_rate/action_model': 9.490932154337855e-05, 'epoch': 0.12})
+[2026-07-28 03:00:15,097][starVLA.training.train_starvla][INFO] - Step 681, Loss: {'train/loss': 0.06788836419582367, 'train/grad_norm_pre_clip': 1.1753193140029907, 'timing/data': 0.0002569202333688736, 'timing/model': 0.6557214250788093, 'learning_rate/qwen_vl_interface': 9.4892074207569e-06, 'learning_rate/action_model': 9.489207420756899e-05, 'epoch': 0.13})
+[2026-07-28 03:00:21,373][starVLA.training.train_starvla][INFO] - Step 682, Loss: {'train/loss': 0.09558238089084625, 'train/grad_norm_pre_clip': 1.8880943059921265, 'timing/data': 0.0002693827264010906, 'timing/model': 0.6884120777249336, 'learning_rate/qwen_vl_interface': 9.487479936400274e-06, 'learning_rate/action_model': 9.487479936400272e-05, 'epoch': 0.13})
+[2026-07-28 03:00:27,202][starVLA.training.train_starvla][INFO] - Step 683, Loss: {'train/loss': 0.07619954645633698, 'train/grad_norm_pre_clip': 1.7999298572540283, 'timing/data': 0.0002644178457558155, 'timing/model': 0.41384083731099963, 'learning_rate/qwen_vl_interface': 9.485749702388923e-06, 'learning_rate/action_model': 9.485749702388923e-05, 'epoch': 0.13})
+[2026-07-28 03:00:33,057][starVLA.training.train_starvla][INFO] - Step 684, Loss: {'train/loss': 0.06281311810016632, 'train/grad_norm_pre_clip': 1.597096562385559, 'timing/data': 0.00028050411492586136, 'timing/model': 0.41635776683688164, 'learning_rate/qwen_vl_interface': 9.484016719845576e-06, 'learning_rate/action_model': 9.484016719845576e-05, 'epoch': 0.13})
+[2026-07-28 03:00:39,004][starVLA.training.train_starvla][INFO] - Step 685, Loss: {'train/loss': 0.06455733627080917, 'train/grad_norm_pre_clip': 1.9906158447265625, 'timing/data': 0.000284356065094471, 'timing/model': 0.4164255219511688, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.13})
+[2026-07-28 03:00:45,043][starVLA.training.train_starvla][INFO] - Step 686, Loss: {'train/loss': 0.07490836828947067, 'train/grad_norm_pre_clip': 1.8068619966506958, 'timing/data': 0.0002751578576862812, 'timing/model': 0.4160228748805821, 'learning_rate/qwen_vl_interface': 9.480542513662734e-06, 'learning_rate/action_model': 9.480542513662734e-05, 'epoch': 0.13})
+[2026-07-28 03:00:51,169][starVLA.training.train_starvla][INFO] - Step 687, Loss: {'train/loss': 0.07536989450454712, 'train/grad_norm_pre_clip': 2.208385467529297, 'timing/data': 0.00025549111887812614, 'timing/model': 0.4136241157539189, 'learning_rate/qwen_vl_interface': 9.478801292277617e-06, 'learning_rate/action_model': 9.478801292277616e-05, 'epoch': 0.13})
+[2026-07-28 03:00:57,266][starVLA.training.train_starvla][INFO] - Step 688, Loss: {'train/loss': 0.09146299213171005, 'train/grad_norm_pre_clip': 2.6101317405700684, 'timing/data': 0.00025389110669493675, 'timing/model': 0.415280899964273, 'learning_rate/qwen_vl_interface': 9.477057326869249e-06, 'learning_rate/action_model': 9.477057326869249e-05, 'epoch': 0.13})
+[2026-07-28 03:01:03,132][starVLA.training.train_starvla][INFO] - Step 689, Loss: {'train/loss': 0.09189080446958542, 'train/grad_norm_pre_clip': 3.166667938232422, 'timing/data': 0.00028022099286317825, 'timing/model': 0.4170139799825847, 'learning_rate/qwen_vl_interface': 9.475310618569274e-06, 'learning_rate/action_model': 9.475310618569274e-05, 'epoch': 0.13})
+[2026-07-28 03:01:09,229][starVLA.training.train_starvla][INFO] - Step 690, Loss: {'train/loss': 0.11879393458366394, 'train/grad_norm_pre_clip': 4.799676418304443, 'timing/data': 0.00033189402893185616, 'timing/model': 0.41544756293296814, 'learning_rate/qwen_vl_interface': 9.473561168511115e-06, 'learning_rate/action_model': 9.473561168511115e-05, 'epoch': 0.13})
+[2026-07-28 03:01:15,122][starVLA.training.train_starvla][INFO] - Step 691, Loss: {'train/loss': 0.10388889908790588, 'train/grad_norm_pre_clip': 3.1103732585906982, 'timing/data': 0.00029856525361537933, 'timing/model': 0.4193817861378193, 'learning_rate/qwen_vl_interface': 9.471808977829967e-06, 'learning_rate/action_model': 9.471808977829967e-05, 'epoch': 0.13})
+[2026-07-28 03:01:21,324][starVLA.training.train_starvla][INFO] - Step 692, Loss: {'train/loss': 0.0652458667755127, 'train/grad_norm_pre_clip': 1.2684766054153442, 'timing/data': 0.00029306625947356224, 'timing/model': 0.7193979229778051, 'learning_rate/qwen_vl_interface': 9.470054047662808e-06, 'learning_rate/action_model': 9.470054047662808e-05, 'epoch': 0.13})
+[2026-07-28 03:01:27,203][starVLA.training.train_starvla][INFO] - Step 693, Loss: {'train/loss': 0.13735902309417725, 'train/grad_norm_pre_clip': 3.8467020988464355, 'timing/data': 0.0002918378449976444, 'timing/model': 0.41763644525781274, 'learning_rate/qwen_vl_interface': 9.468296379148394e-06, 'learning_rate/action_model': 9.468296379148393e-05, 'epoch': 0.13})
+[2026-07-28 03:01:33,305][starVLA.training.train_starvla][INFO] - Step 694, Loss: {'train/loss': 0.10577092319726944, 'train/grad_norm_pre_clip': 4.9100022315979, 'timing/data': 0.00027872109785676, 'timing/model': 0.6539039029739797, 'learning_rate/qwen_vl_interface': 9.466535973427259e-06, 'learning_rate/action_model': 9.466535973427258e-05, 'epoch': 0.13})
+[2026-07-28 03:01:39,172][starVLA.training.train_starvla][INFO] - Step 695, Loss: {'train/loss': 0.12067659199237823, 'train/grad_norm_pre_clip': 5.70363712310791, 'timing/data': 0.00026177987456321716, 'timing/model': 0.41405203100293875, 'learning_rate/qwen_vl_interface': 9.464772831641706e-06, 'learning_rate/action_model': 9.464772831641707e-05, 'epoch': 0.13})
+[2026-07-28 03:01:45,269][starVLA.training.train_starvla][INFO] - Step 696, Loss: {'train/loss': 0.10975059866905212, 'train/grad_norm_pre_clip': 3.6439192295074463, 'timing/data': 0.00025901012122631073, 'timing/model': 0.4139970880933106, 'learning_rate/qwen_vl_interface': 9.463006954935823e-06, 'learning_rate/action_model': 9.463006954935822e-05, 'epoch': 0.13})
+[2026-07-28 03:01:51,135][starVLA.training.train_starvla][INFO] - Step 697, Loss: {'train/loss': 0.10097867250442505, 'train/grad_norm_pre_clip': 2.8440029621124268, 'timing/data': 0.0003098500892519951, 'timing/model': 0.4167557982727885, 'learning_rate/qwen_vl_interface': 9.461238344455468e-06, 'learning_rate/action_model': 9.461238344455468e-05, 'epoch': 0.13})
+[2026-07-28 03:01:57,021][starVLA.training.train_starvla][INFO] - Step 698, Loss: {'train/loss': 0.08297724276781082, 'train/grad_norm_pre_clip': 3.8605129718780518, 'timing/data': 0.0003106929361820221, 'timing/model': 0.41723780427128077, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.13})
+[2026-07-28 03:02:02,890][starVLA.training.train_starvla][INFO] - Step 699, Loss: {'train/loss': 0.08724398910999298, 'train/grad_norm_pre_clip': 2.778378963470459, 'timing/data': 0.00029497500509023666, 'timing/model': 0.4177573509514332, 'learning_rate/qwen_vl_interface': 9.457692926763638e-06, 'learning_rate/action_model': 9.457692926763638e-05, 'epoch': 0.13})
+[2026-07-28 03:02:08,791][starVLA.training.train_starvla][INFO] - Step 700, Loss: {'train/loss': 0.055695101618766785, 'train/grad_norm_pre_clip': 1.8655141592025757, 'timing/data': 0.00025497423484921455, 'timing/model': 0.4290735116228461, 'learning_rate/qwen_vl_interface': 9.455916121852748e-06, 'learning_rate/action_model': 9.455916121852747e-05, 'epoch': 0.13})
+[2026-07-28 03:02:14,658][starVLA.training.train_starvla][INFO] - Step 701, Loss: {'train/loss': 0.10353647917509079, 'train/grad_norm_pre_clip': 2.550462245941162, 'timing/data': 0.0002862638793885708, 'timing/model': 0.4185869172215462, 'learning_rate/qwen_vl_interface': 9.454136587768549e-06, 'learning_rate/action_model': 9.454136587768548e-05, 'epoch': 0.13})
+[2026-07-28 03:02:20,316][starVLA.training.train_starvla][INFO] - Step 702, Loss: {'train/loss': 0.08114521950483322, 'train/grad_norm_pre_clip': 2.453618049621582, 'timing/data': 0.00027689896523952484, 'timing/model': 0.41751692490652204, 'learning_rate/qwen_vl_interface': 9.452354325665763e-06, 'learning_rate/action_model': 9.452354325665763e-05, 'epoch': 0.13})
+[2026-07-28 03:02:26,218][starVLA.training.train_starvla][INFO] - Step 703, Loss: {'train/loss': 0.048922497779130936, 'train/grad_norm_pre_clip': 2.4147658348083496, 'timing/data': 0.00030066678300499916, 'timing/model': 0.6616131220944226, 'learning_rate/qwen_vl_interface': 9.450569336700878e-06, 'learning_rate/action_model': 9.450569336700878e-05, 'epoch': 0.13})
+[2026-07-28 03:02:32,056][starVLA.training.train_starvla][INFO] - Step 704, Loss: {'train/loss': 0.09979639947414398, 'train/grad_norm_pre_clip': 1.8507161140441895, 'timing/data': 0.00027672387659549713, 'timing/model': 0.41607057489454746, 'learning_rate/qwen_vl_interface': 9.448781622032155e-06, 'learning_rate/action_model': 9.448781622032155e-05, 'epoch': 0.13})
+[2026-07-28 03:02:37,916][starVLA.training.train_starvla][INFO] - Step 705, Loss: {'train/loss': 0.10643921047449112, 'train/grad_norm_pre_clip': 2.6711812019348145, 'timing/data': 0.0002910229377448559, 'timing/model': 0.4167754449881613, 'learning_rate/qwen_vl_interface': 9.446991182819624e-06, 'learning_rate/action_model': 9.446991182819623e-05, 'epoch': 0.13})
+[2026-07-28 03:02:43,780][starVLA.training.train_starvla][INFO] - Step 706, Loss: {'train/loss': 0.0634130910038948, 'train/grad_norm_pre_clip': 2.040834903717041, 'timing/data': 0.00043092016130685806, 'timing/model': 0.4169661751948297, 'learning_rate/qwen_vl_interface': 9.44519802022508e-06, 'learning_rate/action_model': 9.44519802022508e-05, 'epoch': 0.13})
+[2026-07-28 03:02:49,848][starVLA.training.train_starvla][INFO] - Step 707, Loss: {'train/loss': 0.05618395656347275, 'train/grad_norm_pre_clip': 1.4640049934387207, 'timing/data': 0.0002645431086421013, 'timing/model': 0.41531427996233106, 'learning_rate/qwen_vl_interface': 9.443402135412088e-06, 'learning_rate/action_model': 9.443402135412088e-05, 'epoch': 0.13})
+[2026-07-28 03:02:55,982][starVLA.training.train_starvla][INFO] - Step 708, Loss: {'train/loss': 0.05510180443525314, 'train/grad_norm_pre_clip': 1.4711527824401855, 'timing/data': 0.0003404691815376282, 'timing/model': 0.6513231685385108, 'learning_rate/qwen_vl_interface': 9.441603529545977e-06, 'learning_rate/action_model': 9.441603529545976e-05, 'epoch': 0.13})
+[2026-07-28 03:03:01,832][starVLA.training.train_starvla][INFO] - Step 709, Loss: {'train/loss': 0.10328125208616257, 'train/grad_norm_pre_clip': 2.5941929817199707, 'timing/data': 0.00025502871721982956, 'timing/model': 0.4146558870561421, 'learning_rate/qwen_vl_interface': 9.439802203793843e-06, 'learning_rate/action_model': 9.439802203793842e-05, 'epoch': 0.13})
+[2026-07-28 03:03:07,808][starVLA.training.train_starvla][INFO] - Step 710, Loss: {'train/loss': 0.07923848927021027, 'train/grad_norm_pre_clip': 2.073427200317383, 'timing/data': 0.00034356117248535156, 'timing/model': 0.5867425017058849, 'learning_rate/qwen_vl_interface': 9.437998159324547e-06, 'learning_rate/action_model': 9.437998159324547e-05, 'epoch': 0.13})
+[2026-07-28 03:03:13,664][starVLA.training.train_starvla][INFO] - Step 711, Loss: {'train/loss': 0.07301260530948639, 'train/grad_norm_pre_clip': 1.5192983150482178, 'timing/data': 0.000276756938546896, 'timing/model': 0.41645889403298497, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.13})
+[2026-07-28 03:03:19,528][starVLA.training.train_starvla][INFO] - Step 712, Loss: {'train/loss': 0.05075957253575325, 'train/grad_norm_pre_clip': 1.4246066808700562, 'timing/data': 0.00028265221044421196, 'timing/model': 0.4168597827665508, 'learning_rate/qwen_vl_interface': 9.434381918918731e-06, 'learning_rate/action_model': 9.434381918918731e-05, 'epoch': 0.13})
+[2026-07-28 03:03:25,403][starVLA.training.train_starvla][INFO] - Step 713, Loss: {'train/loss': 0.09538213163614273, 'train/grad_norm_pre_clip': 2.9552156925201416, 'timing/data': 0.00028094975277781487, 'timing/model': 0.4188700928352773, 'learning_rate/qwen_vl_interface': 9.432569725328753e-06, 'learning_rate/action_model': 9.432569725328753e-05, 'epoch': 0.13})
+[2026-07-28 03:03:31,727][starVLA.training.train_starvla][INFO] - Step 714, Loss: {'train/loss': 0.10012714564800262, 'train/grad_norm_pre_clip': 2.1151559352874756, 'timing/data': 0.00027652690187096596, 'timing/model': 0.6509096268564463, 'learning_rate/qwen_vl_interface': 9.430754817714686e-06, 'learning_rate/action_model': 9.430754817714687e-05, 'epoch': 0.13})
+[2026-07-28 03:03:37,602][starVLA.training.train_starvla][INFO] - Step 715, Loss: {'train/loss': 0.16140149533748627, 'train/grad_norm_pre_clip': 3.9338531494140625, 'timing/data': 0.00028470205143094063, 'timing/model': 0.4149556220509112, 'learning_rate/qwen_vl_interface': 9.42893719725421e-06, 'learning_rate/action_model': 9.42893719725421e-05, 'epoch': 0.13})
+[2026-07-28 03:03:43,508][starVLA.training.train_starvla][INFO] - Step 716, Loss: {'train/loss': 0.11875148117542267, 'train/grad_norm_pre_clip': 4.108004093170166, 'timing/data': 0.0003362875431776047, 'timing/model': 0.4165081079117954, 'learning_rate/qwen_vl_interface': 9.427116865126757e-06, 'learning_rate/action_model': 9.427116865126756e-05, 'epoch': 0.13})
+[2026-07-28 03:03:49,365][starVLA.training.train_starvla][INFO] - Step 717, Loss: {'train/loss': 0.06789898872375488, 'train/grad_norm_pre_clip': 2.2159736156463623, 'timing/data': 0.0003393399529159069, 'timing/model': 0.4191289539448917, 'learning_rate/qwen_vl_interface': 9.42529382251352e-06, 'learning_rate/action_model': 9.42529382251352e-05, 'epoch': 0.13})
+[2026-07-28 03:03:55,465][starVLA.training.train_starvla][INFO] - Step 718, Loss: {'train/loss': 0.09845274686813354, 'train/grad_norm_pre_clip': 3.0153706073760986, 'timing/data': 0.00030917814001441, 'timing/model': 0.41487655229866505, 'learning_rate/qwen_vl_interface': 9.423468070597453e-06, 'learning_rate/action_model': 9.423468070597452e-05, 'epoch': 0.13})
+[2026-07-28 03:04:01,571][starVLA.training.train_starvla][INFO] - Step 719, Loss: {'train/loss': 0.04958363622426987, 'train/grad_norm_pre_clip': 1.0102708339691162, 'timing/data': 0.0003177677281200886, 'timing/model': 0.41593938833102584, 'learning_rate/qwen_vl_interface': 9.421639610563264e-06, 'learning_rate/action_model': 9.421639610563264e-05, 'epoch': 0.13})
+[2026-07-28 03:04:07,425][starVLA.training.train_starvla][INFO] - Step 720, Loss: {'train/loss': 0.12565302848815918, 'train/grad_norm_pre_clip': 4.253074645996094, 'timing/data': 0.0003494713455438614, 'timing/model': 0.41656765388324857, 'learning_rate/qwen_vl_interface': 9.419808443597422e-06, 'learning_rate/action_model': 9.419808443597423e-05, 'epoch': 0.13})
+[2026-07-28 03:04:13,313][starVLA.training.train_starvla][INFO] - Step 721, Loss: {'train/loss': 0.10650892555713654, 'train/grad_norm_pre_clip': 3.280416965484619, 'timing/data': 0.00034990906715393066, 'timing/model': 0.4168313667178154, 'learning_rate/qwen_vl_interface': 9.417974570888152e-06, 'learning_rate/action_model': 9.417974570888153e-05, 'epoch': 0.13})
+[2026-07-28 03:04:18,973][starVLA.training.train_starvla][INFO] - Step 722, Loss: {'train/loss': 0.07538403570652008, 'train/grad_norm_pre_clip': 1.7513954639434814, 'timing/data': 0.0003515048883855343, 'timing/model': 0.41615490382537246, 'learning_rate/qwen_vl_interface': 9.416137993625437e-06, 'learning_rate/action_model': 9.416137993625437e-05, 'epoch': 0.13})
+[2026-07-28 03:04:25,309][starVLA.training.train_starvla][INFO] - Step 723, Loss: {'train/loss': 0.064886674284935, 'train/grad_norm_pre_clip': 2.9223623275756836, 'timing/data': 0.0003213747404515743, 'timing/model': 0.6493216627277434, 'learning_rate/qwen_vl_interface': 9.414298713001007e-06, 'learning_rate/action_model': 9.414298713001006e-05, 'epoch': 0.13})
+[2026-07-28 03:04:31,158][starVLA.training.train_starvla][INFO] - Step 724, Loss: {'train/loss': 0.044576309621334076, 'train/grad_norm_pre_clip': 1.810788869857788, 'timing/data': 0.00032020267099142075, 'timing/model': 0.41195865720510483, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.13})
+[2026-07-28 03:04:37,063][starVLA.training.train_starvla][INFO] - Step 725, Loss: {'train/loss': 0.08327488601207733, 'train/grad_norm_pre_clip': 3.349886417388916, 'timing/data': 0.0003318679518997669, 'timing/model': 0.4155089403502643, 'learning_rate/qwen_vl_interface': 9.410612046442719e-06, 'learning_rate/action_model': 9.410612046442718e-05, 'epoch': 0.13})
+[2026-07-28 03:04:42,924][starVLA.training.train_starvla][INFO] - Step 726, Loss: {'train/loss': 0.08611387014389038, 'train/grad_norm_pre_clip': 1.0664156675338745, 'timing/data': 0.0003204052336513996, 'timing/model': 0.41610111203044653, 'learning_rate/qwen_vl_interface': 9.408764662901097e-06, 'learning_rate/action_model': 9.408764662901098e-05, 'epoch': 0.13})
+[2026-07-28 03:04:48,763][starVLA.training.train_starvla][INFO] - Step 727, Loss: {'train/loss': 0.08564655482769012, 'train/grad_norm_pre_clip': 2.0926055908203125, 'timing/data': 0.00033490872010588646, 'timing/model': 0.41679532965645194, 'learning_rate/qwen_vl_interface': 9.406914580782238e-06, 'learning_rate/action_model': 9.406914580782238e-05, 'epoch': 0.13})
+[2026-07-28 03:04:54,872][starVLA.training.train_starvla][INFO] - Step 728, Loss: {'train/loss': 0.07350964099168777, 'train/grad_norm_pre_clip': 2.418639659881592, 'timing/data': 0.00025230273604393005, 'timing/model': 0.6505269738845527, 'learning_rate/qwen_vl_interface': 9.405061801286638e-06, 'learning_rate/action_model': 9.405061801286638e-05, 'epoch': 0.13})
+[2026-07-28 03:05:00,763][starVLA.training.train_starvla][INFO] - Step 729, Loss: {'train/loss': 0.10231812298297882, 'train/grad_norm_pre_clip': 4.622509002685547, 'timing/data': 0.000254632905125618, 'timing/model': 0.4134670910425484, 'learning_rate/qwen_vl_interface': 9.403206325616546e-06, 'learning_rate/action_model': 9.403206325616545e-05, 'epoch': 0.13})
+[2026-07-28 03:05:06,645][starVLA.training.train_starvla][INFO] - Step 730, Loss: {'train/loss': 0.02758345566689968, 'train/grad_norm_pre_clip': 1.0053476095199585, 'timing/data': 0.0002825879491865635, 'timing/model': 0.4148368942551315, 'learning_rate/qwen_vl_interface': 9.40134815497596e-06, 'learning_rate/action_model': 9.40134815497596e-05, 'epoch': 0.13})
+[2026-07-28 03:05:12,760][starVLA.training.train_starvla][INFO] - Step 731, Loss: {'train/loss': 0.07306163012981415, 'train/grad_norm_pre_clip': 4.052411079406738, 'timing/data': 0.0002635749988257885, 'timing/model': 0.4158944212831557, 'learning_rate/qwen_vl_interface': 9.399487290570629e-06, 'learning_rate/action_model': 9.399487290570629e-05, 'epoch': 0.13})
+[2026-07-28 03:05:18,657][starVLA.training.train_starvla][INFO] - Step 732, Loss: {'train/loss': 0.11125151813030243, 'train/grad_norm_pre_clip': 3.1354286670684814, 'timing/data': 0.00028896285220980644, 'timing/model': 0.4188995398581028, 'learning_rate/qwen_vl_interface': 9.397623733608042e-06, 'learning_rate/action_model': 9.397623733608042e-05, 'epoch': 0.13})
+[2026-07-28 03:05:24,523][starVLA.training.train_starvla][INFO] - Step 733, Loss: {'train/loss': 0.0797296017408371, 'train/grad_norm_pre_clip': 1.8994801044464111, 'timing/data': 0.0004754946567118168, 'timing/model': 0.4174220561981201, 'learning_rate/qwen_vl_interface': 9.395757485297447e-06, 'learning_rate/action_model': 9.395757485297446e-05, 'epoch': 0.13})
+[2026-07-28 03:05:30,390][starVLA.training.train_starvla][INFO] - Step 734, Loss: {'train/loss': 0.11611390858888626, 'train/grad_norm_pre_clip': 1.8844355344772339, 'timing/data': 0.0003717178478837013, 'timing/model': 0.41647569090127945, 'learning_rate/qwen_vl_interface': 9.393888546849828e-06, 'learning_rate/action_model': 9.393888546849828e-05, 'epoch': 0.13})
+[2026-07-28 03:05:36,802][starVLA.training.train_starvla][INFO] - Step 735, Loss: {'train/loss': 0.10063644498586655, 'train/grad_norm_pre_clip': 1.6331285238265991, 'timing/data': 0.00025972677394747734, 'timing/model': 0.6474337298423052, 'learning_rate/qwen_vl_interface': 9.392016919477922e-06, 'learning_rate/action_model': 9.392016919477921e-05, 'epoch': 0.14})
+[2026-07-28 03:05:42,704][starVLA.training.train_starvla][INFO] - Step 736, Loss: {'train/loss': 0.09242890775203705, 'train/grad_norm_pre_clip': 3.9158706665039062, 'timing/data': 0.0002560601569712162, 'timing/model': 0.4173698415979743, 'learning_rate/qwen_vl_interface': 9.390142604396206e-06, 'learning_rate/action_model': 9.390142604396205e-05, 'epoch': 0.14})
+[2026-07-28 03:05:48,624][starVLA.training.train_starvla][INFO] - Step 737, Loss: {'train/loss': 0.09487113356590271, 'train/grad_norm_pre_clip': 1.8513097763061523, 'timing/data': 0.00028561707586050034, 'timing/model': 0.41947178076952696, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.14})
+[2026-07-28 03:05:54,536][starVLA.training.train_starvla][INFO] - Step 738, Loss: {'train/loss': 0.07262498140335083, 'train/grad_norm_pre_clip': 2.524731397628784, 'timing/data': 0.0002958271652460098, 'timing/model': 0.41693501407280564, 'learning_rate/qwen_vl_interface': 9.386385915969975e-06, 'learning_rate/action_model': 9.386385915969976e-05, 'epoch': 0.14})
+[2026-07-28 03:06:00,697][starVLA.training.train_starvla][INFO] - Step 739, Loss: {'train/loss': 0.06325910240411758, 'train/grad_norm_pre_clip': 1.3317246437072754, 'timing/data': 0.000396904069930315, 'timing/model': 0.6695155398920178, 'learning_rate/qwen_vl_interface': 9.384503545063138e-06, 'learning_rate/action_model': 9.384503545063137e-05, 'epoch': 0.14})
+[2026-07-28 03:06:06,596][starVLA.training.train_starvla][INFO] - Step 740, Loss: {'train/loss': 0.07036790996789932, 'train/grad_norm_pre_clip': 1.9146474599838257, 'timing/data': 0.00036326004192233086, 'timing/model': 0.4190991260111332, 'learning_rate/qwen_vl_interface': 9.382618491321837e-06, 'learning_rate/action_model': 9.382618491321837e-05, 'epoch': 0.14})
+[2026-07-28 03:06:12,705][starVLA.training.train_starvla][INFO] - Step 741, Loss: {'train/loss': 0.07785218209028244, 'train/grad_norm_pre_clip': 1.251827359199524, 'timing/data': 0.00024691689759492874, 'timing/model': 0.4154121186584234, 'learning_rate/qwen_vl_interface': 9.380730755969265e-06, 'learning_rate/action_model': 9.380730755969264e-05, 'epoch': 0.14})
+[2026-07-28 03:06:18,593][starVLA.training.train_starvla][INFO] - Step 742, Loss: {'train/loss': 0.06290857493877411, 'train/grad_norm_pre_clip': 1.5977948904037476, 'timing/data': 0.0002819080837070942, 'timing/model': 0.4169664978981018, 'learning_rate/qwen_vl_interface': 9.378840340230353e-06, 'learning_rate/action_model': 9.378840340230352e-05, 'epoch': 0.14})
+[2026-07-28 03:06:24,712][starVLA.training.train_starvla][INFO] - Step 743, Loss: {'train/loss': 0.10728225857019424, 'train/grad_norm_pre_clip': 2.775681734085083, 'timing/data': 0.00026835501194000244, 'timing/model': 0.4151198919862509, 'learning_rate/qwen_vl_interface': 9.376947245331769e-06, 'learning_rate/action_model': 9.376947245331767e-05, 'epoch': 0.14})
+[2026-07-28 03:06:30,573][starVLA.training.train_starvla][INFO] - Step 744, Loss: {'train/loss': 0.062375351786613464, 'train/grad_norm_pre_clip': 1.8652079105377197, 'timing/data': 0.00036154361441731453, 'timing/model': 0.4181189308874309, 'learning_rate/qwen_vl_interface': 9.375051472501922e-06, 'learning_rate/action_model': 9.375051472501923e-05, 'epoch': 0.14})
+[2026-07-28 03:06:36,714][starVLA.training.train_starvla][INFO] - Step 745, Loss: {'train/loss': 0.10872712731361389, 'train/grad_norm_pre_clip': 1.321393609046936, 'timing/data': 0.00028206920251250267, 'timing/model': 0.41976407216861844, 'learning_rate/qwen_vl_interface': 9.373153022970961e-06, 'learning_rate/action_model': 9.373153022970961e-05, 'epoch': 0.14})
+[2026-07-28 03:06:42,633][starVLA.training.train_starvla][INFO] - Step 746, Loss: {'train/loss': 0.09719334542751312, 'train/grad_norm_pre_clip': 3.3071365356445312, 'timing/data': 0.0002914979122579098, 'timing/model': 0.41953073302283883, 'learning_rate/qwen_vl_interface': 9.371251897970768e-06, 'learning_rate/action_model': 9.371251897970768e-05, 'epoch': 0.14})
+[2026-07-28 03:06:48,762][starVLA.training.train_starvla][INFO] - Step 747, Loss: {'train/loss': 0.07473653554916382, 'train/grad_norm_pre_clip': 2.34298038482666, 'timing/data': 0.0002466789446771145, 'timing/model': 0.4147561131976545, 'learning_rate/qwen_vl_interface': 9.369348098734961e-06, 'learning_rate/action_model': 9.369348098734961e-05, 'epoch': 0.14})
+[2026-07-28 03:06:54,891][starVLA.training.train_starvla][INFO] - Step 748, Loss: {'train/loss': 0.058117564767599106, 'train/grad_norm_pre_clip': 2.431900978088379, 'timing/data': 0.0002842498943209648, 'timing/model': 0.4171718875877559, 'learning_rate/qwen_vl_interface': 9.367441626498897e-06, 'learning_rate/action_model': 9.367441626498896e-05, 'epoch': 0.14})
+[2026-07-28 03:07:00,773][starVLA.training.train_starvla][INFO] - Step 749, Loss: {'train/loss': 0.049435704946517944, 'train/grad_norm_pre_clip': 14.755330085754395, 'timing/data': 0.00029116589576005936, 'timing/model': 0.42224577674642205, 'learning_rate/qwen_vl_interface': 9.365532482499663e-06, 'learning_rate/action_model': 9.365532482499662e-05, 'epoch': 0.14})
+[2026-07-28 03:07:06,874][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 750
+[2026-07-28 03:07:11,819][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 750
+[2026-07-28 03:11:36,371][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 750
+[2026-07-28 03:11:36,373][starVLA.training.train_starvla][INFO] - Step 750, Loss: {'train/loss': 0.08797957003116608, 'train/grad_norm_pre_clip': 1.8115043640136719, 'eval/action_loss/seconds': 4.942189701832831, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.06308243423700333, 'eval/latency_3/loss': 0.06308243423700333, 'eval/flappy/loss': 0.06308243423700333, 'eval/flappy/latency_3/loss': 0.06308243423700333, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 199.45500236190856, 'rl_games_eval/mid_train/mean_length': 1627.7, 'rl_games_eval/mid_train/std_reward': 161.18966493830942, 'rl_games_eval/mid_train/std_length': 1299.6861197997, 'rl_games_eval/mid_train/macro_mean_reward': 199.45500236190856, 'rl_games_eval/mid_train/macro_mean_length': 1627.7, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 199.45500236190856, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 1627.7, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 161.18966493830942, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1299.6861197997, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 199.45500236190856, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0003632018342614174, 'timing/model': 0.6551510407589376, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.14})
+[2026-07-28 03:11:42,308][starVLA.training.train_starvla][INFO] - Step 751, Loss: {'train/loss': 0.0934058427810669, 'train/grad_norm_pre_clip': 1.8735677003860474, 'timing/data': 0.00029357708990573883, 'timing/model': 0.41588742891326547, 'learning_rate/qwen_vl_interface': 9.361706184168714e-06, 'learning_rate/action_model': 9.361706184168713e-05, 'epoch': 0.14})
+[2026-07-28 03:11:48,373][starVLA.training.train_starvla][INFO] - Step 752, Loss: {'train/loss': 0.09223182499408722, 'train/grad_norm_pre_clip': 2.4583001136779785, 'timing/data': 0.0002680523321032524, 'timing/model': 0.41431050561368465, 'learning_rate/qwen_vl_interface': 9.359789032319843e-06, 'learning_rate/action_model': 9.359789032319842e-05, 'epoch': 0.14})
+[2026-07-28 03:11:54,443][starVLA.training.train_starvla][INFO] - Step 753, Loss: {'train/loss': 0.040510937571525574, 'train/grad_norm_pre_clip': 0.9940184354782104, 'timing/data': 0.0002643261104822159, 'timing/model': 0.4136805427260697, 'learning_rate/qwen_vl_interface': 9.357869213673484e-06, 'learning_rate/action_model': 9.357869213673485e-05, 'epoch': 0.14})
+[2026-07-28 03:12:00,301][starVLA.training.train_starvla][INFO] - Step 754, Loss: {'train/loss': 0.13626444339752197, 'train/grad_norm_pre_clip': 4.096269607543945, 'timing/data': 0.000370061956346035, 'timing/model': 0.416091650724411, 'learning_rate/qwen_vl_interface': 9.355946729475393e-06, 'learning_rate/action_model': 9.355946729475393e-05, 'epoch': 0.14})
+[2026-07-28 03:12:06,384][starVLA.training.train_starvla][INFO] - Step 755, Loss: {'train/loss': 0.05483660101890564, 'train/grad_norm_pre_clip': 2.8205244541168213, 'timing/data': 0.00026058731600642204, 'timing/model': 0.6517208041623235, 'learning_rate/qwen_vl_interface': 9.354021580973049e-06, 'learning_rate/action_model': 9.354021580973048e-05, 'epoch': 0.14})
+[2026-07-28 03:12:12,439][starVLA.training.train_starvla][INFO] - Step 756, Loss: {'train/loss': 0.062632255256176, 'train/grad_norm_pre_clip': 1.4003318548202515, 'timing/data': 0.00025544408708810806, 'timing/model': 0.4122348576784134, 'learning_rate/qwen_vl_interface': 9.352093769415655e-06, 'learning_rate/action_model': 9.352093769415655e-05, 'epoch': 0.14})
+[2026-07-28 03:12:18,269][starVLA.training.train_starvla][INFO] - Step 757, Loss: {'train/loss': 0.06106656789779663, 'train/grad_norm_pre_clip': 1.0041924715042114, 'timing/data': 0.00027470663189888, 'timing/model': 0.41519576916471124, 'learning_rate/qwen_vl_interface': 9.35016329605415e-06, 'learning_rate/action_model': 9.350163296054151e-05, 'epoch': 0.14})
+[2026-07-28 03:12:24,156][starVLA.training.train_starvla][INFO] - Step 758, Loss: {'train/loss': 0.026967184618115425, 'train/grad_norm_pre_clip': 1.1977174282073975, 'timing/data': 0.0003521377220749855, 'timing/model': 0.41565106995403767, 'learning_rate/qwen_vl_interface': 9.348230162141198e-06, 'learning_rate/action_model': 9.348230162141198e-05, 'epoch': 0.14})
+[2026-07-28 03:12:30,234][starVLA.training.train_starvla][INFO] - Step 759, Loss: {'train/loss': 0.04603224992752075, 'train/grad_norm_pre_clip': 2.1533429622650146, 'timing/data': 0.0002565761096775532, 'timing/model': 0.4147330312989652, 'learning_rate/qwen_vl_interface': 9.346294368931188e-06, 'learning_rate/action_model': 9.346294368931188e-05, 'epoch': 0.14})
+[2026-07-28 03:12:36,382][starVLA.training.train_starvla][INFO] - Step 760, Loss: {'train/loss': 0.037652794271707535, 'train/grad_norm_pre_clip': 1.812495470046997, 'timing/data': 0.00028409669175744057, 'timing/model': 0.7086217738687992, 'learning_rate/qwen_vl_interface': 9.344355917680235e-06, 'learning_rate/action_model': 9.344355917680235e-05, 'epoch': 0.14})
+[2026-07-28 03:12:42,296][starVLA.training.train_starvla][INFO] - Step 761, Loss: {'train/loss': 0.1376267969608307, 'train/grad_norm_pre_clip': 4.4042158126831055, 'timing/data': 0.0002503693103790283, 'timing/model': 0.41335098817944527, 'learning_rate/qwen_vl_interface': 9.342414809646179e-06, 'learning_rate/action_model': 9.342414809646179e-05, 'epoch': 0.14})
+[2026-07-28 03:12:48,188][starVLA.training.train_starvla][INFO] - Step 762, Loss: {'train/loss': 0.10081770271062851, 'train/grad_norm_pre_clip': 2.8392415046691895, 'timing/data': 0.0003595338203012943, 'timing/model': 0.41624417109414935, 'learning_rate/qwen_vl_interface': 9.340471046088581e-06, 'learning_rate/action_model': 9.34047104608858e-05, 'epoch': 0.14})
+[2026-07-28 03:12:54,321][starVLA.training.train_starvla][INFO] - Step 763, Loss: {'train/loss': 0.10521991550922394, 'train/grad_norm_pre_clip': 4.032131195068359, 'timing/data': 0.00025624269619584084, 'timing/model': 0.4143365710042417, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.14})
+[2026-07-28 03:13:00,214][starVLA.training.train_starvla][INFO] - Step 764, Loss: {'train/loss': 0.08215293288230896, 'train/grad_norm_pre_clip': 2.1438441276550293, 'timing/data': 0.0003435462713241577, 'timing/model': 0.417494208086282, 'learning_rate/qwen_vl_interface': 9.336575557449641e-06, 'learning_rate/action_model': 9.336575557449641e-05, 'epoch': 0.14})
+[2026-07-28 03:13:06,174][starVLA.training.train_starvla][INFO] - Step 765, Loss: {'train/loss': 0.04260343685746193, 'train/grad_norm_pre_clip': 1.4370815753936768, 'timing/data': 0.0002777986228466034, 'timing/model': 0.4160189931280911, 'learning_rate/qwen_vl_interface': 9.334623834896039e-06, 'learning_rate/action_model': 9.334623834896038e-05, 'epoch': 0.14})
+[2026-07-28 03:13:12,542][starVLA.training.train_starvla][INFO] - Step 766, Loss: {'train/loss': 0.14002515375614166, 'train/grad_norm_pre_clip': 3.727264165878296, 'timing/data': 0.0003025210462510586, 'timing/model': 0.6487685008905828, 'learning_rate/qwen_vl_interface': 9.332669461874376e-06, 'learning_rate/action_model': 9.332669461874374e-05, 'epoch': 0.14})
+[2026-07-28 03:13:18,447][starVLA.training.train_starvla][INFO] - Step 767, Loss: {'train/loss': 0.15723171830177307, 'train/grad_norm_pre_clip': 5.149575233459473, 'timing/data': 0.0003504040651023388, 'timing/model': 0.4167680302634835, 'learning_rate/qwen_vl_interface': 9.330712439652824e-06, 'learning_rate/action_model': 9.330712439652823e-05, 'epoch': 0.14})
+[2026-07-28 03:13:24,342][starVLA.training.train_starvla][INFO] - Step 768, Loss: {'train/loss': 0.08578182011842728, 'train/grad_norm_pre_clip': 2.8332042694091797, 'timing/data': 0.0002739229239523411, 'timing/model': 0.4169467817991972, 'learning_rate/qwen_vl_interface': 9.328752769501272e-06, 'learning_rate/action_model': 9.328752769501272e-05, 'epoch': 0.14})
+[2026-07-28 03:13:30,223][starVLA.training.train_starvla][INFO] - Step 769, Loss: {'train/loss': 0.11532429605722427, 'train/grad_norm_pre_clip': 2.5758376121520996, 'timing/data': 0.0003415010869503021, 'timing/model': 0.4143184646964073, 'learning_rate/qwen_vl_interface': 9.326790452691332e-06, 'learning_rate/action_model': 9.326790452691332e-05, 'epoch': 0.14})
+[2026-07-28 03:13:36,382][starVLA.training.train_starvla][INFO] - Step 770, Loss: {'train/loss': 0.07786200940608978, 'train/grad_norm_pre_clip': 1.7413674592971802, 'timing/data': 0.0003303070552647114, 'timing/model': 0.41530506080016494, 'learning_rate/qwen_vl_interface': 9.324825490496327e-06, 'learning_rate/action_model': 9.324825490496327e-05, 'epoch': 0.14})
+[2026-07-28 03:13:42,501][starVLA.training.train_starvla][INFO] - Step 771, Loss: {'train/loss': 0.07403310388326645, 'train/grad_norm_pre_clip': 1.3776592016220093, 'timing/data': 0.00036082882434129715, 'timing/model': 0.6550813540816307, 'learning_rate/qwen_vl_interface': 9.322857884191303e-06, 'learning_rate/action_model': 9.322857884191302e-05, 'epoch': 0.14})
+[2026-07-28 03:13:48,354][starVLA.training.train_starvla][INFO] - Step 772, Loss: {'train/loss': 0.06675441563129425, 'train/grad_norm_pre_clip': 3.24139142036438, 'timing/data': 0.00031166383996605873, 'timing/model': 0.4140995447523892, 'learning_rate/qwen_vl_interface': 9.320887635053013e-06, 'learning_rate/action_model': 9.320887635053014e-05, 'epoch': 0.14})
+[2026-07-28 03:13:54,736][starVLA.training.train_starvla][INFO] - Step 773, Loss: {'train/loss': 0.1417200267314911, 'train/grad_norm_pre_clip': 2.7916834354400635, 'timing/data': 0.0003264578990638256, 'timing/model': 0.4142764350399375, 'learning_rate/qwen_vl_interface': 9.318914744359936e-06, 'learning_rate/action_model': 9.318914744359937e-05, 'epoch': 0.14})
+[2026-07-28 03:14:00,886][starVLA.training.train_starvla][INFO] - Step 774, Loss: {'train/loss': 0.0738334208726883, 'train/grad_norm_pre_clip': 1.7493247985839844, 'timing/data': 0.000262437853962183, 'timing/model': 0.4133771019987762, 'learning_rate/qwen_vl_interface': 9.316939213392257e-06, 'learning_rate/action_model': 9.316939213392257e-05, 'epoch': 0.14})
+[2026-07-28 03:14:07,002][starVLA.training.train_starvla][INFO] - Step 775, Loss: {'train/loss': 0.08354192227125168, 'train/grad_norm_pre_clip': 2.8963570594787598, 'timing/data': 0.00026347627863287926, 'timing/model': 0.4334867289289832, 'learning_rate/qwen_vl_interface': 9.314961043431877e-06, 'learning_rate/action_model': 9.314961043431876e-05, 'epoch': 0.14})
+[2026-07-28 03:14:12,889][starVLA.training.train_starvla][INFO] - Step 776, Loss: {'train/loss': 0.06884749978780746, 'train/grad_norm_pre_clip': 0.994733989238739, 'timing/data': 0.00028054602444171906, 'timing/model': 0.41997288167476654, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.14})
+[2026-07-28 03:14:18,858][starVLA.training.train_starvla][INFO] - Step 777, Loss: {'train/loss': 0.09571895003318787, 'train/grad_norm_pre_clip': 1.5205873250961304, 'timing/data': 0.00026978272944688797, 'timing/model': 0.6684925360605121, 'learning_rate/qwen_vl_interface': 9.310996791669175e-06, 'learning_rate/action_model': 9.310996791669175e-05, 'epoch': 0.14})
+[2026-07-28 03:14:24,996][starVLA.training.train_starvla][INFO] - Step 778, Loss: {'train/loss': 0.07641018182039261, 'train/grad_norm_pre_clip': 1.2037880420684814, 'timing/data': 0.00024709664285182953, 'timing/model': 0.4161041700281203, 'learning_rate/qwen_vl_interface': 9.309010712439213e-06, 'learning_rate/action_model': 9.309010712439213e-05, 'epoch': 0.14})
+[2026-07-28 03:14:30,900][starVLA.training.train_starvla][INFO] - Step 779, Loss: {'train/loss': 0.08378678560256958, 'train/grad_norm_pre_clip': 1.3989732265472412, 'timing/data': 0.00034831371158361435, 'timing/model': 0.4197115497663617, 'learning_rate/qwen_vl_interface': 9.307021999361267e-06, 'learning_rate/action_model': 9.307021999361267e-05, 'epoch': 0.14})
+[2026-07-28 03:14:36,811][starVLA.training.train_starvla][INFO] - Step 780, Loss: {'train/loss': 0.09568702429533005, 'train/grad_norm_pre_clip': 1.6845974922180176, 'timing/data': 0.00029063737019896507, 'timing/model': 0.41681061405688524, 'learning_rate/qwen_vl_interface': 9.305030653725794e-06, 'learning_rate/action_model': 9.305030653725794e-05, 'epoch': 0.14})
+[2026-07-28 03:14:42,809][starVLA.training.train_starvla][INFO] - Step 781, Loss: {'train/loss': 0.04570449888706207, 'train/grad_norm_pre_clip': 1.4001655578613281, 'timing/data': 0.0002708300016820431, 'timing/model': 0.4180960808880627, 'learning_rate/qwen_vl_interface': 9.303036676824954e-06, 'learning_rate/action_model': 9.303036676824953e-05, 'epoch': 0.14})
+[2026-07-28 03:14:48,696][starVLA.training.train_starvla][INFO] - Step 782, Loss: {'train/loss': 0.05355058237910271, 'train/grad_norm_pre_clip': 1.2345517873764038, 'timing/data': 0.0002726539969444275, 'timing/model': 0.41787185706198215, 'learning_rate/qwen_vl_interface': 9.301040069952616e-06, 'learning_rate/action_model': 9.301040069952616e-05, 'epoch': 0.14})
+[2026-07-28 03:14:54,812][starVLA.training.train_starvla][INFO] - Step 783, Loss: {'train/loss': 0.07088715583086014, 'train/grad_norm_pre_clip': 3.3040902614593506, 'timing/data': 0.00026034098118543625, 'timing/model': 0.4171629720367491, 'learning_rate/qwen_vl_interface': 9.299040834404362e-06, 'learning_rate/action_model': 9.299040834404361e-05, 'epoch': 0.14})
+[2026-07-28 03:15:00,923][starVLA.training.train_starvla][INFO] - Step 784, Loss: {'train/loss': 0.09239332377910614, 'train/grad_norm_pre_clip': 3.881669282913208, 'timing/data': 0.00025893887504935265, 'timing/model': 0.415586035232991, 'learning_rate/qwen_vl_interface': 9.297038971477468e-06, 'learning_rate/action_model': 9.297038971477466e-05, 'epoch': 0.14})
+[2026-07-28 03:15:06,832][starVLA.training.train_starvla][INFO] - Step 785, Loss: {'train/loss': 0.07434717565774918, 'train/grad_norm_pre_clip': 4.022629261016846, 'timing/data': 0.0002712658606469631, 'timing/model': 0.41779134701937437, 'learning_rate/qwen_vl_interface': 9.295034482470922e-06, 'learning_rate/action_model': 9.295034482470923e-05, 'epoch': 0.14})
+[2026-07-28 03:15:12,750][starVLA.training.train_starvla][INFO] - Step 786, Loss: {'train/loss': 0.08836290240287781, 'train/grad_norm_pre_clip': 3.8501405715942383, 'timing/data': 0.0002731126733124256, 'timing/model': 0.4173386939801276, 'learning_rate/qwen_vl_interface': 9.293027368685418e-06, 'learning_rate/action_model': 9.293027368685419e-05, 'epoch': 0.14})
+[2026-07-28 03:15:18,894][starVLA.training.train_starvla][INFO] - Step 787, Loss: {'train/loss': 0.0748005136847496, 'train/grad_norm_pre_clip': 2.0694122314453125, 'timing/data': 0.000404892023652792, 'timing/model': 0.4165114280767739, 'learning_rate/qwen_vl_interface': 9.29101763142335e-06, 'learning_rate/action_model': 9.29101763142335e-05, 'epoch': 0.14})
+[2026-07-28 03:15:24,818][starVLA.training.train_starvla][INFO] - Step 788, Loss: {'train/loss': 0.14345842599868774, 'train/grad_norm_pre_clip': 2.731043577194214, 'timing/data': 0.0002814917825162411, 'timing/model': 0.4693546248599887, 'learning_rate/qwen_vl_interface': 9.28900527198881e-06, 'learning_rate/action_model': 9.289005271988811e-05, 'epoch': 0.14})
+[2026-07-28 03:15:30,935][starVLA.training.train_starvla][INFO] - Step 789, Loss: {'train/loss': 0.10270923376083374, 'train/grad_norm_pre_clip': 3.909942626953125, 'timing/data': 0.0002601989544928074, 'timing/model': 0.41565225226804614, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.14})
+[2026-07-28 03:15:37,037][starVLA.training.train_starvla][INFO] - Step 790, Loss: {'train/loss': 0.17248031497001648, 'train/grad_norm_pre_clip': 3.755141258239746, 'timing/data': 0.0002819402143359184, 'timing/model': 0.6605090759694576, 'learning_rate/qwen_vl_interface': 9.284972691827226e-06, 'learning_rate/action_model': 9.284972691827224e-05, 'epoch': 0.15})
+[2026-07-28 03:15:43,011][starVLA.training.train_starvla][INFO] - Step 791, Loss: {'train/loss': 0.0759965106844902, 'train/grad_norm_pre_clip': 2.7975292205810547, 'timing/data': 0.0002862820401787758, 'timing/model': 0.4171938570216298, 'learning_rate/qwen_vl_interface': 9.282952473716873e-06, 'learning_rate/action_model': 9.282952473716874e-05, 'epoch': 0.15})
+[2026-07-28 03:15:48,902][starVLA.training.train_starvla][INFO] - Step 792, Loss: {'train/loss': 0.07177821546792984, 'train/grad_norm_pre_clip': 1.7207838296890259, 'timing/data': 0.0003490280359983444, 'timing/model': 0.4176006759516895, 'learning_rate/qwen_vl_interface': 9.280929638667447e-06, 'learning_rate/action_model': 9.280929638667447e-05, 'epoch': 0.15})
+[2026-07-28 03:15:55,048][starVLA.training.train_starvla][INFO] - Step 793, Loss: {'train/loss': 0.10813035070896149, 'train/grad_norm_pre_clip': 3.285444498062134, 'timing/data': 0.00025442056357860565, 'timing/model': 0.4163041738793254, 'learning_rate/qwen_vl_interface': 9.278904187991542e-06, 'learning_rate/action_model': 9.278904187991541e-05, 'epoch': 0.15})
+[2026-07-28 03:16:00,941][starVLA.training.train_starvla][INFO] - Step 794, Loss: {'train/loss': 0.07914471626281738, 'train/grad_norm_pre_clip': 4.135631084442139, 'timing/data': 0.00035881204530596733, 'timing/model': 0.4175904649309814, 'learning_rate/qwen_vl_interface': 9.27687612300345e-06, 'learning_rate/action_model': 9.27687612300345e-05, 'epoch': 0.15})
+[2026-07-28 03:16:07,068][starVLA.training.train_starvla][INFO] - Step 795, Loss: {'train/loss': 0.13084831833839417, 'train/grad_norm_pre_clip': 2.2043910026550293, 'timing/data': 0.0002539539709687233, 'timing/model': 0.4157147682271898, 'learning_rate/qwen_vl_interface': 9.27484544501916e-06, 'learning_rate/action_model': 9.274845445019159e-05, 'epoch': 0.15})
+[2026-07-28 03:16:13,207][starVLA.training.train_starvla][INFO] - Step 796, Loss: {'train/loss': 0.12284715473651886, 'train/grad_norm_pre_clip': 2.4248299598693848, 'timing/data': 0.0002535036765038967, 'timing/model': 0.41715946001932025, 'learning_rate/qwen_vl_interface': 9.27281215535636e-06, 'learning_rate/action_model': 9.272812155356358e-05, 'epoch': 0.15})
+[2026-07-28 03:16:19,119][starVLA.training.train_starvla][INFO] - Step 797, Loss: {'train/loss': 0.1132897436618805, 'train/grad_norm_pre_clip': 2.6659533977508545, 'timing/data': 0.00026736175641417503, 'timing/model': 0.41767385182902217, 'learning_rate/qwen_vl_interface': 9.270776255334425e-06, 'learning_rate/action_model': 9.270776255334425e-05, 'epoch': 0.15})
+[2026-07-28 03:16:25,019][starVLA.training.train_starvla][INFO] - Step 798, Loss: {'train/loss': 0.13933779299259186, 'train/grad_norm_pre_clip': 1.5746897459030151, 'timing/data': 0.00036012986674904823, 'timing/model': 0.41812005499377847, 'learning_rate/qwen_vl_interface': 9.268737746274432e-06, 'learning_rate/action_model': 9.268737746274432e-05, 'epoch': 0.15})
+[2026-07-28 03:16:30,937][starVLA.training.train_starvla][INFO] - Step 799, Loss: {'train/loss': 0.08533799648284912, 'train/grad_norm_pre_clip': 1.4278755187988281, 'timing/data': 0.0002710530534386635, 'timing/model': 0.4191224961541593, 'learning_rate/qwen_vl_interface': 9.266696629499146e-06, 'learning_rate/action_model': 9.266696629499145e-05, 'epoch': 0.15})
+[2026-07-28 03:16:37,066][starVLA.training.train_starvla][INFO] - Step 800, Loss: {'train/loss': 0.11101211607456207, 'train/grad_norm_pre_clip': 2.342465877532959, 'timing/data': 0.00035662855952978134, 'timing/model': 0.6546974070370197, 'learning_rate/qwen_vl_interface': 9.264652906333026e-06, 'learning_rate/action_model': 9.264652906333026e-05, 'epoch': 0.15})
+[2026-07-28 03:16:42,958][starVLA.training.train_starvla][INFO] - Step 801, Loss: {'train/loss': 0.09408314526081085, 'train/grad_norm_pre_clip': 1.7896021604537964, 'timing/data': 0.0003445991314947605, 'timing/model': 0.4172967686317861, 'learning_rate/qwen_vl_interface': 9.26260657810222e-06, 'learning_rate/action_model': 9.262606578102221e-05, 'epoch': 0.15})
+[2026-07-28 03:16:48,885][starVLA.training.train_starvla][INFO] - Step 802, Loss: {'train/loss': 0.04039296135306358, 'train/grad_norm_pre_clip': 0.8686643242835999, 'timing/data': 0.0002726982347667217, 'timing/model': 0.4183682440780103, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.15})
+[2026-07-28 03:16:54,767][starVLA.training.train_starvla][INFO] - Step 803, Loss: {'train/loss': 0.05151866376399994, 'train/grad_norm_pre_clip': 0.8761146664619446, 'timing/data': 0.000275271013379097, 'timing/model': 0.4171714182011783, 'learning_rate/qwen_vl_interface': 9.258506111759608e-06, 'learning_rate/action_model': 9.258506111759609e-05, 'epoch': 0.15})
+[2026-07-28 03:17:00,859][starVLA.training.train_starvla][INFO] - Step 804, Loss: {'train/loss': 0.0468764491379261, 'train/grad_norm_pre_clip': 1.2680186033248901, 'timing/data': 0.00024797674268484116, 'timing/model': 0.4171336321160197, 'learning_rate/qwen_vl_interface': 9.256451976308548e-06, 'learning_rate/action_model': 9.25645197630855e-05, 'epoch': 0.15})
+[2026-07-28 03:17:07,001][starVLA.training.train_starvla][INFO] - Step 805, Loss: {'train/loss': 0.08636626601219177, 'train/grad_norm_pre_clip': 2.1965575218200684, 'timing/data': 0.00027381256222724915, 'timing/model': 0.41918690130114555, 'learning_rate/qwen_vl_interface': 9.2543952411143e-06, 'learning_rate/action_model': 9.254395241114299e-05, 'epoch': 0.15})
+[2026-07-28 03:17:12,937][starVLA.training.train_starvla][INFO] - Step 806, Loss: {'train/loss': 0.151370570063591, 'train/grad_norm_pre_clip': 3.7983152866363525, 'timing/data': 0.0002948869951069355, 'timing/model': 0.4177323658950627, 'learning_rate/qwen_vl_interface': 9.252335907511452e-06, 'learning_rate/action_model': 9.252335907511453e-05, 'epoch': 0.15})
+[2026-07-28 03:17:18,812][starVLA.training.train_starvla][INFO] - Step 807, Loss: {'train/loss': 0.07472165673971176, 'train/grad_norm_pre_clip': 1.6507024765014648, 'timing/data': 0.00026828423142433167, 'timing/model': 0.41847079200670123, 'learning_rate/qwen_vl_interface': 9.250273976836289e-06, 'learning_rate/action_model': 9.250273976836289e-05, 'epoch': 0.15})
+[2026-07-28 03:17:25,142][starVLA.training.train_starvla][INFO] - Step 808, Loss: {'train/loss': 0.06295875459909439, 'train/grad_norm_pre_clip': 1.2865005731582642, 'timing/data': 0.000325529370456934, 'timing/model': 0.4172204569913447, 'learning_rate/qwen_vl_interface': 9.248209450426772e-06, 'learning_rate/action_model': 9.248209450426772e-05, 'epoch': 0.15})
+[2026-07-28 03:17:31,019][starVLA.training.train_starvla][INFO] - Step 809, Loss: {'train/loss': 0.06980820000171661, 'train/grad_norm_pre_clip': 2.535649299621582, 'timing/data': 0.000272626057267189, 'timing/model': 0.4172631739638746, 'learning_rate/qwen_vl_interface': 9.24614232962255e-06, 'learning_rate/action_model': 9.24614232962255e-05, 'epoch': 0.15})
+[2026-07-28 03:17:36,941][starVLA.training.train_starvla][INFO] - Step 810, Loss: {'train/loss': 0.11817154288291931, 'train/grad_norm_pre_clip': 4.01273250579834, 'timing/data': 0.00026792101562023163, 'timing/model': 0.41745619801804423, 'learning_rate/qwen_vl_interface': 9.244072615764957e-06, 'learning_rate/action_model': 9.244072615764956e-05, 'epoch': 0.15})
+[2026-07-28 03:17:43,141][starVLA.training.train_starvla][INFO] - Step 811, Loss: {'train/loss': 0.06146187335252762, 'train/grad_norm_pre_clip': 1.6038048267364502, 'timing/data': 0.0002679852768778801, 'timing/model': 0.41726704640313983, 'learning_rate/qwen_vl_interface': 9.242000310197005e-06, 'learning_rate/action_model': 9.242000310197004e-05, 'epoch': 0.15})
+[2026-07-28 03:17:49,230][starVLA.training.train_starvla][INFO] - Step 812, Loss: {'train/loss': 0.05522509291768074, 'train/grad_norm_pre_clip': 1.698875069618225, 'timing/data': 0.00027466705068945885, 'timing/model': 0.4170692772604525, 'learning_rate/qwen_vl_interface': 9.239925414263392e-06, 'learning_rate/action_model': 9.239925414263392e-05, 'epoch': 0.15})
+[2026-07-28 03:17:54,895][starVLA.training.train_starvla][INFO] - Step 813, Loss: {'train/loss': 0.11641747504472733, 'train/grad_norm_pre_clip': 3.1129114627838135, 'timing/data': 0.0002770228311419487, 'timing/model': 0.41755296709015965, 'learning_rate/qwen_vl_interface': 9.237847929310497e-06, 'learning_rate/action_model': 9.237847929310496e-05, 'epoch': 0.15})
+[2026-07-28 03:18:00,566][starVLA.training.train_starvla][INFO] - Step 814, Loss: {'train/loss': 0.06920818984508514, 'train/grad_norm_pre_clip': 1.228936791419983, 'timing/data': 0.00031775515526533127, 'timing/model': 0.4188155746087432, 'learning_rate/qwen_vl_interface': 9.235767856686375e-06, 'learning_rate/action_model': 9.235767856686375e-05, 'epoch': 0.15})
+[2026-07-28 03:18:06,449][starVLA.training.train_starvla][INFO] - Step 815, Loss: {'train/loss': 0.06573324650526047, 'train/grad_norm_pre_clip': 2.9914472103118896, 'timing/data': 0.0002633221447467804, 'timing/model': 0.6527394140139222, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.15})
+[2026-07-28 03:18:12,683][starVLA.training.train_starvla][INFO] - Step 816, Loss: {'train/loss': 0.09059537947177887, 'train/grad_norm_pre_clip': 3.622300624847412, 'timing/data': 0.0002611991949379444, 'timing/model': 0.45702221104875207, 'learning_rate/qwen_vl_interface': 9.23159995382508e-06, 'learning_rate/action_model': 9.231599953825079e-05, 'epoch': 0.15})
+[2026-07-28 03:18:18,830][starVLA.training.train_starvla][INFO] - Step 817, Loss: {'train/loss': 0.08247837424278259, 'train/grad_norm_pre_clip': 2.1487374305725098, 'timing/data': 0.00027635227888822556, 'timing/model': 0.42434668401256204, 'learning_rate/qwen_vl_interface': 9.22951212629241e-06, 'learning_rate/action_model': 9.229512126292411e-05, 'epoch': 0.15})
+[2026-07-28 03:18:25,016][starVLA.training.train_starvla][INFO] - Step 818, Loss: {'train/loss': 0.1036224514245987, 'train/grad_norm_pre_clip': 1.6684305667877197, 'timing/data': 0.0002760491333901882, 'timing/model': 0.41928426688537, 'learning_rate/qwen_vl_interface': 9.227421716497531e-06, 'learning_rate/action_model': 9.227421716497531e-05, 'epoch': 0.15})
+[2026-07-28 03:18:31,348][starVLA.training.train_starvla][INFO] - Step 819, Loss: {'train/loss': 0.09708300977945328, 'train/grad_norm_pre_clip': 2.204148054122925, 'timing/data': 0.00026531098410487175, 'timing/model': 0.4150994117371738, 'learning_rate/qwen_vl_interface': 9.225328725796879e-06, 'learning_rate/action_model': 9.22532872579688e-05, 'epoch': 0.15})
+[2026-07-28 03:18:37,225][starVLA.training.train_starvla][INFO] - Step 820, Loss: {'train/loss': 0.09058910608291626, 'train/grad_norm_pre_clip': 2.6100292205810547, 'timing/data': 0.0002879300154745579, 'timing/model': 0.4162336443550885, 'learning_rate/qwen_vl_interface': 9.223233155548579e-06, 'learning_rate/action_model': 9.223233155548578e-05, 'epoch': 0.15})
+[2026-07-28 03:18:43,332][starVLA.training.train_starvla][INFO] - Step 821, Loss: {'train/loss': 0.04704709351062775, 'train/grad_norm_pre_clip': 1.2586156129837036, 'timing/data': 0.0002561500295996666, 'timing/model': 0.655124447774142, 'learning_rate/qwen_vl_interface': 9.22113500711242e-06, 'learning_rate/action_model': 9.221135007112419e-05, 'epoch': 0.15})
+[2026-07-28 03:18:49,179][starVLA.training.train_starvla][INFO] - Step 822, Loss: {'train/loss': 0.10550558567047119, 'train/grad_norm_pre_clip': 1.9187953472137451, 'timing/data': 0.0002587428316473961, 'timing/model': 0.4147767759859562, 'learning_rate/qwen_vl_interface': 9.219034281849867e-06, 'learning_rate/action_model': 9.219034281849867e-05, 'epoch': 0.15})
+[2026-07-28 03:18:55,261][starVLA.training.train_starvla][INFO] - Step 823, Loss: {'train/loss': 0.09689384698867798, 'train/grad_norm_pre_clip': 2.2860493659973145, 'timing/data': 0.0003435262478888035, 'timing/model': 0.4133647410199046, 'learning_rate/qwen_vl_interface': 9.216930981124062e-06, 'learning_rate/action_model': 9.216930981124062e-05, 'epoch': 0.15})
+[2026-07-28 03:19:01,137][starVLA.training.train_starvla][INFO] - Step 824, Loss: {'train/loss': 0.09124375879764557, 'train/grad_norm_pre_clip': 2.453585147857666, 'timing/data': 0.00026872195303440094, 'timing/model': 0.41568814497441053, 'learning_rate/qwen_vl_interface': 9.214825106299808e-06, 'learning_rate/action_model': 9.214825106299808e-05, 'epoch': 0.15})
+[2026-07-28 03:19:06,989][starVLA.training.train_starvla][INFO] - Step 825, Loss: {'train/loss': 0.10570086538791656, 'train/grad_norm_pre_clip': 1.8532730340957642, 'timing/data': 0.0002863658592104912, 'timing/model': 0.41625818191096187, 'learning_rate/qwen_vl_interface': 9.212716658743592e-06, 'learning_rate/action_model': 9.212716658743592e-05, 'epoch': 0.15})
+[2026-07-28 03:19:13,306][starVLA.training.train_starvla][INFO] - Step 826, Loss: {'train/loss': 0.0660516694188118, 'train/grad_norm_pre_clip': 1.3169517517089844, 'timing/data': 0.00025269389152526855, 'timing/model': 0.4128501582890749, 'learning_rate/qwen_vl_interface': 9.210605639823558e-06, 'learning_rate/action_model': 9.210605639823557e-05, 'epoch': 0.15})
+[2026-07-28 03:19:19,196][starVLA.training.train_starvla][INFO] - Step 827, Loss: {'train/loss': 0.06137767806649208, 'train/grad_norm_pre_clip': 0.9235149025917053, 'timing/data': 0.0002887677401304245, 'timing/model': 0.41816171258687973, 'learning_rate/qwen_vl_interface': 9.208492050909523e-06, 'learning_rate/action_model': 9.208492050909523e-05, 'epoch': 0.15})
+[2026-07-28 03:19:25,068][starVLA.training.train_starvla][INFO] - Step 828, Loss: {'train/loss': 0.03920440375804901, 'train/grad_norm_pre_clip': 0.9559926390647888, 'timing/data': 0.0002769078128039837, 'timing/model': 0.41489687422290444, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.15})
+[2026-07-28 03:19:31,178][starVLA.training.train_starvla][INFO] - Step 829, Loss: {'train/loss': 0.06328506767749786, 'train/grad_norm_pre_clip': 1.3056012392044067, 'timing/data': 0.00025679776445031166, 'timing/model': 0.4128893599845469, 'learning_rate/qwen_vl_interface': 9.204257168587061e-06, 'learning_rate/action_model': 9.204257168587061e-05, 'epoch': 0.15})
+[2026-07-28 03:19:37,273][starVLA.training.train_starvla][INFO] - Step 830, Loss: {'train/loss': 0.1000927984714508, 'train/grad_norm_pre_clip': 2.6133601665496826, 'timing/data': 0.0002763140946626663, 'timing/model': 0.6517327539622784, 'learning_rate/qwen_vl_interface': 9.2021358779266e-06, 'learning_rate/action_model': 9.202135877926601e-05, 'epoch': 0.15})
+[2026-07-28 03:19:43,117][starVLA.training.train_starvla][INFO] - Step 831, Loss: {'train/loss': 0.08155247569084167, 'train/grad_norm_pre_clip': 3.0754833221435547, 'timing/data': 0.0002865786664187908, 'timing/model': 0.4155762023292482, 'learning_rate/qwen_vl_interface': 9.200012022768077e-06, 'learning_rate/action_model': 9.200012022768077e-05, 'epoch': 0.15})
+[2026-07-28 03:19:48,977][starVLA.training.train_starvla][INFO] - Step 832, Loss: {'train/loss': 0.13323673605918884, 'train/grad_norm_pre_clip': 5.0549139976501465, 'timing/data': 0.00027634669095277786, 'timing/model': 0.4153357446193695, 'learning_rate/qwen_vl_interface': 9.19788560448964e-06, 'learning_rate/action_model': 9.197885604489639e-05, 'epoch': 0.15})
+[2026-07-28 03:19:55,085][starVLA.training.train_starvla][INFO] - Step 833, Loss: {'train/loss': 0.17753873765468597, 'train/grad_norm_pre_clip': 4.78857421875, 'timing/data': 0.0002853446640074253, 'timing/model': 0.6531306360848248, 'learning_rate/qwen_vl_interface': 9.195756624471094e-06, 'learning_rate/action_model': 9.195756624471094e-05, 'epoch': 0.15})
+[2026-07-28 03:20:01,183][starVLA.training.train_starvla][INFO] - Step 834, Loss: {'train/loss': 0.0770934447646141, 'train/grad_norm_pre_clip': 1.9507545232772827, 'timing/data': 0.00025733793154358864, 'timing/model': 0.41525372909381986, 'learning_rate/qwen_vl_interface': 9.193625084093912e-06, 'learning_rate/action_model': 9.193625084093911e-05, 'epoch': 0.15})
+[2026-07-28 03:20:07,033][starVLA.training.train_starvla][INFO] - Step 835, Loss: {'train/loss': 0.051399506628513336, 'train/grad_norm_pre_clip': 1.4160610437393188, 'timing/data': 0.00028064800426363945, 'timing/model': 0.4164181030355394, 'learning_rate/qwen_vl_interface': 9.191490984741227e-06, 'learning_rate/action_model': 9.191490984741227e-05, 'epoch': 0.15})
+[2026-07-28 03:20:12,923][starVLA.training.train_starvla][INFO] - Step 836, Loss: {'train/loss': 0.05700615793466568, 'train/grad_norm_pre_clip': 1.1371718645095825, 'timing/data': 0.0002897917293012142, 'timing/model': 0.41679747588932514, 'learning_rate/qwen_vl_interface': 9.189354327797835e-06, 'learning_rate/action_model': 9.189354327797836e-05, 'epoch': 0.15})
+[2026-07-28 03:20:19,179][starVLA.training.train_starvla][INFO] - Step 837, Loss: {'train/loss': 0.08753176778554916, 'train/grad_norm_pre_clip': 1.5986757278442383, 'timing/data': 0.00026983581483364105, 'timing/model': 0.6530469483695924, 'learning_rate/qwen_vl_interface': 9.18721511465019e-06, 'learning_rate/action_model': 9.187215114650191e-05, 'epoch': 0.15})
+[2026-07-28 03:20:25,210][starVLA.training.train_starvla][INFO] - Step 838, Loss: {'train/loss': 0.08606316894292831, 'train/grad_norm_pre_clip': 2.031593084335327, 'timing/data': 0.0002724328078329563, 'timing/model': 0.4161129058338702, 'learning_rate/qwen_vl_interface': 9.185073346686402e-06, 'learning_rate/action_model': 9.185073346686401e-05, 'epoch': 0.15})
+[2026-07-28 03:20:31,565][starVLA.training.train_starvla][INFO] - Step 839, Loss: {'train/loss': 0.10330086946487427, 'train/grad_norm_pre_clip': 3.7677254676818848, 'timing/data': 0.00025488389655947685, 'timing/model': 0.4134868788532913, 'learning_rate/qwen_vl_interface': 9.182929025296242e-06, 'learning_rate/action_model': 9.182929025296242e-05, 'epoch': 0.15})
+[2026-07-28 03:20:37,433][starVLA.training.train_starvla][INFO] - Step 840, Loss: {'train/loss': 0.10908453166484833, 'train/grad_norm_pre_clip': 2.510054588317871, 'timing/data': 0.0003399220295250416, 'timing/model': 0.41489696130156517, 'learning_rate/qwen_vl_interface': 9.180782151871138e-06, 'learning_rate/action_model': 9.180782151871137e-05, 'epoch': 0.15})
+[2026-07-28 03:20:43,548][starVLA.training.train_starvla][INFO] - Step 841, Loss: {'train/loss': 0.1126822680234909, 'train/grad_norm_pre_clip': 2.143878936767578, 'timing/data': 0.00032485881820321083, 'timing/model': 0.41456776624545455, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.15})
+[2026-07-28 03:20:49,407][starVLA.training.train_starvla][INFO] - Step 842, Loss: {'train/loss': 0.1025526225566864, 'train/grad_norm_pre_clip': 4.5144572257995605, 'timing/data': 0.00029553892090916634, 'timing/model': 0.4171843742951751, 'learning_rate/qwen_vl_interface': 9.176480754490078e-06, 'learning_rate/action_model': 9.176480754490078e-05, 'epoch': 0.15})
+[2026-07-28 03:20:55,515][starVLA.training.train_starvla][INFO] - Step 843, Loss: {'train/loss': 0.05603872984647751, 'train/grad_norm_pre_clip': 3.051044464111328, 'timing/data': 0.00027890317142009735, 'timing/model': 0.4190253601409495, 'learning_rate/qwen_vl_interface': 9.174326233325253e-06, 'learning_rate/action_model': 9.174326233325253e-05, 'epoch': 0.15})
+[2026-07-28 03:21:01,423][starVLA.training.train_starvla][INFO] - Step 844, Loss: {'train/loss': 0.04710245132446289, 'train/grad_norm_pre_clip': 1.3601330518722534, 'timing/data': 0.0002778349444270134, 'timing/model': 0.41554102627560496, 'learning_rate/qwen_vl_interface': 9.17216916570774e-06, 'learning_rate/action_model': 9.172169165707741e-05, 'epoch': 0.16})
+[2026-07-28 03:21:07,535][starVLA.training.train_starvla][INFO] - Step 845, Loss: {'train/loss': 0.11411291360855103, 'train/grad_norm_pre_clip': 1.5115553140640259, 'timing/data': 0.0002579609863460064, 'timing/model': 0.41430986393243074, 'learning_rate/qwen_vl_interface': 9.170009553037239e-06, 'learning_rate/action_model': 9.170009553037238e-05, 'epoch': 0.16})
+[2026-07-28 03:21:13,633][starVLA.training.train_starvla][INFO] - Step 846, Loss: {'train/loss': 0.10568591952323914, 'train/grad_norm_pre_clip': 2.2409892082214355, 'timing/data': 0.00026511307805776596, 'timing/model': 0.41671382496133447, 'learning_rate/qwen_vl_interface': 9.167847396715098e-06, 'learning_rate/action_model': 9.167847396715098e-05, 'epoch': 0.16})
+[2026-07-28 03:21:19,511][starVLA.training.train_starvla][INFO] - Step 847, Loss: {'train/loss': 0.06508038192987442, 'train/grad_norm_pre_clip': 2.4683282375335693, 'timing/data': 0.00028009386733174324, 'timing/model': 0.41645788913592696, 'learning_rate/qwen_vl_interface': 9.165682698144317e-06, 'learning_rate/action_model': 9.165682698144317e-05, 'epoch': 0.16})
+[2026-07-28 03:21:25,607][starVLA.training.train_starvla][INFO] - Step 848, Loss: {'train/loss': 0.06668497622013092, 'train/grad_norm_pre_clip': 2.2489662170410156, 'timing/data': 0.00027421116828918457, 'timing/model': 0.6554096792824566, 'learning_rate/qwen_vl_interface': 9.163515458729545e-06, 'learning_rate/action_model': 9.163515458729545e-05, 'epoch': 0.16})
+[2026-07-28 03:21:31,597][starVLA.training.train_starvla][INFO] - Step 849, Loss: {'train/loss': 0.0824459046125412, 'train/grad_norm_pre_clip': 2.490880012512207, 'timing/data': 0.00029261503368616104, 'timing/model': 0.5826859632506967, 'learning_rate/qwen_vl_interface': 9.16134567987708e-06, 'learning_rate/action_model': 9.16134567987708e-05, 'epoch': 0.16})
+[2026-07-28 03:21:37,445][starVLA.training.train_starvla][INFO] - Step 850, Loss: {'train/loss': 0.07493150234222412, 'train/grad_norm_pre_clip': 2.1015169620513916, 'timing/data': 0.0002856259234249592, 'timing/model': 0.4151996150612831, 'learning_rate/qwen_vl_interface': 9.159173362994868e-06, 'learning_rate/action_model': 9.159173362994869e-05, 'epoch': 0.16})
+[2026-07-28 03:21:43,521][starVLA.training.train_starvla][INFO] - Step 851, Loss: {'train/loss': 0.07362926006317139, 'train/grad_norm_pre_clip': 3.631065845489502, 'timing/data': 0.0002756216563284397, 'timing/model': 0.48500327579677105, 'learning_rate/qwen_vl_interface': 9.156998509492503e-06, 'learning_rate/action_model': 9.156998509492503e-05, 'epoch': 0.16})
+[2026-07-28 03:21:49,441][starVLA.training.train_starvla][INFO] - Step 852, Loss: {'train/loss': 0.08650466054677963, 'train/grad_norm_pre_clip': 1.6669471263885498, 'timing/data': 0.0002791420556604862, 'timing/model': 0.4177083168178797, 'learning_rate/qwen_vl_interface': 9.154821120781221e-06, 'learning_rate/action_model': 9.154821120781222e-05, 'epoch': 0.16})
+[2026-07-28 03:21:55,735][starVLA.training.train_starvla][INFO] - Step 853, Loss: {'train/loss': 0.0664081871509552, 'train/grad_norm_pre_clip': 2.4878859519958496, 'timing/data': 0.0002796989865601063, 'timing/model': 0.6495982590131462, 'learning_rate/qwen_vl_interface': 9.152641198273911e-06, 'learning_rate/action_model': 9.15264119827391e-05, 'epoch': 0.16})
+[2026-07-28 03:22:01,598][starVLA.training.train_starvla][INFO] - Step 854, Loss: {'train/loss': 0.08606703579425812, 'train/grad_norm_pre_clip': 2.027397871017456, 'timing/data': 0.00026392610743641853, 'timing/model': 0.41299392096698284, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.16})
+[2026-07-28 03:22:07,418][starVLA.training.train_starvla][INFO] - Step 855, Loss: {'train/loss': 0.07788065075874329, 'train/grad_norm_pre_clip': 1.2061597108840942, 'timing/data': 0.00028309086337685585, 'timing/model': 0.4154380369000137, 'learning_rate/qwen_vl_interface': 9.148273757530948e-06, 'learning_rate/action_model': 9.148273757530948e-05, 'epoch': 0.16})
+[2026-07-28 03:22:13,244][starVLA.training.train_starvla][INFO] - Step 856, Loss: {'train/loss': 0.10787756741046906, 'train/grad_norm_pre_clip': 1.7064001560211182, 'timing/data': 0.00028835469856858253, 'timing/model': 0.41552708903327584, 'learning_rate/qwen_vl_interface': 9.146086242129283e-06, 'learning_rate/action_model': 9.146086242129282e-05, 'epoch': 0.16})
+[2026-07-28 03:22:19,342][starVLA.training.train_starvla][INFO] - Step 857, Loss: {'train/loss': 0.07826793193817139, 'train/grad_norm_pre_clip': 1.9369536638259888, 'timing/data': 0.0002586948685348034, 'timing/model': 0.41558196768164635, 'learning_rate/qwen_vl_interface': 9.143896198599553e-06, 'learning_rate/action_model': 9.143896198599553e-05, 'epoch': 0.16})
+[2026-07-28 03:22:25,179][starVLA.training.train_starvla][INFO] - Step 858, Loss: {'train/loss': 0.0548539012670517, 'train/grad_norm_pre_clip': 1.678956151008606, 'timing/data': 0.00027812598273158073, 'timing/model': 0.4150379071943462, 'learning_rate/qwen_vl_interface': 9.141703628362858e-06, 'learning_rate/action_model': 9.141703628362858e-05, 'epoch': 0.16})
+[2026-07-28 03:22:31,281][starVLA.training.train_starvla][INFO] - Step 859, Loss: {'train/loss': 0.05715294927358627, 'train/grad_norm_pre_clip': 1.8430125713348389, 'timing/data': 0.0002748379483819008, 'timing/model': 0.6524464068934321, 'learning_rate/qwen_vl_interface': 9.139508532841928e-06, 'learning_rate/action_model': 9.139508532841927e-05, 'epoch': 0.16})
+[2026-07-28 03:22:37,121][starVLA.training.train_starvla][INFO] - Step 860, Loss: {'train/loss': 0.10812996327877045, 'train/grad_norm_pre_clip': 2.485051155090332, 'timing/data': 0.0002596941776573658, 'timing/model': 0.4129891572520137, 'learning_rate/qwen_vl_interface': 9.137310913461139e-06, 'learning_rate/action_model': 9.137310913461139e-05, 'epoch': 0.16})
+[2026-07-28 03:22:42,963][starVLA.training.train_starvla][INFO] - Step 861, Loss: {'train/loss': 0.0828213170170784, 'train/grad_norm_pre_clip': 3.562673568725586, 'timing/data': 0.00027098972350358963, 'timing/model': 0.41391728492453694, 'learning_rate/qwen_vl_interface': 9.135110771646502e-06, 'learning_rate/action_model': 9.135110771646502e-05, 'epoch': 0.16})
+[2026-07-28 03:22:49,046][starVLA.training.train_starvla][INFO] - Step 862, Loss: {'train/loss': 0.10331232845783234, 'train/grad_norm_pre_clip': 4.830122947692871, 'timing/data': 0.0002645859494805336, 'timing/model': 0.6552475448697805, 'learning_rate/qwen_vl_interface': 9.132908108825667e-06, 'learning_rate/action_model': 9.132908108825667e-05, 'epoch': 0.16})
+[2026-07-28 03:22:54,908][starVLA.training.train_starvla][INFO] - Step 863, Loss: {'train/loss': 0.0535719096660614, 'train/grad_norm_pre_clip': 2.852302074432373, 'timing/data': 0.00034958403557538986, 'timing/model': 0.41295463498681784, 'learning_rate/qwen_vl_interface': 9.130702926427915e-06, 'learning_rate/action_model': 9.130702926427914e-05, 'epoch': 0.16})
+[2026-07-28 03:23:01,240][starVLA.training.train_starvla][INFO] - Step 864, Loss: {'train/loss': 0.0932604968547821, 'train/grad_norm_pre_clip': 2.624110221862793, 'timing/data': 0.00025645503774285316, 'timing/model': 0.41409444296732545, 'learning_rate/qwen_vl_interface': 9.128495225884167e-06, 'learning_rate/action_model': 9.128495225884166e-05, 'epoch': 0.16})
+[2026-07-28 03:23:07,054][starVLA.training.train_starvla][INFO] - Step 865, Loss: {'train/loss': 0.08879639208316803, 'train/grad_norm_pre_clip': 1.618680715560913, 'timing/data': 0.00028130412101745605, 'timing/model': 0.41542205307632685, 'learning_rate/qwen_vl_interface': 9.126285008626975e-06, 'learning_rate/action_model': 9.126285008626974e-05, 'epoch': 0.16})
+[2026-07-28 03:23:12,877][starVLA.training.train_starvla][INFO] - Step 866, Loss: {'train/loss': 0.10675711929798126, 'train/grad_norm_pre_clip': 1.6340934038162231, 'timing/data': 0.0003561936318874359, 'timing/model': 0.41747507080435753, 'learning_rate/qwen_vl_interface': 9.124072276090524e-06, 'learning_rate/action_model': 9.124072276090524e-05, 'epoch': 0.16})
+[2026-07-28 03:23:18,963][starVLA.training.train_starvla][INFO] - Step 867, Loss: {'train/loss': 0.07837852835655212, 'train/grad_norm_pre_clip': 2.089707374572754, 'timing/data': 0.0002662278711795807, 'timing/model': 0.41685108467936516, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.16})
+[2026-07-28 03:23:25,118][starVLA.training.train_starvla][INFO] - Step 868, Loss: {'train/loss': 0.08167184144258499, 'train/grad_norm_pre_clip': 3.487912654876709, 'timing/data': 0.0003937268629670143, 'timing/model': 0.414307520724833, 'learning_rate/qwen_vl_interface': 9.119639270924754e-06, 'learning_rate/action_model': 9.119639270924754e-05, 'epoch': 0.16})
+[2026-07-28 03:23:30,976][starVLA.training.train_starvla][INFO] - Step 869, Loss: {'train/loss': 0.12595997750759125, 'train/grad_norm_pre_clip': 3.201748847961426, 'timing/data': 0.00029319338500499725, 'timing/model': 0.415917179081589, 'learning_rate/qwen_vl_interface': 9.11741900117196e-06, 'learning_rate/action_model': 9.11741900117196e-05, 'epoch': 0.16})
+[2026-07-28 03:23:37,411][starVLA.training.train_starvla][INFO] - Step 870, Loss: {'train/loss': 0.10069355368614197, 'train/grad_norm_pre_clip': 2.8529508113861084, 'timing/data': 0.0002604029141366482, 'timing/model': 0.41851002583280206, 'learning_rate/qwen_vl_interface': 9.115196221892969e-06, 'learning_rate/action_model': 9.115196221892969e-05, 'epoch': 0.16})
+[2026-07-28 03:23:43,293][starVLA.training.train_starvla][INFO] - Step 871, Loss: {'train/loss': 0.030315756797790527, 'train/grad_norm_pre_clip': 1.599579095840454, 'timing/data': 0.0002702041529119015, 'timing/model': 0.4140957482159138, 'learning_rate/qwen_vl_interface': 9.112970934530111e-06, 'learning_rate/action_model': 9.112970934530111e-05, 'epoch': 0.16})
+[2026-07-28 03:23:49,182][starVLA.training.train_starvla][INFO] - Step 872, Loss: {'train/loss': 0.06581348180770874, 'train/grad_norm_pre_clip': 1.3193670511245728, 'timing/data': 0.0002788510173559189, 'timing/model': 0.41902260202914476, 'learning_rate/qwen_vl_interface': 9.110743140527353e-06, 'learning_rate/action_model': 9.110743140527352e-05, 'epoch': 0.16})
+[2026-07-28 03:23:55,377][starVLA.training.train_starvla][INFO] - Step 873, Loss: {'train/loss': 0.07141485810279846, 'train/grad_norm_pre_clip': 1.3689813613891602, 'timing/data': 0.00028813304379582405, 'timing/model': 0.7015320360660553, 'learning_rate/qwen_vl_interface': 9.108512841330288e-06, 'learning_rate/action_model': 9.108512841330287e-05, 'epoch': 0.16})
+[2026-07-28 03:24:01,571][starVLA.training.train_starvla][INFO] - Step 874, Loss: {'train/loss': 0.057482700794935226, 'train/grad_norm_pre_clip': 1.9478520154953003, 'timing/data': 0.00025337422266602516, 'timing/model': 0.6488613439723849, 'learning_rate/qwen_vl_interface': 9.106280038386129e-06, 'learning_rate/action_model': 9.10628003838613e-05, 'epoch': 0.16})
+[2026-07-28 03:24:07,419][starVLA.training.train_starvla][INFO] - Step 875, Loss: {'train/loss': 0.08528035134077072, 'train/grad_norm_pre_clip': 1.9043610095977783, 'timing/data': 0.0002490151673555374, 'timing/model': 0.4158028503879905, 'learning_rate/qwen_vl_interface': 9.104044733143723e-06, 'learning_rate/action_model': 9.104044733143723e-05, 'epoch': 0.16})
+[2026-07-28 03:24:13,265][starVLA.training.train_starvla][INFO] - Step 876, Loss: {'train/loss': 0.050887636840343475, 'train/grad_norm_pre_clip': 1.0596352815628052, 'timing/data': 0.00027631595730781555, 'timing/model': 0.4170845546759665, 'learning_rate/qwen_vl_interface': 9.10180692705353e-06, 'learning_rate/action_model': 9.10180692705353e-05, 'epoch': 0.16})
+[2026-07-28 03:24:19,153][starVLA.training.train_starvla][INFO] - Step 877, Loss: {'train/loss': 0.05846378207206726, 'train/grad_norm_pre_clip': 1.3284783363342285, 'timing/data': 0.0002760421484708786, 'timing/model': 0.41809581592679024, 'learning_rate/qwen_vl_interface': 9.099566621567643e-06, 'learning_rate/action_model': 9.099566621567642e-05, 'epoch': 0.16})
+[2026-07-28 03:24:24,804][starVLA.training.train_starvla][INFO] - Step 878, Loss: {'train/loss': 0.08048228174448013, 'train/grad_norm_pre_clip': 1.7077292203903198, 'timing/data': 0.0002827388234436512, 'timing/model': 0.41816672682762146, 'learning_rate/qwen_vl_interface': 9.097323818139767e-06, 'learning_rate/action_model': 9.097323818139767e-05, 'epoch': 0.16})
+[2026-07-28 03:24:30,447][starVLA.training.train_starvla][INFO] - Step 879, Loss: {'train/loss': 0.06545032560825348, 'train/grad_norm_pre_clip': 2.0489695072174072, 'timing/data': 0.0002939170226454735, 'timing/model': 0.41718075377866626, 'learning_rate/qwen_vl_interface': 9.095078518225237e-06, 'learning_rate/action_model': 9.095078518225236e-05, 'epoch': 0.16})
+[2026-07-28 03:24:36,343][starVLA.training.train_starvla][INFO] - Step 880, Loss: {'train/loss': 0.07823583483695984, 'train/grad_norm_pre_clip': 1.9949407577514648, 'timing/data': 0.00026311585679650307, 'timing/model': 0.4146185237914324, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.16})
+[2026-07-28 03:24:42,043][starVLA.training.train_starvla][INFO] - Step 881, Loss: {'train/loss': 0.039740994572639465, 'train/grad_norm_pre_clip': 1.5338547229766846, 'timing/data': 0.00028449902310967445, 'timing/model': 0.4173715398646891, 'learning_rate/qwen_vl_interface': 9.09058043476563e-06, 'learning_rate/action_model': 9.09058043476563e-05, 'epoch': 0.16})
+[2026-07-28 03:24:47,897][starVLA.training.train_starvla][INFO] - Step 882, Loss: {'train/loss': 0.07506901770830154, 'train/grad_norm_pre_clip': 2.472588062286377, 'timing/data': 0.00027615996077656746, 'timing/model': 0.4170477367006242, 'learning_rate/qwen_vl_interface': 9.088327654139311e-06, 'learning_rate/action_model': 9.08832765413931e-05, 'epoch': 0.16})
+[2026-07-28 03:24:53,750][starVLA.training.train_starvla][INFO] - Step 883, Loss: {'train/loss': 0.047357138246297836, 'train/grad_norm_pre_clip': 1.3184292316436768, 'timing/data': 0.0002742619253695011, 'timing/model': 0.41666829120367765, 'learning_rate/qwen_vl_interface': 9.08607238286385e-06, 'learning_rate/action_model': 9.08607238286385e-05, 'epoch': 0.16})
+[2026-07-28 03:25:00,114][starVLA.training.train_starvla][INFO] - Step 884, Loss: {'train/loss': 0.07714787125587463, 'train/grad_norm_pre_clip': 2.836209774017334, 'timing/data': 0.00029359618201851845, 'timing/model': 0.6504908609203994, 'learning_rate/qwen_vl_interface': 9.083814622402667e-06, 'learning_rate/action_model': 9.083814622402666e-05, 'epoch': 0.16})
+[2026-07-28 03:25:05,982][starVLA.training.train_starvla][INFO] - Step 885, Loss: {'train/loss': 0.060835640877485275, 'train/grad_norm_pre_clip': 2.192204475402832, 'timing/data': 0.0003564991056919098, 'timing/model': 0.4177530100569129, 'learning_rate/qwen_vl_interface': 9.0815543742208e-06, 'learning_rate/action_model': 9.081554374220799e-05, 'epoch': 0.16})
+[2026-07-28 03:25:11,871][starVLA.training.train_starvla][INFO] - Step 886, Loss: {'train/loss': 0.15628373622894287, 'train/grad_norm_pre_clip': 3.5272138118743896, 'timing/data': 0.00029013492166996, 'timing/model': 0.4176071756519377, 'learning_rate/qwen_vl_interface': 9.079291639784897e-06, 'learning_rate/action_model': 9.079291639784896e-05, 'epoch': 0.16})
+[2026-07-28 03:25:17,974][starVLA.training.train_starvla][INFO] - Step 887, Loss: {'train/loss': 0.054216403514146805, 'train/grad_norm_pre_clip': 2.420487403869629, 'timing/data': 0.000256143044680357, 'timing/model': 0.4167986633256078, 'learning_rate/qwen_vl_interface': 9.077026420563223e-06, 'learning_rate/action_model': 9.077026420563223e-05, 'epoch': 0.16})
+[2026-07-28 03:25:24,103][starVLA.training.train_starvla][INFO] - Step 888, Loss: {'train/loss': 0.1241074800491333, 'train/grad_norm_pre_clip': 4.689750671386719, 'timing/data': 0.00026700692251324654, 'timing/model': 0.4145344002172351, 'learning_rate/qwen_vl_interface': 9.074758718025652e-06, 'learning_rate/action_model': 9.074758718025652e-05, 'epoch': 0.16})
+[2026-07-28 03:25:29,977][starVLA.training.train_starvla][INFO] - Step 889, Loss: {'train/loss': 0.04957764595746994, 'train/grad_norm_pre_clip': 1.2901486158370972, 'timing/data': 0.0002745199017226696, 'timing/model': 0.416859227232635, 'learning_rate/qwen_vl_interface': 9.072488533643678e-06, 'learning_rate/action_model': 9.072488533643677e-05, 'epoch': 0.16})
+[2026-07-28 03:25:36,002][starVLA.training.train_starvla][INFO] - Step 890, Loss: {'train/loss': 0.04895588010549545, 'train/grad_norm_pre_clip': 1.575329065322876, 'timing/data': 0.00027829501777887344, 'timing/model': 0.5998729020357132, 'learning_rate/qwen_vl_interface': 9.070215868890391e-06, 'learning_rate/action_model': 9.070215868890391e-05, 'epoch': 0.16})
+[2026-07-28 03:25:41,871][starVLA.training.train_starvla][INFO] - Step 891, Loss: {'train/loss': 0.10896524786949158, 'train/grad_norm_pre_clip': 3.3174643516540527, 'timing/data': 0.0003515658900141716, 'timing/model': 0.41694861371070147, 'learning_rate/qwen_vl_interface': 9.067940725240503e-06, 'learning_rate/action_model': 9.067940725240503e-05, 'epoch': 0.16})
+[2026-07-28 03:25:47,998][starVLA.training.train_starvla][INFO] - Step 892, Loss: {'train/loss': 0.07721847295761108, 'train/grad_norm_pre_clip': 2.4596290588378906, 'timing/data': 0.0002666381187736988, 'timing/model': 0.41543338587507606, 'learning_rate/qwen_vl_interface': 9.06566310417033e-06, 'learning_rate/action_model': 9.06566310417033e-05, 'epoch': 0.16})
+[2026-07-28 03:25:53,908][starVLA.training.train_starvla][INFO] - Step 893, Loss: {'train/loss': 0.07181978225708008, 'train/grad_norm_pre_clip': 2.7056124210357666, 'timing/data': 0.0002735522575676441, 'timing/model': 0.4173207329586148, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.16})
+[2026-07-28 03:26:00,014][starVLA.training.train_starvla][INFO] - Step 894, Loss: {'train/loss': 0.08411191403865814, 'train/grad_norm_pre_clip': 1.9350124597549438, 'timing/data': 0.0002771741710603237, 'timing/model': 0.4169484688900411, 'learning_rate/qwen_vl_interface': 9.061100435682429e-06, 'learning_rate/action_model': 9.061100435682429e-05, 'epoch': 0.16})
+[2026-07-28 03:26:06,104][starVLA.training.train_starvla][INFO] - Step 895, Loss: {'train/loss': 0.10521164536476135, 'train/grad_norm_pre_clip': 4.056210517883301, 'timing/data': 0.00028086407110095024, 'timing/model': 0.6554751573130488, 'learning_rate/qwen_vl_interface': 9.058815391225363e-06, 'learning_rate/action_model': 9.058815391225362e-05, 'epoch': 0.16})
+[2026-07-28 03:26:12,038][starVLA.training.train_starvla][INFO] - Step 896, Loss: {'train/loss': 0.07541041821241379, 'train/grad_norm_pre_clip': 1.914170742034912, 'timing/data': 0.00029934896156191826, 'timing/model': 0.41866283724084496, 'learning_rate/qwen_vl_interface': 9.056527875269344e-06, 'learning_rate/action_model': 9.056527875269344e-05, 'epoch': 0.16})
+[2026-07-28 03:26:17,907][starVLA.training.train_starvla][INFO] - Step 897, Loss: {'train/loss': 0.0784156545996666, 'train/grad_norm_pre_clip': 1.778206706047058, 'timing/data': 0.0002894829958677292, 'timing/model': 0.41808920400217175, 'learning_rate/qwen_vl_interface': 9.054237889298714e-06, 'learning_rate/action_model': 9.054237889298714e-05, 'epoch': 0.16})
+[2026-07-28 03:26:24,017][starVLA.training.train_starvla][INFO] - Step 898, Loss: {'train/loss': 0.06778573989868164, 'train/grad_norm_pre_clip': 2.255610466003418, 'timing/data': 0.00025241728872060776, 'timing/model': 0.414984714705497, 'learning_rate/qwen_vl_interface': 9.05194543479942e-06, 'learning_rate/action_model': 9.05194543479942e-05, 'epoch': 0.16})
+[2026-07-28 03:26:29,865][starVLA.training.train_starvla][INFO] - Step 899, Loss: {'train/loss': 0.05247751623392105, 'train/grad_norm_pre_clip': 5.007730960845947, 'timing/data': 0.00027649709954857826, 'timing/model': 0.41680823592469096, 'learning_rate/qwen_vl_interface': 9.049650513259007e-06, 'learning_rate/action_model': 9.049650513259008e-05, 'epoch': 0.17})
+[2026-07-28 03:26:35,987][starVLA.training.train_starvla][INFO] - Step 900, Loss: {'train/loss': 0.09132291376590729, 'train/grad_norm_pre_clip': 5.495256423950195, 'timing/data': 0.00026913732290267944, 'timing/model': 0.6726232338696718, 'learning_rate/qwen_vl_interface': 9.047353126166632e-06, 'learning_rate/action_model': 9.047353126166632e-05, 'epoch': 0.17})
+[2026-07-28 03:26:41,895][starVLA.training.train_starvla][INFO] - Step 901, Loss: {'train/loss': 0.0735473483800888, 'train/grad_norm_pre_clip': 2.9578261375427246, 'timing/data': 0.0002540508285164833, 'timing/model': 0.4149653962813318, 'learning_rate/qwen_vl_interface': 9.045053275013038e-06, 'learning_rate/action_model': 9.045053275013038e-05, 'epoch': 0.17})
+[2026-07-28 03:26:47,757][starVLA.training.train_starvla][INFO] - Step 902, Loss: {'train/loss': 0.112449511885643, 'train/grad_norm_pre_clip': 2.219921350479126, 'timing/data': 0.00028249574825167656, 'timing/model': 0.4170263228006661, 'learning_rate/qwen_vl_interface': 9.042750961290578e-06, 'learning_rate/action_model': 9.042750961290578e-05, 'epoch': 0.17})
+[2026-07-28 03:26:53,913][starVLA.training.train_starvla][INFO] - Step 903, Loss: {'train/loss': 0.046235084533691406, 'train/grad_norm_pre_clip': 1.0016720294952393, 'timing/data': 0.00028079282492399216, 'timing/model': 0.6551209692843258, 'learning_rate/qwen_vl_interface': 9.040446186493195e-06, 'learning_rate/action_model': 9.040446186493194e-05, 'epoch': 0.17})
+[2026-07-28 03:26:59,787][starVLA.training.train_starvla][INFO] - Step 904, Loss: {'train/loss': 0.08577346801757812, 'train/grad_norm_pre_clip': 1.0934385061264038, 'timing/data': 0.0002561979927122593, 'timing/model': 0.4184478987008333, 'learning_rate/qwen_vl_interface': 9.038138952116435e-06, 'learning_rate/action_model': 9.038138952116435e-05, 'epoch': 0.17})
+[2026-07-28 03:27:05,667][starVLA.training.train_starvla][INFO] - Step 905, Loss: {'train/loss': 0.0729513168334961, 'train/grad_norm_pre_clip': 1.3531867265701294, 'timing/data': 0.0003823647275567055, 'timing/model': 0.41601682221516967, 'learning_rate/qwen_vl_interface': 9.035829259657437e-06, 'learning_rate/action_model': 9.035829259657436e-05, 'epoch': 0.17})
+[2026-07-28 03:27:11,841][starVLA.training.train_starvla][INFO] - Step 906, Loss: {'train/loss': 0.0651601105928421, 'train/grad_norm_pre_clip': 2.532474994659424, 'timing/data': 0.00028258468955755234, 'timing/model': 0.7047635898925364, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.17})
+[2026-07-28 03:27:17,864][starVLA.training.train_starvla][INFO] - Step 907, Loss: {'train/loss': 0.06935597956180573, 'train/grad_norm_pre_clip': 2.276590585708618, 'timing/data': 0.00034748390316963196, 'timing/model': 0.4126942832954228, 'learning_rate/qwen_vl_interface': 9.031202506489254e-06, 'learning_rate/action_model': 9.031202506489253e-05, 'epoch': 0.17})
+[2026-07-28 03:27:24,037][starVLA.training.train_starvla][INFO] - Step 908, Loss: {'train/loss': 0.0847032442688942, 'train/grad_norm_pre_clip': 2.353954792022705, 'timing/data': 0.00025724200531840324, 'timing/model': 0.4130240469239652, 'learning_rate/qwen_vl_interface': 9.02888544878232e-06, 'learning_rate/action_model': 9.028885448782319e-05, 'epoch': 0.17})
+[2026-07-28 03:27:29,984][starVLA.training.train_starvla][INFO] - Step 909, Loss: {'train/loss': 0.11847160756587982, 'train/grad_norm_pre_clip': 2.6574299335479736, 'timing/data': 0.0002759094350039959, 'timing/model': 0.4187011909671128, 'learning_rate/qwen_vl_interface': 9.026565938997644e-06, 'learning_rate/action_model': 9.026565938997645e-05, 'epoch': 0.17})
+[2026-07-28 03:27:35,884][starVLA.training.train_starvla][INFO] - Step 910, Loss: {'train/loss': 0.0798555240035057, 'train/grad_norm_pre_clip': 2.2769436836242676, 'timing/data': 0.0002843928523361683, 'timing/model': 0.4192644339054823, 'learning_rate/qwen_vl_interface': 9.024243978640334e-06, 'learning_rate/action_model': 9.024243978640333e-05, 'epoch': 0.17})
+[2026-07-28 03:27:42,009][starVLA.training.train_starvla][INFO] - Step 911, Loss: {'train/loss': 0.09718433022499084, 'train/grad_norm_pre_clip': 2.8147270679473877, 'timing/data': 0.0002467641606926918, 'timing/model': 0.41431574104353786, 'learning_rate/qwen_vl_interface': 9.02191956921708e-06, 'learning_rate/action_model': 9.02191956921708e-05, 'epoch': 0.17})
+[2026-07-28 03:27:47,910][starVLA.training.train_starvla][INFO] - Step 912, Loss: {'train/loss': 0.07867978513240814, 'train/grad_norm_pre_clip': 1.7079616785049438, 'timing/data': 0.00035896291956305504, 'timing/model': 0.41495120665058494, 'learning_rate/qwen_vl_interface': 9.019592712236172e-06, 'learning_rate/action_model': 9.019592712236171e-05, 'epoch': 0.17})
+[2026-07-28 03:27:53,829][starVLA.training.train_starvla][INFO] - Step 913, Loss: {'train/loss': 0.07240381836891174, 'train/grad_norm_pre_clip': 1.084157109260559, 'timing/data': 0.0002876473590731621, 'timing/model': 0.41583995101973414, 'learning_rate/qwen_vl_interface': 9.017263409207478e-06, 'learning_rate/action_model': 9.017263409207478e-05, 'epoch': 0.17})
+[2026-07-28 03:27:59,700][starVLA.training.train_starvla][INFO] - Step 914, Loss: {'train/loss': 0.06874201446771622, 'train/grad_norm_pre_clip': 1.2477055788040161, 'timing/data': 0.0002772696316242218, 'timing/model': 0.4165905308909714, 'learning_rate/qwen_vl_interface': 9.014931661642455e-06, 'learning_rate/action_model': 9.014931661642455e-05, 'epoch': 0.17})
+[2026-07-28 03:28:05,832][starVLA.training.train_starvla][INFO] - Step 915, Loss: {'train/loss': 0.07797221094369888, 'train/grad_norm_pre_clip': 2.6842098236083984, 'timing/data': 0.0002601570449769497, 'timing/model': 0.6474105729721487, 'learning_rate/qwen_vl_interface': 9.012597471054154e-06, 'learning_rate/action_model': 9.012597471054154e-05, 'epoch': 0.17})
+[2026-07-28 03:28:12,060][starVLA.training.train_starvla][INFO] - Step 916, Loss: {'train/loss': 0.07712624967098236, 'train/grad_norm_pre_clip': 1.7522143125534058, 'timing/data': 0.0002795183099806309, 'timing/model': 0.4215427818708122, 'learning_rate/qwen_vl_interface': 9.010260838957206e-06, 'learning_rate/action_model': 9.010260838957205e-05, 'epoch': 0.17})
+[2026-07-28 03:28:18,020][starVLA.training.train_starvla][INFO] - Step 917, Loss: {'train/loss': 0.08280454576015472, 'train/grad_norm_pre_clip': 1.7659375667572021, 'timing/data': 0.00027836067602038383, 'timing/model': 0.4223081539385021, 'learning_rate/qwen_vl_interface': 9.007921766867823e-06, 'learning_rate/action_model': 9.007921766867822e-05, 'epoch': 0.17})
+[2026-07-28 03:28:23,991][starVLA.training.train_starvla][INFO] - Step 918, Loss: {'train/loss': 0.0658755674958229, 'train/grad_norm_pre_clip': 1.1857401132583618, 'timing/data': 0.00028580520302057266, 'timing/model': 0.423544405028224, 'learning_rate/qwen_vl_interface': 9.005580256303804e-06, 'learning_rate/action_model': 9.005580256303803e-05, 'epoch': 0.17})
+[2026-07-28 03:28:29,976][starVLA.training.train_starvla][INFO] - Step 919, Loss: {'train/loss': 0.05193883925676346, 'train/grad_norm_pre_clip': 1.2417943477630615, 'timing/data': 0.00028571905568242073, 'timing/model': 0.42482167901471257, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.17})
+[2026-07-28 03:28:36,152][starVLA.training.train_starvla][INFO] - Step 920, Loss: {'train/loss': 0.07593589276075363, 'train/grad_norm_pre_clip': 2.3496577739715576, 'timing/data': 0.00029897503554821014, 'timing/model': 0.7106727962382138, 'learning_rate/qwen_vl_interface': 9.000889925830964e-06, 'learning_rate/action_model': 9.000889925830965e-05, 'epoch': 0.17})
+[2026-07-28 03:28:42,025][starVLA.training.train_starvla][INFO] - Step 921, Loss: {'train/loss': 0.054733745753765106, 'train/grad_norm_pre_clip': 1.6056922674179077, 'timing/data': 0.0002538743428885937, 'timing/model': 0.4135228148661554, 'learning_rate/qwen_vl_interface': 8.998541108965648e-06, 'learning_rate/action_model': 8.998541108965648e-05, 'epoch': 0.17})
+[2026-07-28 03:28:47,898][starVLA.training.train_starvla][INFO] - Step 922, Loss: {'train/loss': 0.05080205947160721, 'train/grad_norm_pre_clip': 1.334526538848877, 'timing/data': 0.0005017239600419998, 'timing/model': 0.41410526307299733, 'learning_rate/qwen_vl_interface': 8.996189859712704e-06, 'learning_rate/action_model': 8.996189859712703e-05, 'epoch': 0.17})
+[2026-07-28 03:28:53,754][starVLA.training.train_starvla][INFO] - Step 923, Loss: {'train/loss': 0.09662076830863953, 'train/grad_norm_pre_clip': 4.118629455566406, 'timing/data': 0.00027903588488698006, 'timing/model': 0.41593830194324255, 'learning_rate/qwen_vl_interface': 8.993836179597833e-06, 'learning_rate/action_model': 8.993836179597833e-05, 'epoch': 0.17})
+[2026-07-28 03:28:59,590][starVLA.training.train_starvla][INFO] - Step 924, Loss: {'train/loss': 0.07328049838542938, 'train/grad_norm_pre_clip': 2.3250718116760254, 'timing/data': 0.00040268339216709137, 'timing/model': 0.4153637890703976, 'learning_rate/qwen_vl_interface': 8.99148007014831e-06, 'learning_rate/action_model': 8.99148007014831e-05, 'epoch': 0.17})
+[2026-07-28 03:29:05,811][starVLA.training.train_starvla][INFO] - Step 925, Loss: {'train/loss': 0.03804949298501015, 'train/grad_norm_pre_clip': 0.7717561721801758, 'timing/data': 0.000280231237411499, 'timing/model': 0.41551877884194255, 'learning_rate/qwen_vl_interface': 8.989121532892989e-06, 'learning_rate/action_model': 8.989121532892988e-05, 'epoch': 0.17})
+[2026-07-28 03:29:11,673][starVLA.training.train_starvla][INFO] - Step 926, Loss: {'train/loss': 0.026097524911165237, 'train/grad_norm_pre_clip': 1.905531406402588, 'timing/data': 0.0002889540046453476, 'timing/model': 0.41606938000768423, 'learning_rate/qwen_vl_interface': 8.986760569362302e-06, 'learning_rate/action_model': 8.986760569362301e-05, 'epoch': 0.17})
+[2026-07-28 03:29:17,856][starVLA.training.train_starvla][INFO] - Step 927, Loss: {'train/loss': 0.1013629212975502, 'train/grad_norm_pre_clip': 4.5721917152404785, 'timing/data': 0.0002678423188626766, 'timing/model': 0.4132420117966831, 'learning_rate/qwen_vl_interface': 8.984397181088251e-06, 'learning_rate/action_model': 8.984397181088251e-05, 'epoch': 0.17})
+[2026-07-28 03:29:23,964][starVLA.training.train_starvla][INFO] - Step 928, Loss: {'train/loss': 0.048184219747781754, 'train/grad_norm_pre_clip': 1.5187692642211914, 'timing/data': 0.00025981292128562927, 'timing/model': 0.41932562785223126, 'learning_rate/qwen_vl_interface': 8.982031369604412e-06, 'learning_rate/action_model': 8.982031369604411e-05, 'epoch': 0.17})
+[2026-07-28 03:29:29,838][starVLA.training.train_starvla][INFO] - Step 929, Loss: {'train/loss': 0.038753677159547806, 'train/grad_norm_pre_clip': 1.3066723346710205, 'timing/data': 0.00027764029800891876, 'timing/model': 0.41625737119466066, 'learning_rate/qwen_vl_interface': 8.979663136445936e-06, 'learning_rate/action_model': 8.979663136445935e-05, 'epoch': 0.17})
+[2026-07-28 03:29:35,817][starVLA.training.train_starvla][INFO] - Step 930, Loss: {'train/loss': 0.08717332780361176, 'train/grad_norm_pre_clip': 3.075307846069336, 'timing/data': 0.00028899405151605606, 'timing/model': 0.41579719400033355, 'learning_rate/qwen_vl_interface': 8.977292483149542e-06, 'learning_rate/action_model': 8.977292483149542e-05, 'epoch': 0.17})
+[2026-07-28 03:29:42,066][starVLA.training.train_starvla][INFO] - Step 931, Loss: {'train/loss': 0.09235485643148422, 'train/grad_norm_pre_clip': 3.1868832111358643, 'timing/data': 0.00028693629428744316, 'timing/model': 0.8233965500257909, 'learning_rate/qwen_vl_interface': 8.974919411253523e-06, 'learning_rate/action_model': 8.974919411253523e-05, 'epoch': 0.17})
+[2026-07-28 03:29:47,672][starVLA.training.train_starvla][INFO] - Step 932, Loss: {'train/loss': 0.06160219386219978, 'train/grad_norm_pre_clip': 1.8256391286849976, 'timing/data': 0.0002773161977529526, 'timing/model': 0.41509637981653214, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.17})
+[2026-07-28 03:29:53,310][starVLA.training.train_starvla][INFO] - Step 933, Loss: {'train/loss': 0.043537795543670654, 'train/grad_norm_pre_clip': 1.2423990964889526, 'timing/data': 0.00026921089738607407, 'timing/model': 0.41487028216943145, 'learning_rate/qwen_vl_interface': 8.970166017823621e-06, 'learning_rate/action_model': 8.97016601782362e-05, 'epoch': 0.17})
+[2026-07-28 03:29:59,184][starVLA.training.train_starvla][INFO] - Step 934, Loss: {'train/loss': 0.04729625582695007, 'train/grad_norm_pre_clip': 1.7774162292480469, 'timing/data': 0.00036385608837008476, 'timing/model': 0.41562902135774493, 'learning_rate/qwen_vl_interface': 8.96778569937416e-06, 'learning_rate/action_model': 8.96778569937416e-05, 'epoch': 0.17})
+[2026-07-28 03:30:05,100][starVLA.training.train_starvla][INFO] - Step 935, Loss: {'train/loss': 0.05064905062317848, 'train/grad_norm_pre_clip': 1.2978770732879639, 'timing/data': 0.0003250869922339916, 'timing/model': 0.4128069533035159, 'learning_rate/qwen_vl_interface': 8.965402968493922e-06, 'learning_rate/action_model': 8.965402968493922e-05, 'epoch': 0.17})
+[2026-07-28 03:30:10,949][starVLA.training.train_starvla][INFO] - Step 936, Loss: {'train/loss': 0.07924170792102814, 'train/grad_norm_pre_clip': 1.5035039186477661, 'timing/data': 0.0002831658348441124, 'timing/model': 0.4143514521420002, 'learning_rate/qwen_vl_interface': 8.96301782672904e-06, 'learning_rate/action_model': 8.963017826729038e-05, 'epoch': 0.17})
+[2026-07-28 03:30:16,788][starVLA.training.train_starvla][INFO] - Step 937, Loss: {'train/loss': 0.050444431602954865, 'train/grad_norm_pre_clip': 2.985438823699951, 'timing/data': 0.000290874857455492, 'timing/model': 0.4134377599693835, 'learning_rate/qwen_vl_interface': 8.960630275627197e-06, 'learning_rate/action_model': 8.960630275627197e-05, 'epoch': 0.17})
+[2026-07-28 03:30:22,595][starVLA.training.train_starvla][INFO] - Step 938, Loss: {'train/loss': 0.044867124408483505, 'train/grad_norm_pre_clip': 0.7396103739738464, 'timing/data': 0.00027828803285956383, 'timing/model': 0.4144132351502776, 'learning_rate/qwen_vl_interface': 8.958240316737655e-06, 'learning_rate/action_model': 8.958240316737656e-05, 'epoch': 0.17})
+[2026-07-28 03:30:28,649][starVLA.training.train_starvla][INFO] - Step 939, Loss: {'train/loss': 0.08693711459636688, 'train/grad_norm_pre_clip': 4.265085697174072, 'timing/data': 0.0003620698116719723, 'timing/model': 0.6499946960248053, 'learning_rate/qwen_vl_interface': 8.955847951611233e-06, 'learning_rate/action_model': 8.955847951611232e-05, 'epoch': 0.17})
+[2026-07-28 03:30:34,746][starVLA.training.train_starvla][INFO] - Step 940, Loss: {'train/loss': 0.09440334141254425, 'train/grad_norm_pre_clip': 1.9980286359786987, 'timing/data': 0.00025269482284784317, 'timing/model': 0.4162991223856807, 'learning_rate/qwen_vl_interface': 8.953453181800307e-06, 'learning_rate/action_model': 8.953453181800306e-05, 'epoch': 0.17})
+[2026-07-28 03:30:40,627][starVLA.training.train_starvla][INFO] - Step 941, Loss: {'train/loss': 0.05232899263501167, 'train/grad_norm_pre_clip': 1.9401285648345947, 'timing/data': 0.0002539600245654583, 'timing/model': 0.4128229538910091, 'learning_rate/qwen_vl_interface': 8.951056008858819e-06, 'learning_rate/action_model': 8.95105600885882e-05, 'epoch': 0.17})
+[2026-07-28 03:30:46,463][starVLA.training.train_starvla][INFO] - Step 942, Loss: {'train/loss': 0.04222521558403969, 'train/grad_norm_pre_clip': 1.020993947982788, 'timing/data': 0.00027524493634700775, 'timing/model': 0.415915688034147, 'learning_rate/qwen_vl_interface': 8.94865643434227e-06, 'learning_rate/action_model': 8.948656434342269e-05, 'epoch': 0.17})
+[2026-07-28 03:30:52,324][starVLA.training.train_starvla][INFO] - Step 943, Loss: {'train/loss': 0.08120545744895935, 'train/grad_norm_pre_clip': 1.5972081422805786, 'timing/data': 0.0002871369943022728, 'timing/model': 0.41940151527523994, 'learning_rate/qwen_vl_interface': 8.946254459807717e-06, 'learning_rate/action_model': 8.946254459807717e-05, 'epoch': 0.17})
+[2026-07-28 03:30:58,168][starVLA.training.train_starvla][INFO] - Step 944, Loss: {'train/loss': 0.08651398122310638, 'train/grad_norm_pre_clip': 1.440710186958313, 'timing/data': 0.0002855360507965088, 'timing/model': 0.41659268690273166, 'learning_rate/qwen_vl_interface': 8.943850086813773e-06, 'learning_rate/action_model': 8.943850086813773e-05, 'epoch': 0.17})
+[2026-07-28 03:31:04,286][starVLA.training.train_starvla][INFO] - Step 945, Loss: {'train/loss': 0.11479215323925018, 'train/grad_norm_pre_clip': 1.2607852220535278, 'timing/data': 0.00026916805654764175, 'timing/model': 0.4169488619081676, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.17})
+[2026-07-28 03:31:10,391][starVLA.training.train_starvla][INFO] - Step 946, Loss: {'train/loss': 0.07672646641731262, 'train/grad_norm_pre_clip': 2.0277342796325684, 'timing/data': 0.00027868105098605156, 'timing/model': 0.6530813458375633, 'learning_rate/qwen_vl_interface': 8.93903415168996e-06, 'learning_rate/action_model': 8.93903415168996e-05, 'epoch': 0.17})
+[2026-07-28 03:31:16,272][starVLA.training.train_starvla][INFO] - Step 947, Loss: {'train/loss': 0.04949049651622772, 'train/grad_norm_pre_clip': 1.9785438776016235, 'timing/data': 0.00033160019665956497, 'timing/model': 0.4178635170683265, 'learning_rate/qwen_vl_interface': 8.9366225926851e-06, 'learning_rate/action_model': 8.936622592685098e-05, 'epoch': 0.17})
+[2026-07-28 03:31:22,169][starVLA.training.train_starvla][INFO] - Step 948, Loss: {'train/loss': 0.12281771004199982, 'train/grad_norm_pre_clip': 1.3733007907867432, 'timing/data': 0.00027168216183781624, 'timing/model': 0.416184077039361, 'learning_rate/qwen_vl_interface': 8.934208641470863e-06, 'learning_rate/action_model': 8.934208641470863e-05, 'epoch': 0.17})
+[2026-07-28 03:31:28,303][starVLA.training.train_starvla][INFO] - Step 949, Loss: {'train/loss': 0.07057221978902817, 'train/grad_norm_pre_clip': 1.9192230701446533, 'timing/data': 0.0003667715936899185, 'timing/model': 0.4145760308019817, 'learning_rate/qwen_vl_interface': 8.931792299613637e-06, 'learning_rate/action_model': 8.931792299613636e-05, 'epoch': 0.17})
+[2026-07-28 03:31:34,181][starVLA.training.train_starvla][INFO] - Step 950, Loss: {'train/loss': 0.059187810868024826, 'train/grad_norm_pre_clip': 1.9439295530319214, 'timing/data': 0.0002749701961874962, 'timing/model': 0.41750014014542103, 'learning_rate/qwen_vl_interface': 8.92937356868136e-06, 'learning_rate/action_model': 8.92937356868136e-05, 'epoch': 0.17})
+[2026-07-28 03:31:40,290][starVLA.training.train_starvla][INFO] - Step 951, Loss: {'train/loss': 0.05021418631076813, 'train/grad_norm_pre_clip': 0.9999833106994629, 'timing/data': 0.00029704486951231956, 'timing/model': 0.4165904251858592, 'learning_rate/qwen_vl_interface': 8.926952450243522e-06, 'learning_rate/action_model': 8.926952450243522e-05, 'epoch': 0.17})
+[2026-07-28 03:31:46,621][starVLA.training.train_starvla][INFO] - Step 952, Loss: {'train/loss': 0.05739004909992218, 'train/grad_norm_pre_clip': 1.4157463312149048, 'timing/data': 0.0002773357555270195, 'timing/model': 0.6497574569657445, 'learning_rate/qwen_vl_interface': 8.924528945871157e-06, 'learning_rate/action_model': 8.924528945871156e-05, 'epoch': 0.17})
+[2026-07-28 03:31:52,562][starVLA.training.train_starvla][INFO] - Step 953, Loss: {'train/loss': 0.07711140811443329, 'train/grad_norm_pre_clip': 2.7878408432006836, 'timing/data': 0.0004599308595061302, 'timing/model': 0.4169425410218537, 'learning_rate/qwen_vl_interface': 8.922103057136857e-06, 'learning_rate/action_model': 8.922103057136856e-05, 'epoch': 0.18})
+[2026-07-28 03:31:58,422][starVLA.training.train_starvla][INFO] - Step 954, Loss: {'train/loss': 0.055131930857896805, 'train/grad_norm_pre_clip': 2.662822723388672, 'timing/data': 0.0002770223654806614, 'timing/model': 0.4185952669940889, 'learning_rate/qwen_vl_interface': 8.919674785614747e-06, 'learning_rate/action_model': 8.919674785614747e-05, 'epoch': 0.18})
+[2026-07-28 03:32:04,539][starVLA.training.train_starvla][INFO] - Step 955, Loss: {'train/loss': 0.10782851278781891, 'train/grad_norm_pre_clip': 3.156330108642578, 'timing/data': 0.0002619628794491291, 'timing/model': 0.4169366699643433, 'learning_rate/qwen_vl_interface': 8.91724413288051e-06, 'learning_rate/action_model': 8.91724413288051e-05, 'epoch': 0.18})
+[2026-07-28 03:32:10,664][starVLA.training.train_starvla][INFO] - Step 956, Loss: {'train/loss': 0.09361076354980469, 'train/grad_norm_pre_clip': 2.287330150604248, 'timing/data': 0.00025843502953648567, 'timing/model': 0.41636142786592245, 'learning_rate/qwen_vl_interface': 8.91481110051137e-06, 'learning_rate/action_model': 8.91481110051137e-05, 'epoch': 0.18})
+[2026-07-28 03:32:16,535][starVLA.training.train_starvla][INFO] - Step 957, Loss: {'train/loss': 0.03670449182391167, 'train/grad_norm_pre_clip': 1.5635923147201538, 'timing/data': 0.0002841460518538952, 'timing/model': 0.4164813286624849, 'learning_rate/qwen_vl_interface': 8.912375690086097e-06, 'learning_rate/action_model': 8.912375690086096e-05, 'epoch': 0.18})
+[2026-07-28 03:32:22,646][starVLA.training.train_starvla][INFO] - Step 958, Loss: {'train/loss': 0.06703123450279236, 'train/grad_norm_pre_clip': 1.2389442920684814, 'timing/data': 0.00025851279497146606, 'timing/model': 0.6477252719923854, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.18})
+[2026-07-28 03:32:28,748][starVLA.training.train_starvla][INFO] - Step 959, Loss: {'train/loss': 0.058234672993421555, 'train/grad_norm_pre_clip': 1.5986418724060059, 'timing/data': 0.00028108106926083565, 'timing/model': 0.41573949996382, 'learning_rate/qwen_vl_interface': 8.907497741389932e-06, 'learning_rate/action_model': 8.907497741389932e-05, 'epoch': 0.18})
+[2026-07-28 03:32:34,630][starVLA.training.train_starvla][INFO] - Step 960, Loss: {'train/loss': 0.045766979455947876, 'train/grad_norm_pre_clip': 0.9167945981025696, 'timing/data': 0.00037697190418839455, 'timing/model': 0.41779060009866953, 'learning_rate/qwen_vl_interface': 8.905055206284289e-06, 'learning_rate/action_model': 8.905055206284288e-05, 'epoch': 0.18})
+[2026-07-28 03:32:40,552][starVLA.training.train_starvla][INFO] - Step 961, Loss: {'train/loss': 0.06494583189487457, 'train/grad_norm_pre_clip': 1.282022476196289, 'timing/data': 0.0002844040282070637, 'timing/model': 0.41835580999031663, 'learning_rate/qwen_vl_interface': 8.902610299453005e-06, 'learning_rate/action_model': 8.902610299453004e-05, 'epoch': 0.18})
+[2026-07-28 03:32:46,412][starVLA.training.train_starvla][INFO] - Step 962, Loss: {'train/loss': 0.05188153311610222, 'train/grad_norm_pre_clip': 1.0078133344650269, 'timing/data': 0.00035861413925886154, 'timing/model': 0.416399119887501, 'learning_rate/qwen_vl_interface': 8.900163022482554e-06, 'learning_rate/action_model': 8.900163022482553e-05, 'epoch': 0.18})
+[2026-07-28 03:32:52,530][starVLA.training.train_starvla][INFO] - Step 963, Loss: {'train/loss': 0.08356796950101852, 'train/grad_norm_pre_clip': 1.5929454565048218, 'timing/data': 0.00028204498812556267, 'timing/model': 0.6573767918162048, 'learning_rate/qwen_vl_interface': 8.897713376960946e-06, 'learning_rate/action_model': 8.897713376960946e-05, 'epoch': 0.18})
+[2026-07-28 03:32:58,164][starVLA.training.train_starvla][INFO] - Step 964, Loss: {'train/loss': 0.07297427207231522, 'train/grad_norm_pre_clip': 2.376417398452759, 'timing/data': 0.00026827165856957436, 'timing/model': 0.41440154565498233, 'learning_rate/qwen_vl_interface': 8.895261364477731e-06, 'learning_rate/action_model': 8.895261364477731e-05, 'epoch': 0.18})
+[2026-07-28 03:33:03,800][starVLA.training.train_starvla][INFO] - Step 965, Loss: {'train/loss': 0.06033209338784218, 'train/grad_norm_pre_clip': 1.4113517999649048, 'timing/data': 0.0002779671922326088, 'timing/model': 0.4164725230075419, 'learning_rate/qwen_vl_interface': 8.892806986623993e-06, 'learning_rate/action_model': 8.892806986623992e-05, 'epoch': 0.18})
+[2026-07-28 03:33:09,715][starVLA.training.train_starvla][INFO] - Step 966, Loss: {'train/loss': 0.10781846940517426, 'train/grad_norm_pre_clip': 1.8471031188964844, 'timing/data': 0.0003257598727941513, 'timing/model': 0.41613335302099586, 'learning_rate/qwen_vl_interface': 8.890350244992349e-06, 'learning_rate/action_model': 8.890350244992348e-05, 'epoch': 0.18})
+[2026-07-28 03:33:15,595][starVLA.training.train_starvla][INFO] - Step 967, Loss: {'train/loss': 0.060038864612579346, 'train/grad_norm_pre_clip': 1.3201115131378174, 'timing/data': 0.00028614234179258347, 'timing/model': 0.4177964893169701, 'learning_rate/qwen_vl_interface': 8.887891141176954e-06, 'learning_rate/action_model': 8.887891141176954e-05, 'epoch': 0.18})
+[2026-07-28 03:33:21,470][starVLA.training.train_starvla][INFO] - Step 968, Loss: {'train/loss': 0.09611959010362625, 'train/grad_norm_pre_clip': 1.6155674457550049, 'timing/data': 0.00030086422339081764, 'timing/model': 0.41627905797213316, 'learning_rate/qwen_vl_interface': 8.885429676773495e-06, 'learning_rate/action_model': 8.885429676773495e-05, 'epoch': 0.18})
+[2026-07-28 03:33:27,852][starVLA.training.train_starvla][INFO] - Step 969, Loss: {'train/loss': 0.07084749639034271, 'train/grad_norm_pre_clip': 4.179526329040527, 'timing/data': 0.000264693982899189, 'timing/model': 0.6499914550222456, 'learning_rate/qwen_vl_interface': 8.882965853379188e-06, 'learning_rate/action_model': 8.882965853379187e-05, 'epoch': 0.18})
+[2026-07-28 03:33:33,738][starVLA.training.train_starvla][INFO] - Step 970, Loss: {'train/loss': 0.09015465527772903, 'train/grad_norm_pre_clip': 3.4565062522888184, 'timing/data': 0.0003626551479101181, 'timing/model': 0.41477237408980727, 'learning_rate/qwen_vl_interface': 8.88049967259278e-06, 'learning_rate/action_model': 8.880499672592778e-05, 'epoch': 0.18})
+[2026-07-28 03:33:39,646][starVLA.training.train_starvla][INFO] - Step 971, Loss: {'train/loss': 0.04827707260847092, 'train/grad_norm_pre_clip': 0.9674308896064758, 'timing/data': 0.0003604101948440075, 'timing/model': 0.41712317802011967, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.18})
+[2026-07-28 03:33:45,527][starVLA.training.train_starvla][INFO] - Step 972, Loss: {'train/loss': 0.06386396288871765, 'train/grad_norm_pre_clip': 1.6486088037490845, 'timing/data': 0.00037182215601205826, 'timing/model': 0.4169514081440866, 'learning_rate/qwen_vl_interface': 8.875560245246307e-06, 'learning_rate/action_model': 8.875560245246307e-05, 'epoch': 0.18})
+[2026-07-28 03:33:51,642][starVLA.training.train_starvla][INFO] - Step 973, Loss: {'train/loss': 0.08893468976020813, 'train/grad_norm_pre_clip': 1.5853846073150635, 'timing/data': 0.00025812815874814987, 'timing/model': 0.4148601139895618, 'learning_rate/qwen_vl_interface': 8.873087001891381e-06, 'learning_rate/action_model': 8.87308700189138e-05, 'epoch': 0.18})
+[2026-07-28 03:33:57,749][starVLA.training.train_starvla][INFO] - Step 974, Loss: {'train/loss': 0.058483392000198364, 'train/grad_norm_pre_clip': 1.6542450189590454, 'timing/data': 0.00026326999068260193, 'timing/model': 0.41639612428843975, 'learning_rate/qwen_vl_interface': 8.870611407554635e-06, 'learning_rate/action_model': 8.870611407554636e-05, 'epoch': 0.18})
+[2026-07-28 03:34:03,597][starVLA.training.train_starvla][INFO] - Step 975, Loss: {'train/loss': 0.06728994846343994, 'train/grad_norm_pre_clip': 1.432913899421692, 'timing/data': 0.00026674382388591766, 'timing/model': 0.41722823679447174, 'learning_rate/qwen_vl_interface': 8.868133463842456e-06, 'learning_rate/action_model': 8.868133463842457e-05, 'epoch': 0.18})
+[2026-07-28 03:34:09,683][starVLA.training.train_starvla][INFO] - Step 976, Loss: {'train/loss': 0.0774768739938736, 'train/grad_norm_pre_clip': 2.8295915126800537, 'timing/data': 0.0003571780398488045, 'timing/model': 0.6539043020457029, 'learning_rate/qwen_vl_interface': 8.865653172362755e-06, 'learning_rate/action_model': 8.865653172362754e-05, 'epoch': 0.18})
+[2026-07-28 03:34:15,530][starVLA.training.train_starvla][INFO] - Step 977, Loss: {'train/loss': 0.062080249190330505, 'train/grad_norm_pre_clip': 1.362103819847107, 'timing/data': 0.0002766740508377552, 'timing/model': 0.41815871000289917, 'learning_rate/qwen_vl_interface': 8.863170534724962e-06, 'learning_rate/action_model': 8.863170534724961e-05, 'epoch': 0.18})
+[2026-07-28 03:34:21,634][starVLA.training.train_starvla][INFO] - Step 978, Loss: {'train/loss': 0.08607681095600128, 'train/grad_norm_pre_clip': 2.4242758750915527, 'timing/data': 0.0002602189779281616, 'timing/model': 0.4144742041826248, 'learning_rate/qwen_vl_interface': 8.860685552540038e-06, 'learning_rate/action_model': 8.860685552540037e-05, 'epoch': 0.18})
+[2026-07-28 03:34:27,508][starVLA.training.train_starvla][INFO] - Step 979, Loss: {'train/loss': 0.03419636934995651, 'train/grad_norm_pre_clip': 1.1887279748916626, 'timing/data': 0.00027876487001776695, 'timing/model': 0.4168349588289857, 'learning_rate/qwen_vl_interface': 8.858198227420461e-06, 'learning_rate/action_model': 8.85819822742046e-05, 'epoch': 0.18})
+[2026-07-28 03:34:33,393][starVLA.training.train_starvla][INFO] - Step 980, Loss: {'train/loss': 0.05509534478187561, 'train/grad_norm_pre_clip': 1.085446834564209, 'timing/data': 0.00029162224382162094, 'timing/model': 0.41691156569868326, 'learning_rate/qwen_vl_interface': 8.855708560980225e-06, 'learning_rate/action_model': 8.855708560980225e-05, 'epoch': 0.18})
+[2026-07-28 03:34:39,512][starVLA.training.train_starvla][INFO] - Step 981, Loss: {'train/loss': 0.06526879221200943, 'train/grad_norm_pre_clip': 1.462889552116394, 'timing/data': 0.00026936503127217293, 'timing/model': 0.4162321537733078, 'learning_rate/qwen_vl_interface': 8.853216554834852e-06, 'learning_rate/action_model': 8.85321655483485e-05, 'epoch': 0.18})
+[2026-07-28 03:34:45,412][starVLA.training.train_starvla][INFO] - Step 982, Loss: {'train/loss': 0.09732665866613388, 'train/grad_norm_pre_clip': 4.008265018463135, 'timing/data': 0.00028642499819397926, 'timing/model': 0.41631920728832483, 'learning_rate/qwen_vl_interface': 8.850722210601376e-06, 'learning_rate/action_model': 8.850722210601377e-05, 'epoch': 0.18})
+[2026-07-28 03:34:51,552][starVLA.training.train_starvla][INFO] - Step 983, Loss: {'train/loss': 0.09962388128042221, 'train/grad_norm_pre_clip': 2.236056089401245, 'timing/data': 0.0002741217613220215, 'timing/model': 0.659490317106247, 'learning_rate/qwen_vl_interface': 8.848225529898351e-06, 'learning_rate/action_model': 8.84822552989835e-05, 'epoch': 0.18})
+[2026-07-28 03:34:57,428][starVLA.training.train_starvla][INFO] - Step 984, Loss: {'train/loss': 0.05916695296764374, 'train/grad_norm_pre_clip': 1.4252382516860962, 'timing/data': 0.00026126718148589134, 'timing/model': 0.4145422517322004, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.18})
+[2026-07-28 03:35:03,324][starVLA.training.train_starvla][INFO] - Step 985, Loss: {'train/loss': 0.06146102398633957, 'train/grad_norm_pre_clip': 3.1391754150390625, 'timing/data': 0.0002834480255842209, 'timing/model': 0.41861209087073803, 'learning_rate/qwen_vl_interface': 8.84322516556544e-06, 'learning_rate/action_model': 8.84322516556544e-05, 'epoch': 0.18})
+[2026-07-28 03:35:09,438][starVLA.training.train_starvla][INFO] - Step 986, Loss: {'train/loss': 0.08767899870872498, 'train/grad_norm_pre_clip': 2.271275043487549, 'timing/data': 0.0002798279747366905, 'timing/model': 0.6546659711748362, 'learning_rate/qwen_vl_interface': 8.840721485180237e-06, 'learning_rate/action_model': 8.840721485180236e-05, 'epoch': 0.18})
+[2026-07-28 03:35:15,308][starVLA.training.train_starvla][INFO] - Step 987, Loss: {'train/loss': 0.09435130655765533, 'train/grad_norm_pre_clip': 2.1736016273498535, 'timing/data': 0.0002725590020418167, 'timing/model': 0.41640070686116815, 'learning_rate/qwen_vl_interface': 8.838215474814848e-06, 'learning_rate/action_model': 8.838215474814847e-05, 'epoch': 0.18})
+[2026-07-28 03:35:21,244][starVLA.training.train_starvla][INFO] - Step 988, Loss: {'train/loss': 0.06491214036941528, 'train/grad_norm_pre_clip': 0.9307884573936462, 'timing/data': 0.0002887910231947899, 'timing/model': 0.4176955339498818, 'learning_rate/qwen_vl_interface': 8.835707136095392e-06, 'learning_rate/action_model': 8.835707136095391e-05, 'epoch': 0.18})
+[2026-07-28 03:35:27,147][starVLA.training.train_starvla][INFO] - Step 989, Loss: {'train/loss': 0.06745626032352448, 'train/grad_norm_pre_clip': 1.643947958946228, 'timing/data': 0.0002913298085331917, 'timing/model': 0.4172711097635329, 'learning_rate/qwen_vl_interface': 8.833196470649505e-06, 'learning_rate/action_model': 8.833196470649505e-05, 'epoch': 0.18})
+[2026-07-28 03:35:33,269][starVLA.training.train_starvla][INFO] - Step 990, Loss: {'train/loss': 0.08419960737228394, 'train/grad_norm_pre_clip': 1.8275372982025146, 'timing/data': 0.00036633480340242386, 'timing/model': 0.4144354132004082, 'learning_rate/qwen_vl_interface': 8.83068348010633e-06, 'learning_rate/action_model': 8.83068348010633e-05, 'epoch': 0.18})
+[2026-07-28 03:35:39,401][starVLA.training.train_starvla][INFO] - Step 991, Loss: {'train/loss': 0.09415587782859802, 'train/grad_norm_pre_clip': 2.0152056217193604, 'timing/data': 0.00028462987393140793, 'timing/model': 0.4179501929320395, 'learning_rate/qwen_vl_interface': 8.82816816609652e-06, 'learning_rate/action_model': 8.828168166096521e-05, 'epoch': 0.18})
+[2026-07-28 03:35:45,293][starVLA.training.train_starvla][INFO] - Step 992, Loss: {'train/loss': 0.06611491739749908, 'train/grad_norm_pre_clip': 0.9950928688049316, 'timing/data': 0.00029563717544078827, 'timing/model': 0.4180468451231718, 'learning_rate/qwen_vl_interface': 8.825650530252238e-06, 'learning_rate/action_model': 8.825650530252238e-05, 'epoch': 0.18})
+[2026-07-28 03:35:51,012][starVLA.training.train_starvla][INFO] - Step 993, Loss: {'train/loss': 0.07410268485546112, 'train/grad_norm_pre_clip': 1.9648038148880005, 'timing/data': 0.0003000609576702118, 'timing/model': 0.41801081225275993, 'learning_rate/qwen_vl_interface': 8.823130574207144e-06, 'learning_rate/action_model': 8.823130574207143e-05, 'epoch': 0.18})
+[2026-07-28 03:35:57,129][starVLA.training.train_starvla][INFO] - Step 994, Loss: {'train/loss': 0.08419419825077057, 'train/grad_norm_pre_clip': 1.42646324634552, 'timing/data': 0.00026070326566696167, 'timing/model': 0.6483693169429898, 'learning_rate/qwen_vl_interface': 8.82060829959641e-06, 'learning_rate/action_model': 8.820608299596411e-05, 'epoch': 0.18})
+[2026-07-28 03:36:03,047][starVLA.training.train_starvla][INFO] - Step 995, Loss: {'train/loss': 0.0642513558268547, 'train/grad_norm_pre_clip': 1.008699655532837, 'timing/data': 0.0003820466808974743, 'timing/model': 0.42067882185801864, 'learning_rate/qwen_vl_interface': 8.81808370805672e-06, 'learning_rate/action_model': 8.818083708056721e-05, 'epoch': 0.18})
+[2026-07-28 03:36:08,986][starVLA.training.train_starvla][INFO] - Step 996, Loss: {'train/loss': 0.08444415777921677, 'train/grad_norm_pre_clip': 2.543093681335449, 'timing/data': 0.00026801275089383125, 'timing/model': 0.4492026949301362, 'learning_rate/qwen_vl_interface': 8.815556801226251e-06, 'learning_rate/action_model': 8.815556801226251e-05, 'epoch': 0.18})
+[2026-07-28 03:36:15,265][starVLA.training.train_starvla][INFO] - Step 997, Loss: {'train/loss': 0.05216958373785019, 'train/grad_norm_pre_clip': 1.373136043548584, 'timing/data': 0.0003849128261208534, 'timing/model': 0.6315658199600875, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.18})
+[2026-07-28 03:36:21,719][starVLA.training.train_starvla][INFO] - Step 998, Loss: {'train/loss': 0.050845611840486526, 'train/grad_norm_pre_clip': 1.6310418844223022, 'timing/data': 0.0002783569507300854, 'timing/model': 0.43043882912024856, 'learning_rate/qwen_vl_interface': 8.810496048253203e-06, 'learning_rate/action_model': 8.810496048253202e-05, 'epoch': 0.18})
+[2026-07-28 03:36:27,704][starVLA.training.train_starvla][INFO] - Step 999, Loss: {'train/loss': 0.07072535157203674, 'train/grad_norm_pre_clip': 1.4379721879959106, 'timing/data': 0.0002860170789062977, 'timing/model': 0.4173448118381202, 'learning_rate/qwen_vl_interface': 8.80796220539449e-06, 'learning_rate/action_model': 8.80796220539449e-05, 'epoch': 0.18})
+[2026-07-28 03:36:33,564][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1000
+[2026-07-28 03:36:38,599][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1000
+[2026-07-28 03:41:48,249][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1000
+[2026-07-28 03:41:48,250][starVLA.training.train_starvla][INFO] - Step 1000, Loss: {'train/loss': 0.09488687664270401, 'train/grad_norm_pre_clip': 2.35461688041687, 'eval/action_loss/seconds': 5.030141984112561, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.05499304458498955, 'eval/latency_3/loss': 0.05499304458498955, 'eval/flappy/loss': 0.05499304458498955, 'eval/flappy/latency_3/loss': 0.05499304458498955, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 197.2250023353845, 'rl_games_eval/mid_train/mean_length': 1609.65, 'rl_games_eval/mid_train/std_reward': 147.19520162716503, 'rl_games_eval/mid_train/std_length': 1187.3336209760087, 'rl_games_eval/mid_train/macro_mean_reward': 197.2250023353845, 'rl_games_eval/mid_train/macro_mean_length': 1609.65, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 197.2250023353845, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 1609.65, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 147.19520162716503, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1187.3336209760087, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 197.2250023353845, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00043568992987275124, 'timing/model': 0.4167599589563906, 'learning_rate/qwen_vl_interface': 8.805426053812731e-06, 'learning_rate/action_model': 8.805426053812731e-05, 'epoch': 0.18})
+[2026-07-28 03:41:48,252][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1000_state
+[2026-07-28 03:41:48,252][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 03:42:25,479][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1000_state/pytorch_model
+[2026-07-28 03:42:25,479][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1000_state/sampler.bin
+[2026-07-28 03:42:25,479][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1000_state/sampler_1.bin
+[2026-07-28 03:42:25,481][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1000_state/random_states_0.pkl
+[2026-07-28 03:42:34,406][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1000
+[2026-07-28 03:42:34,406][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 03:42:34,428][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 03:42:40,447][starVLA.training.train_starvla][INFO] - Step 1001, Loss: {'train/loss': 0.06638924777507782, 'train/grad_norm_pre_clip': 1.5235618352890015, 'timing/data': 0.00028664804995059967, 'timing/model': 0.4159573162905872, 'learning_rate/qwen_vl_interface': 8.802887595153607e-06, 'learning_rate/action_model': 8.802887595153606e-05, 'epoch': 0.18})
+[2026-07-28 03:42:46,301][starVLA.training.train_starvla][INFO] - Step 1002, Loss: {'train/loss': 0.03789106383919716, 'train/grad_norm_pre_clip': 1.3225764036178589, 'timing/data': 0.0002659768797457218, 'timing/model': 0.4171609287150204, 'learning_rate/qwen_vl_interface': 8.800346831064294e-06, 'learning_rate/action_model': 8.800346831064294e-05, 'epoch': 0.18})
+[2026-07-28 03:42:52,414][starVLA.training.train_starvla][INFO] - Step 1003, Loss: {'train/loss': 0.14622464776039124, 'train/grad_norm_pre_clip': 2.738168239593506, 'timing/data': 0.0003886958584189415, 'timing/model': 0.4172701770439744, 'learning_rate/qwen_vl_interface': 8.797803763193471e-06, 'learning_rate/action_model': 8.797803763193471e-05, 'epoch': 0.18})
+[2026-07-28 03:42:58,553][starVLA.training.train_starvla][INFO] - Step 1004, Loss: {'train/loss': 0.0756051242351532, 'train/grad_norm_pre_clip': 3.3303070068359375, 'timing/data': 0.0002516140230000019, 'timing/model': 0.41342455288395286, 'learning_rate/qwen_vl_interface': 8.795258393191302e-06, 'learning_rate/action_model': 8.795258393191301e-05, 'epoch': 0.18})
+[2026-07-28 03:43:04,195][starVLA.training.train_starvla][INFO] - Step 1005, Loss: {'train/loss': 0.0544993132352829, 'train/grad_norm_pre_clip': 1.3003368377685547, 'timing/data': 0.00028331438079476357, 'timing/model': 0.41617394983768463, 'learning_rate/qwen_vl_interface': 8.792710722709451e-06, 'learning_rate/action_model': 8.792710722709451e-05, 'epoch': 0.18})
+[2026-07-28 03:43:09,845][starVLA.training.train_starvla][INFO] - Step 1006, Loss: {'train/loss': 0.06995035707950592, 'train/grad_norm_pre_clip': 1.586295485496521, 'timing/data': 0.0004484360106289387, 'timing/model': 0.4290553121827543, 'learning_rate/qwen_vl_interface': 8.790160753401076e-06, 'learning_rate/action_model': 8.790160753401077e-05, 'epoch': 0.18})
+[2026-07-28 03:43:15,494][starVLA.training.train_starvla][INFO] - Step 1007, Loss: {'train/loss': 0.1003154069185257, 'train/grad_norm_pre_clip': 4.062765598297119, 'timing/data': 0.0002751210704445839, 'timing/model': 0.4178554490208626, 'learning_rate/qwen_vl_interface': 8.787608486920824e-06, 'learning_rate/action_model': 8.787608486920824e-05, 'epoch': 0.18})
+[2026-07-28 03:43:21,371][starVLA.training.train_starvla][INFO] - Step 1008, Loss: {'train/loss': 0.11458690464496613, 'train/grad_norm_pre_clip': 2.7765583992004395, 'timing/data': 0.0002794289030134678, 'timing/model': 0.6505915438756347, 'learning_rate/qwen_vl_interface': 8.78505392492483e-06, 'learning_rate/action_model': 8.78505392492483e-05, 'epoch': 0.19})
+[2026-07-28 03:43:27,454][starVLA.training.train_starvla][INFO] - Step 1009, Loss: {'train/loss': 0.061782106757164, 'train/grad_norm_pre_clip': 2.364591598510742, 'timing/data': 0.0002614823170006275, 'timing/model': 0.4150815811008215, 'learning_rate/qwen_vl_interface': 8.782497069070725e-06, 'learning_rate/action_model': 8.782497069070725e-05, 'epoch': 0.19})
+[2026-07-28 03:43:33,318][starVLA.training.train_starvla][INFO] - Step 1010, Loss: {'train/loss': 0.08129043132066727, 'train/grad_norm_pre_clip': 3.071397542953491, 'timing/data': 0.0002843211404979229, 'timing/model': 0.41676711989566684, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.19})
+[2026-07-28 03:43:39,188][starVLA.training.train_starvla][INFO] - Step 1011, Loss: {'train/loss': 0.07193811237812042, 'train/grad_norm_pre_clip': 1.4377357959747314, 'timing/data': 0.0002786167897284031, 'timing/model': 0.41848157113417983, 'learning_rate/qwen_vl_interface': 8.777376482426129e-06, 'learning_rate/action_model': 8.777376482426128e-05, 'epoch': 0.19})
+[2026-07-28 03:43:45,296][starVLA.training.train_starvla][INFO] - Step 1012, Loss: {'train/loss': 0.07058421522378922, 'train/grad_norm_pre_clip': 1.2717970609664917, 'timing/data': 0.00026335520669817924, 'timing/model': 0.4149172827601433, 'learning_rate/qwen_vl_interface': 8.774812754958332e-06, 'learning_rate/action_model': 8.774812754958331e-05, 'epoch': 0.19})
+[2026-07-28 03:43:51,387][starVLA.training.train_starvla][INFO] - Step 1013, Loss: {'train/loss': 0.06254638731479645, 'train/grad_norm_pre_clip': 2.6659114360809326, 'timing/data': 0.0002809809520840645, 'timing/model': 0.4168239952996373, 'learning_rate/qwen_vl_interface': 8.772246740277805e-06, 'learning_rate/action_model': 8.772246740277805e-05, 'epoch': 0.19})
+[2026-07-28 03:43:57,268][starVLA.training.train_starvla][INFO] - Step 1014, Loss: {'train/loss': 0.07946314662694931, 'train/grad_norm_pre_clip': 2.9960546493530273, 'timing/data': 0.0002758740447461605, 'timing/model': 0.41686296882107854, 'learning_rate/qwen_vl_interface': 8.769678440049608e-06, 'learning_rate/action_model': 8.769678440049608e-05, 'epoch': 0.19})
+[2026-07-28 03:44:03,141][starVLA.training.train_starvla][INFO] - Step 1015, Loss: {'train/loss': 0.0761881172657013, 'train/grad_norm_pre_clip': 1.6936756372451782, 'timing/data': 0.0002876236103475094, 'timing/model': 0.4180740611627698, 'learning_rate/qwen_vl_interface': 8.767107855940287e-06, 'learning_rate/action_model': 8.767107855940286e-05, 'epoch': 0.19})
+[2026-07-28 03:44:09,483][starVLA.training.train_starvla][INFO] - Step 1016, Loss: {'train/loss': 0.05301123484969139, 'train/grad_norm_pre_clip': 1.662062644958496, 'timing/data': 0.00026746978983283043, 'timing/model': 0.41621800092980266, 'learning_rate/qwen_vl_interface': 8.764534989617861e-06, 'learning_rate/action_model': 8.76453498961786e-05, 'epoch': 0.19})
+[2026-07-28 03:44:15,350][starVLA.training.train_starvla][INFO] - Step 1017, Loss: {'train/loss': 0.04803580790758133, 'train/grad_norm_pre_clip': 1.6233348846435547, 'timing/data': 0.00039297109469771385, 'timing/model': 0.4170110607519746, 'learning_rate/qwen_vl_interface': 8.761959842751837e-06, 'learning_rate/action_model': 8.761959842751837e-05, 'epoch': 0.19})
+[2026-07-28 03:44:21,218][starVLA.training.train_starvla][INFO] - Step 1018, Loss: {'train/loss': 0.0613878034055233, 'train/grad_norm_pre_clip': 2.1345956325531006, 'timing/data': 0.00028456375002861023, 'timing/model': 0.4174908841960132, 'learning_rate/qwen_vl_interface': 8.7593824170132e-06, 'learning_rate/action_model': 8.7593824170132e-05, 'epoch': 0.19})
+[2026-07-28 03:44:27,258][starVLA.training.train_starvla][INFO] - Step 1019, Loss: {'train/loss': 0.044433217495679855, 'train/grad_norm_pre_clip': 1.068658471107483, 'timing/data': 0.0002934481017291546, 'timing/model': 0.8243755889125168, 'learning_rate/qwen_vl_interface': 8.756802714074412e-06, 'learning_rate/action_model': 8.756802714074412e-05, 'epoch': 0.19})
+[2026-07-28 03:44:32,924][starVLA.training.train_starvla][INFO] - Step 1020, Loss: {'train/loss': 0.08003462105989456, 'train/grad_norm_pre_clip': 3.3204004764556885, 'timing/data': 0.0002934420481324196, 'timing/model': 0.41940957168117166, 'learning_rate/qwen_vl_interface': 8.75422073560942e-06, 'learning_rate/action_model': 8.754220735609419e-05, 'epoch': 0.19})
+[2026-07-28 03:44:38,576][starVLA.training.train_starvla][INFO] - Step 1021, Loss: {'train/loss': 0.064293771982193, 'train/grad_norm_pre_clip': 2.6945881843566895, 'timing/data': 0.0002814573235809803, 'timing/model': 0.4144189120270312, 'learning_rate/qwen_vl_interface': 8.751636483293635e-06, 'learning_rate/action_model': 8.751636483293634e-05, 'epoch': 0.19})
+[2026-07-28 03:44:44,465][starVLA.training.train_starvla][INFO] - Step 1022, Loss: {'train/loss': 0.11799921095371246, 'train/grad_norm_pre_clip': 3.6697731018066406, 'timing/data': 0.0002880520187318325, 'timing/model': 0.41988273058086634, 'learning_rate/qwen_vl_interface': 8.749049958803952e-06, 'learning_rate/action_model': 8.749049958803951e-05, 'epoch': 0.19})
+[2026-07-28 03:44:50,532][starVLA.training.train_starvla][INFO] - Step 1023, Loss: {'train/loss': 0.07201149314641953, 'train/grad_norm_pre_clip': 2.6517181396484375, 'timing/data': 0.00025718798860907555, 'timing/model': 0.41267234832048416, 'learning_rate/qwen_vl_interface': 8.74646116381874e-06, 'learning_rate/action_model': 8.74646116381874e-05, 'epoch': 0.19})
+[2026-07-28 03:44:56,631][starVLA.training.train_starvla][INFO] - Step 1024, Loss: {'train/loss': 0.04338684305548668, 'train/grad_norm_pre_clip': 1.9378141164779663, 'timing/data': 0.00028937309980392456, 'timing/model': 0.6590148848481476, 'learning_rate/qwen_vl_interface': 8.74387010001784e-06, 'learning_rate/action_model': 8.743870100017838e-05, 'epoch': 0.19})
+[2026-07-28 03:45:02,523][starVLA.training.train_starvla][INFO] - Step 1025, Loss: {'train/loss': 0.09162500500679016, 'train/grad_norm_pre_clip': 3.394230842590332, 'timing/data': 0.00026900600641965866, 'timing/model': 0.4149970719590783, 'learning_rate/qwen_vl_interface': 8.741276769082562e-06, 'learning_rate/action_model': 8.741276769082562e-05, 'epoch': 0.19})
+[2026-07-28 03:45:08,333][starVLA.training.train_starvla][INFO] - Step 1026, Loss: {'train/loss': 0.05850476026535034, 'train/grad_norm_pre_clip': 1.4316624402999878, 'timing/data': 0.0002806996926665306, 'timing/model': 0.4166660071350634, 'learning_rate/qwen_vl_interface': 8.738681172695694e-06, 'learning_rate/action_model': 8.738681172695694e-05, 'epoch': 0.19})
+[2026-07-28 03:45:14,433][starVLA.training.train_starvla][INFO] - Step 1027, Loss: {'train/loss': 0.058739837259054184, 'train/grad_norm_pre_clip': 1.6000711917877197, 'timing/data': 0.00028481893241405487, 'timing/model': 0.6453256239183247, 'learning_rate/qwen_vl_interface': 8.73608331254149e-06, 'learning_rate/action_model': 8.736083312541489e-05, 'epoch': 0.19})
+[2026-07-28 03:45:20,339][starVLA.training.train_starvla][INFO] - Step 1028, Loss: {'train/loss': 0.0627899169921875, 'train/grad_norm_pre_clip': 2.479370594024658, 'timing/data': 0.00026670191437005997, 'timing/model': 0.4155074446462095, 'learning_rate/qwen_vl_interface': 8.733483190305671e-06, 'learning_rate/action_model': 8.733483190305671e-05, 'epoch': 0.19})
+[2026-07-28 03:45:26,210][starVLA.training.train_starvla][INFO] - Step 1029, Loss: {'train/loss': 0.09757571667432785, 'train/grad_norm_pre_clip': 4.754105567932129, 'timing/data': 0.00027511781081557274, 'timing/model': 0.4168807999230921, 'learning_rate/qwen_vl_interface': 8.730880807675431e-06, 'learning_rate/action_model': 8.73088080767543e-05, 'epoch': 0.19})
+[2026-07-28 03:45:32,352][starVLA.training.train_starvla][INFO] - Step 1030, Loss: {'train/loss': 0.05615663528442383, 'train/grad_norm_pre_clip': 1.5483835935592651, 'timing/data': 0.0005130134522914886, 'timing/model': 0.6536467163823545, 'learning_rate/qwen_vl_interface': 8.728276166339426e-06, 'learning_rate/action_model': 8.728276166339426e-05, 'epoch': 0.19})
+[2026-07-28 03:45:38,216][starVLA.training.train_starvla][INFO] - Step 1031, Loss: {'train/loss': 0.06004209816455841, 'train/grad_norm_pre_clip': 2.072126865386963, 'timing/data': 0.00024359207600355148, 'timing/model': 0.41494796704500914, 'learning_rate/qwen_vl_interface': 8.725669267987779e-06, 'learning_rate/action_model': 8.725669267987778e-05, 'epoch': 0.19})
+[2026-07-28 03:45:44,084][starVLA.training.train_starvla][INFO] - Step 1032, Loss: {'train/loss': 0.08614296466112137, 'train/grad_norm_pre_clip': 2.594409465789795, 'timing/data': 0.00027400068938732147, 'timing/model': 0.4161828439682722, 'learning_rate/qwen_vl_interface': 8.723060114312079e-06, 'learning_rate/action_model': 8.723060114312079e-05, 'epoch': 0.19})
+[2026-07-28 03:45:50,212][starVLA.training.train_starvla][INFO] - Step 1033, Loss: {'train/loss': 0.05894023925065994, 'train/grad_norm_pre_clip': 2.227346181869507, 'timing/data': 0.0002470239996910095, 'timing/model': 0.43824672512710094, 'learning_rate/qwen_vl_interface': 8.720448707005377e-06, 'learning_rate/action_model': 8.720448707005376e-05, 'epoch': 0.19})
+[2026-07-28 03:45:56,156][starVLA.training.train_starvla][INFO] - Step 1034, Loss: {'train/loss': 0.05568006634712219, 'train/grad_norm_pre_clip': 2.1046559810638428, 'timing/data': 0.0002805669791996479, 'timing/model': 0.43617677688598633, 'learning_rate/qwen_vl_interface': 8.717835047762188e-06, 'learning_rate/action_model': 8.717835047762187e-05, 'epoch': 0.19})
+[2026-07-28 03:46:02,445][starVLA.training.train_starvla][INFO] - Step 1035, Loss: {'train/loss': 0.0960206687450409, 'train/grad_norm_pre_clip': 3.175422430038452, 'timing/data': 0.00024536019191145897, 'timing/model': 0.4174149129539728, 'learning_rate/qwen_vl_interface': 8.715219138278486e-06, 'learning_rate/action_model': 8.715219138278486e-05, 'epoch': 0.19})
+[2026-07-28 03:46:08,294][starVLA.training.train_starvla][INFO] - Step 1036, Loss: {'train/loss': 0.07054482400417328, 'train/grad_norm_pre_clip': 1.1998831033706665, 'timing/data': 0.00027746008709073067, 'timing/model': 0.4170322730205953, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.19})
+[2026-07-28 03:46:14,434][starVLA.training.train_starvla][INFO] - Step 1037, Loss: {'train/loss': 0.058505717664957047, 'train/grad_norm_pre_clip': 2.0506691932678223, 'timing/data': 0.0002478337846696377, 'timing/model': 0.41493471106514335, 'learning_rate/qwen_vl_interface': 8.709980575380741e-06, 'learning_rate/action_model': 8.709980575380741e-05, 'epoch': 0.19})
+[2026-07-28 03:46:20,302][starVLA.training.train_starvla][INFO] - Step 1038, Loss: {'train/loss': 0.05539784952998161, 'train/grad_norm_pre_clip': 2.4491662979125977, 'timing/data': 0.00028683291748166084, 'timing/model': 0.4184887418523431, 'learning_rate/qwen_vl_interface': 8.707357925365949e-06, 'learning_rate/action_model': 8.707357925365948e-05, 'epoch': 0.19})
+[2026-07-28 03:46:26,402][starVLA.training.train_starvla][INFO] - Step 1039, Loss: {'train/loss': 0.055511146783828735, 'train/grad_norm_pre_clip': 0.6391252279281616, 'timing/data': 0.000245171133428812, 'timing/model': 0.4164127390831709, 'learning_rate/qwen_vl_interface': 8.704733031909134e-06, 'learning_rate/action_model': 8.704733031909133e-05, 'epoch': 0.19})
+[2026-07-28 03:46:32,491][starVLA.training.train_starvla][INFO] - Step 1040, Loss: {'train/loss': 0.0940525159239769, 'train/grad_norm_pre_clip': 1.3635739088058472, 'timing/data': 0.0002699820324778557, 'timing/model': 0.6517104557715356, 'learning_rate/qwen_vl_interface': 8.70210589671356e-06, 'learning_rate/action_model': 8.702105896713561e-05, 'epoch': 0.19})
+[2026-07-28 03:46:38,388][starVLA.training.train_starvla][INFO] - Step 1041, Loss: {'train/loss': 0.07729630172252655, 'train/grad_norm_pre_clip': 1.310729742050171, 'timing/data': 0.00027139391750097275, 'timing/model': 0.4160744142718613, 'learning_rate/qwen_vl_interface': 8.699476521483952e-06, 'learning_rate/action_model': 8.699476521483952e-05, 'epoch': 0.19})
+[2026-07-28 03:46:44,318][starVLA.training.train_starvla][INFO] - Step 1042, Loss: {'train/loss': 0.07823222875595093, 'train/grad_norm_pre_clip': 2.1368846893310547, 'timing/data': 0.00027544284239411354, 'timing/model': 0.4245869140140712, 'learning_rate/qwen_vl_interface': 8.696844907926478e-06, 'learning_rate/action_model': 8.696844907926477e-05, 'epoch': 0.19})
+[2026-07-28 03:46:50,236][starVLA.training.train_starvla][INFO] - Step 1043, Loss: {'train/loss': 0.08298162370920181, 'train/grad_norm_pre_clip': 1.59860360622406, 'timing/data': 0.0002831481397151947, 'timing/model': 0.41839922592043877, 'learning_rate/qwen_vl_interface': 8.694211057748766e-06, 'learning_rate/action_model': 8.694211057748765e-05, 'epoch': 0.19})
+[2026-07-28 03:46:56,351][starVLA.training.train_starvla][INFO] - Step 1044, Loss: {'train/loss': 0.06817471981048584, 'train/grad_norm_pre_clip': 1.7118914127349854, 'timing/data': 0.00024168938398361206, 'timing/model': 0.6510592889972031, 'learning_rate/qwen_vl_interface': 8.69157497265989e-06, 'learning_rate/action_model': 8.69157497265989e-05, 'epoch': 0.19})
+[2026-07-28 03:47:02,206][starVLA.training.train_starvla][INFO] - Step 1045, Loss: {'train/loss': 0.09366622567176819, 'train/grad_norm_pre_clip': 1.3533810377120972, 'timing/data': 0.0002457480877637863, 'timing/model': 0.414515511598438, 'learning_rate/qwen_vl_interface': 8.68893665437038e-06, 'learning_rate/action_model': 8.688936654370378e-05, 'epoch': 0.19})
+[2026-07-28 03:47:08,038][starVLA.training.train_starvla][INFO] - Step 1046, Loss: {'train/loss': 0.0609012246131897, 'train/grad_norm_pre_clip': 1.1561418771743774, 'timing/data': 0.00027387915179133415, 'timing/model': 0.41673360811546445, 'learning_rate/qwen_vl_interface': 8.68629610459221e-06, 'learning_rate/action_model': 8.686296104592209e-05, 'epoch': 0.19})
+[2026-07-28 03:47:14,126][starVLA.training.train_starvla][INFO] - Step 1047, Loss: {'train/loss': 0.05442637950181961, 'train/grad_norm_pre_clip': 1.0343327522277832, 'timing/data': 0.00025641173124313354, 'timing/model': 0.41597465611994267, 'learning_rate/qwen_vl_interface': 8.683653325038805e-06, 'learning_rate/action_model': 8.683653325038804e-05, 'epoch': 0.19})
+[2026-07-28 03:47:19,987][starVLA.training.train_starvla][INFO] - Step 1048, Loss: {'train/loss': 0.12464253604412079, 'train/grad_norm_pre_clip': 3.0457239151000977, 'timing/data': 0.0002756151370704174, 'timing/model': 0.4151638993062079, 'learning_rate/qwen_vl_interface': 8.681008317425037e-06, 'learning_rate/action_model': 8.681008317425037e-05, 'epoch': 0.19})
+[2026-07-28 03:47:25,854][starVLA.training.train_starvla][INFO] - Step 1049, Loss: {'train/loss': 0.07459724694490433, 'train/grad_norm_pre_clip': 1.6002984046936035, 'timing/data': 0.00027863215655088425, 'timing/model': 0.41550279688090086, 'learning_rate/qwen_vl_interface': 8.678361083467221e-06, 'learning_rate/action_model': 8.67836108346722e-05, 'epoch': 0.19})
+[2026-07-28 03:47:31,874][starVLA.training.train_starvla][INFO] - Step 1050, Loss: {'train/loss': 0.05559055134654045, 'train/grad_norm_pre_clip': 1.233773946762085, 'timing/data': 0.0003748149611055851, 'timing/model': 0.5871348157525063, 'learning_rate/qwen_vl_interface': 8.67571162488312e-06, 'learning_rate/action_model': 8.67571162488312e-05, 'epoch': 0.19})
+[2026-07-28 03:47:37,965][starVLA.training.train_starvla][INFO] - Step 1051, Loss: {'train/loss': 0.053448744118213654, 'train/grad_norm_pre_clip': 1.5933140516281128, 'timing/data': 0.0002444586716592312, 'timing/model': 0.4145082919858396, 'learning_rate/qwen_vl_interface': 8.673059943391937e-06, 'learning_rate/action_model': 8.673059943391936e-05, 'epoch': 0.19})
+[2026-07-28 03:47:43,826][starVLA.training.train_starvla][INFO] - Step 1052, Loss: {'train/loss': 0.06888794153928757, 'train/grad_norm_pre_clip': 1.857395887374878, 'timing/data': 0.0002953396178781986, 'timing/model': 0.4166276780888438, 'learning_rate/qwen_vl_interface': 8.670406040714321e-06, 'learning_rate/action_model': 8.670406040714321e-05, 'epoch': 0.19})
+[2026-07-28 03:47:49,697][starVLA.training.train_starvla][INFO] - Step 1053, Loss: {'train/loss': 0.06351889669895172, 'train/grad_norm_pre_clip': 1.856369972229004, 'timing/data': 0.00027895718812942505, 'timing/model': 0.41776186507195234, 'learning_rate/qwen_vl_interface': 8.667749918572361e-06, 'learning_rate/action_model': 8.667749918572361e-05, 'epoch': 0.19})
+[2026-07-28 03:47:55,794][starVLA.training.train_starvla][INFO] - Step 1054, Loss: {'train/loss': 0.06369893252849579, 'train/grad_norm_pre_clip': 1.0698144435882568, 'timing/data': 0.000276132021099329, 'timing/model': 0.4167983029037714, 'learning_rate/qwen_vl_interface': 8.665091578689584e-06, 'learning_rate/action_model': 8.665091578689584e-05, 'epoch': 0.19})
+[2026-07-28 03:48:01,898][starVLA.training.train_starvla][INFO] - Step 1055, Loss: {'train/loss': 0.08216744661331177, 'train/grad_norm_pre_clip': 1.451377511024475, 'timing/data': 0.0002766791731119156, 'timing/model': 0.654239097610116, 'learning_rate/qwen_vl_interface': 8.662431022790958e-06, 'learning_rate/action_model': 8.662431022790958e-05, 'epoch': 0.19})
+[2026-07-28 03:48:07,757][starVLA.training.train_starvla][INFO] - Step 1056, Loss: {'train/loss': 0.06518819183111191, 'train/grad_norm_pre_clip': 2.311173677444458, 'timing/data': 0.0002782689407467842, 'timing/model': 0.41757636005058885, 'learning_rate/qwen_vl_interface': 8.65976825260289e-06, 'learning_rate/action_model': 8.659768252602888e-05, 'epoch': 0.19})
+[2026-07-28 03:48:13,592][starVLA.training.train_starvla][INFO] - Step 1057, Loss: {'train/loss': 0.09496389329433441, 'train/grad_norm_pre_clip': 1.902683973312378, 'timing/data': 0.0004444969817996025, 'timing/model': 0.4162025470286608, 'learning_rate/qwen_vl_interface': 8.657103269853222e-06, 'learning_rate/action_model': 8.657103269853222e-05, 'epoch': 0.19})
+[2026-07-28 03:48:19,680][starVLA.training.train_starvla][INFO] - Step 1058, Loss: {'train/loss': 0.062375109642744064, 'train/grad_norm_pre_clip': 1.2458630800247192, 'timing/data': 0.00025677913799881935, 'timing/model': 0.41397777711972594, 'learning_rate/qwen_vl_interface': 8.654436076271232e-06, 'learning_rate/action_model': 8.654436076271232e-05, 'epoch': 0.19})
+[2026-07-28 03:48:25,755][starVLA.training.train_starvla][INFO] - Step 1059, Loss: {'train/loss': 0.05671514570713043, 'train/grad_norm_pre_clip': 2.431302070617676, 'timing/data': 0.0002552610822021961, 'timing/model': 0.4130039499141276, 'learning_rate/qwen_vl_interface': 8.651766673587635e-06, 'learning_rate/action_model': 8.651766673587635e-05, 'epoch': 0.19})
+[2026-07-28 03:48:31,401][starVLA.training.train_starvla][INFO] - Step 1060, Loss: {'train/loss': 0.06941305100917816, 'train/grad_norm_pre_clip': 1.711679458618164, 'timing/data': 0.0003648200072348118, 'timing/model': 0.4143735272809863, 'learning_rate/qwen_vl_interface': 8.649095063534576e-06, 'learning_rate/action_model': 8.649095063534576e-05, 'epoch': 0.19})
+[2026-07-28 03:48:37,268][starVLA.training.train_starvla][INFO] - Step 1061, Loss: {'train/loss': 0.06603404879570007, 'train/grad_norm_pre_clip': 1.979347586631775, 'timing/data': 0.000276319682598114, 'timing/model': 0.4166590999811888, 'learning_rate/qwen_vl_interface': 8.646421247845633e-06, 'learning_rate/action_model': 8.646421247845633e-05, 'epoch': 0.19})
+[2026-07-28 03:48:43,216][starVLA.training.train_starvla][INFO] - Step 1062, Loss: {'train/loss': 0.039202168583869934, 'train/grad_norm_pre_clip': 2.4805290699005127, 'timing/data': 0.0003400011919438839, 'timing/model': 0.41974667785689235, 'learning_rate/qwen_vl_interface': 8.643745228255816e-06, 'learning_rate/action_model': 8.643745228255817e-05, 'epoch': 0.2})
+[2026-07-28 03:48:49,566][starVLA.training.train_starvla][INFO] - Step 1063, Loss: {'train/loss': 0.04867782071232796, 'train/grad_norm_pre_clip': 1.4874566793441772, 'timing/data': 0.00026121921837329865, 'timing/model': 0.646270208992064, 'learning_rate/qwen_vl_interface': 8.641067006501568e-06, 'learning_rate/action_model': 8.641067006501567e-05, 'epoch': 0.2})
+[2026-07-28 03:48:55,486][starVLA.training.train_starvla][INFO] - Step 1064, Loss: {'train/loss': 0.08070875704288483, 'train/grad_norm_pre_clip': 2.080556631088257, 'timing/data': 0.0002489890903234482, 'timing/model': 0.4145231293514371, 'learning_rate/qwen_vl_interface': 8.638386584320753e-06, 'learning_rate/action_model': 8.638386584320752e-05, 'epoch': 0.2})
+[2026-07-28 03:49:01,445][starVLA.training.train_starvla][INFO] - Step 1065, Loss: {'train/loss': 0.07364299148321152, 'train/grad_norm_pre_clip': 2.19366717338562, 'timing/data': 0.0002754777669906616, 'timing/model': 0.418916838709265, 'learning_rate/qwen_vl_interface': 8.635703963452673e-06, 'learning_rate/action_model': 8.635703963452673e-05, 'epoch': 0.2})
+[2026-07-28 03:49:07,506][starVLA.training.train_starvla][INFO] - Step 1066, Loss: {'train/loss': 0.05869578942656517, 'train/grad_norm_pre_clip': 1.8727390766143799, 'timing/data': 0.0002750619314610958, 'timing/model': 0.6544962297193706, 'learning_rate/qwen_vl_interface': 8.633019145638048e-06, 'learning_rate/action_model': 8.633019145638048e-05, 'epoch': 0.2})
+[2026-07-28 03:49:13,331][starVLA.training.train_starvla][INFO] - Step 1067, Loss: {'train/loss': 0.05626901984214783, 'train/grad_norm_pre_clip': 36.38765335083008, 'timing/data': 0.000287455040961504, 'timing/model': 0.4161854349076748, 'learning_rate/qwen_vl_interface': 8.630332132619029e-06, 'learning_rate/action_model': 8.630332132619028e-05, 'epoch': 0.2})
+[2026-07-28 03:49:19,549][starVLA.training.train_starvla][INFO] - Step 1068, Loss: {'train/loss': 0.05892579257488251, 'train/grad_norm_pre_clip': 1.751897931098938, 'timing/data': 0.00026720669120550156, 'timing/model': 0.4140451718121767, 'learning_rate/qwen_vl_interface': 8.627642926139187e-06, 'learning_rate/action_model': 8.627642926139187e-05, 'epoch': 0.2})
+[2026-07-28 03:49:25,429][starVLA.training.train_starvla][INFO] - Step 1069, Loss: {'train/loss': 0.04925539717078209, 'train/grad_norm_pre_clip': 1.2697618007659912, 'timing/data': 0.00026390329003334045, 'timing/model': 0.4157006270252168, 'learning_rate/qwen_vl_interface': 8.624951527943522e-06, 'learning_rate/action_model': 8.624951527943522e-05, 'epoch': 0.2})
+[2026-07-28 03:49:31,278][starVLA.training.train_starvla][INFO] - Step 1070, Loss: {'train/loss': 0.06688760221004486, 'train/grad_norm_pre_clip': 1.7440400123596191, 'timing/data': 0.00028578704223036766, 'timing/model': 0.4160763151012361, 'learning_rate/qwen_vl_interface': 8.622257939778453e-06, 'learning_rate/action_model': 8.622257939778453e-05, 'epoch': 0.2})
+[2026-07-28 03:49:37,599][starVLA.training.train_starvla][INFO] - Step 1071, Loss: {'train/loss': 0.04999283328652382, 'train/grad_norm_pre_clip': 1.4262694120407104, 'timing/data': 0.00025198841467499733, 'timing/model': 0.647343086078763, 'learning_rate/qwen_vl_interface': 8.619562163391818e-06, 'learning_rate/action_model': 8.619562163391817e-05, 'epoch': 0.2})
+[2026-07-28 03:49:43,544][starVLA.training.train_starvla][INFO] - Step 1072, Loss: {'train/loss': 0.05788770690560341, 'train/grad_norm_pre_clip': 2.6335370540618896, 'timing/data': 0.000261728186160326, 'timing/model': 0.4134442778304219, 'learning_rate/qwen_vl_interface': 8.616864200532877e-06, 'learning_rate/action_model': 8.616864200532877e-05, 'epoch': 0.2})
+[2026-07-28 03:49:49,384][starVLA.training.train_starvla][INFO] - Step 1073, Loss: {'train/loss': 0.05145134776830673, 'train/grad_norm_pre_clip': 2.5683953762054443, 'timing/data': 0.00026875315234065056, 'timing/model': 0.4161846050992608, 'learning_rate/qwen_vl_interface': 8.61416405295231e-06, 'learning_rate/action_model': 8.61416405295231e-05, 'epoch': 0.2})
+[2026-07-28 03:49:55,221][starVLA.training.train_starvla][INFO] - Step 1074, Loss: {'train/loss': 0.0617157518863678, 'train/grad_norm_pre_clip': 1.7773487567901611, 'timing/data': 0.0002792011946439743, 'timing/model': 0.4150248351506889, 'learning_rate/qwen_vl_interface': 8.611461722402214e-06, 'learning_rate/action_model': 8.611461722402215e-05, 'epoch': 0.2})
+[2026-07-28 03:50:01,312][starVLA.training.train_starvla][INFO] - Step 1075, Loss: {'train/loss': 0.08617660403251648, 'train/grad_norm_pre_clip': 1.9113056659698486, 'timing/data': 0.0002681487239897251, 'timing/model': 0.6514036790467799, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.2})
+[2026-07-28 03:50:07,200][starVLA.training.train_starvla][INFO] - Step 1076, Loss: {'train/loss': 0.07786194980144501, 'train/grad_norm_pre_clip': 1.5548512935638428, 'timing/data': 0.0002647950313985348, 'timing/model': 0.41425955994054675, 'learning_rate/qwen_vl_interface': 8.606050519408901e-06, 'learning_rate/action_model': 8.606050519408901e-05, 'epoch': 0.2})
+[2026-07-28 03:50:13,283][starVLA.training.train_starvla][INFO] - Step 1077, Loss: {'train/loss': 0.055769383907318115, 'train/grad_norm_pre_clip': 1.0496723651885986, 'timing/data': 0.00026114657521247864, 'timing/model': 0.6466819969937205, 'learning_rate/qwen_vl_interface': 8.603341650476953e-06, 'learning_rate/action_model': 8.603341650476952e-05, 'epoch': 0.2})
+[2026-07-28 03:50:19,139][starVLA.training.train_starvla][INFO] - Step 1078, Loss: {'train/loss': 0.0769568383693695, 'train/grad_norm_pre_clip': 3.8670101165771484, 'timing/data': 0.000257197767496109, 'timing/model': 0.4126711809076369, 'learning_rate/qwen_vl_interface': 8.600630605598016e-06, 'learning_rate/action_model': 8.600630605598015e-05, 'epoch': 0.2})
+[2026-07-28 03:50:24,999][starVLA.training.train_starvla][INFO] - Step 1079, Loss: {'train/loss': 0.18193542957305908, 'train/grad_norm_pre_clip': 5.810898780822754, 'timing/data': 0.0002906392328441143, 'timing/model': 0.41815102891996503, 'learning_rate/qwen_vl_interface': 8.597917386531255e-06, 'learning_rate/action_model': 8.597917386531254e-05, 'epoch': 0.2})
+[2026-07-28 03:50:30,860][starVLA.training.train_starvla][INFO] - Step 1080, Loss: {'train/loss': 0.08108364045619965, 'train/grad_norm_pre_clip': 2.4982850551605225, 'timing/data': 0.00027633272111415863, 'timing/model': 0.4166525606997311, 'learning_rate/qwen_vl_interface': 8.595201995037252e-06, 'learning_rate/action_model': 8.595201995037251e-05, 'epoch': 0.2})
+[2026-07-28 03:50:36,996][starVLA.training.train_starvla][INFO] - Step 1081, Loss: {'train/loss': 0.05595667287707329, 'train/grad_norm_pre_clip': 1.4030365943908691, 'timing/data': 0.00025887414813041687, 'timing/model': 0.413557481020689, 'learning_rate/qwen_vl_interface': 8.59248443287799e-06, 'learning_rate/action_model': 8.592484432877989e-05, 'epoch': 0.2})
+[2026-07-28 03:50:43,138][starVLA.training.train_starvla][INFO] - Step 1082, Loss: {'train/loss': 0.057301945984363556, 'train/grad_norm_pre_clip': 1.9236737489700317, 'timing/data': 0.0002586878836154938, 'timing/model': 0.42115592677146196, 'learning_rate/qwen_vl_interface': 8.589764701816868e-06, 'learning_rate/action_model': 8.589764701816868e-05, 'epoch': 0.2})
+[2026-07-28 03:50:49,004][starVLA.training.train_starvla][INFO] - Step 1083, Loss: {'train/loss': 0.06396506726741791, 'train/grad_norm_pre_clip': 1.6890748739242554, 'timing/data': 0.0002723592333495617, 'timing/model': 0.4158489932306111, 'learning_rate/qwen_vl_interface': 8.587042803618693e-06, 'learning_rate/action_model': 8.587042803618693e-05, 'epoch': 0.2})
+[2026-07-28 03:50:54,884][starVLA.training.train_starvla][INFO] - Step 1084, Loss: {'train/loss': 0.032861195504665375, 'train/grad_norm_pre_clip': 1.5318124294281006, 'timing/data': 0.00027053896337747574, 'timing/model': 0.4175464669242501, 'learning_rate/qwen_vl_interface': 8.584318740049672e-06, 'learning_rate/action_model': 8.584318740049672e-05, 'epoch': 0.2})
+[2026-07-28 03:51:00,985][starVLA.training.train_starvla][INFO] - Step 1085, Loss: {'train/loss': 0.049212440848350525, 'train/grad_norm_pre_clip': 1.630801796913147, 'timing/data': 0.00027600303292274475, 'timing/model': 0.41736137494444847, 'learning_rate/qwen_vl_interface': 8.581592512877419e-06, 'learning_rate/action_model': 8.581592512877418e-05, 'epoch': 0.2})
+[2026-07-28 03:51:07,070][starVLA.training.train_starvla][INFO] - Step 1086, Loss: {'train/loss': 0.061640188097953796, 'train/grad_norm_pre_clip': 1.4295852184295654, 'timing/data': 0.00025103939697146416, 'timing/model': 0.41405819496139884, 'learning_rate/qwen_vl_interface': 8.578864123870955e-06, 'learning_rate/action_model': 8.578864123870955e-05, 'epoch': 0.2})
+[2026-07-28 03:51:12,931][starVLA.training.train_starvla][INFO] - Step 1087, Loss: {'train/loss': 0.03419627994298935, 'train/grad_norm_pre_clip': 0.6658267378807068, 'timing/data': 0.00028113601729273796, 'timing/model': 0.4171830671839416, 'learning_rate/qwen_vl_interface': 8.576133574800703e-06, 'learning_rate/action_model': 8.576133574800703e-05, 'epoch': 0.2})
+[2026-07-28 03:51:19,010][starVLA.training.train_starvla][INFO] - Step 1088, Loss: {'train/loss': 0.2042204737663269, 'train/grad_norm_pre_clip': 3.735356330871582, 'timing/data': 0.0002464139834046364, 'timing/model': 0.4150092899799347, 'learning_rate/qwen_vl_interface': 8.573400867438486e-06, 'learning_rate/action_model': 8.573400867438486e-05, 'epoch': 0.2})
+[2026-07-28 03:51:24,870][starVLA.training.train_starvla][INFO] - Step 1089, Loss: {'train/loss': 0.1345718502998352, 'train/grad_norm_pre_clip': 4.174111366271973, 'timing/data': 0.00027335574850440025, 'timing/model': 0.4186158822849393, 'learning_rate/qwen_vl_interface': 8.570666003557528e-06, 'learning_rate/action_model': 8.570666003557528e-05, 'epoch': 0.2})
+[2026-07-28 03:51:30,720][starVLA.training.train_starvla][INFO] - Step 1090, Loss: {'train/loss': 0.09659601747989655, 'train/grad_norm_pre_clip': 3.1869685649871826, 'timing/data': 0.0002717040479183197, 'timing/model': 0.41525293001905084, 'learning_rate/qwen_vl_interface': 8.567928984932452e-06, 'learning_rate/action_model': 8.567928984932452e-05, 'epoch': 0.2})
+[2026-07-28 03:51:36,956][starVLA.training.train_starvla][INFO] - Step 1091, Loss: {'train/loss': 0.06663290411233902, 'train/grad_norm_pre_clip': 3.1774697303771973, 'timing/data': 0.00027689896523952484, 'timing/model': 0.8187622590921819, 'learning_rate/qwen_vl_interface': 8.565189813339278e-06, 'learning_rate/action_model': 8.565189813339278e-05, 'epoch': 0.2})
+[2026-07-28 03:51:42,790][starVLA.training.train_starvla][INFO] - Step 1092, Loss: {'train/loss': 0.12769728899002075, 'train/grad_norm_pre_clip': 3.1858224868774414, 'timing/data': 0.00024859514087438583, 'timing/model': 0.41327947098761797, 'learning_rate/qwen_vl_interface': 8.562448490555429e-06, 'learning_rate/action_model': 8.562448490555429e-05, 'epoch': 0.2})
+[2026-07-28 03:51:48,614][starVLA.training.train_starvla][INFO] - Step 1093, Loss: {'train/loss': 0.10537756979465485, 'train/grad_norm_pre_clip': 2.2880702018737793, 'timing/data': 0.0002757599577307701, 'timing/model': 0.4153009410947561, 'learning_rate/qwen_vl_interface': 8.559705018359716e-06, 'learning_rate/action_model': 8.559705018359716e-05, 'epoch': 0.2})
+[2026-07-28 03:51:54,476][starVLA.training.train_starvla][INFO] - Step 1094, Loss: {'train/loss': 0.10440883040428162, 'train/grad_norm_pre_clip': 4.757055282592773, 'timing/data': 0.0002752956934273243, 'timing/model': 0.4184563299641013, 'learning_rate/qwen_vl_interface': 8.55695939853235e-06, 'learning_rate/action_model': 8.55695939853235e-05, 'epoch': 0.2})
+[2026-07-28 03:52:00,566][starVLA.training.train_starvla][INFO] - Step 1095, Loss: {'train/loss': 0.03596510738134384, 'train/grad_norm_pre_clip': 1.7083792686462402, 'timing/data': 0.00028321193531155586, 'timing/model': 0.4170172228477895, 'learning_rate/qwen_vl_interface': 8.554211632854932e-06, 'learning_rate/action_model': 8.554211632854932e-05, 'epoch': 0.2})
+[2026-07-28 03:52:06,448][starVLA.training.train_starvla][INFO] - Step 1096, Loss: {'train/loss': 0.039340950548648834, 'train/grad_norm_pre_clip': 1.9992612600326538, 'timing/data': 0.00028256792575120926, 'timing/model': 0.41766587318852544, 'learning_rate/qwen_vl_interface': 8.551461723110461e-06, 'learning_rate/action_model': 8.551461723110461e-05, 'epoch': 0.2})
+[2026-07-28 03:52:12,311][starVLA.training.train_starvla][INFO] - Step 1097, Loss: {'train/loss': 0.09726819396018982, 'train/grad_norm_pre_clip': 2.1541473865509033, 'timing/data': 0.00028431881219148636, 'timing/model': 0.4168037539348006, 'learning_rate/qwen_vl_interface': 8.548709671083321e-06, 'learning_rate/action_model': 8.54870967108332e-05, 'epoch': 0.2})
+[2026-07-28 03:52:18,384][starVLA.training.train_starvla][INFO] - Step 1098, Loss: {'train/loss': 0.12519881129264832, 'train/grad_norm_pre_clip': 1.8761335611343384, 'timing/data': 0.0002549341879785061, 'timing/model': 0.4144629458896816, 'learning_rate/qwen_vl_interface': 8.545955478559288e-06, 'learning_rate/action_model': 8.545955478559288e-05, 'epoch': 0.2})
+[2026-07-28 03:52:24,486][starVLA.training.train_starvla][INFO] - Step 1099, Loss: {'train/loss': 0.07239316403865814, 'train/grad_norm_pre_clip': 2.4861185550689697, 'timing/data': 0.0002516149543225765, 'timing/model': 0.41520300414413214, 'learning_rate/qwen_vl_interface': 8.54319914732553e-06, 'learning_rate/action_model': 8.543199147325529e-05, 'epoch': 0.2})
+[2026-07-28 03:52:30,355][starVLA.training.train_starvla][INFO] - Step 1100, Loss: {'train/loss': 0.06342151015996933, 'train/grad_norm_pre_clip': 2.485405206680298, 'timing/data': 0.0002806992270052433, 'timing/model': 0.4176912810653448, 'learning_rate/qwen_vl_interface': 8.540440679170599e-06, 'learning_rate/action_model': 8.540440679170598e-05, 'epoch': 0.2})
+[2026-07-28 03:52:36,232][starVLA.training.train_starvla][INFO] - Step 1101, Loss: {'train/loss': 0.05607295781373978, 'train/grad_norm_pre_clip': 3.0617361068725586, 'timing/data': 0.000276144128292799, 'timing/model': 0.41797640221193433, 'learning_rate/qwen_vl_interface': 8.537680075884433e-06, 'learning_rate/action_model': 8.537680075884433e-05, 'epoch': 0.2})
+[2026-07-28 03:52:42,667][starVLA.training.train_starvla][INFO] - Step 1102, Loss: {'train/loss': 0.09596128761768341, 'train/grad_norm_pre_clip': 2.6592044830322266, 'timing/data': 0.00028416095301508904, 'timing/model': 0.6931351628154516, 'learning_rate/qwen_vl_interface': 8.53491733925836e-06, 'learning_rate/action_model': 8.53491733925836e-05, 'epoch': 0.2})
+[2026-07-28 03:52:48,884][starVLA.training.train_starvla][INFO] - Step 1103, Loss: {'train/loss': 0.07504196465015411, 'train/grad_norm_pre_clip': 3.138678789138794, 'timing/data': 0.00028419308364391327, 'timing/model': 0.4788434146903455, 'learning_rate/qwen_vl_interface': 8.532152471085092e-06, 'learning_rate/action_model': 8.53215247108509e-05, 'epoch': 0.2})
+[2026-07-28 03:52:55,134][starVLA.training.train_starvla][INFO] - Step 1104, Loss: {'train/loss': 0.05868241935968399, 'train/grad_norm_pre_clip': 1.6054294109344482, 'timing/data': 0.0003894893452525139, 'timing/model': 0.6547213448211551, 'learning_rate/qwen_vl_interface': 8.529385473158717e-06, 'learning_rate/action_model': 8.529385473158717e-05, 'epoch': 0.2})
+[2026-07-28 03:53:01,101][starVLA.training.train_starvla][INFO] - Step 1105, Loss: {'train/loss': 0.04658932238817215, 'train/grad_norm_pre_clip': 1.0976605415344238, 'timing/data': 0.0002706330269575119, 'timing/model': 0.414560557808727, 'learning_rate/qwen_vl_interface': 8.526616347274712e-06, 'learning_rate/action_model': 8.526616347274712e-05, 'epoch': 0.2})
+[2026-07-28 03:53:07,279][starVLA.training.train_starvla][INFO] - Step 1106, Loss: {'train/loss': 0.1075412780046463, 'train/grad_norm_pre_clip': 2.1362128257751465, 'timing/data': 0.0003690277226269245, 'timing/model': 0.4158145301043987, 'learning_rate/qwen_vl_interface': 8.523845095229935e-06, 'learning_rate/action_model': 8.523845095229935e-05, 'epoch': 0.2})
+[2026-07-28 03:53:13,206][starVLA.training.train_starvla][INFO] - Step 1107, Loss: {'train/loss': 0.08098186552524567, 'train/grad_norm_pre_clip': 1.6458888053894043, 'timing/data': 0.00035625509917736053, 'timing/model': 0.4164887988008559, 'learning_rate/qwen_vl_interface': 8.521071718822616e-06, 'learning_rate/action_model': 8.521071718822616e-05, 'epoch': 0.2})
+[2026-07-28 03:53:19,060][starVLA.training.train_starvla][INFO] - Step 1108, Loss: {'train/loss': 0.030132662504911423, 'train/grad_norm_pre_clip': 1.8439136743545532, 'timing/data': 0.0003065960481762886, 'timing/model': 0.4226116971112788, 'learning_rate/qwen_vl_interface': 8.518296219852375e-06, 'learning_rate/action_model': 8.518296219852375e-05, 'epoch': 0.2})
+[2026-07-28 03:53:25,180][starVLA.training.train_starvla][INFO] - Step 1109, Loss: {'train/loss': 0.046942535787820816, 'train/grad_norm_pre_clip': 1.1059885025024414, 'timing/data': 0.0002586580812931061, 'timing/model': 0.4130364619195461, 'learning_rate/qwen_vl_interface': 8.515518600120201e-06, 'learning_rate/action_model': 8.5155186001202e-05, 'epoch': 0.2})
+[2026-07-28 03:53:31,270][starVLA.training.train_starvla][INFO] - Step 1110, Loss: {'train/loss': 0.05448801815509796, 'train/grad_norm_pre_clip': 2.3051207065582275, 'timing/data': 0.00026636524125933647, 'timing/model': 0.4137474070303142, 'learning_rate/qwen_vl_interface': 8.512738861428458e-06, 'learning_rate/action_model': 8.512738861428458e-05, 'epoch': 0.2})
+[2026-07-28 03:53:37,133][starVLA.training.train_starvla][INFO] - Step 1111, Loss: {'train/loss': 0.06674505770206451, 'train/grad_norm_pre_clip': 1.8590606451034546, 'timing/data': 0.0003435746766626835, 'timing/model': 0.4172107339836657, 'learning_rate/qwen_vl_interface': 8.509957005580892e-06, 'learning_rate/action_model': 8.509957005580892e-05, 'epoch': 0.2})
+[2026-07-28 03:53:43,200][starVLA.training.train_starvla][INFO] - Step 1112, Loss: {'train/loss': 0.09288956224918365, 'train/grad_norm_pre_clip': 2.573932647705078, 'timing/data': 0.00025381287559866905, 'timing/model': 0.4119698340073228, 'learning_rate/qwen_vl_interface': 8.507173034382617e-06, 'learning_rate/action_model': 8.507173034382617e-05, 'epoch': 0.2})
+[2026-07-28 03:53:49,382][starVLA.training.train_starvla][INFO] - Step 1113, Loss: {'train/loss': 0.04710584133863449, 'train/grad_norm_pre_clip': 0.9349973201751709, 'timing/data': 0.00031550321727991104, 'timing/model': 0.7247411250136793, 'learning_rate/qwen_vl_interface': 8.504386949640124e-06, 'learning_rate/action_model': 8.504386949640124e-05, 'epoch': 0.2})
+[2026-07-28 03:53:55,270][starVLA.training.train_starvla][INFO] - Step 1114, Loss: {'train/loss': 0.07735682278871536, 'train/grad_norm_pre_clip': 1.6736435890197754, 'timing/data': 0.0003390810452401638, 'timing/model': 0.41766789415851235, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.2})
+[2026-07-28 03:54:01,170][starVLA.training.train_starvla][INFO] - Step 1115, Loss: {'train/loss': 0.07847296446561813, 'train/grad_norm_pre_clip': 2.1543209552764893, 'timing/data': 0.0003761737607419491, 'timing/model': 0.4268856467679143, 'learning_rate/qwen_vl_interface': 8.498808446755291e-06, 'learning_rate/action_model': 8.498808446755291e-05, 'epoch': 0.2})
+[2026-07-28 03:54:07,459][starVLA.training.train_starvla][INFO] - Step 1116, Loss: {'train/loss': 0.05851048231124878, 'train/grad_norm_pre_clip': 2.419445514678955, 'timing/data': 0.00036282604560256004, 'timing/model': 0.6518550021573901, 'learning_rate/qwen_vl_interface': 8.49601603223278e-06, 'learning_rate/action_model': 8.49601603223278e-05, 'epoch': 0.2})
+[2026-07-28 03:54:13,351][starVLA.training.train_starvla][INFO] - Step 1117, Loss: {'train/loss': 0.05555993318557739, 'train/grad_norm_pre_clip': 1.0564614534378052, 'timing/data': 0.00027395598590373993, 'timing/model': 0.41753026004880667, 'learning_rate/qwen_vl_interface': 8.493221511405708e-06, 'learning_rate/action_model': 8.493221511405707e-05, 'epoch': 0.21})
+[2026-07-28 03:54:19,495][starVLA.training.train_starvla][INFO] - Step 1118, Loss: {'train/loss': 0.04785224050283432, 'train/grad_norm_pre_clip': 2.0128819942474365, 'timing/data': 0.0004195091314613819, 'timing/model': 0.6892393771559, 'learning_rate/qwen_vl_interface': 8.490424886087407e-06, 'learning_rate/action_model': 8.490424886087407e-05, 'epoch': 0.21})
+[2026-07-28 03:54:25,443][starVLA.training.train_starvla][INFO] - Step 1119, Loss: {'train/loss': 0.04847743362188339, 'train/grad_norm_pre_clip': 1.8772315979003906, 'timing/data': 0.00026386603713035583, 'timing/model': 0.4168548113666475, 'learning_rate/qwen_vl_interface': 8.487626158092578e-06, 'learning_rate/action_model': 8.487626158092579e-05, 'epoch': 0.21})
+[2026-07-28 03:54:31,332][starVLA.training.train_starvla][INFO] - Step 1120, Loss: {'train/loss': 0.07583800703287125, 'train/grad_norm_pre_clip': 1.3519121408462524, 'timing/data': 0.0003718910738825798, 'timing/model': 0.42256471794098616, 'learning_rate/qwen_vl_interface': 8.484825329237287e-06, 'learning_rate/action_model': 8.484825329237287e-05, 'epoch': 0.21})
+[2026-07-28 03:54:37,253][starVLA.training.train_starvla][INFO] - Step 1121, Loss: {'train/loss': 0.07054408639669418, 'train/grad_norm_pre_clip': 1.858963131904602, 'timing/data': 0.0002862359397113323, 'timing/model': 0.4177461643703282, 'learning_rate/qwen_vl_interface': 8.48202240133896e-06, 'learning_rate/action_model': 8.482022401338959e-05, 'epoch': 0.21})
+[2026-07-28 03:54:43,487][starVLA.training.train_starvla][INFO] - Step 1122, Loss: {'train/loss': 0.07648681104183197, 'train/grad_norm_pre_clip': 1.626444935798645, 'timing/data': 0.00035884231328964233, 'timing/model': 0.4716274468228221, 'learning_rate/qwen_vl_interface': 8.479217376216385e-06, 'learning_rate/action_model': 8.479217376216385e-05, 'epoch': 0.21})
+[2026-07-28 03:54:49,653][starVLA.training.train_starvla][INFO] - Step 1123, Loss: {'train/loss': 0.09241064637899399, 'train/grad_norm_pre_clip': 4.558449745178223, 'timing/data': 0.00029040221124887466, 'timing/model': 0.4384703692048788, 'learning_rate/qwen_vl_interface': 8.476410255689714e-06, 'learning_rate/action_model': 8.476410255689713e-05, 'epoch': 0.21})
+[2026-07-28 03:54:55,784][starVLA.training.train_starvla][INFO] - Step 1124, Loss: {'train/loss': 0.07408390939235687, 'train/grad_norm_pre_clip': 1.3521682024002075, 'timing/data': 0.0002743452787399292, 'timing/model': 0.6534215752035379, 'learning_rate/qwen_vl_interface': 8.473601041580457e-06, 'learning_rate/action_model': 8.473601041580457e-05, 'epoch': 0.21})
+[2026-07-28 03:55:01,634][starVLA.training.train_starvla][INFO] - Step 1125, Loss: {'train/loss': 0.07135447859764099, 'train/grad_norm_pre_clip': 2.6704154014587402, 'timing/data': 0.00025672605261206627, 'timing/model': 0.41500859474763274, 'learning_rate/qwen_vl_interface': 8.470789735711483e-06, 'learning_rate/action_model': 8.470789735711483e-05, 'epoch': 0.21})
+[2026-07-28 03:55:07,771][starVLA.training.train_starvla][INFO] - Step 1126, Loss: {'train/loss': 0.06426288187503815, 'train/grad_norm_pre_clip': 0.8035255670547485, 'timing/data': 0.00030715204775333405, 'timing/model': 0.41696733608841896, 'learning_rate/qwen_vl_interface': 8.467976339907015e-06, 'learning_rate/action_model': 8.467976339907016e-05, 'epoch': 0.21})
+[2026-07-28 03:55:13,692][starVLA.training.train_starvla][INFO] - Step 1127, Loss: {'train/loss': 0.0435895211994648, 'train/grad_norm_pre_clip': 1.2600809335708618, 'timing/data': 0.0002704719081521034, 'timing/model': 0.41627225698903203, 'learning_rate/qwen_vl_interface': 8.465160855992639e-06, 'learning_rate/action_model': 8.465160855992638e-05, 'epoch': 0.21})
+[2026-07-28 03:55:19,571][starVLA.training.train_starvla][INFO] - Step 1128, Loss: {'train/loss': 0.04903362691402435, 'train/grad_norm_pre_clip': 1.408608078956604, 'timing/data': 0.000290523748844862, 'timing/model': 0.41775457793846726, 'learning_rate/qwen_vl_interface': 8.462343285795289e-06, 'learning_rate/action_model': 8.462343285795288e-05, 'epoch': 0.21})
+[2026-07-28 03:55:25,672][starVLA.training.train_starvla][INFO] - Step 1129, Loss: {'train/loss': 0.05693955719470978, 'train/grad_norm_pre_clip': 1.6583259105682373, 'timing/data': 0.00026424601674079895, 'timing/model': 0.41722236620262265, 'learning_rate/qwen_vl_interface': 8.459523631143259e-06, 'learning_rate/action_model': 8.459523631143258e-05, 'epoch': 0.21})
+[2026-07-28 03:55:31,552][starVLA.training.train_starvla][INFO] - Step 1130, Loss: {'train/loss': 0.08704348653554916, 'train/grad_norm_pre_clip': 4.904417037963867, 'timing/data': 0.00028487667441368103, 'timing/model': 0.4217738597653806, 'learning_rate/qwen_vl_interface': 8.456701893866188e-06, 'learning_rate/action_model': 8.456701893866187e-05, 'epoch': 0.21})
+[2026-07-28 03:55:37,226][starVLA.training.train_starvla][INFO] - Step 1131, Loss: {'train/loss': 0.0541890487074852, 'train/grad_norm_pre_clip': 1.938467025756836, 'timing/data': 0.00027350476011633873, 'timing/model': 0.42023214418441057, 'learning_rate/qwen_vl_interface': 8.453878075795072e-06, 'learning_rate/action_model': 8.453878075795072e-05, 'epoch': 0.21})
+[2026-07-28 03:55:43,296][starVLA.training.train_starvla][INFO] - Step 1132, Loss: {'train/loss': 0.044891633093357086, 'train/grad_norm_pre_clip': 1.856662631034851, 'timing/data': 0.0003860890865325928, 'timing/model': 0.5863281819038093, 'learning_rate/qwen_vl_interface': 8.451052178762258e-06, 'learning_rate/action_model': 8.451052178762258e-05, 'epoch': 0.21})
+[2026-07-28 03:55:49,398][starVLA.training.train_starvla][INFO] - Step 1133, Loss: {'train/loss': 0.04250228404998779, 'train/grad_norm_pre_clip': 0.9230617880821228, 'timing/data': 0.00025352416560053825, 'timing/model': 0.41664779372513294, 'learning_rate/qwen_vl_interface': 8.448224204601439e-06, 'learning_rate/action_model': 8.448224204601438e-05, 'epoch': 0.21})
+[2026-07-28 03:55:55,257][starVLA.training.train_starvla][INFO] - Step 1134, Loss: {'train/loss': 0.077312171459198, 'train/grad_norm_pre_clip': 1.9454667568206787, 'timing/data': 0.0004026792012155056, 'timing/model': 0.4175810478627682, 'learning_rate/qwen_vl_interface': 8.445394155147655e-06, 'learning_rate/action_model': 8.445394155147653e-05, 'epoch': 0.21})
+[2026-07-28 03:56:01,240][starVLA.training.train_starvla][INFO] - Step 1135, Loss: {'train/loss': 0.05521327257156372, 'train/grad_norm_pre_clip': 1.4090327024459839, 'timing/data': 0.0002766852267086506, 'timing/model': 0.42014492861926556, 'learning_rate/qwen_vl_interface': 8.442562032237295e-06, 'learning_rate/action_model': 8.442562032237296e-05, 'epoch': 0.21})
+[2026-07-28 03:56:07,346][starVLA.training.train_starvla][INFO] - Step 1136, Loss: {'train/loss': 0.07383518666028976, 'train/grad_norm_pre_clip': 2.3845748901367188, 'timing/data': 0.0003529610112309456, 'timing/model': 0.41254745004698634, 'learning_rate/qwen_vl_interface': 8.439727837708094e-06, 'learning_rate/action_model': 8.439727837708095e-05, 'epoch': 0.21})
+[2026-07-28 03:56:13,280][starVLA.training.train_starvla][INFO] - Step 1137, Loss: {'train/loss': 0.09534189105033875, 'train/grad_norm_pre_clip': 2.649055004119873, 'timing/data': 0.00027620280161499977, 'timing/model': 0.42181740887463093, 'learning_rate/qwen_vl_interface': 8.436891573399132e-06, 'learning_rate/action_model': 8.436891573399131e-05, 'epoch': 0.21})
+[2026-07-28 03:56:19,441][starVLA.training.train_starvla][INFO] - Step 1138, Loss: {'train/loss': 0.06577682495117188, 'train/grad_norm_pre_clip': 1.9088523387908936, 'timing/data': 0.0003684568218886852, 'timing/model': 0.6729925656691194, 'learning_rate/qwen_vl_interface': 8.434053241150826e-06, 'learning_rate/action_model': 8.434053241150825e-05, 'epoch': 0.21})
+[2026-07-28 03:56:25,306][starVLA.training.train_starvla][INFO] - Step 1139, Loss: {'train/loss': 0.060292355716228485, 'train/grad_norm_pre_clip': 3.112355947494507, 'timing/data': 0.00026505300775170326, 'timing/model': 0.41562105203047395, 'learning_rate/qwen_vl_interface': 8.43121284280494e-06, 'learning_rate/action_model': 8.431212842804939e-05, 'epoch': 0.21})
+[2026-07-28 03:56:31,462][starVLA.training.train_starvla][INFO] - Step 1140, Loss: {'train/loss': 0.1332738995552063, 'train/grad_norm_pre_clip': 5.809030055999756, 'timing/data': 0.00029724324122071266, 'timing/model': 0.6448960048146546, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.21})
+[2026-07-28 03:56:37,352][starVLA.training.train_starvla][INFO] - Step 1141, Loss: {'train/loss': 0.08861663192510605, 'train/grad_norm_pre_clip': 2.605759382247925, 'timing/data': 0.0002596108242869377, 'timing/model': 0.41632480919361115, 'learning_rate/qwen_vl_interface': 8.425525855194184e-06, 'learning_rate/action_model': 8.425525855194183e-05, 'epoch': 0.21})
+[2026-07-28 03:56:43,247][starVLA.training.train_starvla][INFO] - Step 1142, Loss: {'train/loss': 0.07061181962490082, 'train/grad_norm_pre_clip': 2.029707670211792, 'timing/data': 0.00040119001641869545, 'timing/model': 0.4182196659967303, 'learning_rate/qwen_vl_interface': 8.422679269619538e-06, 'learning_rate/action_model': 8.422679269619538e-05, 'epoch': 0.21})
+[2026-07-28 03:56:49,113][starVLA.training.train_starvla][INFO] - Step 1143, Loss: {'train/loss': 0.08061343431472778, 'train/grad_norm_pre_clip': 2.162447452545166, 'timing/data': 0.00028287293389439583, 'timing/model': 0.4181911153718829, 'learning_rate/qwen_vl_interface': 8.419830625327758e-06, 'learning_rate/action_model': 8.419830625327758e-05, 'epoch': 0.21})
+[2026-07-28 03:56:54,971][starVLA.training.train_starvla][INFO] - Step 1144, Loss: {'train/loss': 0.048326704651117325, 'train/grad_norm_pre_clip': 1.722411036491394, 'timing/data': 0.0002918126992881298, 'timing/model': 0.4157924260944128, 'learning_rate/qwen_vl_interface': 8.416979924167303e-06, 'learning_rate/action_model': 8.416979924167303e-05, 'epoch': 0.21})
+[2026-07-28 03:57:01,038][starVLA.training.train_starvla][INFO] - Step 1145, Loss: {'train/loss': 0.06355266273021698, 'train/grad_norm_pre_clip': 1.5717618465423584, 'timing/data': 0.0002567288465797901, 'timing/model': 0.41388718504458666, 'learning_rate/qwen_vl_interface': 8.414127167987957e-06, 'learning_rate/action_model': 8.414127167987957e-05, 'epoch': 0.21})
+[2026-07-28 03:57:07,146][starVLA.training.train_starvla][INFO] - Step 1146, Loss: {'train/loss': 0.11210182309150696, 'train/grad_norm_pre_clip': 4.652706146240234, 'timing/data': 0.00026219896972179413, 'timing/model': 0.42818547366186976, 'learning_rate/qwen_vl_interface': 8.411272358640843e-06, 'learning_rate/action_model': 8.411272358640844e-05, 'epoch': 0.21})
+[2026-07-28 03:57:13,065][starVLA.training.train_starvla][INFO] - Step 1147, Loss: {'train/loss': 0.054202381521463394, 'train/grad_norm_pre_clip': 1.0440614223480225, 'timing/data': 0.0002835029736161232, 'timing/model': 0.4144083592109382, 'learning_rate/qwen_vl_interface': 8.408415497978418e-06, 'learning_rate/action_model': 8.408415497978418e-05, 'epoch': 0.21})
+[2026-07-28 03:57:18,939][starVLA.training.train_starvla][INFO] - Step 1148, Loss: {'train/loss': 0.07538530230522156, 'train/grad_norm_pre_clip': 0.9076623320579529, 'timing/data': 0.0002809939906001091, 'timing/model': 0.4168112240731716, 'learning_rate/qwen_vl_interface': 8.405556587854468e-06, 'learning_rate/action_model': 8.405556587854467e-05, 'epoch': 0.21})
+[2026-07-28 03:57:25,018][starVLA.training.train_starvla][INFO] - Step 1149, Loss: {'train/loss': 0.04115299880504608, 'train/grad_norm_pre_clip': 1.8906781673431396, 'timing/data': 0.00028119422495365143, 'timing/model': 0.6535456529818475, 'learning_rate/qwen_vl_interface': 8.402695630124106e-06, 'learning_rate/action_model': 8.402695630124106e-05, 'epoch': 0.21})
+[2026-07-28 03:57:31,085][starVLA.training.train_starvla][INFO] - Step 1150, Loss: {'train/loss': 0.0685410127043724, 'train/grad_norm_pre_clip': 2.672501802444458, 'timing/data': 0.00025320379063487053, 'timing/model': 0.4136514849960804, 'learning_rate/qwen_vl_interface': 8.399832626643778e-06, 'learning_rate/action_model': 8.399832626643778e-05, 'epoch': 0.21})
+[2026-07-28 03:57:36,933][starVLA.training.train_starvla][INFO] - Step 1151, Loss: {'train/loss': 0.07368868589401245, 'train/grad_norm_pre_clip': 1.5691519975662231, 'timing/data': 0.0002837930805981159, 'timing/model': 0.4158321749418974, 'learning_rate/qwen_vl_interface': 8.396967579271256e-06, 'learning_rate/action_model': 8.396967579271257e-05, 'epoch': 0.21})
+[2026-07-28 03:57:42,786][starVLA.training.train_starvla][INFO] - Step 1152, Loss: {'train/loss': 0.05961529538035393, 'train/grad_norm_pre_clip': 1.9248664379119873, 'timing/data': 0.00028099631890654564, 'timing/model': 0.41603699466213584, 'learning_rate/qwen_vl_interface': 8.394100489865638e-06, 'learning_rate/action_model': 8.394100489865638e-05, 'epoch': 0.21})
+[2026-07-28 03:57:49,086][starVLA.training.train_starvla][INFO] - Step 1153, Loss: {'train/loss': 0.074787437915802, 'train/grad_norm_pre_clip': 4.339371681213379, 'timing/data': 0.00032863207161426544, 'timing/model': 0.6404683408327401, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.21})
+[2026-07-28 03:57:54,907][starVLA.training.train_starvla][INFO] - Step 1154, Loss: {'train/loss': 0.10459685325622559, 'train/grad_norm_pre_clip': 4.545455455780029, 'timing/data': 0.00028004078194499016, 'timing/model': 0.4161518937908113, 'learning_rate/qwen_vl_interface': 8.388360192398131e-06, 'learning_rate/action_model': 8.38836019239813e-05, 'epoch': 0.21})
+[2026-07-28 03:58:00,557][starVLA.training.train_starvla][INFO] - Step 1155, Loss: {'train/loss': 0.06356054544448853, 'train/grad_norm_pre_clip': 5.003237724304199, 'timing/data': 0.00028176093474030495, 'timing/model': 0.4184625390917063, 'learning_rate/qwen_vl_interface': 8.385486988061058e-06, 'learning_rate/action_model': 8.385486988061058e-05, 'epoch': 0.21})
+[2026-07-28 03:58:06,399][starVLA.training.train_starvla][INFO] - Step 1156, Loss: {'train/loss': 0.05333543196320534, 'train/grad_norm_pre_clip': 1.914495587348938, 'timing/data': 0.0002766796387732029, 'timing/model': 0.41554499976336956, 'learning_rate/qwen_vl_interface': 8.38261174914052e-06, 'learning_rate/action_model': 8.38261174914052e-05, 'epoch': 0.21})
+[2026-07-28 03:58:12,256][starVLA.training.train_starvla][INFO] - Step 1157, Loss: {'train/loss': 0.061705105006694794, 'train/grad_norm_pre_clip': 3.0967838764190674, 'timing/data': 0.00025298725813627243, 'timing/model': 0.4128500809893012, 'learning_rate/qwen_vl_interface': 8.379734477502229e-06, 'learning_rate/action_model': 8.379734477502228e-05, 'epoch': 0.21})
+[2026-07-28 03:58:17,884][starVLA.training.train_starvla][INFO] - Step 1158, Loss: {'train/loss': 0.057175226509571075, 'train/grad_norm_pre_clip': 2.094818353652954, 'timing/data': 0.0002882690168917179, 'timing/model': 0.4167696898803115, 'learning_rate/qwen_vl_interface': 8.376855175013214e-06, 'learning_rate/action_model': 8.376855175013213e-05, 'epoch': 0.21})
+[2026-07-28 03:58:23,765][starVLA.training.train_starvla][INFO] - Step 1159, Loss: {'train/loss': 0.07973835617303848, 'train/grad_norm_pre_clip': 1.9314290285110474, 'timing/data': 0.0003097238950431347, 'timing/model': 0.415134783834219, 'learning_rate/qwen_vl_interface': 8.373973843541822e-06, 'learning_rate/action_model': 8.373973843541822e-05, 'epoch': 0.21})
+[2026-07-28 03:58:29,900][starVLA.training.train_starvla][INFO] - Step 1160, Loss: {'train/loss': 0.07443152368068695, 'train/grad_norm_pre_clip': 2.043480157852173, 'timing/data': 0.0002542482689023018, 'timing/model': 0.4149011867120862, 'learning_rate/qwen_vl_interface': 8.371090484957722e-06, 'learning_rate/action_model': 8.371090484957722e-05, 'epoch': 0.21})
+[2026-07-28 03:58:35,994][starVLA.training.train_starvla][INFO] - Step 1161, Loss: {'train/loss': 0.06071045622229576, 'train/grad_norm_pre_clip': 2.037550687789917, 'timing/data': 0.0002662469632923603, 'timing/model': 0.4148820168338716, 'learning_rate/qwen_vl_interface': 8.368205101131891e-06, 'learning_rate/action_model': 8.368205101131891e-05, 'epoch': 0.21})
+[2026-07-28 03:58:41,832][starVLA.training.train_starvla][INFO] - Step 1162, Loss: {'train/loss': 0.06193099170923233, 'train/grad_norm_pre_clip': 1.543941617012024, 'timing/data': 0.00027266377583146095, 'timing/model': 0.4161830903030932, 'learning_rate/qwen_vl_interface': 8.365317693936624e-06, 'learning_rate/action_model': 8.365317693936623e-05, 'epoch': 0.21})
+[2026-07-28 03:58:47,830][starVLA.training.train_starvla][INFO] - Step 1163, Loss: {'train/loss': 0.03746313601732254, 'train/grad_norm_pre_clip': 1.8884276151657104, 'timing/data': 0.00028143683448433876, 'timing/model': 0.5834122961387038, 'learning_rate/qwen_vl_interface': 8.362428265245526e-06, 'learning_rate/action_model': 8.362428265245526e-05, 'epoch': 0.21})
+[2026-07-28 03:58:53,905][starVLA.training.train_starvla][INFO] - Step 1164, Loss: {'train/loss': 0.0736122578382492, 'train/grad_norm_pre_clip': 1.427762508392334, 'timing/data': 0.00027788709849119186, 'timing/model': 0.6537885800935328, 'learning_rate/qwen_vl_interface': 8.359536816933518e-06, 'learning_rate/action_model': 8.359536816933518e-05, 'epoch': 0.21})
+[2026-07-28 03:58:59,750][starVLA.training.train_starvla][INFO] - Step 1165, Loss: {'train/loss': 0.06881926953792572, 'train/grad_norm_pre_clip': 2.0673468112945557, 'timing/data': 0.0004496891051530838, 'timing/model': 0.41554193710908294, 'learning_rate/qwen_vl_interface': 8.356643350876832e-06, 'learning_rate/action_model': 8.356643350876832e-05, 'epoch': 0.21})
+[2026-07-28 03:59:05,636][starVLA.training.train_starvla][INFO] - Step 1166, Loss: {'train/loss': 0.10747399926185608, 'train/grad_norm_pre_clip': 3.268498420715332, 'timing/data': 0.0002766242250800133, 'timing/model': 0.41597535414621234, 'learning_rate/qwen_vl_interface': 8.353747868953002e-06, 'learning_rate/action_model': 8.353747868953003e-05, 'epoch': 0.21})
+[2026-07-28 03:59:11,731][starVLA.training.train_starvla][INFO] - Step 1167, Loss: {'train/loss': 0.05395563691854477, 'train/grad_norm_pre_clip': 1.7115966081619263, 'timing/data': 0.00026260828599333763, 'timing/model': 0.41533184703439474, 'learning_rate/qwen_vl_interface': 8.350850373040876e-06, 'learning_rate/action_model': 8.350850373040876e-05, 'epoch': 0.21})
+[2026-07-28 03:59:17,977][starVLA.training.train_starvla][INFO] - Step 1168, Loss: {'train/loss': 0.08626154065132141, 'train/grad_norm_pre_clip': 3.7952187061309814, 'timing/data': 0.0002529020421206951, 'timing/model': 0.469575562980026, 'learning_rate/qwen_vl_interface': 8.34795086502061e-06, 'learning_rate/action_model': 8.347950865020609e-05, 'epoch': 0.21})
+[2026-07-28 03:59:23,831][starVLA.training.train_starvla][INFO] - Step 1169, Loss: {'train/loss': 0.05008770897984505, 'train/grad_norm_pre_clip': 1.0320804119110107, 'timing/data': 0.0003885449841618538, 'timing/model': 0.4167870790697634, 'learning_rate/qwen_vl_interface': 8.34504934677366e-06, 'learning_rate/action_model': 8.34504934677366e-05, 'epoch': 0.21})
+[2026-07-28 03:59:29,672][starVLA.training.train_starvla][INFO] - Step 1170, Loss: {'train/loss': 0.05947267264127731, 'train/grad_norm_pre_clip': 1.1746717691421509, 'timing/data': 0.0003626542165875435, 'timing/model': 0.4162317840382457, 'learning_rate/qwen_vl_interface': 8.342145820182795e-06, 'learning_rate/action_model': 8.342145820182793e-05, 'epoch': 0.21})
+[2026-07-28 03:59:35,802][starVLA.training.train_starvla][INFO] - Step 1171, Loss: {'train/loss': 0.05451396480202675, 'train/grad_norm_pre_clip': 1.262088418006897, 'timing/data': 0.0002612229436635971, 'timing/model': 0.41438334388658404, 'learning_rate/qwen_vl_interface': 8.339240287132073e-06, 'learning_rate/action_model': 8.339240287132073e-05, 'epoch': 0.22})
+[2026-07-28 03:59:41,882][starVLA.training.train_starvla][INFO] - Step 1172, Loss: {'train/loss': 0.04908321052789688, 'train/grad_norm_pre_clip': 1.15449059009552, 'timing/data': 0.0002536848187446594, 'timing/model': 0.4130018730647862, 'learning_rate/qwen_vl_interface': 8.336332749506872e-06, 'learning_rate/action_model': 8.336332749506872e-05, 'epoch': 0.22})
+[2026-07-28 03:59:47,733][starVLA.training.train_starvla][INFO] - Step 1173, Loss: {'train/loss': 0.09226205945014954, 'train/grad_norm_pre_clip': 2.917994260787964, 'timing/data': 0.0002816389314830303, 'timing/model': 0.41605659993365407, 'learning_rate/qwen_vl_interface': 8.333423209193857e-06, 'learning_rate/action_model': 8.333423209193856e-05, 'epoch': 0.22})
+[2026-07-28 03:59:53,371][starVLA.training.train_starvla][INFO] - Step 1174, Loss: {'train/loss': 0.06177463382482529, 'train/grad_norm_pre_clip': 1.0176424980163574, 'timing/data': 0.00027245189994573593, 'timing/model': 0.4170770179480314, 'learning_rate/qwen_vl_interface': 8.330511668080997e-06, 'learning_rate/action_model': 8.330511668080997e-05, 'epoch': 0.22})
+[2026-07-28 03:59:59,475][starVLA.training.train_starvla][INFO] - Step 1175, Loss: {'train/loss': 0.056481730192899704, 'train/grad_norm_pre_clip': 2.1971094608306885, 'timing/data': 0.0003356221131980419, 'timing/model': 0.6688079899176955, 'learning_rate/qwen_vl_interface': 8.32759812805756e-06, 'learning_rate/action_model': 8.32759812805756e-05, 'epoch': 0.22})
+[2026-07-28 04:00:05,608][starVLA.training.train_starvla][INFO] - Step 1176, Loss: {'train/loss': 0.054425254464149475, 'train/grad_norm_pre_clip': 1.6213182210922241, 'timing/data': 0.0002630813978612423, 'timing/model': 0.6545268893241882, 'learning_rate/qwen_vl_interface': 8.324682591014111e-06, 'learning_rate/action_model': 8.32468259101411e-05, 'epoch': 0.22})
+[2026-07-28 04:00:11,505][starVLA.training.train_starvla][INFO] - Step 1177, Loss: {'train/loss': 0.07220607995986938, 'train/grad_norm_pre_clip': 1.9855985641479492, 'timing/data': 0.0002552778460085392, 'timing/model': 0.4142019678838551, 'learning_rate/qwen_vl_interface': 8.32176505884251e-06, 'learning_rate/action_model': 8.32176505884251e-05, 'epoch': 0.22})
+[2026-07-28 04:00:17,415][starVLA.training.train_starvla][INFO] - Step 1178, Loss: {'train/loss': 0.035989053547382355, 'train/grad_norm_pre_clip': 1.1886303424835205, 'timing/data': 0.0003388887271285057, 'timing/model': 0.41634146124124527, 'learning_rate/qwen_vl_interface': 8.318845533435913e-06, 'learning_rate/action_model': 8.318845533435913e-05, 'epoch': 0.22})
+[2026-07-28 04:00:23,266][starVLA.training.train_starvla][INFO] - Step 1179, Loss: {'train/loss': 0.04820253327488899, 'train/grad_norm_pre_clip': 1.1794105768203735, 'timing/data': 0.00028673699125647545, 'timing/model': 0.4162090579047799, 'learning_rate/qwen_vl_interface': 8.315924016688767e-06, 'learning_rate/action_model': 8.315924016688766e-05, 'epoch': 0.22})
+[2026-07-28 04:00:29,352][starVLA.training.train_starvla][INFO] - Step 1180, Loss: {'train/loss': 0.03993035852909088, 'train/grad_norm_pre_clip': 1.3023723363876343, 'timing/data': 0.00028037792071700096, 'timing/model': 0.6544473692774773, 'learning_rate/qwen_vl_interface': 8.313000510496811e-06, 'learning_rate/action_model': 8.313000510496811e-05, 'epoch': 0.22})
+[2026-07-28 04:00:35,188][starVLA.training.train_starvla][INFO] - Step 1181, Loss: {'train/loss': 0.08811269700527191, 'train/grad_norm_pre_clip': 1.6155916452407837, 'timing/data': 0.00025756005197763443, 'timing/model': 0.4139837878756225, 'learning_rate/qwen_vl_interface': 8.31007501675708e-06, 'learning_rate/action_model': 8.310075016757079e-05, 'epoch': 0.22})
+[2026-07-28 04:00:41,006][starVLA.training.train_starvla][INFO] - Step 1182, Loss: {'train/loss': 0.04469337314367294, 'train/grad_norm_pre_clip': 0.8700243830680847, 'timing/data': 0.0002823490649461746, 'timing/model': 0.414450213778764, 'learning_rate/qwen_vl_interface': 8.307147537367894e-06, 'learning_rate/action_model': 8.307147537367893e-05, 'epoch': 0.22})
+[2026-07-28 04:00:46,880][starVLA.training.train_starvla][INFO] - Step 1183, Loss: {'train/loss': 0.035908617079257965, 'train/grad_norm_pre_clip': 1.568158745765686, 'timing/data': 0.0003679729998111725, 'timing/model': 0.4167272630147636, 'learning_rate/qwen_vl_interface': 8.30421807422886e-06, 'learning_rate/action_model': 8.30421807422886e-05, 'epoch': 0.22})
+[2026-07-28 04:00:52,959][starVLA.training.train_starvla][INFO] - Step 1184, Loss: {'train/loss': 0.04518125206232071, 'train/grad_norm_pre_clip': 1.5335476398468018, 'timing/data': 0.0002618408761918545, 'timing/model': 0.41668206872418523, 'learning_rate/qwen_vl_interface': 8.301286629240876e-06, 'learning_rate/action_model': 8.301286629240876e-05, 'epoch': 0.22})
+[2026-07-28 04:00:58,919][starVLA.training.train_starvla][INFO] - Step 1185, Loss: {'train/loss': 0.06660369038581848, 'train/grad_norm_pre_clip': 1.682558536529541, 'timing/data': 0.00040351971983909607, 'timing/model': 0.4226172682829201, 'learning_rate/qwen_vl_interface': 8.298353204306127e-06, 'learning_rate/action_model': 8.298353204306127e-05, 'epoch': 0.22})
+[2026-07-28 04:01:05,201][starVLA.training.train_starvla][INFO] - Step 1186, Loss: {'train/loss': 0.08287832885980606, 'train/grad_norm_pre_clip': 1.673640489578247, 'timing/data': 0.000306151807308197, 'timing/model': 0.6553062070161104, 'learning_rate/qwen_vl_interface': 8.295417801328081e-06, 'learning_rate/action_model': 8.29541780132808e-05, 'epoch': 0.22})
+[2026-07-28 04:01:11,058][starVLA.training.train_starvla][INFO] - Step 1187, Loss: {'train/loss': 0.0471525602042675, 'train/grad_norm_pre_clip': 2.1116418838500977, 'timing/data': 0.00034582894295454025, 'timing/model': 0.41416336689144373, 'learning_rate/qwen_vl_interface': 8.292480422211486e-06, 'learning_rate/action_model': 8.292480422211485e-05, 'epoch': 0.22})
+[2026-07-28 04:01:17,148][starVLA.training.train_starvla][INFO] - Step 1188, Loss: {'train/loss': 0.0560171976685524, 'train/grad_norm_pre_clip': 1.8339605331420898, 'timing/data': 0.00028876401484012604, 'timing/model': 0.6484638387337327, 'learning_rate/qwen_vl_interface': 8.289541068862377e-06, 'learning_rate/action_model': 8.289541068862377e-05, 'epoch': 0.22})
+[2026-07-28 04:01:22,997][starVLA.training.train_starvla][INFO] - Step 1189, Loss: {'train/loss': 0.09005527943372726, 'train/grad_norm_pre_clip': 2.662992477416992, 'timing/data': 0.0003169327974319458, 'timing/model': 0.4122959980741143, 'learning_rate/qwen_vl_interface': 8.286599743188068e-06, 'learning_rate/action_model': 8.286599743188069e-05, 'epoch': 0.22})
+[2026-07-28 04:01:28,830][starVLA.training.train_starvla][INFO] - Step 1190, Loss: {'train/loss': 0.07088887691497803, 'train/grad_norm_pre_clip': 2.999617099761963, 'timing/data': 0.00038070185109972954, 'timing/model': 0.41578551614657044, 'learning_rate/qwen_vl_interface': 8.283656447097154e-06, 'learning_rate/action_model': 8.283656447097153e-05, 'epoch': 0.22})
+[2026-07-28 04:01:34,930][starVLA.training.train_starvla][INFO] - Step 1191, Loss: {'train/loss': 0.04414530470967293, 'train/grad_norm_pre_clip': 2.116953134536743, 'timing/data': 0.0002460302785038948, 'timing/model': 0.4158652722835541, 'learning_rate/qwen_vl_interface': 8.280711182499507e-06, 'learning_rate/action_model': 8.280711182499506e-05, 'epoch': 0.22})
+[2026-07-28 04:01:40,761][starVLA.training.train_starvla][INFO] - Step 1192, Loss: {'train/loss': 0.0508151650428772, 'train/grad_norm_pre_clip': 1.8526583909988403, 'timing/data': 0.0005652070976793766, 'timing/model': 0.4172959220595658, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.22})
+[2026-07-28 04:01:46,662][starVLA.training.train_starvla][INFO] - Step 1193, Loss: {'train/loss': 0.0715760886669159, 'train/grad_norm_pre_clip': 1.2009762525558472, 'timing/data': 0.00027688825502991676, 'timing/model': 0.4179713651537895, 'learning_rate/qwen_vl_interface': 8.274814755429892e-06, 'learning_rate/action_model': 8.274814755429891e-05, 'epoch': 0.22})
+[2026-07-28 04:01:52,862][starVLA.training.train_starvla][INFO] - Step 1194, Loss: {'train/loss': 0.06532410532236099, 'train/grad_norm_pre_clip': 1.7499912977218628, 'timing/data': 0.00026589399203658104, 'timing/model': 0.4142147130332887, 'learning_rate/qwen_vl_interface': 8.271863596784053e-06, 'learning_rate/action_model': 8.271863596784051e-05, 'epoch': 0.22})
+[2026-07-28 04:01:59,340][starVLA.training.train_starvla][INFO] - Step 1195, Loss: {'train/loss': 0.04692084342241287, 'train/grad_norm_pre_clip': 1.7334675788879395, 'timing/data': 0.00026005832478404045, 'timing/model': 0.4142712140455842, 'learning_rate/qwen_vl_interface': 8.268910477283732e-06, 'learning_rate/action_model': 8.26891047728373e-05, 'epoch': 0.22})
+[2026-07-28 04:02:05,320][starVLA.training.train_starvla][INFO] - Step 1196, Loss: {'train/loss': 0.038528844714164734, 'train/grad_norm_pre_clip': 1.7617367506027222, 'timing/data': 0.00034005288034677505, 'timing/model': 0.41736356168985367, 'learning_rate/qwen_vl_interface': 8.265955398845177e-06, 'learning_rate/action_model': 8.265955398845177e-05, 'epoch': 0.22})
+[2026-07-28 04:02:11,158][starVLA.training.train_starvla][INFO] - Step 1197, Loss: {'train/loss': 0.093578040599823, 'train/grad_norm_pre_clip': 2.4704413414001465, 'timing/data': 0.00028725387528538704, 'timing/model': 0.4169339989311993, 'learning_rate/qwen_vl_interface': 8.26299836338591e-06, 'learning_rate/action_model': 8.26299836338591e-05, 'epoch': 0.22})
+[2026-07-28 04:02:17,269][starVLA.training.train_starvla][INFO] - Step 1198, Loss: {'train/loss': 0.048755623400211334, 'train/grad_norm_pre_clip': 0.7729883790016174, 'timing/data': 0.0002761818468570709, 'timing/model': 0.41602418292313814, 'learning_rate/qwen_vl_interface': 8.260039372824715e-06, 'learning_rate/action_model': 8.260039372824715e-05, 'epoch': 0.22})
+[2026-07-28 04:02:23,132][starVLA.training.train_starvla][INFO] - Step 1199, Loss: {'train/loss': 0.058400772511959076, 'train/grad_norm_pre_clip': 3.94968581199646, 'timing/data': 0.00029904209077358246, 'timing/model': 0.4168075700290501, 'learning_rate/qwen_vl_interface': 8.257078429081652e-06, 'learning_rate/action_model': 8.257078429081651e-05, 'epoch': 0.22})
+[2026-07-28 04:02:29,293][starVLA.training.train_starvla][INFO] - Step 1200, Loss: {'train/loss': 0.08896102011203766, 'train/grad_norm_pre_clip': 3.6254115104675293, 'timing/data': 0.00028075603768229485, 'timing/model': 0.6612313562072814, 'learning_rate/qwen_vl_interface': 8.254115534078041e-06, 'learning_rate/action_model': 8.254115534078042e-05, 'epoch': 0.22})
+[2026-07-28 04:02:35,385][starVLA.training.train_starvla][INFO] - Step 1201, Loss: {'train/loss': 0.12553724646568298, 'train/grad_norm_pre_clip': 2.456674814224243, 'timing/data': 0.00027813296765089035, 'timing/model': 0.4159967452287674, 'learning_rate/qwen_vl_interface': 8.25115068973648e-06, 'learning_rate/action_model': 8.25115068973648e-05, 'epoch': 0.22})
+[2026-07-28 04:02:41,245][starVLA.training.train_starvla][INFO] - Step 1202, Loss: {'train/loss': 0.060115158557891846, 'train/grad_norm_pre_clip': 1.147925853729248, 'timing/data': 0.00026820413768291473, 'timing/model': 0.416030284948647, 'learning_rate/qwen_vl_interface': 8.248183897980815e-06, 'learning_rate/action_model': 8.248183897980815e-05, 'epoch': 0.22})
+[2026-07-28 04:02:47,134][starVLA.training.train_starvla][INFO] - Step 1203, Loss: {'train/loss': 0.1360015869140625, 'train/grad_norm_pre_clip': 3.048722982406616, 'timing/data': 0.0002778046764433384, 'timing/model': 0.41615720838308334, 'learning_rate/qwen_vl_interface': 8.24521516073617e-06, 'learning_rate/action_model': 8.24521516073617e-05, 'epoch': 0.22})
+[2026-07-28 04:02:53,168][starVLA.training.train_starvla][INFO] - Step 1204, Loss: {'train/loss': 0.0726802870631218, 'train/grad_norm_pre_clip': 1.3959938287734985, 'timing/data': 0.00028304383158683777, 'timing/model': 0.5867878305725753, 'learning_rate/qwen_vl_interface': 8.24224447992893e-06, 'learning_rate/action_model': 8.242244479928928e-05, 'epoch': 0.22})
+[2026-07-28 04:02:59,046][starVLA.training.train_starvla][INFO] - Step 1205, Loss: {'train/loss': 0.05510587990283966, 'train/grad_norm_pre_clip': 1.997887134552002, 'timing/data': 0.0003211800940334797, 'timing/model': 0.41738517628982663, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.22})
+[2026-07-28 04:03:04,928][starVLA.training.train_starvla][INFO] - Step 1206, Loss: {'train/loss': 0.035520005971193314, 'train/grad_norm_pre_clip': 1.5111594200134277, 'timing/data': 0.0003046328201889992, 'timing/model': 0.41636788798496127, 'learning_rate/qwen_vl_interface': 8.236297295338474e-06, 'learning_rate/action_model': 8.236297295338474e-05, 'epoch': 0.22})
+[2026-07-28 04:03:11,074][starVLA.training.train_starvla][INFO] - Step 1207, Loss: {'train/loss': 0.06428783386945724, 'train/grad_norm_pre_clip': 2.2803585529327393, 'timing/data': 0.00025271298363804817, 'timing/model': 0.41484535206109285, 'learning_rate/qwen_vl_interface': 8.233320795414327e-06, 'learning_rate/action_model': 8.233320795414327e-05, 'epoch': 0.22})
+[2026-07-28 04:03:17,183][starVLA.training.train_starvla][INFO] - Step 1208, Loss: {'train/loss': 0.04653450846672058, 'train/grad_norm_pre_clip': 0.9037924408912659, 'timing/data': 0.000265651848167181, 'timing/model': 0.4155673789791763, 'learning_rate/qwen_vl_interface': 8.230342359645705e-06, 'learning_rate/action_model': 8.230342359645705e-05, 'epoch': 0.22})
+[2026-07-28 04:03:23,048][starVLA.training.train_starvla][INFO] - Step 1209, Loss: {'train/loss': 0.06525718420743942, 'train/grad_norm_pre_clip': 1.7614519596099854, 'timing/data': 0.00042826728895306587, 'timing/model': 0.41506640892475843, 'learning_rate/qwen_vl_interface': 8.227361989965285e-06, 'learning_rate/action_model': 8.227361989965284e-05, 'epoch': 0.22})
+[2026-07-28 04:03:28,960][starVLA.training.train_starvla][INFO] - Step 1210, Loss: {'train/loss': 0.08938974142074585, 'train/grad_norm_pre_clip': 2.0819292068481445, 'timing/data': 0.00029558222740888596, 'timing/model': 0.41719830268993974, 'learning_rate/qwen_vl_interface': 8.22437968830699e-06, 'learning_rate/action_model': 8.22437968830699e-05, 'epoch': 0.22})
+[2026-07-28 04:03:35,035][starVLA.training.train_starvla][INFO] - Step 1211, Loss: {'train/loss': 0.07079125940799713, 'train/grad_norm_pre_clip': 2.809528112411499, 'timing/data': 0.00027147913351655006, 'timing/model': 0.6523010460659862, 'learning_rate/qwen_vl_interface': 8.221395456606012e-06, 'learning_rate/action_model': 8.221395456606011e-05, 'epoch': 0.22})
+[2026-07-28 04:03:41,122][starVLA.training.train_starvla][INFO] - Step 1212, Loss: {'train/loss': 0.0864851102232933, 'train/grad_norm_pre_clip': 2.268393039703369, 'timing/data': 0.0002587209455668926, 'timing/model': 0.41573842195793986, 'learning_rate/qwen_vl_interface': 8.21840929679878e-06, 'learning_rate/action_model': 8.218409296798778e-05, 'epoch': 0.22})
+[2026-07-28 04:03:47,009][starVLA.training.train_starvla][INFO] - Step 1213, Loss: {'train/loss': 0.08708111196756363, 'train/grad_norm_pre_clip': 2.3578178882598877, 'timing/data': 0.0003947080112993717, 'timing/model': 0.41707510501146317, 'learning_rate/qwen_vl_interface': 8.215421210822982e-06, 'learning_rate/action_model': 8.215421210822982e-05, 'epoch': 0.22})
+[2026-07-28 04:03:52,946][starVLA.training.train_starvla][INFO] - Step 1214, Loss: {'train/loss': 0.13330654799938202, 'train/grad_norm_pre_clip': 6.180759429931641, 'timing/data': 0.0002958383411169052, 'timing/model': 0.417071171104908, 'learning_rate/qwen_vl_interface': 8.212431200617557e-06, 'learning_rate/action_model': 8.212431200617556e-05, 'epoch': 0.22})
+[2026-07-28 04:03:59,068][starVLA.training.train_starvla][INFO] - Step 1215, Loss: {'train/loss': 0.05380810424685478, 'train/grad_norm_pre_clip': 1.862118124961853, 'timing/data': 0.00026945024728775024, 'timing/model': 0.41287245228886604, 'learning_rate/qwen_vl_interface': 8.209439268122689e-06, 'learning_rate/action_model': 8.209439268122688e-05, 'epoch': 0.22})
+[2026-07-28 04:04:04,978][starVLA.training.train_starvla][INFO] - Step 1216, Loss: {'train/loss': 0.05832378566265106, 'train/grad_norm_pre_clip': 1.701461911201477, 'timing/data': 0.00027996813878417015, 'timing/model': 0.41641730768606067, 'learning_rate/qwen_vl_interface': 8.206445415279808e-06, 'learning_rate/action_model': 8.206445415279807e-05, 'epoch': 0.22})
+[2026-07-28 04:04:11,097][starVLA.training.train_starvla][INFO] - Step 1217, Loss: {'train/loss': 0.08265351504087448, 'train/grad_norm_pre_clip': 1.9962562322616577, 'timing/data': 0.00028607388958334923, 'timing/model': 0.650960183236748, 'learning_rate/qwen_vl_interface': 8.203449644031595e-06, 'learning_rate/action_model': 8.203449644031596e-05, 'epoch': 0.22})
+[2026-07-28 04:04:16,974][starVLA.training.train_starvla][INFO] - Step 1218, Loss: {'train/loss': 0.12746834754943848, 'train/grad_norm_pre_clip': 6.88646125793457, 'timing/data': 0.00026091374456882477, 'timing/model': 0.41483031725510955, 'learning_rate/qwen_vl_interface': 8.200451956321974e-06, 'learning_rate/action_model': 8.200451956321974e-05, 'epoch': 0.22})
+[2026-07-28 04:04:23,081][starVLA.training.train_starvla][INFO] - Step 1219, Loss: {'train/loss': 0.05319654941558838, 'train/grad_norm_pre_clip': 2.0727508068084717, 'timing/data': 0.00025610579177737236, 'timing/model': 0.41387408087030053, 'learning_rate/qwen_vl_interface': 8.197452354096112e-06, 'learning_rate/action_model': 8.197452354096112e-05, 'epoch': 0.22})
+[2026-07-28 04:04:28,974][starVLA.training.train_starvla][INFO] - Step 1220, Loss: {'train/loss': 0.10826970636844635, 'train/grad_norm_pre_clip': 3.211824655532837, 'timing/data': 0.000283082015812397, 'timing/model': 0.41700471425428987, 'learning_rate/qwen_vl_interface': 8.194450839300418e-06, 'learning_rate/action_model': 8.194450839300417e-05, 'epoch': 0.22})
+[2026-07-28 04:04:34,809][starVLA.training.train_starvla][INFO] - Step 1221, Loss: {'train/loss': 0.05729581415653229, 'train/grad_norm_pre_clip': 1.7192714214324951, 'timing/data': 0.00027907872572541237, 'timing/model': 0.4174477490596473, 'learning_rate/qwen_vl_interface': 8.191447413882539e-06, 'learning_rate/action_model': 8.191447413882539e-05, 'epoch': 0.22})
+[2026-07-28 04:04:41,119][starVLA.training.train_starvla][INFO] - Step 1222, Loss: {'train/loss': 0.08309076726436615, 'train/grad_norm_pre_clip': 1.7814643383026123, 'timing/data': 0.00026025110855698586, 'timing/model': 0.6525521161966026, 'learning_rate/qwen_vl_interface': 8.188442079791369e-06, 'learning_rate/action_model': 8.188442079791369e-05, 'epoch': 0.22})
+[2026-07-28 04:04:46,976][starVLA.training.train_starvla][INFO] - Step 1223, Loss: {'train/loss': 0.05684381723403931, 'train/grad_norm_pre_clip': 1.0794017314910889, 'timing/data': 0.0002865111455321312, 'timing/model': 0.416429809294641, 'learning_rate/qwen_vl_interface': 8.185434838977036e-06, 'learning_rate/action_model': 8.185434838977037e-05, 'epoch': 0.22})
+[2026-07-28 04:04:52,881][starVLA.training.train_starvla][INFO] - Step 1224, Loss: {'train/loss': 0.06267842650413513, 'train/grad_norm_pre_clip': 1.0869228839874268, 'timing/data': 0.0002950122579932213, 'timing/model': 0.4164415430277586, 'learning_rate/qwen_vl_interface': 8.182425693390905e-06, 'learning_rate/action_model': 8.182425693390905e-05, 'epoch': 0.22})
+[2026-07-28 04:04:59,015][starVLA.training.train_starvla][INFO] - Step 1225, Loss: {'train/loss': 0.06314976513385773, 'train/grad_norm_pre_clip': 1.5881719589233398, 'timing/data': 0.00026382505893707275, 'timing/model': 0.41854610899463296, 'learning_rate/qwen_vl_interface': 8.179414644985582e-06, 'learning_rate/action_model': 8.179414644985582e-05, 'epoch': 0.23})
+[2026-07-28 04:05:04,849][starVLA.training.train_starvla][INFO] - Step 1226, Loss: {'train/loss': 0.06092318147420883, 'train/grad_norm_pre_clip': 2.0180537700653076, 'timing/data': 0.0002872627228498459, 'timing/model': 0.4161852980032563, 'learning_rate/qwen_vl_interface': 8.176401695714897e-06, 'learning_rate/action_model': 8.176401695714895e-05, 'epoch': 0.23})
+[2026-07-28 04:05:10,952][starVLA.training.train_starvla][INFO] - Step 1227, Loss: {'train/loss': 0.09351252019405365, 'train/grad_norm_pre_clip': 4.663522243499756, 'timing/data': 0.0002865460701286793, 'timing/model': 0.6565872579813004, 'learning_rate/qwen_vl_interface': 8.173386847533922e-06, 'learning_rate/action_model': 8.173386847533921e-05, 'epoch': 0.23})
+[2026-07-28 04:05:16,812][starVLA.training.train_starvla][INFO] - Step 1228, Loss: {'train/loss': 0.08168298006057739, 'train/grad_norm_pre_clip': 2.651991605758667, 'timing/data': 0.00027077505365014076, 'timing/model': 0.4132056678645313, 'learning_rate/qwen_vl_interface': 8.170370102398962e-06, 'learning_rate/action_model': 8.17037010239896e-05, 'epoch': 0.23})
+[2026-07-28 04:05:22,948][starVLA.training.train_starvla][INFO] - Step 1229, Loss: {'train/loss': 0.06577969342470169, 'train/grad_norm_pre_clip': 2.4945285320281982, 'timing/data': 0.0003078519366681576, 'timing/model': 0.6504562650807202, 'learning_rate/qwen_vl_interface': 8.167351462267548e-06, 'learning_rate/action_model': 8.167351462267548e-05, 'epoch': 0.23})
+[2026-07-28 04:05:28,826][starVLA.training.train_starvla][INFO] - Step 1230, Loss: {'train/loss': 0.0690820962190628, 'train/grad_norm_pre_clip': 1.5207513570785522, 'timing/data': 0.0002631908282637596, 'timing/model': 0.41447844123467803, 'learning_rate/qwen_vl_interface': 8.164330929098444e-06, 'learning_rate/action_model': 8.164330929098443e-05, 'epoch': 0.23})
+[2026-07-28 04:05:34,670][starVLA.training.train_starvla][INFO] - Step 1231, Loss: {'train/loss': 0.03792983293533325, 'train/grad_norm_pre_clip': 0.9266226887702942, 'timing/data': 0.00028419075533747673, 'timing/model': 0.4174851207062602, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.23})
+[2026-07-28 04:05:40,528][starVLA.training.train_starvla][INFO] - Step 1232, Loss: {'train/loss': 0.08459502458572388, 'train/grad_norm_pre_clip': 2.9637155532836914, 'timing/data': 0.0002737618051469326, 'timing/model': 0.4174863244406879, 'learning_rate/qwen_vl_interface': 8.158284191488355e-06, 'learning_rate/action_model': 8.158284191488355e-05, 'epoch': 0.23})
+[2026-07-28 04:05:46,633][starVLA.training.train_starvla][INFO] - Step 1233, Loss: {'train/loss': 0.050114184617996216, 'train/grad_norm_pre_clip': 1.481396198272705, 'timing/data': 0.00029224809259176254, 'timing/model': 0.6572079020552337, 'learning_rate/qwen_vl_interface': 8.155257990971032e-06, 'learning_rate/action_model': 8.155257990971031e-05, 'epoch': 0.23})
+[2026-07-28 04:05:52,500][starVLA.training.train_starvla][INFO] - Step 1234, Loss: {'train/loss': 0.07444743067026138, 'train/grad_norm_pre_clip': 2.168046712875366, 'timing/data': 0.0002580448053777218, 'timing/model': 0.41357310907915235, 'learning_rate/qwen_vl_interface': 8.152229905263341e-06, 'learning_rate/action_model': 8.152229905263342e-05, 'epoch': 0.23})
+[2026-07-28 04:05:58,591][starVLA.training.train_starvla][INFO] - Step 1235, Loss: {'train/loss': 0.06989187002182007, 'train/grad_norm_pre_clip': 1.2272908687591553, 'timing/data': 0.0002624313347041607, 'timing/model': 0.414846648927778, 'learning_rate/qwen_vl_interface': 8.149199936330175e-06, 'learning_rate/action_model': 8.149199936330174e-05, 'epoch': 0.23})
+[2026-07-28 04:06:04,471][starVLA.training.train_starvla][INFO] - Step 1236, Loss: {'train/loss': 0.02184678614139557, 'train/grad_norm_pre_clip': 0.7613663673400879, 'timing/data': 0.0004290160723030567, 'timing/model': 0.4165931320749223, 'learning_rate/qwen_vl_interface': 8.146168086137646e-06, 'learning_rate/action_model': 8.146168086137646e-05, 'epoch': 0.23})
+[2026-07-28 04:06:10,363][starVLA.training.train_starvla][INFO] - Step 1237, Loss: {'train/loss': 0.07269695401191711, 'train/grad_norm_pre_clip': 1.6161834001541138, 'timing/data': 0.0002888268791139126, 'timing/model': 0.41920592496171594, 'learning_rate/qwen_vl_interface': 8.143134356653087e-06, 'learning_rate/action_model': 8.143134356653087e-05, 'epoch': 0.23})
+[2026-07-28 04:06:16,480][starVLA.training.train_starvla][INFO] - Step 1238, Loss: {'train/loss': 0.06096959114074707, 'train/grad_norm_pre_clip': 1.6506153345108032, 'timing/data': 0.0002687140367925167, 'timing/model': 0.41537894681096077, 'learning_rate/qwen_vl_interface': 8.140098749845056e-06, 'learning_rate/action_model': 8.140098749845055e-05, 'epoch': 0.23})
+[2026-07-28 04:06:22,547][starVLA.training.train_starvla][INFO] - Step 1239, Loss: {'train/loss': 0.06283069401979446, 'train/grad_norm_pre_clip': 0.8391585350036621, 'timing/data': 0.00028274906799197197, 'timing/model': 0.41608400689437985, 'learning_rate/qwen_vl_interface': 8.137061267683323e-06, 'learning_rate/action_model': 8.137061267683323e-05, 'epoch': 0.23})
+[2026-07-28 04:06:28,409][starVLA.training.train_starvla][INFO] - Step 1240, Loss: {'train/loss': 0.1045677661895752, 'train/grad_norm_pre_clip': 2.0009899139404297, 'timing/data': 0.0002704281359910965, 'timing/model': 0.41713350638747215, 'learning_rate/qwen_vl_interface': 8.134021912138879e-06, 'learning_rate/action_model': 8.134021912138879e-05, 'epoch': 0.23})
+[2026-07-28 04:06:34,052][starVLA.training.train_starvla][INFO] - Step 1241, Loss: {'train/loss': 0.06534209102392197, 'train/grad_norm_pre_clip': 0.9107331037521362, 'timing/data': 0.00028468575328588486, 'timing/model': 0.4166992618702352, 'learning_rate/qwen_vl_interface': 8.130980685183927e-06, 'learning_rate/action_model': 8.130980685183926e-05, 'epoch': 0.23})
+[2026-07-28 04:06:40,158][starVLA.training.train_starvla][INFO] - Step 1242, Loss: {'train/loss': 0.043754905462265015, 'train/grad_norm_pre_clip': 0.8170472979545593, 'timing/data': 0.0002998928539454937, 'timing/model': 0.4170467178337276, 'learning_rate/qwen_vl_interface': 8.127937588791887e-06, 'learning_rate/action_model': 8.127937588791886e-05, 'epoch': 0.23})
+[2026-07-28 04:06:45,844][starVLA.training.train_starvla][INFO] - Step 1243, Loss: {'train/loss': 0.06654026359319687, 'train/grad_norm_pre_clip': 0.924860417842865, 'timing/data': 0.0003587799146771431, 'timing/model': 0.4164119050838053, 'learning_rate/qwen_vl_interface': 8.12489262493739e-06, 'learning_rate/action_model': 8.12489262493739e-05, 'epoch': 0.23})
+[2026-07-28 04:06:51,492][starVLA.training.train_starvla][INFO] - Step 1244, Loss: {'train/loss': 0.08044879138469696, 'train/grad_norm_pre_clip': 2.0422630310058594, 'timing/data': 0.0002915468066930771, 'timing/model': 0.4168679122813046, 'learning_rate/qwen_vl_interface': 8.121845795596282e-06, 'learning_rate/action_model': 8.121845795596282e-05, 'epoch': 0.23})
+[2026-07-28 04:06:57,495][starVLA.training.train_starvla][INFO] - Step 1245, Loss: {'train/loss': 0.054539185017347336, 'train/grad_norm_pre_clip': 1.7897441387176514, 'timing/data': 0.00028537027537822723, 'timing/model': 0.5854285177774727, 'learning_rate/qwen_vl_interface': 8.118797102745614e-06, 'learning_rate/action_model': 8.118797102745615e-05, 'epoch': 0.23})
+[2026-07-28 04:07:03,367][starVLA.training.train_starvla][INFO] - Step 1246, Loss: {'train/loss': 0.028039783239364624, 'train/grad_norm_pre_clip': 0.8122386932373047, 'timing/data': 0.00044605834409594536, 'timing/model': 0.41656764689832926, 'learning_rate/qwen_vl_interface': 8.115746548363655e-06, 'learning_rate/action_model': 8.115746548363655e-05, 'epoch': 0.23})
+[2026-07-28 04:07:09,472][starVLA.training.train_starvla][INFO] - Step 1247, Loss: {'train/loss': 0.05733940377831459, 'train/grad_norm_pre_clip': 1.253969669342041, 'timing/data': 0.00029208511114120483, 'timing/model': 0.6581306657753885, 'learning_rate/qwen_vl_interface': 8.112694134429873e-06, 'learning_rate/action_model': 8.112694134429873e-05, 'epoch': 0.23})
+[2026-07-28 04:07:15,362][starVLA.training.train_starvla][INFO] - Step 1248, Loss: {'train/loss': 0.043532535433769226, 'train/grad_norm_pre_clip': 2.290653944015503, 'timing/data': 0.0002753143198788166, 'timing/model': 0.4158811969682574, 'learning_rate/qwen_vl_interface': 8.109639862924947e-06, 'learning_rate/action_model': 8.109639862924945e-05, 'epoch': 0.23})
+[2026-07-28 04:07:21,479][starVLA.training.train_starvla][INFO] - Step 1249, Loss: {'train/loss': 0.05012078210711479, 'train/grad_norm_pre_clip': 1.4634733200073242, 'timing/data': 0.0002929503098130226, 'timing/model': 0.416681915987283, 'learning_rate/qwen_vl_interface': 8.10658373583076e-06, 'learning_rate/action_model': 8.106583735830761e-05, 'epoch': 0.23})
+[2026-07-28 04:07:27,152][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1250
+[2026-07-28 04:07:32,069][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1250
+[2026-07-28 04:15:17,719][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1250
+[2026-07-28 04:15:17,720][starVLA.training.train_starvla][INFO] - Step 1250, Loss: {'train/loss': 0.052660875022411346, 'train/grad_norm_pre_clip': 1.7787514925003052, 'eval/action_loss/seconds': 4.913996264804155, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.06315884739160538, 'eval/latency_3/loss': 0.06315884739160538, 'eval/flappy/loss': 0.06315884739160538, 'eval/flappy/latency_3/loss': 0.06315884739160538, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 388.7600045718253, 'rl_games_eval/mid_train/mean_length': 3150.55, 'rl_games_eval/mid_train/std_reward': 108.89891494959997, 'rl_games_eval/mid_train/std_length': 876.8407195722608, 'rl_games_eval/mid_train/macro_mean_reward': 388.7600045718253, 'rl_games_eval/mid_train/macro_mean_length': 3150.55, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 388.7600045718253, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 3150.55, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 108.89891494959997, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 876.8407195722608, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 388.7600045718253, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000270988792181015, 'timing/model': 0.41608677012845874, 'learning_rate/qwen_vl_interface': 8.1035257551304e-06, 'learning_rate/action_model': 8.1035257551304e-05, 'epoch': 0.23})
+[2026-07-28 04:15:24,039][starVLA.training.train_starvla][INFO] - Step 1251, Loss: {'train/loss': 0.06103690713644028, 'train/grad_norm_pre_clip': 2.45633602142334, 'timing/data': 0.00026085786521434784, 'timing/model': 0.43317597918212414, 'learning_rate/qwen_vl_interface': 8.100465922808158e-06, 'learning_rate/action_model': 8.100465922808159e-05, 'epoch': 0.23})
+[2026-07-28 04:15:30,123][starVLA.training.train_starvla][INFO] - Step 1252, Loss: {'train/loss': 0.062440067529678345, 'train/grad_norm_pre_clip': 1.8919798135757446, 'timing/data': 0.00025663338601589203, 'timing/model': 0.41475135181099176, 'learning_rate/qwen_vl_interface': 8.097404240849527e-06, 'learning_rate/action_model': 8.097404240849527e-05, 'epoch': 0.23})
+[2026-07-28 04:15:35,958][starVLA.training.train_starvla][INFO] - Step 1253, Loss: {'train/loss': 0.04135281965136528, 'train/grad_norm_pre_clip': 0.8836658596992493, 'timing/data': 0.00027254922315478325, 'timing/model': 0.41460186801850796, 'learning_rate/qwen_vl_interface': 8.094340711241198e-06, 'learning_rate/action_model': 8.094340711241198e-05, 'epoch': 0.23})
+[2026-07-28 04:15:41,811][starVLA.training.train_starvla][INFO] - Step 1254, Loss: {'train/loss': 0.06909023225307465, 'train/grad_norm_pre_clip': 1.6811050176620483, 'timing/data': 0.0002638231962919235, 'timing/model': 0.41537573002278805, 'learning_rate/qwen_vl_interface': 8.09127533597106e-06, 'learning_rate/action_model': 8.091275335971061e-05, 'epoch': 0.23})
+[2026-07-28 04:15:47,909][starVLA.training.train_starvla][INFO] - Step 1255, Loss: {'train/loss': 0.0567975714802742, 'train/grad_norm_pre_clip': 2.287179946899414, 'timing/data': 0.00026916712522506714, 'timing/model': 0.6487759877927601, 'learning_rate/qwen_vl_interface': 8.088208117028205e-06, 'learning_rate/action_model': 8.088208117028206e-05, 'epoch': 0.23})
+[2026-07-28 04:15:53,750][starVLA.training.train_starvla][INFO] - Step 1256, Loss: {'train/loss': 0.08199075609445572, 'train/grad_norm_pre_clip': 1.9928714036941528, 'timing/data': 0.0002535278908908367, 'timing/model': 0.4135850891470909, 'learning_rate/qwen_vl_interface': 8.085139056402918e-06, 'learning_rate/action_model': 8.085139056402917e-05, 'epoch': 0.23})
+[2026-07-28 04:15:59,827][starVLA.training.train_starvla][INFO] - Step 1257, Loss: {'train/loss': 0.056653957813978195, 'train/grad_norm_pre_clip': 2.9253084659576416, 'timing/data': 0.00026005785912275314, 'timing/model': 0.41410537995398045, 'learning_rate/qwen_vl_interface': 8.082068156086675e-06, 'learning_rate/action_model': 8.082068156086675e-05, 'epoch': 0.23})
+[2026-07-28 04:16:05,448][starVLA.training.train_starvla][INFO] - Step 1258, Loss: {'train/loss': 0.09881168603897095, 'train/grad_norm_pre_clip': 2.527690887451172, 'timing/data': 0.00027840631082654, 'timing/model': 0.4153246241621673, 'learning_rate/qwen_vl_interface': 8.078995418072156e-06, 'learning_rate/action_model': 8.078995418072157e-05, 'epoch': 0.23})
+[2026-07-28 04:16:11,086][starVLA.training.train_starvla][INFO] - Step 1259, Loss: {'train/loss': 0.05871883034706116, 'train/grad_norm_pre_clip': 1.1887896060943604, 'timing/data': 0.000276023056358099, 'timing/model': 0.4155385410413146, 'learning_rate/qwen_vl_interface': 8.075920844353223e-06, 'learning_rate/action_model': 8.075920844353223e-05, 'epoch': 0.23})
+[2026-07-28 04:16:17,156][starVLA.training.train_starvla][INFO] - Step 1260, Loss: {'train/loss': 0.03913091868162155, 'train/grad_norm_pre_clip': 1.7827550172805786, 'timing/data': 0.00028629088774323463, 'timing/model': 0.6522583849728107, 'learning_rate/qwen_vl_interface': 8.072844436924934e-06, 'learning_rate/action_model': 8.072844436924934e-05, 'epoch': 0.23})
+[2026-07-28 04:16:23,267][starVLA.training.train_starvla][INFO] - Step 1261, Loss: {'train/loss': 0.037221670150756836, 'train/grad_norm_pre_clip': 1.0285367965698242, 'timing/data': 0.00026248395442962646, 'timing/model': 0.41505929082632065, 'learning_rate/qwen_vl_interface': 8.069766197783537e-06, 'learning_rate/action_model': 8.069766197783537e-05, 'epoch': 0.23})
+[2026-07-28 04:16:29,128][starVLA.training.train_starvla][INFO] - Step 1262, Loss: {'train/loss': 0.04689671844244003, 'train/grad_norm_pre_clip': 2.1279845237731934, 'timing/data': 0.0002576680853962898, 'timing/model': 0.41370048094540834, 'learning_rate/qwen_vl_interface': 8.066686128926468e-06, 'learning_rate/action_model': 8.066686128926468e-05, 'epoch': 0.23})
+[2026-07-28 04:16:35,032][starVLA.training.train_starvla][INFO] - Step 1263, Loss: {'train/loss': 0.06021696701645851, 'train/grad_norm_pre_clip': 1.733017086982727, 'timing/data': 0.0002940320409834385, 'timing/model': 0.4168980489484966, 'learning_rate/qwen_vl_interface': 8.063604232352352e-06, 'learning_rate/action_model': 8.06360423235235e-05, 'epoch': 0.23})
+[2026-07-28 04:16:40,855][starVLA.training.train_starvla][INFO] - Step 1264, Loss: {'train/loss': 0.05397447198629379, 'train/grad_norm_pre_clip': 1.738135814666748, 'timing/data': 0.00027436530217528343, 'timing/model': 0.41559908678755164, 'learning_rate/qwen_vl_interface': 8.060520510060996e-06, 'learning_rate/action_model': 8.060520510060995e-05, 'epoch': 0.23})
+[2026-07-28 04:16:46,941][starVLA.training.train_starvla][INFO] - Step 1265, Loss: {'train/loss': 0.06259986758232117, 'train/grad_norm_pre_clip': 1.4123096466064453, 'timing/data': 0.0002697957679629326, 'timing/model': 0.6468766792677343, 'learning_rate/qwen_vl_interface': 8.057434964053391e-06, 'learning_rate/action_model': 8.057434964053392e-05, 'epoch': 0.23})
+[2026-07-28 04:16:52,832][starVLA.training.train_starvla][INFO] - Step 1266, Loss: {'train/loss': 0.05400621145963669, 'train/grad_norm_pre_clip': 2.413472890853882, 'timing/data': 0.00036113616079092026, 'timing/model': 0.4128585229627788, 'learning_rate/qwen_vl_interface': 8.05434759633172e-06, 'learning_rate/action_model': 8.054347596331721e-05, 'epoch': 0.23})
+[2026-07-28 04:16:58,666][starVLA.training.train_starvla][INFO] - Step 1267, Loss: {'train/loss': 0.030602172017097473, 'train/grad_norm_pre_clip': 0.8497866988182068, 'timing/data': 0.00027019111439585686, 'timing/model': 0.4166488549672067, 'learning_rate/qwen_vl_interface': 8.05125840889934e-06, 'learning_rate/action_model': 8.05125840889934e-05, 'epoch': 0.23})
+[2026-07-28 04:17:04,758][starVLA.training.train_starvla][INFO] - Step 1268, Loss: {'train/loss': 0.08473901450634003, 'train/grad_norm_pre_clip': 3.46478009223938, 'timing/data': 0.0002767890691757202, 'timing/model': 0.647006178740412, 'learning_rate/qwen_vl_interface': 8.048167403760792e-06, 'learning_rate/action_model': 8.048167403760792e-05, 'epoch': 0.23})
+[2026-07-28 04:17:10,607][starVLA.training.train_starvla][INFO] - Step 1269, Loss: {'train/loss': 0.03058343566954136, 'train/grad_norm_pre_clip': 0.9928650856018066, 'timing/data': 0.0002656676806509495, 'timing/model': 0.4164573671296239, 'learning_rate/qwen_vl_interface': 8.045074582921797e-06, 'learning_rate/action_model': 8.045074582921797e-05, 'epoch': 0.23})
+[2026-07-28 04:17:16,478][starVLA.training.train_starvla][INFO] - Step 1270, Loss: {'train/loss': 0.051696859300136566, 'train/grad_norm_pre_clip': 1.5533483028411865, 'timing/data': 0.00030806707218289375, 'timing/model': 0.41637625405564904, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.23})
+[2026-07-28 04:17:22,620][starVLA.training.train_starvla][INFO] - Step 1271, Loss: {'train/loss': 0.08949929475784302, 'train/grad_norm_pre_clip': 2.3875715732574463, 'timing/data': 0.0002816440537571907, 'timing/model': 0.652472279034555, 'learning_rate/qwen_vl_interface': 8.038883502171223e-06, 'learning_rate/action_model': 8.038883502171223e-05, 'epoch': 0.23})
+[2026-07-28 04:17:28,475][starVLA.training.train_starvla][INFO] - Step 1272, Loss: {'train/loss': 0.09009694308042526, 'train/grad_norm_pre_clip': 1.6983225345611572, 'timing/data': 0.00025197211652994156, 'timing/model': 0.41519738594070077, 'learning_rate/qwen_vl_interface': 8.035785246276973e-06, 'learning_rate/action_model': 8.035785246276973e-05, 'epoch': 0.23})
+[2026-07-28 04:17:34,332][starVLA.training.train_starvla][INFO] - Step 1273, Loss: {'train/loss': 0.05387718230485916, 'train/grad_norm_pre_clip': 1.1610208749771118, 'timing/data': 0.00029426999390125275, 'timing/model': 0.4172859089449048, 'learning_rate/qwen_vl_interface': 8.03268518271692e-06, 'learning_rate/action_model': 8.032685182716919e-05, 'epoch': 0.23})
+[2026-07-28 04:17:40,202][starVLA.training.train_starvla][INFO] - Step 1274, Loss: {'train/loss': 0.09412543475627899, 'train/grad_norm_pre_clip': 1.499042272567749, 'timing/data': 0.00028239982202649117, 'timing/model': 0.41695048566907644, 'learning_rate/qwen_vl_interface': 8.029583313502664e-06, 'learning_rate/action_model': 8.029583313502664e-05, 'epoch': 0.23})
+[2026-07-28 04:17:46,297][starVLA.training.train_starvla][INFO] - Step 1275, Loss: {'train/loss': 0.0669451430439949, 'train/grad_norm_pre_clip': 2.4238357543945312, 'timing/data': 0.0002612336538732052, 'timing/model': 0.41578712966293097, 'learning_rate/qwen_vl_interface': 8.026479640646974e-06, 'learning_rate/action_model': 8.026479640646973e-05, 'epoch': 0.23})
+[2026-07-28 04:17:52,392][starVLA.training.train_starvla][INFO] - Step 1276, Loss: {'train/loss': 0.07532378286123276, 'train/grad_norm_pre_clip': 1.550768494606018, 'timing/data': 0.0002599838189780712, 'timing/model': 0.4155016331933439, 'learning_rate/qwen_vl_interface': 8.023374166163787e-06, 'learning_rate/action_model': 8.023374166163787e-05, 'epoch': 0.23})
+[2026-07-28 04:17:58,238][starVLA.training.train_starvla][INFO] - Step 1277, Loss: {'train/loss': 0.07168115675449371, 'train/grad_norm_pre_clip': 1.6648255586624146, 'timing/data': 0.00027801981195807457, 'timing/model': 0.4248130959458649, 'learning_rate/qwen_vl_interface': 8.020266892068214e-06, 'learning_rate/action_model': 8.020266892068214e-05, 'epoch': 0.23})
+[2026-07-28 04:18:04,267][starVLA.training.train_starvla][INFO] - Step 1278, Loss: {'train/loss': 0.06811799108982086, 'train/grad_norm_pre_clip': 1.4961042404174805, 'timing/data': 0.00026733800768852234, 'timing/model': 0.5819075386971235, 'learning_rate/qwen_vl_interface': 8.017157820376532e-06, 'learning_rate/action_model': 8.017157820376532e-05, 'epoch': 0.23})
+[2026-07-28 04:18:10,132][starVLA.training.train_starvla][INFO] - Step 1279, Loss: {'train/loss': 0.06875724345445633, 'train/grad_norm_pre_clip': 1.9871907234191895, 'timing/data': 0.000279831700026989, 'timing/model': 0.4165524812415242, 'learning_rate/qwen_vl_interface': 8.014046953106183e-06, 'learning_rate/action_model': 8.014046953106182e-05, 'epoch': 0.23})
+[2026-07-28 04:18:16,235][starVLA.training.train_starvla][INFO] - Step 1280, Loss: {'train/loss': 0.07005241513252258, 'train/grad_norm_pre_clip': 1.8893370628356934, 'timing/data': 0.00025290343910455704, 'timing/model': 0.41477779718115926, 'learning_rate/qwen_vl_interface': 8.010934292275775e-06, 'learning_rate/action_model': 8.010934292275775e-05, 'epoch': 0.24})
+[2026-07-28 04:18:22,142][starVLA.training.train_starvla][INFO] - Step 1281, Loss: {'train/loss': 0.059889137744903564, 'train/grad_norm_pre_clip': 1.4415572881698608, 'timing/data': 0.0002827378921210766, 'timing/model': 0.4162095310166478, 'learning_rate/qwen_vl_interface': 8.007819839905079e-06, 'learning_rate/action_model': 8.007819839905078e-05, 'epoch': 0.24})
+[2026-07-28 04:18:28,211][starVLA.training.train_starvla][INFO] - Step 1282, Loss: {'train/loss': 0.053700126707553864, 'train/grad_norm_pre_clip': 1.7403881549835205, 'timing/data': 0.0002647237852215767, 'timing/model': 0.4147483850829303, 'learning_rate/qwen_vl_interface': 8.004703598015032e-06, 'learning_rate/action_model': 8.004703598015031e-05, 'epoch': 0.24})
+[2026-07-28 04:18:34,081][starVLA.training.train_starvla][INFO] - Step 1283, Loss: {'train/loss': 0.08754652738571167, 'train/grad_norm_pre_clip': 2.1007440090179443, 'timing/data': 0.00028607482090592384, 'timing/model': 0.4155375137925148, 'learning_rate/qwen_vl_interface': 8.001585568627724e-06, 'learning_rate/action_model': 8.001585568627725e-05, 'epoch': 0.24})
+[2026-07-28 04:18:39,937][starVLA.training.train_starvla][INFO] - Step 1284, Loss: {'train/loss': 0.055424656718969345, 'train/grad_norm_pre_clip': 1.6884135007858276, 'timing/data': 0.00027766916900873184, 'timing/model': 0.4170003938488662, 'learning_rate/qwen_vl_interface': 7.998465753766418e-06, 'learning_rate/action_model': 7.998465753766418e-05, 'epoch': 0.24})
+[2026-07-28 04:18:46,013][starVLA.training.train_starvla][INFO] - Step 1285, Loss: {'train/loss': 0.059269361197948456, 'train/grad_norm_pre_clip': 1.9040900468826294, 'timing/data': 0.000275603961199522, 'timing/model': 0.6529655703343451, 'learning_rate/qwen_vl_interface': 7.995344155455523e-06, 'learning_rate/action_model': 7.995344155455522e-05, 'epoch': 0.24})
+[2026-07-28 04:18:52,089][starVLA.training.train_starvla][INFO] - Step 1286, Loss: {'train/loss': 0.0627172440290451, 'train/grad_norm_pre_clip': 2.5090954303741455, 'timing/data': 0.00025862595066428185, 'timing/model': 0.41430696472525597, 'learning_rate/qwen_vl_interface': 7.992220775720609e-06, 'learning_rate/action_model': 7.992220775720609e-05, 'epoch': 0.24})
+[2026-07-28 04:18:57,923][starVLA.training.train_starvla][INFO] - Step 1287, Loss: {'train/loss': 0.11053808033466339, 'train/grad_norm_pre_clip': 1.969773292541504, 'timing/data': 0.0002789958380162716, 'timing/model': 0.414829274173826, 'learning_rate/qwen_vl_interface': 7.989095616588409e-06, 'learning_rate/action_model': 7.989095616588408e-05, 'epoch': 0.24})
+[2026-07-28 04:19:03,779][starVLA.training.train_starvla][INFO] - Step 1288, Loss: {'train/loss': 0.07246401906013489, 'train/grad_norm_pre_clip': 2.0266788005828857, 'timing/data': 0.0002803266979753971, 'timing/model': 0.419581756927073, 'learning_rate/qwen_vl_interface': 7.985968680086799e-06, 'learning_rate/action_model': 7.985968680086798e-05, 'epoch': 0.24})
+[2026-07-28 04:19:09,414][starVLA.training.train_starvla][INFO] - Step 1289, Loss: {'train/loss': 0.07589063793420792, 'train/grad_norm_pre_clip': 1.9192811250686646, 'timing/data': 0.0002768789418041706, 'timing/model': 0.41492520412430167, 'learning_rate/qwen_vl_interface': 7.982839968244816e-06, 'learning_rate/action_model': 7.982839968244816e-05, 'epoch': 0.24})
+[2026-07-28 04:19:15,068][starVLA.training.train_starvla][INFO] - Step 1290, Loss: {'train/loss': 0.07701382040977478, 'train/grad_norm_pre_clip': 1.4457613229751587, 'timing/data': 0.00027944473549723625, 'timing/model': 0.41686537535861135, 'learning_rate/qwen_vl_interface': 7.97970948309265e-06, 'learning_rate/action_model': 7.979709483092649e-05, 'epoch': 0.24})
+[2026-07-28 04:19:20,936][starVLA.training.train_starvla][INFO] - Step 1291, Loss: {'train/loss': 0.06545108556747437, 'train/grad_norm_pre_clip': 1.0926061868667603, 'timing/data': 0.000277599785476923, 'timing/model': 0.649478537030518, 'learning_rate/qwen_vl_interface': 7.976577226661635e-06, 'learning_rate/action_model': 7.976577226661634e-05, 'epoch': 0.24})
+[2026-07-28 04:19:26,766][starVLA.training.train_starvla][INFO] - Step 1292, Loss: {'train/loss': 0.07025662064552307, 'train/grad_norm_pre_clip': 2.7633986473083496, 'timing/data': 0.0002585737966001034, 'timing/model': 0.4140312196686864, 'learning_rate/qwen_vl_interface': 7.973443200984259e-06, 'learning_rate/action_model': 7.973443200984259e-05, 'epoch': 0.24})
+[2026-07-28 04:19:32,396][starVLA.training.train_starvla][INFO] - Step 1293, Loss: {'train/loss': 0.07364676892757416, 'train/grad_norm_pre_clip': 2.0941104888916016, 'timing/data': 0.0002779001370072365, 'timing/model': 0.41652656020596623, 'learning_rate/qwen_vl_interface': 7.970307408094158e-06, 'learning_rate/action_model': 7.970307408094157e-05, 'epoch': 0.24})
+[2026-07-28 04:19:38,253][starVLA.training.train_starvla][INFO] - Step 1294, Loss: {'train/loss': 0.0479932576417923, 'train/grad_norm_pre_clip': 2.107326030731201, 'timing/data': 0.0003538290038704872, 'timing/model': 0.41665412299335003, 'learning_rate/qwen_vl_interface': 7.967169850026113e-06, 'learning_rate/action_model': 7.967169850026113e-05, 'epoch': 0.24})
+[2026-07-28 04:19:44,102][starVLA.training.train_starvla][INFO] - Step 1295, Loss: {'train/loss': 0.04072117060422897, 'train/grad_norm_pre_clip': 1.053727149963379, 'timing/data': 0.00027864379808306694, 'timing/model': 0.41602636221796274, 'learning_rate/qwen_vl_interface': 7.964030528816055e-06, 'learning_rate/action_model': 7.964030528816055e-05, 'epoch': 0.24})
+[2026-07-28 04:19:50,420][starVLA.training.train_starvla][INFO] - Step 1296, Loss: {'train/loss': 0.09683027118444443, 'train/grad_norm_pre_clip': 2.2652289867401123, 'timing/data': 0.00025880197063088417, 'timing/model': 0.6492648017592728, 'learning_rate/qwen_vl_interface': 7.960889446501052e-06, 'learning_rate/action_model': 7.960889446501052e-05, 'epoch': 0.24})
+[2026-07-28 04:19:56,275][starVLA.training.train_starvla][INFO] - Step 1297, Loss: {'train/loss': 0.07291808724403381, 'train/grad_norm_pre_clip': 2.889824390411377, 'timing/data': 0.0002773371525108814, 'timing/model': 0.41694425279274583, 'learning_rate/qwen_vl_interface': 7.957746605119317e-06, 'learning_rate/action_model': 7.957746605119318e-05, 'epoch': 0.24})
+[2026-07-28 04:20:02,143][starVLA.training.train_starvla][INFO] - Step 1298, Loss: {'train/loss': 0.03258989006280899, 'train/grad_norm_pre_clip': 1.3274266719818115, 'timing/data': 0.0002785748802125454, 'timing/model': 0.4155730721540749, 'learning_rate/qwen_vl_interface': 7.954602006710209e-06, 'learning_rate/action_model': 7.954602006710209e-05, 'epoch': 0.24})
+[2026-07-28 04:20:08,253][starVLA.training.train_starvla][INFO] - Step 1299, Loss: {'train/loss': 0.03023514151573181, 'train/grad_norm_pre_clip': 0.764378547668457, 'timing/data': 0.0002577630802989006, 'timing/model': 0.4159551430493593, 'learning_rate/qwen_vl_interface': 7.951455653314222e-06, 'learning_rate/action_model': 7.951455653314221e-05, 'epoch': 0.24})
+[2026-07-28 04:20:14,345][starVLA.training.train_starvla][INFO] - Step 1300, Loss: {'train/loss': 0.07910368591547012, 'train/grad_norm_pre_clip': 1.7908915281295776, 'timing/data': 0.0002667950466275215, 'timing/model': 0.41397941764444113, 'learning_rate/qwen_vl_interface': 7.94830754697299e-06, 'learning_rate/action_model': 7.94830754697299e-05, 'epoch': 0.24})
+[2026-07-28 04:20:20,222][starVLA.training.train_starvla][INFO] - Step 1301, Loss: {'train/loss': 0.08283556997776031, 'train/grad_norm_pre_clip': 3.0496299266815186, 'timing/data': 0.000265561044216156, 'timing/model': 0.41513295704498887, 'learning_rate/qwen_vl_interface': 7.945157689729287e-06, 'learning_rate/action_model': 7.945157689729287e-05, 'epoch': 0.24})
+[2026-07-28 04:20:26,083][starVLA.training.train_starvla][INFO] - Step 1302, Loss: {'train/loss': 0.04956912249326706, 'train/grad_norm_pre_clip': 1.072389006614685, 'timing/data': 0.00027790991589426994, 'timing/model': 0.41633519995957613, 'learning_rate/qwen_vl_interface': 7.94200608362702e-06, 'learning_rate/action_model': 7.94200608362702e-05, 'epoch': 0.24})
+[2026-07-28 04:20:32,201][starVLA.training.train_starvla][INFO] - Step 1303, Loss: {'train/loss': 0.1224857047200203, 'train/grad_norm_pre_clip': 2.3123562335968018, 'timing/data': 0.00027276715263724327, 'timing/model': 0.649565305095166, 'learning_rate/qwen_vl_interface': 7.938852730711232e-06, 'learning_rate/action_model': 7.938852730711231e-05, 'epoch': 0.24})
+[2026-07-28 04:20:38,067][starVLA.training.train_starvla][INFO] - Step 1304, Loss: {'train/loss': 0.04007364809513092, 'train/grad_norm_pre_clip': 0.9899662137031555, 'timing/data': 0.0003278600051999092, 'timing/model': 0.41428234335035086, 'learning_rate/qwen_vl_interface': 7.935697633028099e-06, 'learning_rate/action_model': 7.935697633028099e-05, 'epoch': 0.24})
+[2026-07-28 04:20:43,900][starVLA.training.train_starvla][INFO] - Step 1305, Loss: {'train/loss': 0.046759601682424545, 'train/grad_norm_pre_clip': 1.8765095472335815, 'timing/data': 0.00028489669784903526, 'timing/model': 0.41566091077402234, 'learning_rate/qwen_vl_interface': 7.932540792624931e-06, 'learning_rate/action_model': 7.93254079262493e-05, 'epoch': 0.24})
+[2026-07-28 04:20:49,758][starVLA.training.train_starvla][INFO] - Step 1306, Loss: {'train/loss': 0.04823121428489685, 'train/grad_norm_pre_clip': 1.443830966949463, 'timing/data': 0.00025642383843660355, 'timing/model': 0.4154519089497626, 'learning_rate/qwen_vl_interface': 7.929382211550166e-06, 'learning_rate/action_model': 7.929382211550165e-05, 'epoch': 0.24})
+[2026-07-28 04:20:55,861][starVLA.training.train_starvla][INFO] - Step 1307, Loss: {'train/loss': 0.05744659900665283, 'train/grad_norm_pre_clip': 1.9757328033447266, 'timing/data': 0.00027605192735791206, 'timing/model': 0.6721738930791616, 'learning_rate/qwen_vl_interface': 7.926221891853373e-06, 'learning_rate/action_model': 7.926221891853372e-05, 'epoch': 0.24})
+[2026-07-28 04:21:01,716][starVLA.training.train_starvla][INFO] - Step 1308, Loss: {'train/loss': 0.06838946044445038, 'train/grad_norm_pre_clip': 1.5786432027816772, 'timing/data': 0.00028026383370161057, 'timing/model': 0.4159837760962546, 'learning_rate/qwen_vl_interface': 7.923059835585251e-06, 'learning_rate/action_model': 7.923059835585251e-05, 'epoch': 0.24})
+[2026-07-28 04:21:07,811][starVLA.training.train_starvla][INFO] - Step 1309, Loss: {'train/loss': 0.07122386246919632, 'train/grad_norm_pre_clip': 2.6359729766845703, 'timing/data': 0.0003330092877149582, 'timing/model': 0.41354556614533067, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.24})
+[2026-07-28 04:21:13,683][starVLA.training.train_starvla][INFO] - Step 1310, Loss: {'train/loss': 0.03594767674803734, 'train/grad_norm_pre_clip': 1.1925137042999268, 'timing/data': 0.00026230188086628914, 'timing/model': 0.4173517576418817, 'learning_rate/qwen_vl_interface': 7.916730521543436e-06, 'learning_rate/action_model': 7.916730521543436e-05, 'epoch': 0.24})
+[2026-07-28 04:21:19,538][starVLA.training.train_starvla][INFO] - Step 1311, Loss: {'train/loss': 0.05381573736667633, 'train/grad_norm_pre_clip': 2.3864190578460693, 'timing/data': 0.0002801711671054363, 'timing/model': 0.4162375838495791, 'learning_rate/qwen_vl_interface': 7.913563267876765e-06, 'learning_rate/action_model': 7.913563267876766e-05, 'epoch': 0.24})
+[2026-07-28 04:21:25,617][starVLA.training.train_starvla][INFO] - Step 1312, Loss: {'train/loss': 0.05927973985671997, 'train/grad_norm_pre_clip': 1.2358710765838623, 'timing/data': 0.00030233897268772125, 'timing/model': 0.664669135119766, 'learning_rate/qwen_vl_interface': 7.910394285852811e-06, 'learning_rate/action_model': 7.910394285852811e-05, 'epoch': 0.24})
+[2026-07-28 04:21:31,699][starVLA.training.train_starvla][INFO] - Step 1313, Loss: {'train/loss': 0.08850184082984924, 'train/grad_norm_pre_clip': 2.7369027137756348, 'timing/data': 0.0002625272609293461, 'timing/model': 0.41480185417458415, 'learning_rate/qwen_vl_interface': 7.907223577527886e-06, 'learning_rate/action_model': 7.907223577527886e-05, 'epoch': 0.24})
+[2026-07-28 04:21:37,593][starVLA.training.train_starvla][INFO] - Step 1314, Loss: {'train/loss': 0.11773272603750229, 'train/grad_norm_pre_clip': 3.5503671169281006, 'timing/data': 0.00027994485571980476, 'timing/model': 0.4165014438331127, 'learning_rate/qwen_vl_interface': 7.904051144959432e-06, 'learning_rate/action_model': 7.904051144959431e-05, 'epoch': 0.24})
+[2026-07-28 04:21:43,459][starVLA.training.train_starvla][INFO] - Step 1315, Loss: {'train/loss': 0.06070542335510254, 'train/grad_norm_pre_clip': 2.001801013946533, 'timing/data': 0.000282960943877697, 'timing/model': 0.41614481760188937, 'learning_rate/qwen_vl_interface': 7.900876990206002e-06, 'learning_rate/action_model': 7.900876990206002e-05, 'epoch': 0.24})
+[2026-07-28 04:21:49,544][starVLA.training.train_starvla][INFO] - Step 1316, Loss: {'train/loss': 0.05223235487937927, 'train/grad_norm_pre_clip': 1.9981225728988647, 'timing/data': 0.00027279602363705635, 'timing/model': 0.41451719915494323, 'learning_rate/qwen_vl_interface': 7.897701115327276e-06, 'learning_rate/action_model': 7.897701115327276e-05, 'epoch': 0.24})
+[2026-07-28 04:21:55,438][starVLA.training.train_starvla][INFO] - Step 1317, Loss: {'train/loss': 0.047772955149412155, 'train/grad_norm_pre_clip': 0.7514815926551819, 'timing/data': 0.00028876401484012604, 'timing/model': 0.41898017982020974, 'learning_rate/qwen_vl_interface': 7.894523522384043e-06, 'learning_rate/action_model': 7.894523522384042e-05, 'epoch': 0.24})
+[2026-07-28 04:22:01,521][starVLA.training.train_starvla][INFO] - Step 1318, Loss: {'train/loss': 0.07076424360275269, 'train/grad_norm_pre_clip': 2.1589407920837402, 'timing/data': 0.00025937240570783615, 'timing/model': 0.6527462871745229, 'learning_rate/qwen_vl_interface': 7.891344213438205e-06, 'learning_rate/action_model': 7.891344213438204e-05, 'epoch': 0.24})
+[2026-07-28 04:22:07,519][starVLA.training.train_starvla][INFO] - Step 1319, Loss: {'train/loss': 0.06690357625484467, 'train/grad_norm_pre_clip': 1.6185801029205322, 'timing/data': 0.0002687661908566952, 'timing/model': 0.5817248062230647, 'learning_rate/qwen_vl_interface': 7.888163190552783e-06, 'learning_rate/action_model': 7.888163190552783e-05, 'epoch': 0.24})
+[2026-07-28 04:22:13,369][starVLA.training.train_starvla][INFO] - Step 1320, Loss: {'train/loss': 0.052953578531742096, 'train/grad_norm_pre_clip': 1.4979392290115356, 'timing/data': 0.00034736795350909233, 'timing/model': 0.4158101831562817, 'learning_rate/qwen_vl_interface': 7.88498045579191e-06, 'learning_rate/action_model': 7.88498045579191e-05, 'epoch': 0.24})
+[2026-07-28 04:22:19,235][starVLA.training.train_starvla][INFO] - Step 1321, Loss: {'train/loss': 0.062291525304317474, 'train/grad_norm_pre_clip': 1.8142638206481934, 'timing/data': 0.0002914681099355221, 'timing/model': 0.4161771866492927, 'learning_rate/qwen_vl_interface': 7.881796011220827e-06, 'learning_rate/action_model': 7.881796011220827e-05, 'epoch': 0.24})
+[2026-07-28 04:22:25,106][starVLA.training.train_starvla][INFO] - Step 1322, Loss: {'train/loss': 0.06951209157705307, 'train/grad_norm_pre_clip': 1.3735829591751099, 'timing/data': 0.00026792287826538086, 'timing/model': 0.4160627508535981, 'learning_rate/qwen_vl_interface': 7.878609858905885e-06, 'learning_rate/action_model': 7.878609858905886e-05, 'epoch': 0.24})
+[2026-07-28 04:22:31,196][starVLA.training.train_starvla][INFO] - Step 1323, Loss: {'train/loss': 0.05682423338294029, 'train/grad_norm_pre_clip': 2.6067967414855957, 'timing/data': 0.00026438990607857704, 'timing/model': 0.4142848029732704, 'learning_rate/qwen_vl_interface': 7.875422000914544e-06, 'learning_rate/action_model': 7.875422000914544e-05, 'epoch': 0.24})
+[2026-07-28 04:22:37,024][starVLA.training.train_starvla][INFO] - Step 1324, Loss: {'train/loss': 0.0573689341545105, 'train/grad_norm_pre_clip': 2.806889533996582, 'timing/data': 0.00028576096519827843, 'timing/model': 0.41648849099874496, 'learning_rate/qwen_vl_interface': 7.872232439315371e-06, 'learning_rate/action_model': 7.87223243931537e-05, 'epoch': 0.24})
+[2026-07-28 04:22:42,884][starVLA.training.train_starvla][INFO] - Step 1325, Loss: {'train/loss': 0.04430336132645607, 'train/grad_norm_pre_clip': 1.0762461423873901, 'timing/data': 0.0002704290673136711, 'timing/model': 0.4166954951360822, 'learning_rate/qwen_vl_interface': 7.869041176178038e-06, 'learning_rate/action_model': 7.869041176178037e-05, 'epoch': 0.24})
+[2026-07-28 04:22:48,809][starVLA.training.train_starvla][INFO] - Step 1326, Loss: {'train/loss': 0.05571775138378143, 'train/grad_norm_pre_clip': 1.1705573797225952, 'timing/data': 0.0002833828330039978, 'timing/model': 0.4196038800291717, 'learning_rate/qwen_vl_interface': 7.865848213573316e-06, 'learning_rate/action_model': 7.865848213573317e-05, 'epoch': 0.24})
+[2026-07-28 04:22:54,920][starVLA.training.train_starvla][INFO] - Step 1327, Loss: {'train/loss': 0.08501636981964111, 'train/grad_norm_pre_clip': 1.4009004831314087, 'timing/data': 0.0004162769764661789, 'timing/model': 0.6445322232320905, 'learning_rate/qwen_vl_interface': 7.862653553573092e-06, 'learning_rate/action_model': 7.862653553573092e-05, 'epoch': 0.24})
+[2026-07-28 04:23:00,781][starVLA.training.train_starvla][INFO] - Step 1328, Loss: {'train/loss': 0.04133202135562897, 'train/grad_norm_pre_clip': 1.65904700756073, 'timing/data': 0.0002580038271844387, 'timing/model': 0.4136499152518809, 'learning_rate/qwen_vl_interface': 7.859457198250342e-06, 'learning_rate/action_model': 7.859457198250342e-05, 'epoch': 0.24})
+[2026-07-28 04:23:06,658][starVLA.training.train_starvla][INFO] - Step 1329, Loss: {'train/loss': 0.05589735135436058, 'train/grad_norm_pre_clip': 2.2592830657958984, 'timing/data': 0.0002906979061663151, 'timing/model': 0.4164863689802587, 'learning_rate/qwen_vl_interface': 7.856259149679146e-06, 'learning_rate/action_model': 7.856259149679146e-05, 'epoch': 0.24})
+[2026-07-28 04:23:12,542][starVLA.training.train_starvla][INFO] - Step 1330, Loss: {'train/loss': 0.0695958286523819, 'train/grad_norm_pre_clip': 1.6124916076660156, 'timing/data': 0.0002879449166357517, 'timing/model': 0.4166853451170027, 'learning_rate/qwen_vl_interface': 7.853059409934685e-06, 'learning_rate/action_model': 7.853059409934685e-05, 'epoch': 0.24})
+[2026-07-28 04:23:18,466][starVLA.training.train_starvla][INFO] - Step 1331, Loss: {'train/loss': 0.07406668365001678, 'train/grad_norm_pre_clip': 1.8513081073760986, 'timing/data': 0.0003681671805679798, 'timing/model': 0.41693736193701625, 'learning_rate/qwen_vl_interface': 7.849857981093231e-06, 'learning_rate/action_model': 7.84985798109323e-05, 'epoch': 0.24})
+[2026-07-28 04:23:24,721][starVLA.training.train_starvla][INFO] - Step 1332, Loss: {'train/loss': 0.0813850685954094, 'train/grad_norm_pre_clip': 2.5385513305664062, 'timing/data': 0.000282221008092165, 'timing/model': 0.6647698790766299, 'learning_rate/qwen_vl_interface': 7.84665486523216e-06, 'learning_rate/action_model': 7.846654865232159e-05, 'epoch': 0.24})
+[2026-07-28 04:23:30,952][starVLA.training.train_starvla][INFO] - Step 1333, Loss: {'train/loss': 0.057583484798669815, 'train/grad_norm_pre_clip': 1.7213388681411743, 'timing/data': 0.0002586641348898411, 'timing/model': 0.4154908610507846, 'learning_rate/qwen_vl_interface': 7.84345006442994e-06, 'learning_rate/action_model': 7.843450064429939e-05, 'epoch': 0.24})
+[2026-07-28 04:23:36,706][starVLA.training.train_starvla][INFO] - Step 1334, Loss: {'train/loss': 0.053959861397743225, 'train/grad_norm_pre_clip': 2.0108253955841064, 'timing/data': 0.00028757471591234207, 'timing/model': 0.4144901963882148, 'learning_rate/qwen_vl_interface': 7.840243580766129e-06, 'learning_rate/action_model': 7.840243580766128e-05, 'epoch': 0.25})
+[2026-07-28 04:23:42,571][starVLA.training.train_starvla][INFO] - Step 1335, Loss: {'train/loss': 0.1081247478723526, 'train/grad_norm_pre_clip': 2.873648166656494, 'timing/data': 0.0002649109810590744, 'timing/model': 0.41439661011099815, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 0.25})
+[2026-07-28 04:23:48,430][starVLA.training.train_starvla][INFO] - Step 1336, Loss: {'train/loss': 0.056643806397914886, 'train/grad_norm_pre_clip': 2.112046241760254, 'timing/data': 0.0002732668071985245, 'timing/model': 0.414769824128598, 'learning_rate/qwen_vl_interface': 7.833825573177435e-06, 'learning_rate/action_model': 7.833825573177436e-05, 'epoch': 0.25})
+[2026-07-28 04:23:54,582][starVLA.training.train_starvla][INFO] - Step 1337, Loss: {'train/loss': 0.06920182704925537, 'train/grad_norm_pre_clip': 1.2348891496658325, 'timing/data': 0.00024667009711265564, 'timing/model': 0.41300252778455615, 'learning_rate/qwen_vl_interface': 7.83061405341713e-06, 'learning_rate/action_model': 7.83061405341713e-05, 'epoch': 0.25})
+[2026-07-28 04:24:00,477][starVLA.training.train_starvla][INFO] - Step 1338, Loss: {'train/loss': 0.06882576644420624, 'train/grad_norm_pre_clip': 0.9016134142875671, 'timing/data': 0.00028099119663238525, 'timing/model': 0.42426752811297774, 'learning_rate/qwen_vl_interface': 7.827400859124384e-06, 'learning_rate/action_model': 7.827400859124384e-05, 'epoch': 0.25})
+[2026-07-28 04:24:06,650][starVLA.training.train_starvla][INFO] - Step 1339, Loss: {'train/loss': 0.0718456357717514, 'train/grad_norm_pre_clip': 0.863964319229126, 'timing/data': 0.00026214634999632835, 'timing/model': 0.4146252889186144, 'learning_rate/qwen_vl_interface': 7.8241859923842e-06, 'learning_rate/action_model': 7.8241859923842e-05, 'epoch': 0.25})
+[2026-07-28 04:24:12,830][starVLA.training.train_starvla][INFO] - Step 1340, Loss: {'train/loss': 0.0644407793879509, 'train/grad_norm_pre_clip': 2.5560688972473145, 'timing/data': 0.00027984194457530975, 'timing/model': 0.679427944123745, 'learning_rate/qwen_vl_interface': 7.820969455282676e-06, 'learning_rate/action_model': 7.820969455282675e-05, 'epoch': 0.25})
+[2026-07-28 04:24:18,698][starVLA.training.train_starvla][INFO] - Step 1341, Loss: {'train/loss': 0.057296790182590485, 'train/grad_norm_pre_clip': 1.8063416481018066, 'timing/data': 0.0002735401503741741, 'timing/model': 0.4159567770548165, 'learning_rate/qwen_vl_interface': 7.817751249906982e-06, 'learning_rate/action_model': 7.817751249906983e-05, 'epoch': 0.25})
+[2026-07-28 04:24:24,552][starVLA.training.train_starvla][INFO] - Step 1342, Loss: {'train/loss': 0.035880036652088165, 'train/grad_norm_pre_clip': 2.2720229625701904, 'timing/data': 0.00026497989892959595, 'timing/model': 0.4152076388709247, 'learning_rate/qwen_vl_interface': 7.814531378345381e-06, 'learning_rate/action_model': 7.814531378345381e-05, 'epoch': 0.25})
+[2026-07-28 04:24:30,646][starVLA.training.train_starvla][INFO] - Step 1343, Loss: {'train/loss': 0.037680674344301224, 'train/grad_norm_pre_clip': 1.3186577558517456, 'timing/data': 0.00027869967743754387, 'timing/model': 0.65102693811059, 'learning_rate/qwen_vl_interface': 7.811309842687213e-06, 'learning_rate/action_model': 7.811309842687212e-05, 'epoch': 0.25})
+[2026-07-28 04:24:36,696][starVLA.training.train_starvla][INFO] - Step 1344, Loss: {'train/loss': 0.03592879697680473, 'train/grad_norm_pre_clip': 1.055508017539978, 'timing/data': 0.0002622818574309349, 'timing/model': 0.41377021884545684, 'learning_rate/qwen_vl_interface': 7.808086645022895e-06, 'learning_rate/action_model': 7.808086645022895e-05, 'epoch': 0.25})
+[2026-07-28 04:24:42,548][starVLA.training.train_starvla][INFO] - Step 1345, Loss: {'train/loss': 0.06732822954654694, 'train/grad_norm_pre_clip': 2.0336525440216064, 'timing/data': 0.0002927025780081749, 'timing/model': 0.4153073690831661, 'learning_rate/qwen_vl_interface': 7.804861787443924e-06, 'learning_rate/action_model': 7.804861787443923e-05, 'epoch': 0.25})
+[2026-07-28 04:24:48,426][starVLA.training.train_starvla][INFO] - Step 1346, Loss: {'train/loss': 0.041125088930130005, 'train/grad_norm_pre_clip': 1.5044158697128296, 'timing/data': 0.00028893398120999336, 'timing/model': 0.41666107065975666, 'learning_rate/qwen_vl_interface': 7.801635272042879e-06, 'learning_rate/action_model': 7.801635272042877e-05, 'epoch': 0.25})
+[2026-07-28 04:24:54,726][starVLA.training.train_starvla][INFO] - Step 1347, Loss: {'train/loss': 0.09241641312837601, 'train/grad_norm_pre_clip': 2.9403321743011475, 'timing/data': 0.00026599597185850143, 'timing/model': 0.4154522870667279, 'learning_rate/qwen_vl_interface': 7.798407100913407e-06, 'learning_rate/action_model': 7.798407100913407e-05, 'epoch': 0.25})
+[2026-07-28 04:25:00,347][starVLA.training.train_starvla][INFO] - Step 1348, Loss: {'train/loss': 0.09720338881015778, 'train/grad_norm_pre_clip': 2.830028772354126, 'timing/data': 0.00028455397114157677, 'timing/model': 0.4159337147139013, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.25})
+[2026-07-28 04:25:05,983][starVLA.training.train_starvla][INFO] - Step 1349, Loss: {'train/loss': 0.05089162290096283, 'train/grad_norm_pre_clip': 1.6438546180725098, 'timing/data': 0.0002830629236996174, 'timing/model': 0.41688095312565565, 'learning_rate/qwen_vl_interface': 7.79194579984916e-06, 'learning_rate/action_model': 7.791945799849161e-05, 'epoch': 0.25})
+[2026-07-28 04:25:11,987][starVLA.training.train_starvla][INFO] - Step 1350, Loss: {'train/loss': 0.049280375242233276, 'train/grad_norm_pre_clip': 1.4992057085037231, 'timing/data': 0.00028524501249194145, 'timing/model': 0.5871305521577597, 'learning_rate/qwen_vl_interface': 7.788712674107054e-06, 'learning_rate/action_model': 7.788712674107055e-05, 'epoch': 0.25})
+[2026-07-28 04:25:18,072][starVLA.training.train_starvla][INFO] - Step 1351, Loss: {'train/loss': 0.0498453825712204, 'train/grad_norm_pre_clip': 1.7446342706680298, 'timing/data': 0.0002503800205886364, 'timing/model': 0.4140567728318274, 'learning_rate/qwen_vl_interface': 7.785477901021859e-06, 'learning_rate/action_model': 7.785477901021859e-05, 'epoch': 0.25})
+[2026-07-28 04:25:23,929][starVLA.training.train_starvla][INFO] - Step 1352, Loss: {'train/loss': 0.05468015372753143, 'train/grad_norm_pre_clip': 2.4428155422210693, 'timing/data': 0.000281662680208683, 'timing/model': 0.41600439697504044, 'learning_rate/qwen_vl_interface': 7.78224148269258e-06, 'learning_rate/action_model': 7.78224148269258e-05, 'epoch': 0.25})
+[2026-07-28 04:25:29,800][starVLA.training.train_starvla][INFO] - Step 1353, Loss: {'train/loss': 0.038694508373737335, 'train/grad_norm_pre_clip': 1.5443321466445923, 'timing/data': 0.0002670893445611, 'timing/model': 0.4308557310141623, 'learning_rate/qwen_vl_interface': 7.779003421219293e-06, 'learning_rate/action_model': 7.779003421219293e-05, 'epoch': 0.25})
+[2026-07-28 04:25:36,114][starVLA.training.train_starvla][INFO] - Step 1354, Loss: {'train/loss': 0.0433860719203949, 'train/grad_norm_pre_clip': 1.2814655303955078, 'timing/data': 0.00048322929069399834, 'timing/model': 0.6638018488883972, 'learning_rate/qwen_vl_interface': 7.775763718703144e-06, 'learning_rate/action_model': 7.775763718703144e-05, 'epoch': 0.25})
+[2026-07-28 04:25:41,968][starVLA.training.train_starvla][INFO] - Step 1355, Loss: {'train/loss': 0.0777226984500885, 'train/grad_norm_pre_clip': 2.801913261413574, 'timing/data': 0.0002868799492716789, 'timing/model': 0.4149128431454301, 'learning_rate/qwen_vl_interface': 7.772522377246339e-06, 'learning_rate/action_model': 7.77252237724634e-05, 'epoch': 0.25})
+[2026-07-28 04:25:47,825][starVLA.training.train_starvla][INFO] - Step 1356, Loss: {'train/loss': 0.04002712666988373, 'train/grad_norm_pre_clip': 1.0735805034637451, 'timing/data': 0.0002797199413180351, 'timing/model': 0.4165350808762014, 'learning_rate/qwen_vl_interface': 7.769279398952152e-06, 'learning_rate/action_model': 7.769279398952151e-05, 'epoch': 0.25})
+[2026-07-28 04:25:53,680][starVLA.training.train_starvla][INFO] - Step 1357, Loss: {'train/loss': 0.0638759434223175, 'train/grad_norm_pre_clip': 1.838139295578003, 'timing/data': 0.00028321798890829086, 'timing/model': 0.41615782398730516, 'learning_rate/qwen_vl_interface': 7.766034785924914e-06, 'learning_rate/action_model': 7.766034785924912e-05, 'epoch': 0.25})
+[2026-07-28 04:25:59,740][starVLA.training.train_starvla][INFO] - Step 1358, Loss: {'train/loss': 0.022532671689987183, 'train/grad_norm_pre_clip': 0.5998775362968445, 'timing/data': 0.00025702500715851784, 'timing/model': 0.4138438617810607, 'learning_rate/qwen_vl_interface': 7.762788540270018e-06, 'learning_rate/action_model': 7.762788540270018e-05, 'epoch': 0.25})
+[2026-07-28 04:26:05,875][starVLA.training.train_starvla][INFO] - Step 1359, Loss: {'train/loss': 0.040492527186870575, 'train/grad_norm_pre_clip': 1.4398523569107056, 'timing/data': 0.0002767941914498806, 'timing/model': 0.6582699986174703, 'learning_rate/qwen_vl_interface': 7.759540664093919e-06, 'learning_rate/action_model': 7.759540664093919e-05, 'epoch': 0.25})
+[2026-07-28 04:26:11,743][starVLA.training.train_starvla][INFO] - Step 1360, Loss: {'train/loss': 0.08756230026483536, 'train/grad_norm_pre_clip': 1.6293691396713257, 'timing/data': 0.0002544308081269264, 'timing/model': 0.41422354569658637, 'learning_rate/qwen_vl_interface': 7.756291159504132e-06, 'learning_rate/action_model': 7.756291159504132e-05, 'epoch': 0.25})
+[2026-07-28 04:26:17,597][starVLA.training.train_starvla][INFO] - Step 1361, Loss: {'train/loss': 0.05750438570976257, 'train/grad_norm_pre_clip': 1.4191758632659912, 'timing/data': 0.00027812691405415535, 'timing/model': 0.41848356602713466, 'learning_rate/qwen_vl_interface': 7.75304002860922e-06, 'learning_rate/action_model': 7.75304002860922e-05, 'epoch': 0.25})
+[2026-07-28 04:26:23,476][starVLA.training.train_starvla][INFO] - Step 1362, Loss: {'train/loss': 0.06962744146585464, 'train/grad_norm_pre_clip': 1.7150570154190063, 'timing/data': 0.0002846168354153633, 'timing/model': 0.4161443170160055, 'learning_rate/qwen_vl_interface': 7.749787273518807e-06, 'learning_rate/action_model': 7.749787273518807e-05, 'epoch': 0.25})
+[2026-07-28 04:26:29,566][starVLA.training.train_starvla][INFO] - Step 1363, Loss: {'train/loss': 0.05191522836685181, 'train/grad_norm_pre_clip': 1.4634782075881958, 'timing/data': 0.0002871840260922909, 'timing/model': 0.6521934741176665, 'learning_rate/qwen_vl_interface': 7.746532896343575e-06, 'learning_rate/action_model': 7.746532896343575e-05, 'epoch': 0.25})
+[2026-07-28 04:26:35,416][starVLA.training.train_starvla][INFO] - Step 1364, Loss: {'train/loss': 0.04332683980464935, 'train/grad_norm_pre_clip': 1.2011425495147705, 'timing/data': 0.0002564270980656147, 'timing/model': 0.4180669658817351, 'learning_rate/qwen_vl_interface': 7.743276899195249e-06, 'learning_rate/action_model': 7.743276899195249e-05, 'epoch': 0.25})
+[2026-07-28 04:26:41,520][starVLA.training.train_starvla][INFO] - Step 1365, Loss: {'train/loss': 0.06519506871700287, 'train/grad_norm_pre_clip': 4.387174606323242, 'timing/data': 0.00029645301401615143, 'timing/model': 0.6511725587770343, 'learning_rate/qwen_vl_interface': 7.740019284186612e-06, 'learning_rate/action_model': 7.740019284186612e-05, 'epoch': 0.25})
+[2026-07-28 04:26:47,384][starVLA.training.train_starvla][INFO] - Step 1366, Loss: {'train/loss': 0.07854032516479492, 'train/grad_norm_pre_clip': 3.7965750694274902, 'timing/data': 0.00025138678029179573, 'timing/model': 0.41394607815891504, 'learning_rate/qwen_vl_interface': 7.736760053431496e-06, 'learning_rate/action_model': 7.736760053431495e-05, 'epoch': 0.25})
+[2026-07-28 04:26:53,220][starVLA.training.train_starvla][INFO] - Step 1367, Loss: {'train/loss': 0.01576695404946804, 'train/grad_norm_pre_clip': 0.5435638427734375, 'timing/data': 0.00035325484350323677, 'timing/model': 0.4168088808655739, 'learning_rate/qwen_vl_interface': 7.73349920904478e-06, 'learning_rate/action_model': 7.73349920904478e-05, 'epoch': 0.25})
+[2026-07-28 04:26:59,301][starVLA.training.train_starvla][INFO] - Step 1368, Loss: {'train/loss': 0.04293045401573181, 'train/grad_norm_pre_clip': 1.276350498199463, 'timing/data': 0.00026108231395483017, 'timing/model': 0.41390554793179035, 'learning_rate/qwen_vl_interface': 7.73023675314239e-06, 'learning_rate/action_model': 7.73023675314239e-05, 'epoch': 0.25})
+[2026-07-28 04:27:05,191][starVLA.training.train_starvla][INFO] - Step 1369, Loss: {'train/loss': 0.04071229323744774, 'train/grad_norm_pre_clip': 0.9135720729827881, 'timing/data': 0.0002800440415740013, 'timing/model': 0.4169457722455263, 'learning_rate/qwen_vl_interface': 7.726972687841298e-06, 'learning_rate/action_model': 7.726972687841299e-05, 'epoch': 0.25})
+[2026-07-28 04:27:11,287][starVLA.training.train_starvla][INFO] - Step 1370, Loss: {'train/loss': 0.03617682307958603, 'train/grad_norm_pre_clip': 1.4082754850387573, 'timing/data': 0.00026289140805602074, 'timing/model': 0.4150786017999053, 'learning_rate/qwen_vl_interface': 7.723707015259522e-06, 'learning_rate/action_model': 7.723707015259522e-05, 'epoch': 0.25})
+[2026-07-28 04:27:17,361][starVLA.training.train_starvla][INFO] - Step 1371, Loss: {'train/loss': 0.10352876782417297, 'train/grad_norm_pre_clip': 3.7178854942321777, 'timing/data': 0.00026571983471512794, 'timing/model': 0.4153989013284445, 'learning_rate/qwen_vl_interface': 7.72043973751612e-06, 'learning_rate/action_model': 7.720439737516119e-05, 'epoch': 0.25})
+[2026-07-28 04:27:23,239][starVLA.training.train_starvla][INFO] - Step 1372, Loss: {'train/loss': 0.049775466322898865, 'train/grad_norm_pre_clip': 1.4727619886398315, 'timing/data': 0.00027146562933921814, 'timing/model': 0.4315982679836452, 'learning_rate/qwen_vl_interface': 7.717170856731192e-06, 'learning_rate/action_model': 7.717170856731193e-05, 'epoch': 0.25})
+[2026-07-28 04:27:29,124][starVLA.training.train_starvla][INFO] - Step 1373, Loss: {'train/loss': 0.06567052751779556, 'train/grad_norm_pre_clip': 2.3411478996276855, 'timing/data': 0.00026680296286940575, 'timing/model': 0.41871433705091476, 'learning_rate/qwen_vl_interface': 7.713900375025883e-06, 'learning_rate/action_model': 7.713900375025882e-05, 'epoch': 0.25})
+[2026-07-28 04:27:35,227][starVLA.training.train_starvla][INFO] - Step 1374, Loss: {'train/loss': 0.03787000849843025, 'train/grad_norm_pre_clip': 1.0364289283752441, 'timing/data': 0.00026282062754034996, 'timing/model': 0.41607574466615915, 'learning_rate/qwen_vl_interface': 7.71062829452237e-06, 'learning_rate/action_model': 7.710628294522369e-05, 'epoch': 0.25})
+[2026-07-28 04:27:41,315][starVLA.training.train_starvla][INFO] - Step 1375, Loss: {'train/loss': 0.06462178379297256, 'train/grad_norm_pre_clip': 2.1864125728607178, 'timing/data': 0.0002683298662304878, 'timing/model': 0.6472392077557743, 'learning_rate/qwen_vl_interface': 7.707354617343871e-06, 'learning_rate/action_model': 7.707354617343871e-05, 'epoch': 0.25})
+[2026-07-28 04:27:47,173][starVLA.training.train_starvla][INFO] - Step 1376, Loss: {'train/loss': 0.11371000111103058, 'train/grad_norm_pre_clip': 2.76230788230896, 'timing/data': 0.00026127416640520096, 'timing/model': 0.41397620271891356, 'learning_rate/qwen_vl_interface': 7.704079345614639e-06, 'learning_rate/action_model': 7.704079345614639e-05, 'epoch': 0.25})
+[2026-07-28 04:27:53,032][starVLA.training.train_starvla][INFO] - Step 1377, Loss: {'train/loss': 0.09831051528453827, 'train/grad_norm_pre_clip': 3.4037106037139893, 'timing/data': 0.0002644644118845463, 'timing/model': 0.41678157588467, 'learning_rate/qwen_vl_interface': 7.700802481459964e-06, 'learning_rate/action_model': 7.700802481459964e-05, 'epoch': 0.25})
+[2026-07-28 04:27:58,896][starVLA.training.train_starvla][INFO] - Step 1378, Loss: {'train/loss': 0.0705745667219162, 'train/grad_norm_pre_clip': 1.8024473190307617, 'timing/data': 0.0002759341150522232, 'timing/model': 0.4175012200139463, 'learning_rate/qwen_vl_interface': 7.697524027006168e-06, 'learning_rate/action_model': 7.697524027006167e-05, 'epoch': 0.25})
+[2026-07-28 04:28:05,170][starVLA.training.train_starvla][INFO] - Step 1379, Loss: {'train/loss': 0.0353686660528183, 'train/grad_norm_pre_clip': 1.1425256729125977, 'timing/data': 0.0002744961529970169, 'timing/model': 0.6486493637785316, 'learning_rate/qwen_vl_interface': 7.694243984380603e-06, 'learning_rate/action_model': 7.694243984380603e-05, 'epoch': 0.25})
+[2026-07-28 04:28:11,038][starVLA.training.train_starvla][INFO] - Step 1380, Loss: {'train/loss': 0.0524238683283329, 'train/grad_norm_pre_clip': 1.6834906339645386, 'timing/data': 0.0002535199746489525, 'timing/model': 0.4131140219978988, 'learning_rate/qwen_vl_interface': 7.690962355711651e-06, 'learning_rate/action_model': 7.690962355711652e-05, 'epoch': 0.25})
+[2026-07-28 04:28:17,080][starVLA.training.train_starvla][INFO] - Step 1381, Loss: {'train/loss': 0.06537079811096191, 'train/grad_norm_pre_clip': 1.8907464742660522, 'timing/data': 0.00027471408247947693, 'timing/model': 0.41265362687408924, 'learning_rate/qwen_vl_interface': 7.68767914312873e-06, 'learning_rate/action_model': 7.68767914312873e-05, 'epoch': 0.25})
+[2026-07-28 04:28:22,993][starVLA.training.train_starvla][INFO] - Step 1382, Loss: {'train/loss': 0.06466478109359741, 'train/grad_norm_pre_clip': 2.2650511264801025, 'timing/data': 0.0002654953859746456, 'timing/model': 0.41713323909789324, 'learning_rate/qwen_vl_interface': 7.684394348762278e-06, 'learning_rate/action_model': 7.684394348762277e-05, 'epoch': 0.25})
+[2026-07-28 04:28:28,637][starVLA.training.train_starvla][INFO] - Step 1383, Loss: {'train/loss': 0.07179087400436401, 'train/grad_norm_pre_clip': 1.823795199394226, 'timing/data': 0.00029156729578971863, 'timing/model': 0.41563504841178656, 'learning_rate/qwen_vl_interface': 7.681107974743763e-06, 'learning_rate/action_model': 7.681107974743764e-05, 'epoch': 0.25})
+[2026-07-28 04:28:34,648][starVLA.training.train_starvla][INFO] - Step 1384, Loss: {'train/loss': 0.06619097292423248, 'train/grad_norm_pre_clip': 1.996942162513733, 'timing/data': 0.000268629752099514, 'timing/model': 0.4166360879316926, 'learning_rate/qwen_vl_interface': 7.677820023205676e-06, 'learning_rate/action_model': 7.677820023205677e-05, 'epoch': 0.25})
+[2026-07-28 04:28:40,747][starVLA.training.train_starvla][INFO] - Step 1385, Loss: {'train/loss': 0.041034042835235596, 'train/grad_norm_pre_clip': 1.0214731693267822, 'timing/data': 0.00029114028438925743, 'timing/model': 0.4156022910028696, 'learning_rate/qwen_vl_interface': 7.674530496281536e-06, 'learning_rate/action_model': 7.674530496281536e-05, 'epoch': 0.25})
+[2026-07-28 04:28:46,844][starVLA.training.train_starvla][INFO] - Step 1386, Loss: {'train/loss': 0.06463190913200378, 'train/grad_norm_pre_clip': 1.1100589036941528, 'timing/data': 0.00025748321786522865, 'timing/model': 0.41379303578287363, 'learning_rate/qwen_vl_interface': 7.67123939610588e-06, 'learning_rate/action_model': 7.67123939610588e-05, 'epoch': 0.25})
+[2026-07-28 04:28:52,710][starVLA.training.train_starvla][INFO] - Step 1387, Loss: {'train/loss': 0.054934628307819366, 'train/grad_norm_pre_clip': 0.9524854421615601, 'timing/data': 0.0002753189764916897, 'timing/model': 0.41630866611376405, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.25})
+[2026-07-28 04:28:58,787][starVLA.training.train_starvla][INFO] - Step 1388, Loss: {'train/loss': 0.044808849692344666, 'train/grad_norm_pre_clip': 1.4459861516952515, 'timing/data': 0.0002637510187923908, 'timing/model': 0.41343055199831724, 'learning_rate/qwen_vl_interface': 7.664652484543271e-06, 'learning_rate/action_model': 7.664652484543271e-05, 'epoch': 0.25})
+[2026-07-28 04:29:04,646][starVLA.training.train_starvla][INFO] - Step 1389, Loss: {'train/loss': 0.04081631451845169, 'train/grad_norm_pre_clip': 1.4438650608062744, 'timing/data': 0.00028262799605727196, 'timing/model': 0.41614909283816814, 'learning_rate/qwen_vl_interface': 7.661356677430499e-06, 'learning_rate/action_model': 7.661356677430499e-05, 'epoch': 0.26})
+[2026-07-28 04:29:10,750][starVLA.training.train_starvla][INFO] - Step 1390, Loss: {'train/loss': 0.04306400194764137, 'train/grad_norm_pre_clip': 1.678674578666687, 'timing/data': 0.0003067697398364544, 'timing/model': 0.6527199847623706, 'learning_rate/qwen_vl_interface': 7.658059305614556e-06, 'learning_rate/action_model': 7.658059305614555e-05, 'epoch': 0.26})
+[2026-07-28 04:29:16,738][starVLA.training.train_starvla][INFO] - Step 1391, Loss: {'train/loss': 0.056899555027484894, 'train/grad_norm_pre_clip': 1.853110432624817, 'timing/data': 0.00026919180527329445, 'timing/model': 0.5843780771829188, 'learning_rate/qwen_vl_interface': 7.654760371235073e-06, 'learning_rate/action_model': 7.654760371235074e-05, 'epoch': 0.26})
+[2026-07-28 04:29:22,596][starVLA.training.train_starvla][INFO] - Step 1392, Loss: {'train/loss': 0.059551045298576355, 'train/grad_norm_pre_clip': 1.8919140100479126, 'timing/data': 0.00028798170387744904, 'timing/model': 0.4161244258284569, 'learning_rate/qwen_vl_interface': 7.651459876432694e-06, 'learning_rate/action_model': 7.651459876432694e-05, 'epoch': 0.26})
+[2026-07-28 04:29:28,453][starVLA.training.train_starvla][INFO] - Step 1393, Loss: {'train/loss': 0.05695067346096039, 'train/grad_norm_pre_clip': 1.4587751626968384, 'timing/data': 0.00026706233620643616, 'timing/model': 0.4159750263206661, 'learning_rate/qwen_vl_interface': 7.648157823349074e-06, 'learning_rate/action_model': 7.648157823349074e-05, 'epoch': 0.26})
+[2026-07-28 04:29:34,540][starVLA.training.train_starvla][INFO] - Step 1394, Loss: {'train/loss': 0.10228000581264496, 'train/grad_norm_pre_clip': 6.401208400726318, 'timing/data': 0.0002514738589525223, 'timing/model': 0.41349736927077174, 'learning_rate/qwen_vl_interface': 7.64485421412688e-06, 'learning_rate/action_model': 7.644854214126879e-05, 'epoch': 0.26})
+[2026-07-28 04:29:40,662][starVLA.training.train_starvla][INFO] - Step 1395, Loss: {'train/loss': 0.04304563254117966, 'train/grad_norm_pre_clip': 1.3817909955978394, 'timing/data': 0.00030411314219236374, 'timing/model': 0.4168617529794574, 'learning_rate/qwen_vl_interface': 7.641549050909784e-06, 'learning_rate/action_model': 7.641549050909784e-05, 'epoch': 0.26})
+[2026-07-28 04:29:46,533][starVLA.training.train_starvla][INFO] - Step 1396, Loss: {'train/loss': 0.03921672701835632, 'train/grad_norm_pre_clip': 0.6975347399711609, 'timing/data': 0.00026889704167842865, 'timing/model': 0.41603750735521317, 'learning_rate/qwen_vl_interface': 7.638242335842476e-06, 'learning_rate/action_model': 7.638242335842476e-05, 'epoch': 0.26})
+[2026-07-28 04:29:52,385][starVLA.training.train_starvla][INFO] - Step 1397, Loss: {'train/loss': 0.09776104986667633, 'train/grad_norm_pre_clip': 0.9825574159622192, 'timing/data': 0.0002604657784104347, 'timing/model': 0.41520152194425464, 'learning_rate/qwen_vl_interface': 7.634934071070643e-06, 'learning_rate/action_model': 7.634934071070643e-05, 'epoch': 0.26})
+[2026-07-28 04:29:58,444][starVLA.training.train_starvla][INFO] - Step 1398, Loss: {'train/loss': 0.036268237978219986, 'train/grad_norm_pre_clip': 1.0847383737564087, 'timing/data': 0.00025572115555405617, 'timing/model': 0.4136259383521974, 'learning_rate/qwen_vl_interface': 7.631624258740985e-06, 'learning_rate/action_model': 7.631624258740984e-05, 'epoch': 0.26})
+[2026-07-28 04:30:04,530][starVLA.training.train_starvla][INFO] - Step 1399, Loss: {'train/loss': 0.047112252563238144, 'train/grad_norm_pre_clip': 1.5143028497695923, 'timing/data': 0.00024985335767269135, 'timing/model': 0.41495123505592346, 'learning_rate/qwen_vl_interface': 7.628312901001203e-06, 'learning_rate/action_model': 7.628312901001204e-05, 'epoch': 0.26})
+[2026-07-28 04:30:10,383][starVLA.training.train_starvla][INFO] - Step 1400, Loss: {'train/loss': 0.0617985874414444, 'train/grad_norm_pre_clip': 1.208808422088623, 'timing/data': 0.00027862610295414925, 'timing/model': 0.416818433906883, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 0.26})
+[2026-07-28 04:30:16,456][starVLA.training.train_starvla][INFO] - Step 1401, Loss: {'train/loss': 0.016776790842413902, 'train/grad_norm_pre_clip': 0.45354002714157104, 'timing/data': 0.00028823595494031906, 'timing/model': 0.6576471817679703, 'learning_rate/qwen_vl_interface': 7.621685557887083e-06, 'learning_rate/action_model': 7.621685557887083e-05, 'epoch': 0.26})
+[2026-07-28 04:30:22,300][starVLA.training.train_starvla][INFO] - Step 1402, Loss: {'train/loss': 0.05419706553220749, 'train/grad_norm_pre_clip': 0.9241971969604492, 'timing/data': 0.0002714758738875389, 'timing/model': 0.41493123304098845, 'learning_rate/qwen_vl_interface': 7.618369576813157e-06, 'learning_rate/action_model': 7.618369576813157e-05, 'epoch': 0.26})
+[2026-07-28 04:30:28,153][starVLA.training.train_starvla][INFO] - Step 1403, Loss: {'train/loss': 0.05523303896188736, 'train/grad_norm_pre_clip': 2.861121416091919, 'timing/data': 0.00027923472225666046, 'timing/model': 0.4160084458999336, 'learning_rate/qwen_vl_interface': 7.615052058929924e-06, 'learning_rate/action_model': 7.615052058929923e-05, 'epoch': 0.26})
+[2026-07-28 04:30:34,235][starVLA.training.train_starvla][INFO] - Step 1404, Loss: {'train/loss': 0.04316668212413788, 'train/grad_norm_pre_clip': 0.8866640329360962, 'timing/data': 0.00027526309713721275, 'timing/model': 0.650456827133894, 'learning_rate/qwen_vl_interface': 7.611733006390087e-06, 'learning_rate/action_model': 7.611733006390086e-05, 'epoch': 0.26})
+[2026-07-28 04:30:40,069][starVLA.training.train_starvla][INFO] - Step 1405, Loss: {'train/loss': 0.06033872812986374, 'train/grad_norm_pre_clip': 2.9190258979797363, 'timing/data': 0.0002607707865536213, 'timing/model': 0.4142100941389799, 'learning_rate/qwen_vl_interface': 7.608412421347344e-06, 'learning_rate/action_model': 7.608412421347344e-05, 'epoch': 0.26})
+[2026-07-28 04:30:46,161][starVLA.training.train_starvla][INFO] - Step 1406, Loss: {'train/loss': 0.07467779517173767, 'train/grad_norm_pre_clip': 1.2107609510421753, 'timing/data': 0.0003247731365263462, 'timing/model': 0.6572236833162606, 'learning_rate/qwen_vl_interface': 7.605090305956384e-06, 'learning_rate/action_model': 7.605090305956383e-05, 'epoch': 0.26})
+[2026-07-28 04:30:52,011][starVLA.training.train_starvla][INFO] - Step 1407, Loss: {'train/loss': 0.029045604169368744, 'train/grad_norm_pre_clip': 1.58022141456604, 'timing/data': 0.00025739800184965134, 'timing/model': 0.4140563551336527, 'learning_rate/qwen_vl_interface': 7.6017666623728954e-06, 'learning_rate/action_model': 7.601766662372896e-05, 'epoch': 0.26})
+[2026-07-28 04:30:57,820][starVLA.training.train_starvla][INFO] - Step 1408, Loss: {'train/loss': 0.043952472507953644, 'train/grad_norm_pre_clip': 1.447800636291504, 'timing/data': 0.0005372310988605022, 'timing/model': 0.41555286990478635, 'learning_rate/qwen_vl_interface': 7.598441492753551e-06, 'learning_rate/action_model': 7.598441492753551e-05, 'epoch': 0.26})
+[2026-07-28 04:31:03,907][starVLA.training.train_starvla][INFO] - Step 1409, Loss: {'train/loss': 0.04985298961400986, 'train/grad_norm_pre_clip': 1.8239860534667969, 'timing/data': 0.0002581826411187649, 'timing/model': 0.4133233791217208, 'learning_rate/qwen_vl_interface': 7.595114799256018e-06, 'learning_rate/action_model': 7.595114799256017e-05, 'epoch': 0.26})
+[2026-07-28 04:31:09,750][starVLA.training.train_starvla][INFO] - Step 1410, Loss: {'train/loss': 0.03925967589020729, 'train/grad_norm_pre_clip': 0.6061420440673828, 'timing/data': 0.00027184607461094856, 'timing/model': 0.4152747089974582, 'learning_rate/qwen_vl_interface': 7.591786584038954e-06, 'learning_rate/action_model': 7.591786584038954e-05, 'epoch': 0.26})
+[2026-07-28 04:31:15,590][starVLA.training.train_starvla][INFO] - Step 1411, Loss: {'train/loss': 0.07420410215854645, 'train/grad_norm_pre_clip': 1.2425799369812012, 'timing/data': 0.000262577086687088, 'timing/model': 0.41698189079761505, 'learning_rate/qwen_vl_interface': 7.588456849261999e-06, 'learning_rate/action_model': 7.588456849261998e-05, 'epoch': 0.26})
+[2026-07-28 04:31:21,660][starVLA.training.train_starvla][INFO] - Step 1412, Loss: {'train/loss': 0.07159950584173203, 'train/grad_norm_pre_clip': 2.3890087604522705, 'timing/data': 0.00032934825867414474, 'timing/model': 0.6452173041179776, 'learning_rate/qwen_vl_interface': 7.585125597085784e-06, 'learning_rate/action_model': 7.585125597085783e-05, 'epoch': 0.26})
+[2026-07-28 04:31:27,501][starVLA.training.train_starvla][INFO] - Step 1413, Loss: {'train/loss': 0.06955507397651672, 'train/grad_norm_pre_clip': 2.266984462738037, 'timing/data': 0.000249532051384449, 'timing/model': 0.41343063907697797, 'learning_rate/qwen_vl_interface': 7.58179282967192e-06, 'learning_rate/action_model': 7.58179282967192e-05, 'epoch': 0.26})
+[2026-07-28 04:31:33,342][starVLA.training.train_starvla][INFO] - Step 1414, Loss: {'train/loss': 0.038593851029872894, 'train/grad_norm_pre_clip': 1.0695115327835083, 'timing/data': 0.0002711308188736439, 'timing/model': 0.41666733799502254, 'learning_rate/qwen_vl_interface': 7.578458549183005e-06, 'learning_rate/action_model': 7.578458549183005e-05, 'epoch': 0.26})
+[2026-07-28 04:31:39,185][starVLA.training.train_starvla][INFO] - Step 1415, Loss: {'train/loss': 0.08123249560594559, 'train/grad_norm_pre_clip': 2.702941417694092, 'timing/data': 0.00027928873896598816, 'timing/model': 0.41468540811911225, 'learning_rate/qwen_vl_interface': 7.575122757782617e-06, 'learning_rate/action_model': 7.575122757782617e-05, 'epoch': 0.26})
+[2026-07-28 04:31:45,290][starVLA.training.train_starvla][INFO] - Step 1416, Loss: {'train/loss': 0.08166911453008652, 'train/grad_norm_pre_clip': 2.0175914764404297, 'timing/data': 0.0002735690213739872, 'timing/model': 0.6461751386523247, 'learning_rate/qwen_vl_interface': 7.571785457635316e-06, 'learning_rate/action_model': 7.571785457635317e-05, 'epoch': 0.26})
+[2026-07-28 04:31:51,440][starVLA.training.train_starvla][INFO] - Step 1417, Loss: {'train/loss': 0.08691483736038208, 'train/grad_norm_pre_clip': 1.5381443500518799, 'timing/data': 0.0002448637969791889, 'timing/model': 0.4140539267100394, 'learning_rate/qwen_vl_interface': 7.568446650906637e-06, 'learning_rate/action_model': 7.568446650906636e-05, 'epoch': 0.26})
+[2026-07-28 04:31:57,251][starVLA.training.train_starvla][INFO] - Step 1418, Loss: {'train/loss': 0.02713688835501671, 'train/grad_norm_pre_clip': 1.0273832082748413, 'timing/data': 0.00025947205722332, 'timing/model': 0.4169469950720668, 'learning_rate/qwen_vl_interface': 7.565106339763099e-06, 'learning_rate/action_model': 7.565106339763099e-05, 'epoch': 0.26})
+[2026-07-28 04:32:02,897][starVLA.training.train_starvla][INFO] - Step 1419, Loss: {'train/loss': 0.06434288620948792, 'train/grad_norm_pre_clip': 1.1906384229660034, 'timing/data': 0.00027501583099365234, 'timing/model': 0.418477485422045, 'learning_rate/qwen_vl_interface': 7.561764526372191e-06, 'learning_rate/action_model': 7.561764526372191e-05, 'epoch': 0.26})
+[2026-07-28 04:32:08,536][starVLA.training.train_starvla][INFO] - Step 1420, Loss: {'train/loss': 0.021839594468474388, 'train/grad_norm_pre_clip': 0.7558386921882629, 'timing/data': 0.0002808677963912487, 'timing/model': 0.4175164932385087, 'learning_rate/qwen_vl_interface': 7.5584212129023796e-06, 'learning_rate/action_model': 7.558421212902379e-05, 'epoch': 0.26})
+[2026-07-28 04:32:14,813][starVLA.training.train_starvla][INFO] - Step 1421, Loss: {'train/loss': 0.042458683252334595, 'train/grad_norm_pre_clip': 0.9831687808036804, 'timing/data': 0.00027335062623023987, 'timing/model': 0.4150189608335495, 'learning_rate/qwen_vl_interface': 7.555076401523106e-06, 'learning_rate/action_model': 7.555076401523105e-05, 'epoch': 0.26})
+[2026-07-28 04:32:20,976][starVLA.training.train_starvla][INFO] - Step 1422, Loss: {'train/loss': 0.03468755632638931, 'train/grad_norm_pre_clip': 1.5377991199493408, 'timing/data': 0.00033841608092188835, 'timing/model': 0.4126850310713053, 'learning_rate/qwen_vl_interface': 7.551730094404782e-06, 'learning_rate/action_model': 7.551730094404781e-05, 'epoch': 0.26})
+[2026-07-28 04:32:26,841][starVLA.training.train_starvla][INFO] - Step 1423, Loss: {'train/loss': 0.06153709441423416, 'train/grad_norm_pre_clip': 1.7350759506225586, 'timing/data': 0.0003787982277572155, 'timing/model': 0.4158981190994382, 'learning_rate/qwen_vl_interface': 7.5483822937187876e-06, 'learning_rate/action_model': 7.548382293718787e-05, 'epoch': 0.26})
+[2026-07-28 04:32:32,724][starVLA.training.train_starvla][INFO] - Step 1424, Loss: {'train/loss': 0.06393464654684067, 'train/grad_norm_pre_clip': 1.5425212383270264, 'timing/data': 0.0003095921128988266, 'timing/model': 0.41408306499943137, 'learning_rate/qwen_vl_interface': 7.545033001637478e-06, 'learning_rate/action_model': 7.545033001637479e-05, 'epoch': 0.26})
+[2026-07-28 04:32:38,592][starVLA.training.train_starvla][INFO] - Step 1425, Loss: {'train/loss': 0.04862731695175171, 'train/grad_norm_pre_clip': 1.4100443124771118, 'timing/data': 0.00029153376817703247, 'timing/model': 0.41565119102597237, 'learning_rate/qwen_vl_interface': 7.54168222033417e-06, 'learning_rate/action_model': 7.54168222033417e-05, 'epoch': 0.26})
+[2026-07-28 04:32:44,749][starVLA.training.train_starvla][INFO] - Step 1426, Loss: {'train/loss': 0.06275111436843872, 'train/grad_norm_pre_clip': 1.9558855295181274, 'timing/data': 0.00039424607530236244, 'timing/model': 0.7037713029421866, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.26})
+[2026-07-28 04:32:50,629][starVLA.training.train_starvla][INFO] - Step 1427, Loss: {'train/loss': 0.022449972108006477, 'train/grad_norm_pre_clip': 1.2724213600158691, 'timing/data': 0.0002688108943402767, 'timing/model': 0.41599795827642083, 'learning_rate/qwen_vl_interface': 7.534976198759667e-06, 'learning_rate/action_model': 7.534976198759666e-05, 'epoch': 0.26})
+[2026-07-28 04:32:56,509][starVLA.training.train_starvla][INFO] - Step 1428, Loss: {'train/loss': 0.029529914259910583, 'train/grad_norm_pre_clip': 1.8981919288635254, 'timing/data': 0.00048432406038045883, 'timing/model': 0.4195191818289459, 'learning_rate/qwen_vl_interface': 7.531620962839935e-06, 'learning_rate/action_model': 7.531620962839935e-05, 'epoch': 0.26})
+[2026-07-28 04:33:02,668][starVLA.training.train_starvla][INFO] - Step 1429, Loss: {'train/loss': 0.07216425985097885, 'train/grad_norm_pre_clip': 2.4738900661468506, 'timing/data': 0.00026639318093657494, 'timing/model': 0.4161388650536537, 'learning_rate/qwen_vl_interface': 7.5282642464011325e-06, 'learning_rate/action_model': 7.528264246401133e-05, 'epoch': 0.26})
+[2026-07-28 04:33:08,588][starVLA.training.train_starvla][INFO] - Step 1430, Loss: {'train/loss': 0.038254816085100174, 'train/grad_norm_pre_clip': 1.5499154329299927, 'timing/data': 0.0002740751951932907, 'timing/model': 0.4209775160998106, 'learning_rate/qwen_vl_interface': 7.524906051621397e-06, 'learning_rate/action_model': 7.524906051621396e-05, 'epoch': 0.26})
+[2026-07-28 04:33:14,451][starVLA.training.train_starvla][INFO] - Step 1431, Loss: {'train/loss': 0.0842963308095932, 'train/grad_norm_pre_clip': 3.2321481704711914, 'timing/data': 0.00027162907645106316, 'timing/model': 0.41525028785690665, 'learning_rate/qwen_vl_interface': 7.521546380679822e-06, 'learning_rate/action_model': 7.521546380679822e-05, 'epoch': 0.26})
+[2026-07-28 04:33:20,497][starVLA.training.train_starvla][INFO] - Step 1432, Loss: {'train/loss': 0.07136593759059906, 'train/grad_norm_pre_clip': 3.3288729190826416, 'timing/data': 0.0003619398921728134, 'timing/model': 0.5856923731043935, 'learning_rate/qwen_vl_interface': 7.518185235756461e-06, 'learning_rate/action_model': 7.518185235756461e-05, 'epoch': 0.26})
+[2026-07-28 04:33:26,583][starVLA.training.train_starvla][INFO] - Step 1433, Loss: {'train/loss': 0.03411068394780159, 'train/grad_norm_pre_clip': 1.4575998783111572, 'timing/data': 0.00026134075596928596, 'timing/model': 0.41404469730332494, 'learning_rate/qwen_vl_interface': 7.514822619032328e-06, 'learning_rate/action_model': 7.514822619032327e-05, 'epoch': 0.26})
+[2026-07-28 04:33:32,466][starVLA.training.train_starvla][INFO] - Step 1434, Loss: {'train/loss': 0.08588658273220062, 'train/grad_norm_pre_clip': 2.818162679672241, 'timing/data': 0.00039192521944642067, 'timing/model': 0.4168418934568763, 'learning_rate/qwen_vl_interface': 7.511458532689385e-06, 'learning_rate/action_model': 7.511458532689385e-05, 'epoch': 0.26})
+[2026-07-28 04:33:38,392][starVLA.training.train_starvla][INFO] - Step 1435, Loss: {'train/loss': 0.044179484248161316, 'train/grad_norm_pre_clip': 1.9509905576705933, 'timing/data': 0.0005048909224569798, 'timing/model': 0.4171144189313054, 'learning_rate/qwen_vl_interface': 7.508092978910553e-06, 'learning_rate/action_model': 7.508092978910552e-05, 'epoch': 0.26})
+[2026-07-28 04:33:44,551][starVLA.training.train_starvla][INFO] - Step 1436, Loss: {'train/loss': 0.0526384636759758, 'train/grad_norm_pre_clip': 1.7885240316390991, 'timing/data': 0.0002893097698688507, 'timing/model': 0.41614071372896433, 'learning_rate/qwen_vl_interface': 7.5047259598797015e-06, 'learning_rate/action_model': 7.504725959879701e-05, 'epoch': 0.26})
+[2026-07-28 04:33:50,812][starVLA.training.train_starvla][INFO] - Step 1437, Loss: {'train/loss': 0.059592895209789276, 'train/grad_norm_pre_clip': 4.356028079986572, 'timing/data': 0.00030743610113859177, 'timing/model': 0.7096836781129241, 'learning_rate/qwen_vl_interface': 7.501357477781654e-06, 'learning_rate/action_model': 7.501357477781654e-05, 'epoch': 0.26})
+[2026-07-28 04:33:56,715][starVLA.training.train_starvla][INFO] - Step 1438, Loss: {'train/loss': 0.06266804784536362, 'train/grad_norm_pre_clip': 2.4306061267852783, 'timing/data': 0.00029452424496412277, 'timing/model': 0.4173833900131285, 'learning_rate/qwen_vl_interface': 7.497987534802185e-06, 'learning_rate/action_model': 7.497987534802184e-05, 'epoch': 0.26})
+[2026-07-28 04:34:02,597][starVLA.training.train_starvla][INFO] - Step 1439, Loss: {'train/loss': 0.07313632220029831, 'train/grad_norm_pre_clip': 3.1214613914489746, 'timing/data': 0.00030109379440546036, 'timing/model': 0.41802626475691795, 'learning_rate/qwen_vl_interface': 7.494616133128008e-06, 'learning_rate/action_model': 7.494616133128008e-05, 'epoch': 0.26})
+[2026-07-28 04:34:08,706][starVLA.training.train_starvla][INFO] - Step 1440, Loss: {'train/loss': 0.04536227881908417, 'train/grad_norm_pre_clip': 2.0916645526885986, 'timing/data': 0.00028699496760964394, 'timing/model': 0.41605680203065276, 'learning_rate/qwen_vl_interface': 7.49124327494679e-06, 'learning_rate/action_model': 7.49124327494679e-05, 'epoch': 0.26})
+[2026-07-28 04:34:14,895][starVLA.training.train_starvla][INFO] - Step 1441, Loss: {'train/loss': 0.04229004681110382, 'train/grad_norm_pre_clip': 1.2098191976547241, 'timing/data': 0.00027827685698866844, 'timing/model': 0.4169074180535972, 'learning_rate/qwen_vl_interface': 7.487868962447147e-06, 'learning_rate/action_model': 7.487868962447147e-05, 'epoch': 0.26})
+[2026-07-28 04:34:20,557][starVLA.training.train_starvla][INFO] - Step 1442, Loss: {'train/loss': 0.04572300985455513, 'train/grad_norm_pre_clip': 1.4442214965820312, 'timing/data': 0.0003310828469693661, 'timing/model': 0.41674160910770297, 'learning_rate/qwen_vl_interface': 7.48449319781863e-06, 'learning_rate/action_model': 7.484493197818629e-05, 'epoch': 0.26})
+[2026-07-28 04:34:26,442][starVLA.training.train_starvla][INFO] - Step 1443, Loss: {'train/loss': 0.03279880806803703, 'train/grad_norm_pre_clip': 1.6008142232894897, 'timing/data': 0.00029332004487514496, 'timing/model': 0.41830924013629556, 'learning_rate/qwen_vl_interface': 7.481115983251735e-06, 'learning_rate/action_model': 7.481115983251735e-05, 'epoch': 0.27})
+[2026-07-28 04:34:32,323][starVLA.training.train_starvla][INFO] - Step 1444, Loss: {'train/loss': 0.054525651037693024, 'train/grad_norm_pre_clip': 1.5016593933105469, 'timing/data': 0.0002989047206938267, 'timing/model': 0.4164835256524384, 'learning_rate/qwen_vl_interface': 7.477737320937902e-06, 'learning_rate/action_model': 7.477737320937902e-05, 'epoch': 0.27})
+[2026-07-28 04:34:38,723][starVLA.training.train_starvla][INFO] - Step 1445, Loss: {'train/loss': 0.05245119333267212, 'train/grad_norm_pre_clip': 1.327805757522583, 'timing/data': 0.0002742726355791092, 'timing/model': 0.6480092480778694, 'learning_rate/qwen_vl_interface': 7.474357213069507e-06, 'learning_rate/action_model': 7.474357213069507e-05, 'epoch': 0.27})
+[2026-07-28 04:34:44,619][starVLA.training.train_starvla][INFO] - Step 1446, Loss: {'train/loss': 0.07128705084323883, 'train/grad_norm_pre_clip': 2.6055517196655273, 'timing/data': 0.00027313921600580215, 'timing/model': 0.4143796903081238, 'learning_rate/qwen_vl_interface': 7.470975661839868e-06, 'learning_rate/action_model': 7.470975661839868e-05, 'epoch': 0.27})
+[2026-07-28 04:34:50,509][starVLA.training.train_starvla][INFO] - Step 1447, Loss: {'train/loss': 0.06413733959197998, 'train/grad_norm_pre_clip': 1.0053459405899048, 'timing/data': 0.0002793408930301666, 'timing/model': 0.41690453700721264, 'learning_rate/qwen_vl_interface': 7.4675926694432336e-06, 'learning_rate/action_model': 7.467592669443233e-05, 'epoch': 0.27})
+[2026-07-28 04:34:56,615][starVLA.training.train_starvla][INFO] - Step 1448, Loss: {'train/loss': 0.04817748814821243, 'train/grad_norm_pre_clip': 1.0928484201431274, 'timing/data': 0.0002752849832177162, 'timing/model': 0.6591302738524973, 'learning_rate/qwen_vl_interface': 7.4642082380747905e-06, 'learning_rate/action_model': 7.46420823807479e-05, 'epoch': 0.27})
+[2026-07-28 04:35:02,496][starVLA.training.train_starvla][INFO] - Step 1449, Loss: {'train/loss': 0.05651750788092613, 'train/grad_norm_pre_clip': 1.1467368602752686, 'timing/data': 0.00028910301625728607, 'timing/model': 0.4480935428291559, 'learning_rate/qwen_vl_interface': 7.460822369930664e-06, 'learning_rate/action_model': 7.460822369930664e-05, 'epoch': 0.27})
+[2026-07-28 04:35:08,629][starVLA.training.train_starvla][INFO] - Step 1450, Loss: {'train/loss': 0.06296998262405396, 'train/grad_norm_pre_clip': 1.2354481220245361, 'timing/data': 0.0003324029967188835, 'timing/model': 0.4150256011635065, 'learning_rate/qwen_vl_interface': 7.457435067207901e-06, 'learning_rate/action_model': 7.4574350672079e-05, 'epoch': 0.27})
+[2026-07-28 04:35:14,544][starVLA.training.train_starvla][INFO] - Step 1451, Loss: {'train/loss': 0.05479327589273453, 'train/grad_norm_pre_clip': 1.966086745262146, 'timing/data': 0.00030059507116675377, 'timing/model': 0.41602054610848427, 'learning_rate/qwen_vl_interface': 7.454046332104491e-06, 'learning_rate/action_model': 7.45404633210449e-05, 'epoch': 0.27})
+[2026-07-28 04:35:20,397][starVLA.training.train_starvla][INFO] - Step 1452, Loss: {'train/loss': 0.06376366317272186, 'train/grad_norm_pre_clip': 1.4845603704452515, 'timing/data': 0.00035901786759495735, 'timing/model': 0.4187920270487666, 'learning_rate/qwen_vl_interface': 7.450656166819344e-06, 'learning_rate/action_model': 7.450656166819344e-05, 'epoch': 0.27})
+[2026-07-28 04:35:26,730][starVLA.training.train_starvla][INFO] - Step 1453, Loss: {'train/loss': 0.0727677047252655, 'train/grad_norm_pre_clip': 2.44051194190979, 'timing/data': 0.0002604159526526928, 'timing/model': 0.6559169148094952, 'learning_rate/qwen_vl_interface': 7.447264573552303e-06, 'learning_rate/action_model': 7.447264573552302e-05, 'epoch': 0.27})
+[2026-07-28 04:35:32,608][starVLA.training.train_starvla][INFO] - Step 1454, Loss: {'train/loss': 0.08836868405342102, 'train/grad_norm_pre_clip': 1.9424371719360352, 'timing/data': 0.00026318384334445, 'timing/model': 0.41479938430711627, 'learning_rate/qwen_vl_interface': 7.443871554504135e-06, 'learning_rate/action_model': 7.443871554504134e-05, 'epoch': 0.27})
+[2026-07-28 04:35:38,464][starVLA.training.train_starvla][INFO] - Step 1455, Loss: {'train/loss': 0.09055222570896149, 'train/grad_norm_pre_clip': 2.0419187545776367, 'timing/data': 0.00028157420456409454, 'timing/model': 0.4177883439697325, 'learning_rate/qwen_vl_interface': 7.440477111876535e-06, 'learning_rate/action_model': 7.440477111876535e-05, 'epoch': 0.27})
+[2026-07-28 04:35:44,337][starVLA.training.train_starvla][INFO] - Step 1456, Loss: {'train/loss': 0.037469282746315, 'train/grad_norm_pre_clip': 1.5004945993423462, 'timing/data': 0.0003064079210162163, 'timing/model': 0.41678940411657095, 'learning_rate/qwen_vl_interface': 7.437081247872118e-06, 'learning_rate/action_model': 7.437081247872118e-05, 'epoch': 0.27})
+[2026-07-28 04:35:50,500][starVLA.training.train_starvla][INFO] - Step 1457, Loss: {'train/loss': 0.0489397831261158, 'train/grad_norm_pre_clip': 1.1211488246917725, 'timing/data': 0.00029443902894854546, 'timing/model': 0.6506733191199601, 'learning_rate/qwen_vl_interface': 7.433683964694425e-06, 'learning_rate/action_model': 7.433683964694425e-05, 'epoch': 0.27})
+[2026-07-28 04:35:56,367][starVLA.training.train_starvla][INFO] - Step 1458, Loss: {'train/loss': 0.06823350489139557, 'train/grad_norm_pre_clip': 1.2976950407028198, 'timing/data': 0.0002605295740067959, 'timing/model': 0.414403953589499, 'learning_rate/qwen_vl_interface': 7.430285264547913e-06, 'learning_rate/action_model': 7.430285264547913e-05, 'epoch': 0.27})
+[2026-07-28 04:36:02,470][starVLA.training.train_starvla][INFO] - Step 1459, Loss: {'train/loss': 0.06975148618221283, 'train/grad_norm_pre_clip': 1.213076114654541, 'timing/data': 0.0002984199672937393, 'timing/model': 0.6597842983901501, 'learning_rate/qwen_vl_interface': 7.426885149637964e-06, 'learning_rate/action_model': 7.426885149637965e-05, 'epoch': 0.27})
+[2026-07-28 04:36:08,347][starVLA.training.train_starvla][INFO] - Step 1460, Loss: {'train/loss': 0.037493083626031876, 'train/grad_norm_pre_clip': 1.0514235496520996, 'timing/data': 0.00026734964922070503, 'timing/model': 0.41392535204067826, 'learning_rate/qwen_vl_interface': 7.42348362217088e-06, 'learning_rate/action_model': 7.423483622170879e-05, 'epoch': 0.27})
+[2026-07-28 04:36:14,236][starVLA.training.train_starvla][INFO] - Step 1461, Loss: {'train/loss': 0.0687505304813385, 'train/grad_norm_pre_clip': 2.2340822219848633, 'timing/data': 0.00029236311092972755, 'timing/model': 0.4202754069119692, 'learning_rate/qwen_vl_interface': 7.420080684353865e-06, 'learning_rate/action_model': 7.420080684353865e-05, 'epoch': 0.27})
+[2026-07-28 04:36:20,143][starVLA.training.train_starvla][INFO] - Step 1462, Loss: {'train/loss': 0.031938109546899796, 'train/grad_norm_pre_clip': 1.4672223329544067, 'timing/data': 0.0002856752835214138, 'timing/model': 0.4168787728995085, 'learning_rate/qwen_vl_interface': 7.416676338395057e-06, 'learning_rate/action_model': 7.416676338395057e-05, 'epoch': 0.27})
+[2026-07-28 04:36:26,239][starVLA.training.train_starvla][INFO] - Step 1463, Loss: {'train/loss': 0.10450667142868042, 'train/grad_norm_pre_clip': 1.8258178234100342, 'timing/data': 0.00026703299954533577, 'timing/model': 0.41748176282271743, 'learning_rate/qwen_vl_interface': 7.413270586503496e-06, 'learning_rate/action_model': 7.413270586503496e-05, 'epoch': 0.27})
+[2026-07-28 04:36:32,333][starVLA.training.train_starvla][INFO] - Step 1464, Loss: {'train/loss': 0.06468024104833603, 'train/grad_norm_pre_clip': 5.026353359222412, 'timing/data': 0.00026225531473755836, 'timing/model': 0.4148603971116245, 'learning_rate/qwen_vl_interface': 7.409863430889137e-06, 'learning_rate/action_model': 7.409863430889136e-05, 'epoch': 0.27})
+[2026-07-28 04:36:38,176][starVLA.training.train_starvla][INFO] - Step 1465, Loss: {'train/loss': 0.04307322949171066, 'train/grad_norm_pre_clip': 2.443103790283203, 'timing/data': 0.0002783401869237423, 'timing/model': 0.4169076350517571, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.27})
+[2026-07-28 04:36:44,028][starVLA.training.train_starvla][INFO] - Step 1466, Loss: {'train/loss': 0.0299028679728508, 'train/grad_norm_pre_clip': 1.45672607421875, 'timing/data': 0.00031513068825006485, 'timing/model': 0.41729968320578337, 'learning_rate/qwen_vl_interface': 7.403044917336404e-06, 'learning_rate/action_model': 7.403044917336404e-05, 'epoch': 0.27})
+[2026-07-28 04:36:50,184][starVLA.training.train_starvla][INFO] - Step 1467, Loss: {'train/loss': 0.06032577157020569, 'train/grad_norm_pre_clip': 2.2381162643432617, 'timing/data': 0.0002822037786245346, 'timing/model': 0.4179022451862693, 'learning_rate/qwen_vl_interface': 7.399633563822487e-06, 'learning_rate/action_model': 7.399633563822487e-05, 'epoch': 0.27})
+[2026-07-28 04:36:56,299][starVLA.training.train_starvla][INFO] - Step 1468, Loss: {'train/loss': 0.06344743818044662, 'train/grad_norm_pre_clip': 1.607241153717041, 'timing/data': 0.00026180781424045563, 'timing/model': 0.41629392793402076, 'learning_rate/qwen_vl_interface': 7.396220815434691e-06, 'learning_rate/action_model': 7.39622081543469e-05, 'epoch': 0.27})
+[2026-07-28 04:37:02,166][starVLA.training.train_starvla][INFO] - Step 1469, Loss: {'train/loss': 0.050891704857349396, 'train/grad_norm_pre_clip': 1.523481011390686, 'timing/data': 0.0002939840778708458, 'timing/model': 0.41653035394847393, 'learning_rate/qwen_vl_interface': 7.392806674387506e-06, 'learning_rate/action_model': 7.392806674387506e-05, 'epoch': 0.27})
+[2026-07-28 04:37:08,269][starVLA.training.train_starvla][INFO] - Step 1470, Loss: {'train/loss': 0.03917780891060829, 'train/grad_norm_pre_clip': 0.7665842175483704, 'timing/data': 0.0002583591267466545, 'timing/model': 0.41673634899780154, 'learning_rate/qwen_vl_interface': 7.389391142896336e-06, 'learning_rate/action_model': 7.389391142896335e-05, 'epoch': 0.27})
+[2026-07-28 04:37:14,179][starVLA.training.train_starvla][INFO] - Step 1471, Loss: {'train/loss': 0.03427523374557495, 'train/grad_norm_pre_clip': 1.5069283246994019, 'timing/data': 0.00027622003108263016, 'timing/model': 0.41811236878857017, 'learning_rate/qwen_vl_interface': 7.38597422317748e-06, 'learning_rate/action_model': 7.385974223177479e-05, 'epoch': 0.27})
+[2026-07-28 04:37:20,063][starVLA.training.train_starvla][INFO] - Step 1472, Loss: {'train/loss': 0.052916303277015686, 'train/grad_norm_pre_clip': 1.508620262145996, 'timing/data': 0.0002726009115576744, 'timing/model': 0.41558791836723685, 'learning_rate/qwen_vl_interface': 7.382555917448141e-06, 'learning_rate/action_model': 7.382555917448141e-05, 'epoch': 0.27})
+[2026-07-28 04:37:26,319][starVLA.training.train_starvla][INFO] - Step 1473, Loss: {'train/loss': 0.05923859775066376, 'train/grad_norm_pre_clip': 1.3921163082122803, 'timing/data': 0.0002898643724620342, 'timing/model': 0.8262004149146378, 'learning_rate/qwen_vl_interface': 7.379136227926418e-06, 'learning_rate/action_model': 7.379136227926418e-05, 'epoch': 0.27})
+[2026-07-28 04:37:32,207][starVLA.training.train_starvla][INFO] - Step 1474, Loss: {'train/loss': 0.05688239261507988, 'train/grad_norm_pre_clip': 1.9690290689468384, 'timing/data': 0.00037418585270643234, 'timing/model': 0.41458992194384336, 'learning_rate/qwen_vl_interface': 7.3757151568313144e-06, 'learning_rate/action_model': 7.375715156831315e-05, 'epoch': 0.27})
+[2026-07-28 04:37:38,072][starVLA.training.train_starvla][INFO] - Step 1475, Loss: {'train/loss': 0.06263737380504608, 'train/grad_norm_pre_clip': 2.475616931915283, 'timing/data': 0.0002858066000044346, 'timing/model': 0.4169063060544431, 'learning_rate/qwen_vl_interface': 7.3722927063827235e-06, 'learning_rate/action_model': 7.372292706382723e-05, 'epoch': 0.27})
+[2026-07-28 04:37:43,981][starVLA.training.train_starvla][INFO] - Step 1476, Loss: {'train/loss': 0.045048899948596954, 'train/grad_norm_pre_clip': 2.96652889251709, 'timing/data': 0.00027695391327142715, 'timing/model': 0.4162005800753832, 'learning_rate/qwen_vl_interface': 7.368868878801437e-06, 'learning_rate/action_model': 7.368868878801437e-05, 'epoch': 0.27})
+[2026-07-28 04:37:50,085][starVLA.training.train_starvla][INFO] - Step 1477, Loss: {'train/loss': 0.05105520039796829, 'train/grad_norm_pre_clip': 1.5808929204940796, 'timing/data': 0.00025567878037691116, 'timing/model': 0.41468186490237713, 'learning_rate/qwen_vl_interface': 7.365443676309139e-06, 'learning_rate/action_model': 7.365443676309139e-05, 'epoch': 0.27})
+[2026-07-28 04:37:55,969][starVLA.training.train_starvla][INFO] - Step 1478, Loss: {'train/loss': 0.05948200076818466, 'train/grad_norm_pre_clip': 1.6990206241607666, 'timing/data': 0.0002820109948515892, 'timing/model': 0.41668475419282913, 'learning_rate/qwen_vl_interface': 7.362017101128406e-06, 'learning_rate/action_model': 7.362017101128406e-05, 'epoch': 0.27})
+[2026-07-28 04:38:01,851][starVLA.training.train_starvla][INFO] - Step 1479, Loss: {'train/loss': 0.03390558809041977, 'train/grad_norm_pre_clip': 1.1271864175796509, 'timing/data': 0.00028523197397589684, 'timing/model': 0.421262021176517, 'learning_rate/qwen_vl_interface': 7.358589155482707e-06, 'learning_rate/action_model': 7.358589155482707e-05, 'epoch': 0.27})
+[2026-07-28 04:38:07,966][starVLA.training.train_starvla][INFO] - Step 1480, Loss: {'train/loss': 0.07860372960567474, 'train/grad_norm_pre_clip': 1.9860972166061401, 'timing/data': 0.0002862238325178623, 'timing/model': 0.41391072096303105, 'learning_rate/qwen_vl_interface': 7.355159841596397e-06, 'learning_rate/action_model': 7.355159841596397e-05, 'epoch': 0.27})
+[2026-07-28 04:38:14,078][starVLA.training.train_starvla][INFO] - Step 1481, Loss: {'train/loss': 0.06394357234239578, 'train/grad_norm_pre_clip': 2.394510269165039, 'timing/data': 0.00027065398171544075, 'timing/model': 0.6504763863049448, 'learning_rate/qwen_vl_interface': 7.351729161694722e-06, 'learning_rate/action_model': 7.351729161694722e-05, 'epoch': 0.27})
+[2026-07-28 04:38:19,955][starVLA.training.train_starvla][INFO] - Step 1482, Loss: {'train/loss': 0.11794282495975494, 'train/grad_norm_pre_clip': 3.6055450439453125, 'timing/data': 0.0002817451022565365, 'timing/model': 0.4140115864574909, 'learning_rate/qwen_vl_interface': 7.348297118003811e-06, 'learning_rate/action_model': 7.348297118003811e-05, 'epoch': 0.27})
+[2026-07-28 04:38:25,834][starVLA.training.train_starvla][INFO] - Step 1483, Loss: {'train/loss': 0.07745477557182312, 'train/grad_norm_pre_clip': 1.6998659372329712, 'timing/data': 0.00028871605172753334, 'timing/model': 0.41736909886822104, 'learning_rate/qwen_vl_interface': 7.3448637127506815e-06, 'learning_rate/action_model': 7.344863712750682e-05, 'epoch': 0.27})
+[2026-07-28 04:38:31,957][starVLA.training.train_starvla][INFO] - Step 1484, Loss: {'train/loss': 0.03879985958337784, 'train/grad_norm_pre_clip': 1.008048176765442, 'timing/data': 0.00028494978323578835, 'timing/model': 0.6584328422322869, 'learning_rate/qwen_vl_interface': 7.341428948163232e-06, 'learning_rate/action_model': 7.341428948163231e-05, 'epoch': 0.27})
+[2026-07-28 04:38:37,848][starVLA.training.train_starvla][INFO] - Step 1485, Loss: {'train/loss': 0.07274627685546875, 'train/grad_norm_pre_clip': 1.3173463344573975, 'timing/data': 0.0002833832986652851, 'timing/model': 0.4181033680215478, 'learning_rate/qwen_vl_interface': 7.337992826470245e-06, 'learning_rate/action_model': 7.337992826470245e-05, 'epoch': 0.27})
+[2026-07-28 04:38:43,679][starVLA.training.train_starvla][INFO] - Step 1486, Loss: {'train/loss': 0.0599963553249836, 'train/grad_norm_pre_clip': 1.2060290575027466, 'timing/data': 0.00029153795912861824, 'timing/model': 0.41663931403309107, 'learning_rate/qwen_vl_interface': 7.334555349901381e-06, 'learning_rate/action_model': 7.33455534990138e-05, 'epoch': 0.27})
+[2026-07-28 04:38:49,805][starVLA.training.train_starvla][INFO] - Step 1487, Loss: {'train/loss': 0.06998910754919052, 'train/grad_norm_pre_clip': 1.9090100526809692, 'timing/data': 0.00027145491912961006, 'timing/model': 0.4150979658588767, 'learning_rate/qwen_vl_interface': 7.331116520687184e-06, 'learning_rate/action_model': 7.331116520687184e-05, 'epoch': 0.27})
+[2026-07-28 04:38:55,882][starVLA.training.train_starvla][INFO] - Step 1488, Loss: {'train/loss': 0.05177413299679756, 'train/grad_norm_pre_clip': 2.8266828060150146, 'timing/data': 0.00026465440168976784, 'timing/model': 0.4140105121769011, 'learning_rate/qwen_vl_interface': 7.32767634105907e-06, 'learning_rate/action_model': 7.32767634105907e-05, 'epoch': 0.27})
+[2026-07-28 04:39:01,746][starVLA.training.train_starvla][INFO] - Step 1489, Loss: {'train/loss': 0.05327098071575165, 'train/grad_norm_pre_clip': 0.6620838642120361, 'timing/data': 0.00048804283142089844, 'timing/model': 0.4162349351681769, 'learning_rate/qwen_vl_interface': 7.324234813249334e-06, 'learning_rate/action_model': 7.324234813249335e-05, 'epoch': 0.27})
+[2026-07-28 04:39:07,605][starVLA.training.train_starvla][INFO] - Step 1490, Loss: {'train/loss': 0.03372698649764061, 'train/grad_norm_pre_clip': 1.1191902160644531, 'timing/data': 0.00028106383979320526, 'timing/model': 0.41526011005043983, 'learning_rate/qwen_vl_interface': 7.320791939491149e-06, 'learning_rate/action_model': 7.320791939491149e-05, 'epoch': 0.27})
+[2026-07-28 04:39:13,697][starVLA.training.train_starvla][INFO] - Step 1491, Loss: {'train/loss': 0.05039522424340248, 'train/grad_norm_pre_clip': 0.9434020519256592, 'timing/data': 0.00026642903685569763, 'timing/model': 0.4143795892596245, 'learning_rate/qwen_vl_interface': 7.317347722018557e-06, 'learning_rate/action_model': 7.317347722018557e-05, 'epoch': 0.27})
+[2026-07-28 04:39:19,785][starVLA.training.train_starvla][INFO] - Step 1492, Loss: {'train/loss': 0.043742433190345764, 'train/grad_norm_pre_clip': 1.273788332939148, 'timing/data': 0.0002622739411890507, 'timing/model': 0.41391210490837693, 'learning_rate/qwen_vl_interface': 7.313902163066472e-06, 'learning_rate/action_model': 7.313902163066471e-05, 'epoch': 0.27})
+[2026-07-28 04:39:25,648][starVLA.training.train_starvla][INFO] - Step 1493, Loss: {'train/loss': 0.05823540687561035, 'train/grad_norm_pre_clip': 2.090074300765991, 'timing/data': 0.0002868608571588993, 'timing/model': 0.416067386046052, 'learning_rate/qwen_vl_interface': 7.3104552648706805e-06, 'learning_rate/action_model': 7.310455264870681e-05, 'epoch': 0.27})
+[2026-07-28 04:39:31,725][starVLA.training.train_starvla][INFO] - Step 1494, Loss: {'train/loss': 0.05970863625407219, 'train/grad_norm_pre_clip': 2.3746073246002197, 'timing/data': 0.00027911411598324776, 'timing/model': 0.6484919497743249, 'learning_rate/qwen_vl_interface': 7.307007029667839e-06, 'learning_rate/action_model': 7.30700702966784e-05, 'epoch': 0.27})
+[2026-07-28 04:39:37,799][starVLA.training.train_starvla][INFO] - Step 1495, Loss: {'train/loss': 0.08441208302974701, 'train/grad_norm_pre_clip': 1.2199125289916992, 'timing/data': 0.00028503546491265297, 'timing/model': 0.6559849279001355, 'learning_rate/qwen_vl_interface': 7.303557459695468e-06, 'learning_rate/action_model': 7.303557459695469e-05, 'epoch': 0.27})
+[2026-07-28 04:39:43,648][starVLA.training.train_starvla][INFO] - Step 1496, Loss: {'train/loss': 0.05229488015174866, 'train/grad_norm_pre_clip': 0.8930912613868713, 'timing/data': 0.00029126694425940514, 'timing/model': 0.4165578633546829, 'learning_rate/qwen_vl_interface': 7.300106557191956e-06, 'learning_rate/action_model': 7.300106557191956e-05, 'epoch': 0.27})
+[2026-07-28 04:39:49,521][starVLA.training.train_starvla][INFO] - Step 1497, Loss: {'train/loss': 0.06275778263807297, 'train/grad_norm_pre_clip': 2.0295279026031494, 'timing/data': 0.00029587186872959137, 'timing/model': 0.41618199506774545, 'learning_rate/qwen_vl_interface': 7.296654324396557e-06, 'learning_rate/action_model': 7.296654324396556e-05, 'epoch': 0.27})
+[2026-07-28 04:39:55,798][starVLA.training.train_starvla][INFO] - Step 1498, Loss: {'train/loss': 0.056601665914058685, 'train/grad_norm_pre_clip': 2.8354151248931885, 'timing/data': 0.0002804221585392952, 'timing/model': 0.4511857395991683, 'learning_rate/qwen_vl_interface': 7.2932007635493875e-06, 'learning_rate/action_model': 7.293200763549387e-05, 'epoch': 0.28})
+[2026-07-28 04:40:01,677][starVLA.training.train_starvla][INFO] - Step 1499, Loss: {'train/loss': 0.06852812319993973, 'train/grad_norm_pre_clip': 1.983077049255371, 'timing/data': 0.00027656136080622673, 'timing/model': 0.4156367159448564, 'learning_rate/qwen_vl_interface': 7.289745876891427e-06, 'learning_rate/action_model': 7.289745876891426e-05, 'epoch': 0.28})
+[2026-07-28 04:40:07,799][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1500
+[2026-07-28 04:40:12,741][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1500
+[2026-07-28 04:45:01,064][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1500
+[2026-07-28 04:45:01,065][starVLA.training.train_starvla][INFO] - Step 1500, Loss: {'train/loss': 0.02413015067577362, 'train/grad_norm_pre_clip': 0.990699291229248, 'eval/action_loss/seconds': 4.939320440869778, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.06526307016611099, 'eval/latency_3/loss': 0.06526307016611099, 'eval/flappy/loss': 0.06526307016611099, 'eval/flappy/latency_3/loss': 0.06526307016611099, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 198.610002348572, 'rl_games_eval/mid_train/mean_length': 1618.8, 'rl_games_eval/mid_train/std_reward': 156.37761499774624, 'rl_games_eval/mid_train/std_length': 1261.2708115230448, 'rl_games_eval/mid_train/macro_mean_reward': 198.610002348572, 'rl_games_eval/mid_train/macro_mean_length': 1618.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 198.610002348572, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 1618.8, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 156.37761499774624, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1261.2708115230448, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 198.610002348572, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0002819439396262169, 'timing/model': 0.6584390574134886, 'learning_rate/qwen_vl_interface': 7.286289666664508e-06, 'learning_rate/action_model': 7.286289666664508e-05, 'epoch': 0.28})
+[2026-07-28 04:45:01,067][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1500_state
+[2026-07-28 04:45:01,067][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 04:45:38,371][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1500_state/pytorch_model
+[2026-07-28 04:45:38,371][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1500_state/sampler.bin
+[2026-07-28 04:45:38,372][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1500_state/sampler_1.bin
+[2026-07-28 04:45:38,378][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1500_state/random_states_0.pkl
+[2026-07-28 04:45:49,576][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 1500
+[2026-07-28 04:45:49,576][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 04:45:49,600][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 04:45:55,367][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.038575612008571625, 'train/grad_norm_pre_clip': 1.3452095985412598, 'timing/data': 0.0002976651303470135, 'timing/model': 0.41726063936948776, 'learning_rate/qwen_vl_interface': 7.282832135111334e-06, 'learning_rate/action_model': 7.282832135111334e-05, 'epoch': 0.28})
+[2026-07-28 04:46:01,697][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.07413332909345627, 'train/grad_norm_pre_clip': 2.5863773822784424, 'timing/data': 0.00026241596788167953, 'timing/model': 0.6440892424434423, 'learning_rate/qwen_vl_interface': 7.2793732844754584e-06, 'learning_rate/action_model': 7.279373284475459e-05, 'epoch': 0.28})
+[2026-07-28 04:46:07,555][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.13495658338069916, 'train/grad_norm_pre_clip': 6.379332542419434, 'timing/data': 0.0002554594539105892, 'timing/model': 0.4138174611143768, 'learning_rate/qwen_vl_interface': 7.275913117001291e-06, 'learning_rate/action_model': 7.27591311700129e-05, 'epoch': 0.28})
+[2026-07-28 04:46:13,420][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.05843164399266243, 'train/grad_norm_pre_clip': 3.6809473037719727, 'timing/data': 0.0002837688662111759, 'timing/model': 0.41624001087620854, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.28})
+[2026-07-28 04:46:19,511][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.055893510580062866, 'train/grad_norm_pre_clip': 1.1476550102233887, 'timing/data': 0.00028053298592567444, 'timing/model': 0.655609262175858, 'learning_rate/qwen_vl_interface': 7.268988840519994e-06, 'learning_rate/action_model': 7.268988840519993e-05, 'epoch': 0.28})
+[2026-07-28 04:46:25,370][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.08441649377346039, 'train/grad_norm_pre_clip': 2.467776298522949, 'timing/data': 0.0002708490937948227, 'timing/model': 0.41640471713617444, 'learning_rate/qwen_vl_interface': 7.265524736005951e-06, 'learning_rate/action_model': 7.26552473600595e-05, 'epoch': 0.28})
+[2026-07-28 04:46:31,212][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.05794312059879303, 'train/grad_norm_pre_clip': 1.1389840841293335, 'timing/data': 0.00026046065613627434, 'timing/model': 0.4161409200169146, 'learning_rate/qwen_vl_interface': 7.262059323639792e-06, 'learning_rate/action_model': 7.262059323639792e-05, 'epoch': 0.28})
+[2026-07-28 04:46:37,304][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.08964527398347855, 'train/grad_norm_pre_clip': 3.0214879512786865, 'timing/data': 0.0002526971511542797, 'timing/model': 0.41544799599796534, 'learning_rate/qwen_vl_interface': 7.25859260567018e-06, 'learning_rate/action_model': 7.258592605670179e-05, 'epoch': 0.28})
+[2026-07-28 04:46:43,377][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.050356894731521606, 'train/grad_norm_pre_clip': 1.7825160026550293, 'timing/data': 0.00025662127882242203, 'timing/model': 0.41684631211683154, 'learning_rate/qwen_vl_interface': 7.2551245843466334e-06, 'learning_rate/action_model': 7.255124584346633e-05, 'epoch': 0.28})
+[2026-07-28 04:46:49,419][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.0820896252989769, 'train/grad_norm_pre_clip': 2.190089225769043, 'timing/data': 0.00027478672564029694, 'timing/model': 0.4737694077193737, 'learning_rate/qwen_vl_interface': 7.251655261919515e-06, 'learning_rate/action_model': 7.251655261919515e-05, 'epoch': 0.28})
+[2026-07-28 04:46:55,277][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.045511484146118164, 'train/grad_norm_pre_clip': 0.7747396230697632, 'timing/data': 0.00028208503499627113, 'timing/model': 0.41640721494331956, 'learning_rate/qwen_vl_interface': 7.248184640640027e-06, 'learning_rate/action_model': 7.248184640640026e-05, 'epoch': 0.28})
+[2026-07-28 04:47:01,438][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.05528080090880394, 'train/grad_norm_pre_clip': 2.0842950344085693, 'timing/data': 0.000250284094363451, 'timing/model': 0.46395373810082674, 'learning_rate/qwen_vl_interface': 7.24471272276022e-06, 'learning_rate/action_model': 7.24471272276022e-05, 'epoch': 0.28})
+[2026-07-28 04:47:07,543][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.07520315051078796, 'train/grad_norm_pre_clip': 1.8746793270111084, 'timing/data': 0.000253050122410059, 'timing/model': 0.4147021062672138, 'learning_rate/qwen_vl_interface': 7.241239510532985e-06, 'learning_rate/action_model': 7.241239510532984e-05, 'epoch': 0.28})
+[2026-07-28 04:47:13,416][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.07572852075099945, 'train/grad_norm_pre_clip': 1.1203348636627197, 'timing/data': 0.0002701692283153534, 'timing/model': 0.4165013190358877, 'learning_rate/qwen_vl_interface': 7.237765006212049e-06, 'learning_rate/action_model': 7.237765006212048e-05, 'epoch': 0.28})
+[2026-07-28 04:47:19,504][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.08122369647026062, 'train/grad_norm_pre_clip': 1.620551347732544, 'timing/data': 0.00027575530111789703, 'timing/model': 0.6487307348288596, 'learning_rate/qwen_vl_interface': 7.2342892120519805e-06, 'learning_rate/action_model': 7.23428921205198e-05, 'epoch': 0.28})
+[2026-07-28 04:47:25,556][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.06290090829133987, 'train/grad_norm_pre_clip': 0.7817625999450684, 'timing/data': 0.0004946775734424591, 'timing/model': 0.656604191288352, 'learning_rate/qwen_vl_interface': 7.230812130308187e-06, 'learning_rate/action_model': 7.230812130308186e-05, 'epoch': 0.28})
+[2026-07-28 04:47:31,426][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.06261695176362991, 'train/grad_norm_pre_clip': 1.8792452812194824, 'timing/data': 0.000256961677223444, 'timing/model': 0.4155847323127091, 'learning_rate/qwen_vl_interface': 7.227333763236907e-06, 'learning_rate/action_model': 7.227333763236906e-05, 'epoch': 0.28})
+[2026-07-28 04:47:37,299][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.06553800404071808, 'train/grad_norm_pre_clip': 1.056931734085083, 'timing/data': 0.00027378275990486145, 'timing/model': 0.4153040945529938, 'learning_rate/qwen_vl_interface': 7.223854113095214e-06, 'learning_rate/action_model': 7.223854113095214e-05, 'epoch': 0.28})
+[2026-07-28 04:47:43,422][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.044091202318668365, 'train/grad_norm_pre_clip': 1.10828697681427, 'timing/data': 0.0002634171396493912, 'timing/model': 0.4154565539211035, 'learning_rate/qwen_vl_interface': 7.220373182141021e-06, 'learning_rate/action_model': 7.22037318214102e-05, 'epoch': 0.28})
+[2026-07-28 04:47:49,271][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.05110640451312065, 'train/grad_norm_pre_clip': 1.2945382595062256, 'timing/data': 0.00028233835473656654, 'timing/model': 0.4167186999693513, 'learning_rate/qwen_vl_interface': 7.216890972633057e-06, 'learning_rate/action_model': 7.216890972633057e-05, 'epoch': 0.28})
+[2026-07-28 04:47:55,435][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.0525115430355072, 'train/grad_norm_pre_clip': 1.8557980060577393, 'timing/data': 0.00028271693736314774, 'timing/model': 0.6594069269485772, 'learning_rate/qwen_vl_interface': 7.213407486830899e-06, 'learning_rate/action_model': 7.213407486830899e-05, 'epoch': 0.28})
+[2026-07-28 04:48:01,528][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.04000598564743996, 'train/grad_norm_pre_clip': 1.273476481437683, 'timing/data': 0.0002600802108645439, 'timing/model': 0.41349558625370264, 'learning_rate/qwen_vl_interface': 7.209922726994937e-06, 'learning_rate/action_model': 7.209922726994936e-05, 'epoch': 0.28})
+[2026-07-28 04:48:07,378][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.04339087754487991, 'train/grad_norm_pre_clip': 1.2439786195755005, 'timing/data': 0.0002855239436030388, 'timing/model': 0.41663204273208976, 'learning_rate/qwen_vl_interface': 7.206436695386396e-06, 'learning_rate/action_model': 7.206436695386396e-05, 'epoch': 0.28})
+[2026-07-28 04:48:13,283][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.04071415960788727, 'train/grad_norm_pre_clip': 1.2699567079544067, 'timing/data': 0.00029128510504961014, 'timing/model': 0.41829804610460997, 'learning_rate/qwen_vl_interface': 7.202949394267326e-06, 'learning_rate/action_model': 7.202949394267326e-05, 'epoch': 0.28})
+[2026-07-28 04:48:19,308][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.03523357957601547, 'train/grad_norm_pre_clip': 1.49154794216156, 'timing/data': 0.00028508203104138374, 'timing/model': 0.5863203648477793, 'learning_rate/qwen_vl_interface': 7.199460825900596e-06, 'learning_rate/action_model': 7.199460825900596e-05, 'epoch': 0.28})
+[2026-07-28 04:48:25,173][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.06199813261628151, 'train/grad_norm_pre_clip': 1.1992268562316895, 'timing/data': 0.0002922811545431614, 'timing/model': 0.4176551727578044, 'learning_rate/qwen_vl_interface': 7.1959709925499e-06, 'learning_rate/action_model': 7.1959709925499e-05, 'epoch': 0.28})
+[2026-07-28 04:48:31,299][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.039073534309864044, 'train/grad_norm_pre_clip': 0.8722118735313416, 'timing/data': 0.0002855798229575157, 'timing/model': 0.656603405252099, 'learning_rate/qwen_vl_interface': 7.1924798964797545e-06, 'learning_rate/action_model': 7.192479896479754e-05, 'epoch': 0.28})
+[2026-07-28 04:48:37,181][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.03483736142516136, 'train/grad_norm_pre_clip': 0.8756455183029175, 'timing/data': 0.00025787483900785446, 'timing/model': 0.41439704969525337, 'learning_rate/qwen_vl_interface': 7.188987539955495e-06, 'learning_rate/action_model': 7.188987539955494e-05, 'epoch': 0.28})
+[2026-07-28 04:48:43,306][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.06206400319933891, 'train/grad_norm_pre_clip': 2.803161144256592, 'timing/data': 0.00027724122628569603, 'timing/model': 0.4208869356662035, 'learning_rate/qwen_vl_interface': 7.185493925243271e-06, 'learning_rate/action_model': 7.185493925243271e-05, 'epoch': 0.28})
+[2026-07-28 04:48:49,198][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.05056913197040558, 'train/grad_norm_pre_clip': 2.515425205230713, 'timing/data': 0.0002937810495495796, 'timing/model': 0.41977316373959184, 'learning_rate/qwen_vl_interface': 7.181999054610053e-06, 'learning_rate/action_model': 7.181999054610052e-05, 'epoch': 0.28})
+[2026-07-28 04:48:55,373][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.059325456619262695, 'train/grad_norm_pre_clip': 2.6553049087524414, 'timing/data': 0.00027465587481856346, 'timing/model': 0.41538555501028895, 'learning_rate/qwen_vl_interface': 7.178502930323622e-06, 'learning_rate/action_model': 7.178502930323622e-05, 'epoch': 0.28})
+[2026-07-28 04:49:01,700][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.07811225205659866, 'train/grad_norm_pre_clip': 1.5956891775131226, 'timing/data': 0.0002518771216273308, 'timing/model': 0.41393902292475104, 'learning_rate/qwen_vl_interface': 7.175005554652577e-06, 'learning_rate/action_model': 7.175005554652578e-05, 'epoch': 0.28})
+[2026-07-28 04:49:07,946][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.027296334505081177, 'train/grad_norm_pre_clip': 1.66745924949646, 'timing/data': 0.00026194704696536064, 'timing/model': 0.417981896083802, 'learning_rate/qwen_vl_interface': 7.171506929866328e-06, 'learning_rate/action_model': 7.171506929866328e-05, 'epoch': 0.28})
+[2026-07-28 04:49:13,834][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.044912081211805344, 'train/grad_norm_pre_clip': 1.5726674795150757, 'timing/data': 0.00027479417622089386, 'timing/model': 0.4192528943531215, 'learning_rate/qwen_vl_interface': 7.168007058235095e-06, 'learning_rate/action_model': 7.168007058235095e-05, 'epoch': 0.28})
+[2026-07-28 04:49:19,743][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.038232315331697464, 'train/grad_norm_pre_clip': 1.0370674133300781, 'timing/data': 0.0002880571410059929, 'timing/model': 0.4184703570790589, 'learning_rate/qwen_vl_interface': 7.164505942029905e-06, 'learning_rate/action_model': 7.164505942029904e-05, 'epoch': 0.28})
+[2026-07-28 04:49:25,900][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.05198561027646065, 'train/grad_norm_pre_clip': 2.271437883377075, 'timing/data': 0.0002677440643310547, 'timing/model': 0.4142980449832976, 'learning_rate/qwen_vl_interface': 7.161003583522594e-06, 'learning_rate/action_model': 7.161003583522593e-05, 'epoch': 0.28})
+[2026-07-28 04:49:31,832][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.05158770829439163, 'train/grad_norm_pre_clip': 0.6498441696166992, 'timing/data': 0.000380881130695343, 'timing/model': 0.41774581698700786, 'learning_rate/qwen_vl_interface': 7.157499984985804e-06, 'learning_rate/action_model': 7.157499984985803e-05, 'epoch': 0.28})
+[2026-07-28 04:49:37,966][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.03401927649974823, 'train/grad_norm_pre_clip': 1.1293303966522217, 'timing/data': 0.0002736411988735199, 'timing/model': 0.6479724873788655, 'learning_rate/qwen_vl_interface': 7.153995148692985e-06, 'learning_rate/action_model': 7.153995148692985e-05, 'epoch': 0.28})
+[2026-07-28 04:49:43,899][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.03971591219305992, 'train/grad_norm_pre_clip': 1.3457111120224, 'timing/data': 0.0002464097924530506, 'timing/model': 0.4167550108395517, 'learning_rate/qwen_vl_interface': 7.150489076918383e-06, 'learning_rate/action_model': 7.150489076918384e-05, 'epoch': 0.28})
+[2026-07-28 04:49:49,794][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.04165608435869217, 'train/grad_norm_pre_clip': 1.1367764472961426, 'timing/data': 0.0002893637865781784, 'timing/model': 0.41931261494755745, 'learning_rate/qwen_vl_interface': 7.1469817719370535e-06, 'learning_rate/action_model': 7.146981771937053e-05, 'epoch': 0.28})
+[2026-07-28 04:49:55,925][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.021132249385118484, 'train/grad_norm_pre_clip': 0.5593448877334595, 'timing/data': 0.0002743462100625038, 'timing/model': 0.6615655478090048, 'learning_rate/qwen_vl_interface': 7.1434732360248505e-06, 'learning_rate/action_model': 7.14347323602485e-05, 'epoch': 0.28})
+[2026-07-28 04:50:01,803][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.07274146378040314, 'train/grad_norm_pre_clip': 1.478376865386963, 'timing/data': 0.000250017736107111, 'timing/model': 0.4146762993186712, 'learning_rate/qwen_vl_interface': 7.139963471458418e-06, 'learning_rate/action_model': 7.139963471458418e-05, 'epoch': 0.28})
+[2026-07-28 04:50:07,882][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.04853362217545509, 'train/grad_norm_pre_clip': 1.079114556312561, 'timing/data': 0.000268016941845417, 'timing/model': 0.41420785896480083, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.28})
+[2026-07-28 04:50:13,541][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.036001160740852356, 'train/grad_norm_pre_clip': 1.5931367874145508, 'timing/data': 0.00026491517201066017, 'timing/model': 0.4177966960705817, 'learning_rate/qwen_vl_interface': 7.132940265473466e-06, 'learning_rate/action_model': 7.132940265473466e-05, 'epoch': 0.28})
+[2026-07-28 04:50:19,406][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.04864747077226639, 'train/grad_norm_pre_clip': 1.8384109735488892, 'timing/data': 0.00029475707560777664, 'timing/model': 0.41496217995882034, 'learning_rate/qwen_vl_interface': 7.129426828612227e-06, 'learning_rate/action_model': 7.129426828612227e-05, 'epoch': 0.28})
+[2026-07-28 04:50:25,500][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.05574417486786842, 'train/grad_norm_pre_clip': 1.2408146858215332, 'timing/data': 0.0002480819821357727, 'timing/model': 0.4131773114204407, 'learning_rate/qwen_vl_interface': 7.125912172211326e-06, 'learning_rate/action_model': 7.125912172211325e-05, 'epoch': 0.28})
+[2026-07-28 04:50:31,365][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.03827238455414772, 'train/grad_norm_pre_clip': 1.4852639436721802, 'timing/data': 0.0002584671601653099, 'timing/model': 0.4164445549249649, 'learning_rate/qwen_vl_interface': 7.122396298551381e-06, 'learning_rate/action_model': 7.12239629855138e-05, 'epoch': 0.28})
+[2026-07-28 04:50:37,474][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.06410428136587143, 'train/grad_norm_pre_clip': 1.8297423124313354, 'timing/data': 0.00024930620566010475, 'timing/model': 0.4152107611298561, 'learning_rate/qwen_vl_interface': 7.118879209913806e-06, 'learning_rate/action_model': 7.118879209913805e-05, 'epoch': 0.28})
+[2026-07-28 04:50:43,360][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.05431080237030983, 'train/grad_norm_pre_clip': 1.2033523321151733, 'timing/data': 0.0003339899703860283, 'timing/model': 0.41760019212961197, 'learning_rate/qwen_vl_interface': 7.115360908580803e-06, 'learning_rate/action_model': 7.115360908580802e-05, 'epoch': 0.28})
+[2026-07-28 04:50:49,462][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.04456634074449539, 'train/grad_norm_pre_clip': 0.9759104251861572, 'timing/data': 0.0003425651229918003, 'timing/model': 0.650470154825598, 'learning_rate/qwen_vl_interface': 7.1118413968353575e-06, 'learning_rate/action_model': 7.111841396835357e-05, 'epoch': 0.28})
+[2026-07-28 04:50:55,337][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.028372999280691147, 'train/grad_norm_pre_clip': 0.8434085249900818, 'timing/data': 0.00025430694222450256, 'timing/model': 0.4141393597237766, 'learning_rate/qwen_vl_interface': 7.108320676961242e-06, 'learning_rate/action_model': 7.108320676961242e-05, 'epoch': 0.28})
+[2026-07-28 04:51:01,450][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.045148372650146484, 'train/grad_norm_pre_clip': 1.0910608768463135, 'timing/data': 0.00028412695974111557, 'timing/model': 0.6543619600124657, 'learning_rate/qwen_vl_interface': 7.104798751243018e-06, 'learning_rate/action_model': 7.104798751243019e-05, 'epoch': 0.29})
+[2026-07-28 04:51:07,346][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.046544209122657776, 'train/grad_norm_pre_clip': 1.0730351209640503, 'timing/data': 0.0002826307900249958, 'timing/model': 0.4165415447205305, 'learning_rate/qwen_vl_interface': 7.1012756219660195e-06, 'learning_rate/action_model': 7.101275621966019e-05, 'epoch': 0.29})
+[2026-07-28 04:51:13,301][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.0450114905834198, 'train/grad_norm_pre_clip': 1.4754701852798462, 'timing/data': 0.00027701398357748985, 'timing/model': 0.4171046200208366, 'learning_rate/qwen_vl_interface': 7.09775129141637e-06, 'learning_rate/action_model': 7.097751291416369e-05, 'epoch': 0.29})
+[2026-07-28 04:51:19,238][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.0317152738571167, 'train/grad_norm_pre_clip': 1.5395827293395996, 'timing/data': 0.0002761459909379482, 'timing/model': 0.4163711699657142, 'learning_rate/qwen_vl_interface': 7.094225761880966e-06, 'learning_rate/action_model': 7.094225761880967e-05, 'epoch': 0.29})
+[2026-07-28 04:51:25,375][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.06167957931756973, 'train/grad_norm_pre_clip': 1.5377277135849, 'timing/data': 0.00025076884776353836, 'timing/model': 0.4148595309816301, 'learning_rate/qwen_vl_interface': 7.0906990356474895e-06, 'learning_rate/action_model': 7.090699035647489e-05, 'epoch': 0.29})
+[2026-07-28 04:51:31,037][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.044478222727775574, 'train/grad_norm_pre_clip': 1.7060892581939697, 'timing/data': 0.00027055200189352036, 'timing/model': 0.4183351439423859, 'learning_rate/qwen_vl_interface': 7.087171115004393e-06, 'learning_rate/action_model': 7.087171115004393e-05, 'epoch': 0.29})
+[2026-07-28 04:51:36,952][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.05807943642139435, 'train/grad_norm_pre_clip': 3.2589428424835205, 'timing/data': 0.0002772589214146137, 'timing/model': 0.4179133069701493, 'learning_rate/qwen_vl_interface': 7.083642002240905e-06, 'learning_rate/action_model': 7.083642002240905e-05, 'epoch': 0.29})
+[2026-07-28 04:51:42,894][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.04849688336253166, 'train/grad_norm_pre_clip': 1.2202050685882568, 'timing/data': 0.00028034206479787827, 'timing/model': 0.419709833804518, 'learning_rate/qwen_vl_interface': 7.080111699647031e-06, 'learning_rate/action_model': 7.080111699647031e-05, 'epoch': 0.29})
+[2026-07-28 04:51:49,086][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.0368681401014328, 'train/grad_norm_pre_clip': 0.8485981822013855, 'timing/data': 0.0002486039884388447, 'timing/model': 0.4167215288616717, 'learning_rate/qwen_vl_interface': 7.076580209513542e-06, 'learning_rate/action_model': 7.076580209513542e-05, 'epoch': 0.29})
+[2026-07-28 04:51:54,829][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.056583110243082047, 'train/grad_norm_pre_clip': 2.31298565864563, 'timing/data': 0.00028523802757263184, 'timing/model': 0.41729932418093085, 'learning_rate/qwen_vl_interface': 7.073047534131987e-06, 'learning_rate/action_model': 7.073047534131986e-05, 'epoch': 0.29})
+[2026-07-28 04:52:00,742][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.04557601362466812, 'train/grad_norm_pre_clip': 1.8450970649719238, 'timing/data': 0.00030283816158771515, 'timing/model': 0.4207621179521084, 'learning_rate/qwen_vl_interface': 7.069513675794679e-06, 'learning_rate/action_model': 7.069513675794679e-05, 'epoch': 0.29})
+[2026-07-28 04:52:06,889][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.033168260008096695, 'train/grad_norm_pre_clip': 0.9110642075538635, 'timing/data': 0.0002597370184957981, 'timing/model': 0.4177307481877506, 'learning_rate/qwen_vl_interface': 7.065978636794704e-06, 'learning_rate/action_model': 7.065978636794704e-05, 'epoch': 0.29})
+[2026-07-28 04:52:13,023][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.03484002500772476, 'train/grad_norm_pre_clip': 0.885230302810669, 'timing/data': 0.000255512073636055, 'timing/model': 0.41893713735044, 'learning_rate/qwen_vl_interface': 7.062442419425904e-06, 'learning_rate/action_model': 7.062442419425903e-05, 'epoch': 0.29})
+[2026-07-28 04:52:18,907][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.06829681992530823, 'train/grad_norm_pre_clip': 1.858641505241394, 'timing/data': 0.00028340378776192665, 'timing/model': 0.4191634929738939, 'learning_rate/qwen_vl_interface': 7.058905025982895e-06, 'learning_rate/action_model': 7.058905025982894e-05, 'epoch': 0.29})
+[2026-07-28 04:52:24,997][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.05229064077138901, 'train/grad_norm_pre_clip': 1.8728467226028442, 'timing/data': 0.0003984668292105198, 'timing/model': 0.5851108347997069, 'learning_rate/qwen_vl_interface': 7.055366458761056e-06, 'learning_rate/action_model': 7.055366458761056e-05, 'epoch': 0.29})
+[2026-07-28 04:52:31,187][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.04413983225822449, 'train/grad_norm_pre_clip': 2.160053253173828, 'timing/data': 0.0002896171063184738, 'timing/model': 0.6642130450345576, 'learning_rate/qwen_vl_interface': 7.051826720056522e-06, 'learning_rate/action_model': 7.051826720056521e-05, 'epoch': 0.29})
+[2026-07-28 04:52:37,091][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.02146846428513527, 'train/grad_norm_pre_clip': 0.6306691765785217, 'timing/data': 0.0002794191241264343, 'timing/model': 0.4187914398498833, 'learning_rate/qwen_vl_interface': 7.04828581216619e-06, 'learning_rate/action_model': 7.04828581216619e-05, 'epoch': 0.29})
+[2026-07-28 04:52:43,008][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.06371358782052994, 'train/grad_norm_pre_clip': 1.175372838973999, 'timing/data': 0.000285114161670208, 'timing/model': 0.41947576217353344, 'learning_rate/qwen_vl_interface': 7.04474373738772e-06, 'learning_rate/action_model': 7.04474373738772e-05, 'epoch': 0.29})
+[2026-07-28 04:52:49,131][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.05655089020729065, 'train/grad_norm_pre_clip': 1.573127269744873, 'timing/data': 0.0005097053945064545, 'timing/model': 0.41849951911717653, 'learning_rate/qwen_vl_interface': 7.041200498019523e-06, 'learning_rate/action_model': 7.041200498019523e-05, 'epoch': 0.29})
+[2026-07-28 04:52:55,311][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.10430841147899628, 'train/grad_norm_pre_clip': 4.117564678192139, 'timing/data': 0.0003935028798878193, 'timing/model': 0.419803979806602, 'learning_rate/qwen_vl_interface': 7.037656096360771e-06, 'learning_rate/action_model': 7.037656096360771e-05, 'epoch': 0.29})
+[2026-07-28 04:53:01,260][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.03366749733686447, 'train/grad_norm_pre_clip': 1.7577275037765503, 'timing/data': 0.0002839239314198494, 'timing/model': 0.41851137997582555, 'learning_rate/qwen_vl_interface': 7.0341105347113844e-06, 'learning_rate/action_model': 7.034110534711384e-05, 'epoch': 0.29})
+[2026-07-28 04:53:07,190][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.06799200922250748, 'train/grad_norm_pre_clip': 1.5550413131713867, 'timing/data': 0.0002705371007323265, 'timing/model': 0.4158451878465712, 'learning_rate/qwen_vl_interface': 7.030563815372042e-06, 'learning_rate/action_model': 7.030563815372042e-05, 'epoch': 0.29})
+[2026-07-28 04:53:13,328][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.04734271764755249, 'train/grad_norm_pre_clip': 0.7952618598937988, 'timing/data': 0.00025150785222649574, 'timing/model': 0.4156569908373058, 'learning_rate/qwen_vl_interface': 7.027015940644173e-06, 'learning_rate/action_model': 7.027015940644173e-05, 'epoch': 0.29})
+[2026-07-28 04:53:19,457][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.039618197828531265, 'train/grad_norm_pre_clip': 1.1544359922409058, 'timing/data': 0.00024383515119552612, 'timing/model': 0.41385162388905883, 'learning_rate/qwen_vl_interface': 7.02346691282995e-06, 'learning_rate/action_model': 7.02346691282995e-05, 'epoch': 0.29})
+[2026-07-28 04:53:25,287][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.05348823219537735, 'train/grad_norm_pre_clip': 1.6584160327911377, 'timing/data': 0.00026959972456097603, 'timing/model': 0.41507853427901864, 'learning_rate/qwen_vl_interface': 7.019916734232304e-06, 'learning_rate/action_model': 7.019916734232303e-05, 'epoch': 0.29})
+[2026-07-28 04:53:30,924][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.0392761267721653, 'train/grad_norm_pre_clip': 1.5282671451568604, 'timing/data': 0.0002556340768933296, 'timing/model': 0.416612749453634, 'learning_rate/qwen_vl_interface': 7.0163654071549035e-06, 'learning_rate/action_model': 7.016365407154904e-05, 'epoch': 0.29})
+[2026-07-28 04:53:37,058][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.033828191459178925, 'train/grad_norm_pre_clip': 0.8271478414535522, 'timing/data': 0.00031908508390188217, 'timing/model': 0.6724778590723872, 'learning_rate/qwen_vl_interface': 7.012812933902167e-06, 'learning_rate/action_model': 7.012812933902167e-05, 'epoch': 0.29})
+[2026-07-28 04:53:43,196][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.05448777973651886, 'train/grad_norm_pre_clip': 1.2361828088760376, 'timing/data': 0.0002743508666753769, 'timing/model': 0.6516794390045106, 'learning_rate/qwen_vl_interface': 7.009259316779258e-06, 'learning_rate/action_model': 7.009259316779257e-05, 'epoch': 0.29})
+[2026-07-28 04:53:49,066][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.0878947377204895, 'train/grad_norm_pre_clip': 1.629123330116272, 'timing/data': 0.00025869719684123993, 'timing/model': 0.41565171210095286, 'learning_rate/qwen_vl_interface': 7.005704558092078e-06, 'learning_rate/action_model': 7.005704558092078e-05, 'epoch': 0.29})
+[2026-07-28 04:53:55,209][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.0421023815870285, 'train/grad_norm_pre_clip': 1.6281224489212036, 'timing/data': 0.0002856329083442688, 'timing/model': 0.4747609170153737, 'learning_rate/qwen_vl_interface': 7.002148660147271e-06, 'learning_rate/action_model': 7.002148660147271e-05, 'epoch': 0.29})
+[2026-07-28 04:54:01,125][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.10205145180225372, 'train/grad_norm_pre_clip': 2.726253032684326, 'timing/data': 0.0002762819640338421, 'timing/model': 0.41699767718091607, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.29})
+[2026-07-28 04:54:07,419][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.07356320321559906, 'train/grad_norm_pre_clip': 1.8267269134521484, 'timing/data': 0.0002855858765542507, 'timing/model': 0.6501962128095329, 'learning_rate/qwen_vl_interface': 6.99503345571505e-06, 'learning_rate/action_model': 6.99503345571505e-05, 'epoch': 0.29})
+[2026-07-28 04:54:13,510][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.05210184305906296, 'train/grad_norm_pre_clip': 1.727946162223816, 'timing/data': 0.00025396374985575676, 'timing/model': 0.4161989511922002, 'learning_rate/qwen_vl_interface': 6.991474153844614e-06, 'learning_rate/action_model': 6.991474153844613e-05, 'epoch': 0.29})
+[2026-07-28 04:54:19,721][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.03967607393860817, 'train/grad_norm_pre_clip': 0.7313718199729919, 'timing/data': 0.00027718301862478256, 'timing/model': 0.4158071018755436, 'learning_rate/qwen_vl_interface': 6.98791372195051e-06, 'learning_rate/action_model': 6.98791372195051e-05, 'epoch': 0.29})
+[2026-07-28 04:54:25,722][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.06737802177667618, 'train/grad_norm_pre_clip': 2.0188889503479004, 'timing/data': 0.0002693459391593933, 'timing/model': 0.4141208059154451, 'learning_rate/qwen_vl_interface': 6.984352162343056e-06, 'learning_rate/action_model': 6.984352162343057e-05, 'epoch': 0.29})
+[2026-07-28 04:54:32,001][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.03303644061088562, 'train/grad_norm_pre_clip': 0.967461884021759, 'timing/data': 0.00033988989889621735, 'timing/model': 0.41416637878865004, 'learning_rate/qwen_vl_interface': 6.980789477333316e-06, 'learning_rate/action_model': 6.980789477333316e-05, 'epoch': 0.29})
+[2026-07-28 04:54:38,134][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.04648103564977646, 'train/grad_norm_pre_clip': 1.144563913345337, 'timing/data': 0.00026897015050053596, 'timing/model': 0.4151954553090036, 'learning_rate/qwen_vl_interface': 6.977225669233074e-06, 'learning_rate/action_model': 6.977225669233074e-05, 'epoch': 0.29})
+[2026-07-28 04:54:44,509][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.046383753418922424, 'train/grad_norm_pre_clip': 0.7671251893043518, 'timing/data': 0.0002625701017677784, 'timing/model': 0.6504649729467928, 'learning_rate/qwen_vl_interface': 6.973660740354848e-06, 'learning_rate/action_model': 6.973660740354848e-05, 'epoch': 0.29})
+[2026-07-28 04:54:50,481][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.06278051435947418, 'train/grad_norm_pre_clip': 1.4230077266693115, 'timing/data': 0.0003256858326494694, 'timing/model': 0.41191147407516837, 'learning_rate/qwen_vl_interface': 6.970094693011881e-06, 'learning_rate/action_model': 6.97009469301188e-05, 'epoch': 0.29})
+[2026-07-28 04:54:56,580][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.08119231462478638, 'train/grad_norm_pre_clip': 1.6267203092575073, 'timing/data': 0.00025798194110393524, 'timing/model': 0.641167308203876, 'learning_rate/qwen_vl_interface': 6.966527529518142e-06, 'learning_rate/action_model': 6.966527529518142e-05, 'epoch': 0.29})
+[2026-07-28 04:55:02,428][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.04505595564842224, 'train/grad_norm_pre_clip': 1.3189005851745605, 'timing/data': 0.0002555358223617077, 'timing/model': 0.4125878410413861, 'learning_rate/qwen_vl_interface': 6.962959252188325e-06, 'learning_rate/action_model': 6.962959252188325e-05, 'epoch': 0.29})
+[2026-07-28 04:55:08,290][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.04036654159426689, 'train/grad_norm_pre_clip': 1.0291693210601807, 'timing/data': 0.0003687632270157337, 'timing/model': 0.41751555213704705, 'learning_rate/qwen_vl_interface': 6.9593898633378476e-06, 'learning_rate/action_model': 6.959389863337847e-05, 'epoch': 0.29})
+[2026-07-28 04:55:14,429][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.03964368253946304, 'train/grad_norm_pre_clip': 1.0212304592132568, 'timing/data': 0.0002481951378285885, 'timing/model': 0.41281544603407383, 'learning_rate/qwen_vl_interface': 6.955819365282846e-06, 'learning_rate/action_model': 6.955819365282845e-05, 'epoch': 0.29})
+[2026-07-28 04:55:20,282][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.026887744665145874, 'train/grad_norm_pre_clip': 0.8183910846710205, 'timing/data': 0.0003284919075667858, 'timing/model': 0.41465865913778543, 'learning_rate/qwen_vl_interface': 6.952247760340176e-06, 'learning_rate/action_model': 6.952247760340176e-05, 'epoch': 0.29})
+[2026-07-28 04:55:26,125][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.07015920430421829, 'train/grad_norm_pre_clip': 2.0773346424102783, 'timing/data': 0.000274749007076025, 'timing/model': 0.41481243167072535, 'learning_rate/qwen_vl_interface': 6.948675050827418e-06, 'learning_rate/action_model': 6.948675050827419e-05, 'epoch': 0.29})
+[2026-07-28 04:55:32,194][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.0371328666806221, 'train/grad_norm_pre_clip': 1.026075839996338, 'timing/data': 0.0002498440444469452, 'timing/model': 0.4123742408119142, 'learning_rate/qwen_vl_interface': 6.945101239062862e-06, 'learning_rate/action_model': 6.945101239062863e-05, 'epoch': 0.29})
+[2026-07-28 04:55:38,284][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.04880714416503906, 'train/grad_norm_pre_clip': 1.7883775234222412, 'timing/data': 0.00024969736114144325, 'timing/model': 0.4145615450106561, 'learning_rate/qwen_vl_interface': 6.941526327365513e-06, 'learning_rate/action_model': 6.941526327365513e-05, 'epoch': 0.29})
+[2026-07-28 04:55:44,134][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.06377588957548141, 'train/grad_norm_pre_clip': 1.6914384365081787, 'timing/data': 0.00028056418523192406, 'timing/model': 0.41737574292346835, 'learning_rate/qwen_vl_interface': 6.937950318055098e-06, 'learning_rate/action_model': 6.937950318055098e-05, 'epoch': 0.29})
+[2026-07-28 04:55:50,042][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.08851610124111176, 'train/grad_norm_pre_clip': 2.248615264892578, 'timing/data': 0.0002596266567707062, 'timing/model': 0.41532836901023984, 'learning_rate/qwen_vl_interface': 6.934373213452045e-06, 'learning_rate/action_model': 6.934373213452044e-05, 'epoch': 0.29})
+[2026-07-28 04:55:56,144][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.04242616146802902, 'train/grad_norm_pre_clip': 1.1598670482635498, 'timing/data': 0.0002812962047755718, 'timing/model': 0.41468859417364, 'learning_rate/qwen_vl_interface': 6.930795015877502e-06, 'learning_rate/action_model': 6.930795015877502e-05, 'epoch': 0.29})
+[2026-07-28 04:56:02,021][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.039123885333538055, 'train/grad_norm_pre_clip': 1.9417388439178467, 'timing/data': 0.0002678120508790016, 'timing/model': 0.4143290529027581, 'learning_rate/qwen_vl_interface': 6.927215727653321e-06, 'learning_rate/action_model': 6.927215727653321e-05, 'epoch': 0.29})
+[2026-07-28 04:56:08,162][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.04411657154560089, 'train/grad_norm_pre_clip': 1.3916925191879272, 'timing/data': 0.0002748738043010235, 'timing/model': 0.6587153389118612, 'learning_rate/qwen_vl_interface': 6.9236353511020625e-06, 'learning_rate/action_model': 6.923635351102063e-05, 'epoch': 0.29})
+[2026-07-28 04:56:14,254][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.05526237562298775, 'train/grad_norm_pre_clip': 0.9172955751419067, 'timing/data': 0.0002464400604367256, 'timing/model': 0.4146587341092527, 'learning_rate/qwen_vl_interface': 6.920053888546995e-06, 'learning_rate/action_model': 6.920053888546994e-05, 'epoch': 0.29})
+[2026-07-28 04:56:20,134][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.05288117378950119, 'train/grad_norm_pre_clip': 1.6835747957229614, 'timing/data': 0.00027152569964528084, 'timing/model': 0.4148014257661998, 'learning_rate/qwen_vl_interface': 6.916471342312093e-06, 'learning_rate/action_model': 6.916471342312093e-05, 'epoch': 0.29})
+[2026-07-28 04:56:26,009][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.06224726140499115, 'train/grad_norm_pre_clip': 1.534077525138855, 'timing/data': 0.00037723174318671227, 'timing/model': 0.4152652486227453, 'learning_rate/qwen_vl_interface': 6.9128877147220274e-06, 'learning_rate/action_model': 6.912887714722027e-05, 'epoch': 0.29})
+[2026-07-28 04:56:32,109][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.06799742579460144, 'train/grad_norm_pre_clip': 1.420969009399414, 'timing/data': 0.00030406564474105835, 'timing/model': 0.5822199713438749, 'learning_rate/qwen_vl_interface': 6.909303008102179e-06, 'learning_rate/action_model': 6.909303008102179e-05, 'epoch': 0.3})
+[2026-07-28 04:56:37,976][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.05418878793716431, 'train/grad_norm_pre_clip': 1.1145631074905396, 'timing/data': 0.00027282489463686943, 'timing/model': 0.41649059671908617, 'learning_rate/qwen_vl_interface': 6.905717224778623e-06, 'learning_rate/action_model': 6.905717224778622e-05, 'epoch': 0.3})
+[2026-07-28 04:56:43,830][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.03796866908669472, 'train/grad_norm_pre_clip': 0.5089591145515442, 'timing/data': 0.000272147823125124, 'timing/model': 0.4156278590671718, 'learning_rate/qwen_vl_interface': 6.902130367078135e-06, 'learning_rate/action_model': 6.902130367078134e-05, 'epoch': 0.3})
+[2026-07-28 04:56:49,962][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.07082545757293701, 'train/grad_norm_pre_clip': 2.00181245803833, 'timing/data': 0.00033486727625131607, 'timing/model': 0.41332150623202324, 'learning_rate/qwen_vl_interface': 6.8985424373281895e-06, 'learning_rate/action_model': 6.89854243732819e-05, 'epoch': 0.3})
+[2026-07-28 04:56:56,088][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.10029943287372589, 'train/grad_norm_pre_clip': 1.3299058675765991, 'timing/data': 0.00026513123884797096, 'timing/model': 0.4140913509763777, 'learning_rate/qwen_vl_interface': 6.894953437856954e-06, 'learning_rate/action_model': 6.894953437856953e-05, 'epoch': 0.3})
+[2026-07-28 04:57:01,956][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.056724004447460175, 'train/grad_norm_pre_clip': 1.7510532140731812, 'timing/data': 0.00027162209153175354, 'timing/model': 0.41555530624464154, 'learning_rate/qwen_vl_interface': 6.891363370993292e-06, 'learning_rate/action_model': 6.891363370993293e-05, 'epoch': 0.3})
+[2026-07-28 04:57:07,814][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.04840433597564697, 'train/grad_norm_pre_clip': 1.6389273405075073, 'timing/data': 0.0002629552036523819, 'timing/model': 0.4152241167612374, 'learning_rate/qwen_vl_interface': 6.88777223906676e-06, 'learning_rate/action_model': 6.88777223906676e-05, 'epoch': 0.3})
+[2026-07-28 04:57:13,908][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.06466846168041229, 'train/grad_norm_pre_clip': 1.0166834592819214, 'timing/data': 0.0002723722718656063, 'timing/model': 0.6534700249321759, 'learning_rate/qwen_vl_interface': 6.8841800444076e-06, 'learning_rate/action_model': 6.8841800444076e-05, 'epoch': 0.3})
+[2026-07-28 04:57:20,046][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.04064587876200676, 'train/grad_norm_pre_clip': 1.4773951768875122, 'timing/data': 0.0002558608539402485, 'timing/model': 0.41709413193166256, 'learning_rate/qwen_vl_interface': 6.8805867893467525e-06, 'learning_rate/action_model': 6.880586789346752e-05, 'epoch': 0.3})
+[2026-07-28 04:57:25,931][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.05378426983952522, 'train/grad_norm_pre_clip': 1.101915717124939, 'timing/data': 0.00036186911165714264, 'timing/model': 0.4160297950729728, 'learning_rate/qwen_vl_interface': 6.87699247621584e-06, 'learning_rate/action_model': 6.87699247621584e-05, 'epoch': 0.3})
+[2026-07-28 04:57:31,824][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.03260621428489685, 'train/grad_norm_pre_clip': 0.5796162486076355, 'timing/data': 0.0002706451341509819, 'timing/model': 0.41430630488321185, 'learning_rate/qwen_vl_interface': 6.873397107347176e-06, 'learning_rate/action_model': 6.873397107347176e-05, 'epoch': 0.3})
+[2026-07-28 04:57:37,934][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.052852727472782135, 'train/grad_norm_pre_clip': 1.0537301301956177, 'timing/data': 0.0002512098290026188, 'timing/model': 0.4130445830523968, 'learning_rate/qwen_vl_interface': 6.869800685073752e-06, 'learning_rate/action_model': 6.869800685073752e-05, 'epoch': 0.3})
+[2026-07-28 04:57:43,789][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.05519542098045349, 'train/grad_norm_pre_clip': 1.9852626323699951, 'timing/data': 0.00033672014251351357, 'timing/model': 0.41460662614554167, 'learning_rate/qwen_vl_interface': 6.86620321172925e-06, 'learning_rate/action_model': 6.86620321172925e-05, 'epoch': 0.3})
+[2026-07-28 04:57:49,868][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.07829821109771729, 'train/grad_norm_pre_clip': 2.18058443069458, 'timing/data': 0.00026615383103489876, 'timing/model': 0.64391651423648, 'learning_rate/qwen_vl_interface': 6.862604689648029e-06, 'learning_rate/action_model': 6.862604689648029e-05, 'epoch': 0.3})
+[2026-07-28 04:57:55,726][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.03256482630968094, 'train/grad_norm_pre_clip': 1.3038966655731201, 'timing/data': 0.00024809129536151886, 'timing/model': 0.4119938062503934, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.3})
+[2026-07-28 04:58:01,843][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.06571584939956665, 'train/grad_norm_pre_clip': 1.5742281675338745, 'timing/data': 0.00024662911891937256, 'timing/model': 0.4120489410124719, 'learning_rate/qwen_vl_interface': 6.855404508616285e-06, 'learning_rate/action_model': 6.855404508616285e-05, 'epoch': 0.3})
+[2026-07-28 04:58:08,037][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.05243750661611557, 'train/grad_norm_pre_clip': 1.2630066871643066, 'timing/data': 0.00031188083812594414, 'timing/model': 0.41844631591811776, 'learning_rate/qwen_vl_interface': 6.851802854337881e-06, 'learning_rate/action_model': 6.851802854337881e-05, 'epoch': 0.3})
+[2026-07-28 04:58:14,008][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.0788927972316742, 'train/grad_norm_pre_clip': 4.281049728393555, 'timing/data': 0.0007235207594931126, 'timing/model': 0.41324902419000864, 'learning_rate/qwen_vl_interface': 6.848200160666993e-06, 'learning_rate/action_model': 6.848200160666993e-05, 'epoch': 0.3})
+[2026-07-28 04:58:20,364][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.08587951958179474, 'train/grad_norm_pre_clip': 2.7860465049743652, 'timing/data': 0.000247288029640913, 'timing/model': 0.648381834384054, 'learning_rate/qwen_vl_interface': 6.844596429941374e-06, 'learning_rate/action_model': 6.844596429941373e-05, 'epoch': 0.3})
+[2026-07-28 04:58:26,238][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.038202837109565735, 'train/grad_norm_pre_clip': 2.038203001022339, 'timing/data': 0.0002768137492239475, 'timing/model': 0.41550407791510224, 'learning_rate/qwen_vl_interface': 6.8409916644994425e-06, 'learning_rate/action_model': 6.840991664499442e-05, 'epoch': 0.3})
+[2026-07-28 04:58:32,081][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.06375709176063538, 'train/grad_norm_pre_clip': 1.8166968822479248, 'timing/data': 0.00025577982887625694, 'timing/model': 0.41502135805785656, 'learning_rate/qwen_vl_interface': 6.837385866680294e-06, 'learning_rate/action_model': 6.837385866680293e-05, 'epoch': 0.3})
+[2026-07-28 04:58:38,154][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.06373946368694305, 'train/grad_norm_pre_clip': 1.6330232620239258, 'timing/data': 0.000250163022428751, 'timing/model': 0.41469103330746293, 'learning_rate/qwen_vl_interface': 6.833779038823691e-06, 'learning_rate/action_model': 6.833779038823691e-05, 'epoch': 0.3})
+[2026-07-28 04:58:44,038][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.03449522703886032, 'train/grad_norm_pre_clip': 2.0191972255706787, 'timing/data': 0.00027748895809054375, 'timing/model': 0.4221293949522078, 'learning_rate/qwen_vl_interface': 6.830171183270063e-06, 'learning_rate/action_model': 6.830171183270064e-05, 'epoch': 0.3})
+[2026-07-28 04:58:50,154][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.07087947428226471, 'train/grad_norm_pre_clip': 2.550269842147827, 'timing/data': 0.0002690521068871021, 'timing/model': 0.6508902739733458, 'learning_rate/qwen_vl_interface': 6.826562302360511e-06, 'learning_rate/action_model': 6.82656230236051e-05, 'epoch': 0.3})
+[2026-07-28 04:58:56,048][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.055872540920972824, 'train/grad_norm_pre_clip': 2.629476308822632, 'timing/data': 0.00024021603167057037, 'timing/model': 0.41272640926763415, 'learning_rate/qwen_vl_interface': 6.822952398436797e-06, 'learning_rate/action_model': 6.822952398436797e-05, 'epoch': 0.3})
+[2026-07-28 04:59:02,179][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.08194206655025482, 'train/grad_norm_pre_clip': 2.5512988567352295, 'timing/data': 0.00027154432609677315, 'timing/model': 0.6535855899564922, 'learning_rate/qwen_vl_interface': 6.81934147384135e-06, 'learning_rate/action_model': 6.81934147384135e-05, 'epoch': 0.3})
+[2026-07-28 04:59:08,093][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.1117960512638092, 'train/grad_norm_pre_clip': 3.4384877681732178, 'timing/data': 0.0003313361667096615, 'timing/model': 0.4133833469823003, 'learning_rate/qwen_vl_interface': 6.815729530917257e-06, 'learning_rate/action_model': 6.815729530917256e-05, 'epoch': 0.3})
+[2026-07-28 04:59:14,019][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.05320967361330986, 'train/grad_norm_pre_clip': 3.5204827785491943, 'timing/data': 0.00028115930035710335, 'timing/model': 0.42225643899291754, 'learning_rate/qwen_vl_interface': 6.812116572008271e-06, 'learning_rate/action_model': 6.812116572008271e-05, 'epoch': 0.3})
+[2026-07-28 04:59:19,920][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.03451186418533325, 'train/grad_norm_pre_clip': 1.20654296875, 'timing/data': 0.00025776494294404984, 'timing/model': 0.4206183631904423, 'learning_rate/qwen_vl_interface': 6.808502599458803e-06, 'learning_rate/action_model': 6.808502599458803e-05, 'epoch': 0.3})
+[2026-07-28 04:59:26,049][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.06612455099821091, 'train/grad_norm_pre_clip': 2.566455841064453, 'timing/data': 0.00027076201513409615, 'timing/model': 0.6498218029737473, 'learning_rate/qwen_vl_interface': 6.8048876156139154e-06, 'learning_rate/action_model': 6.804887615613915e-05, 'epoch': 0.3})
+[2026-07-28 04:59:31,928][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.056203071027994156, 'train/grad_norm_pre_clip': 1.8600287437438965, 'timing/data': 0.00025372905656695366, 'timing/model': 0.41253732657060027, 'learning_rate/qwen_vl_interface': 6.801271622819337e-06, 'learning_rate/action_model': 6.801271622819336e-05, 'epoch': 0.3})
+[2026-07-28 04:59:38,065][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.04542318731546402, 'train/grad_norm_pre_clip': 2.3180065155029297, 'timing/data': 0.00026191817596554756, 'timing/model': 0.41260351287201047, 'learning_rate/qwen_vl_interface': 6.797654623421443e-06, 'learning_rate/action_model': 6.797654623421444e-05, 'epoch': 0.3})
+[2026-07-28 04:59:43,941][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.03813536465167999, 'train/grad_norm_pre_clip': 0.8506457209587097, 'timing/data': 0.0002710539847612381, 'timing/model': 0.42290632985532284, 'learning_rate/qwen_vl_interface': 6.794036619767269e-06, 'learning_rate/action_model': 6.79403661976727e-05, 'epoch': 0.3})
+[2026-07-28 04:59:49,864][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.05353051424026489, 'train/grad_norm_pre_clip': 0.9303241968154907, 'timing/data': 0.00027458788827061653, 'timing/model': 0.42244751611724496, 'learning_rate/qwen_vl_interface': 6.790417614204494e-06, 'learning_rate/action_model': 6.790417614204494e-05, 'epoch': 0.3})
+[2026-07-28 04:59:55,993][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.04878878965973854, 'train/grad_norm_pre_clip': 1.7120749950408936, 'timing/data': 0.00026342133060097694, 'timing/model': 0.41360124899074435, 'learning_rate/qwen_vl_interface': 6.786797609081453e-06, 'learning_rate/action_model': 6.786797609081453e-05, 'epoch': 0.3})
+[2026-07-28 05:00:02,125][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.0343426838517189, 'train/grad_norm_pre_clip': 1.5729947090148926, 'timing/data': 0.00027342792600393295, 'timing/model': 0.4238904397934675, 'learning_rate/qwen_vl_interface': 6.783176606747129e-06, 'learning_rate/action_model': 6.78317660674713e-05, 'epoch': 0.3})
+[2026-07-28 05:00:08,068][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.05264418572187424, 'train/grad_norm_pre_clip': 0.6533450484275818, 'timing/data': 0.000283682718873024, 'timing/model': 0.4246660298667848, 'learning_rate/qwen_vl_interface': 6.77955460955115e-06, 'learning_rate/action_model': 6.77955460955115e-05, 'epoch': 0.3})
+[2026-07-28 05:00:13,754][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.04800180345773697, 'train/grad_norm_pre_clip': 1.6339536905288696, 'timing/data': 0.00032995687797665596, 'timing/model': 0.4231255850754678, 'learning_rate/qwen_vl_interface': 6.775931619843791e-06, 'learning_rate/action_model': 6.77593161984379e-05, 'epoch': 0.3})
+[2026-07-28 05:00:19,904][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.03364485502243042, 'train/grad_norm_pre_clip': 1.586127758026123, 'timing/data': 0.00025207875296473503, 'timing/model': 0.4121519820764661, 'learning_rate/qwen_vl_interface': 6.7723076399759715e-06, 'learning_rate/action_model': 6.772307639975972e-05, 'epoch': 0.3})
+[2026-07-28 05:00:25,574][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.04251111298799515, 'train/grad_norm_pre_clip': 1.3234903812408447, 'timing/data': 0.00025834981352090836, 'timing/model': 0.4151589940302074, 'learning_rate/qwen_vl_interface': 6.7686826722992516e-06, 'learning_rate/action_model': 6.768682672299252e-05, 'epoch': 0.3})
+[2026-07-28 05:00:31,225][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.10688267648220062, 'train/grad_norm_pre_clip': 1.9022166728973389, 'timing/data': 0.00027297576889395714, 'timing/model': 0.4147548913024366, 'learning_rate/qwen_vl_interface': 6.765056719165833e-06, 'learning_rate/action_model': 6.765056719165834e-05, 'epoch': 0.3})
+[2026-07-28 05:00:37,271][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.03994842246174812, 'train/grad_norm_pre_clip': 1.625003457069397, 'timing/data': 0.00028787460178136826, 'timing/model': 0.5837717391550541, 'learning_rate/qwen_vl_interface': 6.761429782928558e-06, 'learning_rate/action_model': 6.761429782928557e-05, 'epoch': 0.3})
+[2026-07-28 05:00:43,167][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.10006604343652725, 'train/grad_norm_pre_clip': 3.0164599418640137, 'timing/data': 0.00028174323961138725, 'timing/model': 0.41833500005304813, 'learning_rate/qwen_vl_interface': 6.757801865940908e-06, 'learning_rate/action_model': 6.757801865940908e-05, 'epoch': 0.3})
+[2026-07-28 05:00:49,298][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.12248503416776657, 'train/grad_norm_pre_clip': 3.0992541313171387, 'timing/data': 0.0002956087701022625, 'timing/model': 0.6547673060558736, 'learning_rate/qwen_vl_interface': 6.754172970556996e-06, 'learning_rate/action_model': 6.754172970556996e-05, 'epoch': 0.3})
+[2026-07-28 05:00:55,188][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.053778596222400665, 'train/grad_norm_pre_clip': 1.6035981178283691, 'timing/data': 0.00040761707350611687, 'timing/model': 0.4145128056406975, 'learning_rate/qwen_vl_interface': 6.750543099131573e-06, 'learning_rate/action_model': 6.750543099131573e-05, 'epoch': 0.3})
+[2026-07-28 05:01:01,283][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.03093712404370308, 'train/grad_norm_pre_clip': 0.6858956813812256, 'timing/data': 0.0002841879613697529, 'timing/model': 0.4153885650448501, 'learning_rate/qwen_vl_interface': 6.746912254020025e-06, 'learning_rate/action_model': 6.746912254020025e-05, 'epoch': 0.3})
+[2026-07-28 05:01:06,928][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.031254999339580536, 'train/grad_norm_pre_clip': 1.108640193939209, 'timing/data': 0.00029452797025442123, 'timing/model': 0.4158467832021415, 'learning_rate/qwen_vl_interface': 6.743280437578366e-06, 'learning_rate/action_model': 6.743280437578366e-05, 'epoch': 0.3})
+[2026-07-28 05:01:12,803][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.11515218019485474, 'train/grad_norm_pre_clip': 2.6109166145324707, 'timing/data': 0.00033481698483228683, 'timing/model': 0.41571636917069554, 'learning_rate/qwen_vl_interface': 6.7396476521632456e-06, 'learning_rate/action_model': 6.739647652163246e-05, 'epoch': 0.3})
+[2026-07-28 05:01:18,672][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.07352195680141449, 'train/grad_norm_pre_clip': 3.046442985534668, 'timing/data': 0.00028251390904188156, 'timing/model': 0.4169205389916897, 'learning_rate/qwen_vl_interface': 6.736013900131934e-06, 'learning_rate/action_model': 6.736013900131934e-05, 'epoch': 0.3})
+[2026-07-28 05:01:24,750][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.06589247286319733, 'train/grad_norm_pre_clip': 1.4486713409423828, 'timing/data': 0.000253364909440279, 'timing/model': 0.41854483215138316, 'learning_rate/qwen_vl_interface': 6.732379183842339e-06, 'learning_rate/action_model': 6.73237918384234e-05, 'epoch': 0.3})
+[2026-07-28 05:01:30,848][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.0595349557697773, 'train/grad_norm_pre_clip': 1.2227036952972412, 'timing/data': 0.00025349436327815056, 'timing/model': 0.41384317679330707, 'learning_rate/qwen_vl_interface': 6.728743505652984e-06, 'learning_rate/action_model': 6.728743505652983e-05, 'epoch': 0.3})
+[2026-07-28 05:01:36,490][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.04658685624599457, 'train/grad_norm_pre_clip': 1.105409860610962, 'timing/data': 0.000291645061224699, 'timing/model': 0.4152935538440943, 'learning_rate/qwen_vl_interface': 6.725106867923023e-06, 'learning_rate/action_model': 6.725106867923023e-05, 'epoch': 0.3})
+[2026-07-28 05:01:42,365][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.05640875920653343, 'train/grad_norm_pre_clip': 1.260441780090332, 'timing/data': 0.0003623371012508869, 'timing/model': 0.41592788975685835, 'learning_rate/qwen_vl_interface': 6.721469273012229e-06, 'learning_rate/action_model': 6.721469273012229e-05, 'epoch': 0.3})
+[2026-07-28 05:01:48,078][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.033905062824487686, 'train/grad_norm_pre_clip': 0.8619915246963501, 'timing/data': 0.00027557695284485817, 'timing/model': 0.4164583287201822, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.3})
+[2026-07-28 05:01:53,980][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.03835318610072136, 'train/grad_norm_pre_clip': 0.9395606517791748, 'timing/data': 0.00027743028476834297, 'timing/model': 0.6536315660923719, 'learning_rate/qwen_vl_interface': 6.714191221090352e-06, 'learning_rate/action_model': 6.714191221090352e-05, 'epoch': 0.31})
+[2026-07-28 05:01:59,853][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.045871078968048096, 'train/grad_norm_pre_clip': 2.080014228820801, 'timing/data': 0.00027135759592056274, 'timing/model': 0.413560395129025, 'learning_rate/qwen_vl_interface': 6.710550768801916e-06, 'learning_rate/action_model': 6.710550768801915e-05, 'epoch': 0.31})
+[2026-07-28 05:02:05,721][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.0704755187034607, 'train/grad_norm_pre_clip': 1.8588505983352661, 'timing/data': 0.0002753492444753647, 'timing/model': 0.4162114807404578, 'learning_rate/qwen_vl_interface': 6.706909368777944e-06, 'learning_rate/action_model': 6.706909368777944e-05, 'epoch': 0.31})
+[2026-07-28 05:02:11,593][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.05682515352964401, 'train/grad_norm_pre_clip': 1.6078839302062988, 'timing/data': 0.00028291111811995506, 'timing/model': 0.41620638128370047, 'learning_rate/qwen_vl_interface': 6.703267023381299e-06, 'learning_rate/action_model': 6.7032670233813e-05, 'epoch': 0.31})
+[2026-07-28 05:02:17,712][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.042330559343099594, 'train/grad_norm_pre_clip': 1.0053973197937012, 'timing/data': 0.0002649887464940548, 'timing/model': 0.4131108019500971, 'learning_rate/qwen_vl_interface': 6.69962373497546e-06, 'learning_rate/action_model': 6.69962373497546e-05, 'epoch': 0.31})
+[2026-07-28 05:02:23,835][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.04393298923969269, 'train/grad_norm_pre_clip': 1.028351902961731, 'timing/data': 0.00029072025790810585, 'timing/model': 0.41375855868682265, 'learning_rate/qwen_vl_interface': 6.69597950592452e-06, 'learning_rate/action_model': 6.69597950592452e-05, 'epoch': 0.31})
+[2026-07-28 05:02:29,768][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.08297888189554214, 'train/grad_norm_pre_clip': 2.129319429397583, 'timing/data': 0.00027282116934657097, 'timing/model': 0.41488028410822153, 'learning_rate/qwen_vl_interface': 6.692334338593175e-06, 'learning_rate/action_model': 6.692334338593175e-05, 'epoch': 0.31})
+[2026-07-28 05:02:35,633][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.0839940756559372, 'train/grad_norm_pre_clip': 1.9480305910110474, 'timing/data': 0.0002738870680332184, 'timing/model': 0.4159550191834569, 'learning_rate/qwen_vl_interface': 6.688688235346736e-06, 'learning_rate/action_model': 6.688688235346736e-05, 'epoch': 0.31})
+[2026-07-28 05:02:41,927][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.06165783107280731, 'train/grad_norm_pre_clip': 1.7441434860229492, 'timing/data': 0.000245444942265749, 'timing/model': 0.4156603920273483, 'learning_rate/qwen_vl_interface': 6.685041198551125e-06, 'learning_rate/action_model': 6.685041198551124e-05, 'epoch': 0.31})
+[2026-07-28 05:02:47,837][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.042450446635484695, 'train/grad_norm_pre_clip': 0.7524213194847107, 'timing/data': 0.0002909279428422451, 'timing/model': 0.41574364295229316, 'learning_rate/qwen_vl_interface': 6.681393230572857e-06, 'learning_rate/action_model': 6.681393230572856e-05, 'epoch': 0.31})
+[2026-07-28 05:02:53,514][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.061654090881347656, 'train/grad_norm_pre_clip': 1.6087552309036255, 'timing/data': 0.0002765711396932602, 'timing/model': 0.4158395561389625, 'learning_rate/qwen_vl_interface': 6.677744333779063e-06, 'learning_rate/action_model': 6.677744333779062e-05, 'epoch': 0.31})
+[2026-07-28 05:02:59,657][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.056441061198711395, 'train/grad_norm_pre_clip': 1.2372733354568481, 'timing/data': 0.0002894611097872257, 'timing/model': 0.6532458779402077, 'learning_rate/qwen_vl_interface': 6.674094510537471e-06, 'learning_rate/action_model': 6.674094510537471e-05, 'epoch': 0.31})
+[2026-07-28 05:03:05,810][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.05200641602277756, 'train/grad_norm_pre_clip': 0.9388198256492615, 'timing/data': 0.00028597889468073845, 'timing/model': 0.6466938788071275, 'learning_rate/qwen_vl_interface': 6.670443763216412e-06, 'learning_rate/action_model': 6.670443763216411e-05, 'epoch': 0.31})
+[2026-07-28 05:03:11,699][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.07604406028985977, 'train/grad_norm_pre_clip': 3.448169231414795, 'timing/data': 0.00024486659094691277, 'timing/model': 0.41321735782548785, 'learning_rate/qwen_vl_interface': 6.666792094184817e-06, 'learning_rate/action_model': 6.666792094184818e-05, 'epoch': 0.31})
+[2026-07-28 05:03:17,591][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.06006380170583725, 'train/grad_norm_pre_clip': 2.322406530380249, 'timing/data': 0.0002772440202534199, 'timing/model': 0.416016620118171, 'learning_rate/qwen_vl_interface': 6.663139505812215e-06, 'learning_rate/action_model': 6.663139505812215e-05, 'epoch': 0.31})
+[2026-07-28 05:03:23,512][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.04143260419368744, 'train/grad_norm_pre_clip': 1.6690877676010132, 'timing/data': 0.00031123682856559753, 'timing/model': 0.4278409667313099, 'learning_rate/qwen_vl_interface': 6.659486000468728e-06, 'learning_rate/action_model': 6.659486000468728e-05, 'epoch': 0.31})
+[2026-07-28 05:03:29,617][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.051038824021816254, 'train/grad_norm_pre_clip': 0.8591411113739014, 'timing/data': 0.0002701450139284134, 'timing/model': 0.6529426402412355, 'learning_rate/qwen_vl_interface': 6.6558315805250805e-06, 'learning_rate/action_model': 6.65583158052508e-05, 'epoch': 0.31})
+[2026-07-28 05:03:35,595][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.05379851907491684, 'train/grad_norm_pre_clip': 0.9642283916473389, 'timing/data': 0.00040881428867578506, 'timing/model': 0.41926667280495167, 'learning_rate/qwen_vl_interface': 6.652176248352581e-06, 'learning_rate/action_model': 6.65217624835258e-05, 'epoch': 0.31})
+[2026-07-28 05:03:41,841][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.054020896553993225, 'train/grad_norm_pre_clip': 0.9721719026565552, 'timing/data': 0.000253116711974144, 'timing/model': 0.41413306491449475, 'learning_rate/qwen_vl_interface': 6.64852000632314e-06, 'learning_rate/action_model': 6.648520006323139e-05, 'epoch': 0.31})
+[2026-07-28 05:03:47,725][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.086857870221138, 'train/grad_norm_pre_clip': 2.0944876670837402, 'timing/data': 0.00028204405680298805, 'timing/model': 0.4167795879766345, 'learning_rate/qwen_vl_interface': 6.644862856809247e-06, 'learning_rate/action_model': 6.644862856809247e-05, 'epoch': 0.31})
+[2026-07-28 05:03:53,369][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.05343698337674141, 'train/grad_norm_pre_clip': 2.3393030166625977, 'timing/data': 0.0002697790041565895, 'timing/model': 0.4165731347166002, 'learning_rate/qwen_vl_interface': 6.64120480218399e-06, 'learning_rate/action_model': 6.64120480218399e-05, 'epoch': 0.31})
+[2026-07-28 05:03:59,246][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.03195329010486603, 'train/grad_norm_pre_clip': 1.0959076881408691, 'timing/data': 0.0002654870040714741, 'timing/model': 0.4171482720412314, 'learning_rate/qwen_vl_interface': 6.637545844821043e-06, 'learning_rate/action_model': 6.637545844821043e-05, 'epoch': 0.31})
+[2026-07-28 05:04:05,579][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.041704703122377396, 'train/grad_norm_pre_clip': 1.5470694303512573, 'timing/data': 0.0003568721003830433, 'timing/model': 0.6520684380084276, 'learning_rate/qwen_vl_interface': 6.633885987094661e-06, 'learning_rate/action_model': 6.633885987094661e-05, 'epoch': 0.31})
+[2026-07-28 05:04:11,438][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.032223135232925415, 'train/grad_norm_pre_clip': 0.938531756401062, 'timing/data': 0.00024316413328051567, 'timing/model': 0.4131534853950143, 'learning_rate/qwen_vl_interface': 6.630225231379687e-06, 'learning_rate/action_model': 6.630225231379687e-05, 'epoch': 0.31})
+[2026-07-28 05:04:17,305][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.05581672489643097, 'train/grad_norm_pre_clip': 1.5427467823028564, 'timing/data': 0.0002866610884666443, 'timing/model': 0.4146040789783001, 'learning_rate/qwen_vl_interface': 6.626563580051546e-06, 'learning_rate/action_model': 6.626563580051546e-05, 'epoch': 0.31})
+[2026-07-28 05:04:23,400][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.06576552242040634, 'train/grad_norm_pre_clip': 1.503235101699829, 'timing/data': 0.00027603935450315475, 'timing/model': 0.41468291403725743, 'learning_rate/qwen_vl_interface': 6.622901035486241e-06, 'learning_rate/action_model': 6.622901035486241e-05, 'epoch': 0.31})
+[2026-07-28 05:04:29,303][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.019179334864020348, 'train/grad_norm_pre_clip': 0.6640846729278564, 'timing/data': 0.0002671987749636173, 'timing/model': 0.41691811801865697, 'learning_rate/qwen_vl_interface': 6.61923760006036e-06, 'learning_rate/action_model': 6.61923760006036e-05, 'epoch': 0.31})
+[2026-07-28 05:04:35,437][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.06593519449234009, 'train/grad_norm_pre_clip': 1.2874276638031006, 'timing/data': 0.0002533290535211563, 'timing/model': 0.4156507980078459, 'learning_rate/qwen_vl_interface': 6.615573276151068e-06, 'learning_rate/action_model': 6.615573276151068e-05, 'epoch': 0.31})
+[2026-07-28 05:04:41,481][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.03045848198235035, 'train/grad_norm_pre_clip': 1.109546422958374, 'timing/data': 0.00027350522577762604, 'timing/model': 0.58188739977777, 'learning_rate/qwen_vl_interface': 6.6119080661361e-06, 'learning_rate/action_model': 6.6119080661361e-05, 'epoch': 0.31})
+[2026-07-28 05:04:47,351][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.08601237088441849, 'train/grad_norm_pre_clip': 1.7319809198379517, 'timing/data': 0.0002682260237634182, 'timing/model': 0.4163201376795769, 'learning_rate/qwen_vl_interface': 6.608241972393774e-06, 'learning_rate/action_model': 6.608241972393774e-05, 'epoch': 0.31})
+[2026-07-28 05:04:53,217][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.04622545838356018, 'train/grad_norm_pre_clip': 1.9428908824920654, 'timing/data': 0.00028916820883750916, 'timing/model': 0.4160676607862115, 'learning_rate/qwen_vl_interface': 6.60457499730298e-06, 'learning_rate/action_model': 6.60457499730298e-05, 'epoch': 0.31})
+[2026-07-28 05:04:59,322][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.04897844046354294, 'train/grad_norm_pre_clip': 1.9375497102737427, 'timing/data': 0.00025272835046052933, 'timing/model': 0.41591332014650106, 'learning_rate/qwen_vl_interface': 6.600907143243176e-06, 'learning_rate/action_model': 6.600907143243176e-05, 'epoch': 0.31})
+[2026-07-28 05:05:05,440][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.03688221797347069, 'train/grad_norm_pre_clip': 0.9896755814552307, 'timing/data': 0.00028129108250141144, 'timing/model': 0.41659924387931824, 'learning_rate/qwen_vl_interface': 6.5972384125943945e-06, 'learning_rate/action_model': 6.597238412594394e-05, 'epoch': 0.31})
+[2026-07-28 05:05:11,297][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.028927065432071686, 'train/grad_norm_pre_clip': 1.6908724308013916, 'timing/data': 0.00027254363521933556, 'timing/model': 0.41640865709632635, 'learning_rate/qwen_vl_interface': 6.593568807737233e-06, 'learning_rate/action_model': 6.593568807737233e-05, 'epoch': 0.31})
+[2026-07-28 05:05:17,176][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.04743018001317978, 'train/grad_norm_pre_clip': 1.7627580165863037, 'timing/data': 0.00027225539088249207, 'timing/model': 0.4139638291671872, 'learning_rate/qwen_vl_interface': 6.5898983310528584e-06, 'learning_rate/action_model': 6.589898331052858e-05, 'epoch': 0.31})
+[2026-07-28 05:05:23,031][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.034771326929330826, 'train/grad_norm_pre_clip': 1.3304671049118042, 'timing/data': 0.00028987834230065346, 'timing/model': 0.4151972187682986, 'learning_rate/qwen_vl_interface': 6.586226984923004e-06, 'learning_rate/action_model': 6.586226984923004e-05, 'epoch': 0.31})
+[2026-07-28 05:05:29,373][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.06693345308303833, 'train/grad_norm_pre_clip': 3.1402530670166016, 'timing/data': 0.00025650300085544586, 'timing/model': 0.6625340157188475, 'learning_rate/qwen_vl_interface': 6.582554771729967e-06, 'learning_rate/action_model': 6.582554771729966e-05, 'epoch': 0.31})
+[2026-07-28 05:05:35,265][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.06295087933540344, 'train/grad_norm_pre_clip': 1.6607626676559448, 'timing/data': 0.00024399394169449806, 'timing/model': 0.4133933437988162, 'learning_rate/qwen_vl_interface': 6.578881693856605e-06, 'learning_rate/action_model': 6.578881693856604e-05, 'epoch': 0.31})
+[2026-07-28 05:05:41,145][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.07062982767820358, 'train/grad_norm_pre_clip': 2.334761619567871, 'timing/data': 0.0003983532078564167, 'timing/model': 0.41584052704274654, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.31})
+[2026-07-28 05:05:47,032][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.05010414123535156, 'train/grad_norm_pre_clip': 1.4322214126586914, 'timing/data': 0.0002628457732498646, 'timing/model': 0.4181403541006148, 'learning_rate/qwen_vl_interface': 6.571532953603151e-06, 'learning_rate/action_model': 6.571532953603151e-05, 'epoch': 0.31})
+[2026-07-28 05:05:52,901][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.043494369834661484, 'train/grad_norm_pre_clip': 0.9870210289955139, 'timing/data': 0.0002723950892686844, 'timing/model': 0.4152630972675979, 'learning_rate/qwen_vl_interface': 6.567857295991577e-06, 'learning_rate/action_model': 6.567857295991576e-05, 'epoch': 0.31})
+[2026-07-28 05:05:59,006][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.05912204831838608, 'train/grad_norm_pre_clip': 1.2313493490219116, 'timing/data': 0.00026480481028556824, 'timing/model': 0.6512057879008353, 'learning_rate/qwen_vl_interface': 6.56418078323671e-06, 'learning_rate/action_model': 6.56418078323671e-05, 'epoch': 0.31})
+[2026-07-28 05:06:04,882][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.028483625501394272, 'train/grad_norm_pre_clip': 0.4987749457359314, 'timing/data': 0.0002547409385442734, 'timing/model': 0.4144730609841645, 'learning_rate/qwen_vl_interface': 6.5605034177242e-06, 'learning_rate/action_model': 6.5605034177242e-05, 'epoch': 0.31})
+[2026-07-28 05:06:10,987][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.04750174283981323, 'train/grad_norm_pre_clip': 1.605912208557129, 'timing/data': 0.00026260362938046455, 'timing/model': 0.4141698502935469, 'learning_rate/qwen_vl_interface': 6.556825201840252e-06, 'learning_rate/action_model': 6.556825201840252e-05, 'epoch': 0.31})
+[2026-07-28 05:06:16,858][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.06720270216464996, 'train/grad_norm_pre_clip': 2.0737521648406982, 'timing/data': 0.0004543750546872616, 'timing/model': 0.41612496227025986, 'learning_rate/qwen_vl_interface': 6.553146137971618e-06, 'learning_rate/action_model': 6.553146137971618e-05, 'epoch': 0.31})
+[2026-07-28 05:06:22,831][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.04231501370668411, 'train/grad_norm_pre_clip': 1.0674573183059692, 'timing/data': 0.0002819760702550411, 'timing/model': 0.4163063350133598, 'learning_rate/qwen_vl_interface': 6.549466228505603e-06, 'learning_rate/action_model': 6.549466228505603e-05, 'epoch': 0.31})
+[2026-07-28 05:06:29,228][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.02692096121609211, 'train/grad_norm_pre_clip': 1.8388830423355103, 'timing/data': 0.00026133889332413673, 'timing/model': 0.41339887911453843, 'learning_rate/qwen_vl_interface': 6.545785475830061e-06, 'learning_rate/action_model': 6.54578547583006e-05, 'epoch': 0.31})
+[2026-07-28 05:06:35,517][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.033259160816669464, 'train/grad_norm_pre_clip': 1.01405668258667, 'timing/data': 0.0002676290459930897, 'timing/model': 0.6523866197094321, 'learning_rate/qwen_vl_interface': 6.54210388233339e-06, 'learning_rate/action_model': 6.542103882333391e-05, 'epoch': 0.31})
+[2026-07-28 05:06:41,683][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.041973777115345, 'train/grad_norm_pre_clip': 2.5853488445281982, 'timing/data': 0.0002838759683072567, 'timing/model': 0.416253206320107, 'learning_rate/qwen_vl_interface': 6.538421450404538e-06, 'learning_rate/action_model': 6.538421450404538e-05, 'epoch': 0.31})
+[2026-07-28 05:06:47,802][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.05801645666360855, 'train/grad_norm_pre_clip': 1.262810468673706, 'timing/data': 0.0002521779388189316, 'timing/model': 0.4137384449131787, 'learning_rate/qwen_vl_interface': 6.534738182432999e-06, 'learning_rate/action_model': 6.534738182432999e-05, 'epoch': 0.31})
+[2026-07-28 05:06:53,910][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.0730535089969635, 'train/grad_norm_pre_clip': 1.4691637754440308, 'timing/data': 0.0002742782235145569, 'timing/model': 0.4161731470376253, 'learning_rate/qwen_vl_interface': 6.531054080808798e-06, 'learning_rate/action_model': 6.531054080808798e-05, 'epoch': 0.31})
+[2026-07-28 05:07:00,029][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.04851733148097992, 'train/grad_norm_pre_clip': 1.0509430170059204, 'timing/data': 0.0002490859478712082, 'timing/model': 0.41366257797926664, 'learning_rate/qwen_vl_interface': 6.5273691479225124e-06, 'learning_rate/action_model': 6.527369147922511e-05, 'epoch': 0.31})
+[2026-07-28 05:07:05,935][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.07428760826587677, 'train/grad_norm_pre_clip': 1.6569781303405762, 'timing/data': 0.00026832614094018936, 'timing/model': 0.4155062627978623, 'learning_rate/qwen_vl_interface': 6.523683386165257e-06, 'learning_rate/action_model': 6.523683386165256e-05, 'epoch': 0.31})
+[2026-07-28 05:07:12,080][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.05580694228410721, 'train/grad_norm_pre_clip': 1.8881474733352661, 'timing/data': 0.0002682385966181755, 'timing/model': 0.6523146680556238, 'learning_rate/qwen_vl_interface': 6.519996797928677e-06, 'learning_rate/action_model': 6.519996797928677e-05, 'epoch': 0.31})
+[2026-07-28 05:07:17,977][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.04513360932469368, 'train/grad_norm_pre_clip': 2.08241605758667, 'timing/data': 0.0002421736717224121, 'timing/model': 0.41483876621350646, 'learning_rate/qwen_vl_interface': 6.5163093856049664e-06, 'learning_rate/action_model': 6.516309385604967e-05, 'epoch': 0.32})
+[2026-07-28 05:07:24,147][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.042647525668144226, 'train/grad_norm_pre_clip': 2.239135265350342, 'timing/data': 0.00024310126900672913, 'timing/model': 0.41318343579769135, 'learning_rate/qwen_vl_interface': 6.512621151586842e-06, 'learning_rate/action_model': 6.512621151586842e-05, 'epoch': 0.32})
+[2026-07-28 05:07:30,003][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.072952039539814, 'train/grad_norm_pre_clip': 0.9198880195617676, 'timing/data': 0.00025688810274004936, 'timing/model': 0.41679249377921224, 'learning_rate/qwen_vl_interface': 6.508932098267561e-06, 'learning_rate/action_model': 6.50893209826756e-05, 'epoch': 0.32})
+[2026-07-28 05:07:35,866][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.05490494146943092, 'train/grad_norm_pre_clip': 1.4330960512161255, 'timing/data': 0.0002684001810848713, 'timing/model': 0.4145284681580961, 'learning_rate/qwen_vl_interface': 6.505242228040908e-06, 'learning_rate/action_model': 6.505242228040907e-05, 'epoch': 0.32})
+[2026-07-28 05:07:41,726][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.021606571972370148, 'train/grad_norm_pre_clip': 0.948352038860321, 'timing/data': 0.0002681901678442955, 'timing/model': 0.4152138321660459, 'learning_rate/qwen_vl_interface': 6.501551543301203e-06, 'learning_rate/action_model': 6.501551543301203e-05, 'epoch': 0.32})
+[2026-07-28 05:07:48,067][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.06644529849290848, 'train/grad_norm_pre_clip': 2.5770092010498047, 'timing/data': 0.00023745698854327202, 'timing/model': 0.6465103821828961, 'learning_rate/qwen_vl_interface': 6.497860046443288e-06, 'learning_rate/action_model': 6.497860046443288e-05, 'epoch': 0.32})
+[2026-07-28 05:07:54,003][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.06824877858161926, 'train/grad_norm_pre_clip': 1.7582385540008545, 'timing/data': 0.00025569507852196693, 'timing/model': 0.41458380687981844, 'learning_rate/qwen_vl_interface': 6.494167739862538e-06, 'learning_rate/action_model': 6.494167739862538e-05, 'epoch': 0.32})
+[2026-07-28 05:07:59,867][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.08258682489395142, 'train/grad_norm_pre_clip': 2.5875089168548584, 'timing/data': 0.00027024419978260994, 'timing/model': 0.41610136022791266, 'learning_rate/qwen_vl_interface': 6.490474625954849e-06, 'learning_rate/action_model': 6.490474625954849e-05, 'epoch': 0.32})
+[2026-07-28 05:08:05,764][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.05759749561548233, 'train/grad_norm_pre_clip': 2.405282974243164, 'timing/data': 0.0002807588316500187, 'timing/model': 0.41615427611395717, 'learning_rate/qwen_vl_interface': 6.486780707116642e-06, 'learning_rate/action_model': 6.486780707116642e-05, 'epoch': 0.32})
+[2026-07-28 05:08:11,890][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.03661920130252838, 'train/grad_norm_pre_clip': 1.4333815574645996, 'timing/data': 0.00027212733402848244, 'timing/model': 0.41550065390765667, 'learning_rate/qwen_vl_interface': 6.483085985744864e-06, 'learning_rate/action_model': 6.483085985744863e-05, 'epoch': 0.32})
+[2026-07-28 05:08:17,769][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.02355112135410309, 'train/grad_norm_pre_clip': 1.4471684694290161, 'timing/data': 0.0002634958364069462, 'timing/model': 0.41583056934177876, 'learning_rate/qwen_vl_interface': 6.479390464236974e-06, 'learning_rate/action_model': 6.479390464236974e-05, 'epoch': 0.32})
+[2026-07-28 05:08:23,894][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.08682198822498322, 'train/grad_norm_pre_clip': 1.8854665756225586, 'timing/data': 0.0002695857547223568, 'timing/model': 0.658897927030921, 'learning_rate/qwen_vl_interface': 6.475694144990961e-06, 'learning_rate/action_model': 6.475694144990961e-05, 'epoch': 0.32})
+[2026-07-28 05:08:29,962][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.03370557352900505, 'train/grad_norm_pre_clip': 1.50483238697052, 'timing/data': 0.00024791061878204346, 'timing/model': 0.41411478305235505, 'learning_rate/qwen_vl_interface': 6.471997030405326e-06, 'learning_rate/action_model': 6.471997030405325e-05, 'epoch': 0.32})
+[2026-07-28 05:08:35,604][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.0585080161690712, 'train/grad_norm_pre_clip': 1.384347677230835, 'timing/data': 0.0002790340222418308, 'timing/model': 0.417405242100358, 'learning_rate/qwen_vl_interface': 6.468299122879082e-06, 'learning_rate/action_model': 6.468299122879081e-05, 'epoch': 0.32})
+[2026-07-28 05:08:41,258][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.059760890901088715, 'train/grad_norm_pre_clip': 1.2681357860565186, 'timing/data': 0.0002745669335126877, 'timing/model': 0.4160338509827852, 'learning_rate/qwen_vl_interface': 6.464600424811765e-06, 'learning_rate/action_model': 6.464600424811764e-05, 'epoch': 0.32})
+[2026-07-28 05:08:47,319][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.05360719561576843, 'train/grad_norm_pre_clip': 2.719446897506714, 'timing/data': 0.0002597589045763016, 'timing/model': 0.583522574044764, 'learning_rate/qwen_vl_interface': 6.460900938603418e-06, 'learning_rate/action_model': 6.460900938603418e-05, 'epoch': 0.32})
+[2026-07-28 05:08:53,497][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.04394078999757767, 'train/grad_norm_pre_clip': 1.0150375366210938, 'timing/data': 0.00031570205464959145, 'timing/model': 0.6499159862287343, 'learning_rate/qwen_vl_interface': 6.457200666654601e-06, 'learning_rate/action_model': 6.457200666654601e-05, 'epoch': 0.32})
+[2026-07-28 05:08:59,386][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.047222819179296494, 'train/grad_norm_pre_clip': 1.4559447765350342, 'timing/data': 0.0005273339338600636, 'timing/model': 0.4192008660174906, 'learning_rate/qwen_vl_interface': 6.453499611366376e-06, 'learning_rate/action_model': 6.453499611366376e-05, 'epoch': 0.32})
+[2026-07-28 05:09:05,307][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.06337273120880127, 'train/grad_norm_pre_clip': 2.168607473373413, 'timing/data': 0.00031760474666953087, 'timing/model': 0.4202790013514459, 'learning_rate/qwen_vl_interface': 6.449797775140321e-06, 'learning_rate/action_model': 6.44979777514032e-05, 'epoch': 0.32})
+[2026-07-28 05:09:11,524][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.039811767637729645, 'train/grad_norm_pre_clip': 0.7916927337646484, 'timing/data': 0.0003553759306669235, 'timing/model': 0.4158860887400806, 'learning_rate/qwen_vl_interface': 6.446095160378514e-06, 'learning_rate/action_model': 6.446095160378514e-05, 'epoch': 0.32})
+[2026-07-28 05:09:17,818][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.07251426577568054, 'train/grad_norm_pre_clip': 1.2775686979293823, 'timing/data': 0.00029664114117622375, 'timing/model': 0.41372830560430884, 'learning_rate/qwen_vl_interface': 6.442391769483544e-06, 'learning_rate/action_model': 6.442391769483544e-05, 'epoch': 0.32})
+[2026-07-28 05:09:23,702][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.07137897610664368, 'train/grad_norm_pre_clip': 1.105110764503479, 'timing/data': 0.0003217160701751709, 'timing/model': 0.4144513597711921, 'learning_rate/qwen_vl_interface': 6.4386876048585e-06, 'learning_rate/action_model': 6.4386876048585e-05, 'epoch': 0.32})
+[2026-07-28 05:09:29,601][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.03693254664540291, 'train/grad_norm_pre_clip': 0.6859668493270874, 'timing/data': 0.0003405180759727955, 'timing/model': 0.4156352300196886, 'learning_rate/qwen_vl_interface': 6.434982668906975e-06, 'learning_rate/action_model': 6.434982668906974e-05, 'epoch': 0.32})
+[2026-07-28 05:09:35,934][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.03359207510948181, 'train/grad_norm_pre_clip': 0.7371345162391663, 'timing/data': 0.0002992739900946617, 'timing/model': 0.6513231131248176, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.32})
+[2026-07-28 05:09:41,585][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.015931345522403717, 'train/grad_norm_pre_clip': 0.8909711837768555, 'timing/data': 0.00033876998350024223, 'timing/model': 0.41687544621527195, 'learning_rate/qwen_vl_interface': 6.427570492641351e-06, 'learning_rate/action_model': 6.427570492641351e-05, 'epoch': 0.32})
+[2026-07-28 05:09:47,454][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.11002488434314728, 'train/grad_norm_pre_clip': 3.0400636196136475, 'timing/data': 0.0003297128714621067, 'timing/model': 0.4145101513713598, 'learning_rate/qwen_vl_interface': 6.423863257136931e-06, 'learning_rate/action_model': 6.42386325713693e-05, 'epoch': 0.32})
+[2026-07-28 05:09:53,323][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.04437992721796036, 'train/grad_norm_pre_clip': 1.8950047492980957, 'timing/data': 0.00032834894955158234, 'timing/model': 0.4144045142456889, 'learning_rate/qwen_vl_interface': 6.42015525992539e-06, 'learning_rate/action_model': 6.42015525992539e-05, 'epoch': 0.32})
+[2026-07-28 05:09:59,187][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.05471830815076828, 'train/grad_norm_pre_clip': 1.5549261569976807, 'timing/data': 0.0003246050328016281, 'timing/model': 0.4147180779837072, 'learning_rate/qwen_vl_interface': 6.416446503412803e-06, 'learning_rate/action_model': 6.416446503412803e-05, 'epoch': 0.32})
+[2026-07-28 05:10:05,279][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.04932830482721329, 'train/grad_norm_pre_clip': 1.5717229843139648, 'timing/data': 0.0003135763108730316, 'timing/model': 0.6513146166689694, 'learning_rate/qwen_vl_interface': 6.412736990005746e-06, 'learning_rate/action_model': 6.412736990005746e-05, 'epoch': 0.32})
+[2026-07-28 05:10:10,913][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.018422970548272133, 'train/grad_norm_pre_clip': 0.787499725818634, 'timing/data': 0.0002998160198330879, 'timing/model': 0.41341834515333176, 'learning_rate/qwen_vl_interface': 6.409026722111279e-06, 'learning_rate/action_model': 6.409026722111278e-05, 'epoch': 0.32})
+[2026-07-28 05:10:16,778][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.03872283548116684, 'train/grad_norm_pre_clip': 1.3766934871673584, 'timing/data': 0.0003227037377655506, 'timing/model': 0.4144319505430758, 'learning_rate/qwen_vl_interface': 6.405315702136955e-06, 'learning_rate/action_model': 6.405315702136954e-05, 'epoch': 0.32})
+[2026-07-28 05:10:22,654][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.0359724797308445, 'train/grad_norm_pre_clip': 1.1232914924621582, 'timing/data': 0.0003358391113579273, 'timing/model': 0.41478641983121634, 'learning_rate/qwen_vl_interface': 6.401603932490815e-06, 'learning_rate/action_model': 6.401603932490815e-05, 'epoch': 0.32})
+[2026-07-28 05:10:28,516][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.026376662775874138, 'train/grad_norm_pre_clip': 0.9498509168624878, 'timing/data': 0.00027604494243860245, 'timing/model': 0.41429911786690354, 'learning_rate/qwen_vl_interface': 6.3978914155813865e-06, 'learning_rate/action_model': 6.397891415581386e-05, 'epoch': 0.32})
+[2026-07-28 05:10:34,654][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.06952164322137833, 'train/grad_norm_pre_clip': 1.6367695331573486, 'timing/data': 0.0002642180770635605, 'timing/model': 0.41532120388001204, 'learning_rate/qwen_vl_interface': 6.394178153817682e-06, 'learning_rate/action_model': 6.394178153817682e-05, 'epoch': 0.32})
+[2026-07-28 05:10:40,754][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.03919856995344162, 'train/grad_norm_pre_clip': 1.6243690252304077, 'timing/data': 0.00026387302204966545, 'timing/model': 0.6491362629458308, 'learning_rate/qwen_vl_interface': 6.390464149609193e-06, 'learning_rate/action_model': 6.390464149609193e-05, 'epoch': 0.32})
+[2026-07-28 05:10:46,639][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1750
+[2026-07-28 05:10:51,641][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1750
+[2026-07-28 05:17:27,031][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1750
+[2026-07-28 05:17:27,032][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.07233187556266785, 'train/grad_norm_pre_clip': 2.2522330284118652, 'eval/action_loss/seconds': 4.99908640794456, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.06377995759248734, 'eval/latency_3/loss': 0.06377995759248734, 'eval/flappy/loss': 0.06377995759248734, 'eval/flappy/latency_3/loss': 0.06377995759248734, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 256.76500303186475, 'rl_games_eval/mid_train/mean_length': 2089.45, 'rl_games_eval/mid_train/std_reward': 147.19110632602846, 'rl_games_eval/mid_train/std_length': 1185.6933614978202, 'rl_games_eval/mid_train/macro_mean_reward': 256.76500303186475, 'rl_games_eval/mid_train/macro_mean_length': 2089.45, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 256.76500303186475, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2089.45, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 147.19110632602846, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1185.6933614978202, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 256.76500303186475, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0002589179202914238, 'timing/model': 0.4143509631976485, 'learning_rate/qwen_vl_interface': 6.386749405365901e-06, 'learning_rate/action_model': 6.3867494053659e-05, 'epoch': 0.32})
+[2026-07-28 05:17:33,034][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.16221590340137482, 'train/grad_norm_pre_clip': 3.779989719390869, 'timing/data': 0.00027658184990286827, 'timing/model': 0.4177316720597446, 'learning_rate/qwen_vl_interface': 6.3830339234982596e-06, 'learning_rate/action_model': 6.38303392349826e-05, 'epoch': 0.32})
+[2026-07-28 05:17:39,147][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.10585127770900726, 'train/grad_norm_pre_clip': 3.9185609817504883, 'timing/data': 0.0002649943344295025, 'timing/model': 0.41636916110292077, 'learning_rate/qwen_vl_interface': 6.3793177064172075e-06, 'learning_rate/action_model': 6.379317706417207e-05, 'epoch': 0.32})
+[2026-07-28 05:17:45,024][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.06594755500555038, 'train/grad_norm_pre_clip': 2.0107581615448, 'timing/data': 0.00025006476789712906, 'timing/model': 0.41677093505859375, 'learning_rate/qwen_vl_interface': 6.375600756534157e-06, 'learning_rate/action_model': 6.375600756534157e-05, 'epoch': 0.32})
+[2026-07-28 05:17:51,139][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.04962526261806488, 'train/grad_norm_pre_clip': 1.9109643697738647, 'timing/data': 0.00025399308651685715, 'timing/model': 0.415510434191674, 'learning_rate/qwen_vl_interface': 6.371883076260996e-06, 'learning_rate/action_model': 6.371883076260997e-05, 'epoch': 0.32})
+[2026-07-28 05:17:57,014][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.041634418070316315, 'train/grad_norm_pre_clip': 2.088300943374634, 'timing/data': 0.00028272438794374466, 'timing/model': 0.4159384611994028, 'learning_rate/qwen_vl_interface': 6.368164668010087e-06, 'learning_rate/action_model': 6.368164668010087e-05, 'epoch': 0.32})
+[2026-07-28 05:18:03,117][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.06316481530666351, 'train/grad_norm_pre_clip': 2.9615256786346436, 'timing/data': 0.00024512410163879395, 'timing/model': 0.4140158952213824, 'learning_rate/qwen_vl_interface': 6.364445534194263e-06, 'learning_rate/action_model': 6.364445534194263e-05, 'epoch': 0.32})
+[2026-07-28 05:18:09,260][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.03066830150783062, 'train/grad_norm_pre_clip': 1.495374083518982, 'timing/data': 0.00027523841708898544, 'timing/model': 0.6581884860061109, 'learning_rate/qwen_vl_interface': 6.360725677226834e-06, 'learning_rate/action_model': 6.360725677226834e-05, 'epoch': 0.32})
+[2026-07-28 05:18:15,156][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.04244992136955261, 'train/grad_norm_pre_clip': 1.2142484188079834, 'timing/data': 0.00034034112468361855, 'timing/model': 0.415983596816659, 'learning_rate/qwen_vl_interface': 6.357005099521569e-06, 'learning_rate/action_model': 6.357005099521569e-05, 'epoch': 0.32})
+[2026-07-28 05:18:21,027][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.04230585694313049, 'train/grad_norm_pre_clip': 2.209674596786499, 'timing/data': 0.0004153097979724407, 'timing/model': 0.41591763217002153, 'learning_rate/qwen_vl_interface': 6.353283803492717e-06, 'learning_rate/action_model': 6.353283803492717e-05, 'epoch': 0.32})
+[2026-07-28 05:18:26,900][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.042633190751075745, 'train/grad_norm_pre_clip': 0.8102138638496399, 'timing/data': 0.00028114160522818565, 'timing/model': 0.4165104399435222, 'learning_rate/qwen_vl_interface': 6.349561791554981e-06, 'learning_rate/action_model': 6.349561791554981e-05, 'epoch': 0.32})
+[2026-07-28 05:18:33,004][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.06589129567146301, 'train/grad_norm_pre_clip': 2.8475289344787598, 'timing/data': 0.0003447355702519417, 'timing/model': 0.6484524551779032, 'learning_rate/qwen_vl_interface': 6.345839066123536e-06, 'learning_rate/action_model': 6.345839066123535e-05, 'epoch': 0.32})
+[2026-07-28 05:18:39,135][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.05155087262392044, 'train/grad_norm_pre_clip': 1.010096549987793, 'timing/data': 0.00025478703901171684, 'timing/model': 0.6688024969771504, 'learning_rate/qwen_vl_interface': 6.342115629614019e-06, 'learning_rate/action_model': 6.342115629614019e-05, 'epoch': 0.32})
+[2026-07-28 05:18:45,016][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.02842910960316658, 'train/grad_norm_pre_clip': 0.816267728805542, 'timing/data': 0.00029348116368055344, 'timing/model': 0.41379484301432967, 'learning_rate/qwen_vl_interface': 6.338391484442527e-06, 'learning_rate/action_model': 6.338391484442526e-05, 'epoch': 0.32})
+[2026-07-28 05:18:50,942][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.04204285144805908, 'train/grad_norm_pre_clip': 0.8009113669395447, 'timing/data': 0.00027848081663250923, 'timing/model': 0.4166439878754318, 'learning_rate/qwen_vl_interface': 6.334666633025616e-06, 'learning_rate/action_model': 6.334666633025615e-05, 'epoch': 0.32})
+[2026-07-28 05:18:56,833][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.058425650000572205, 'train/grad_norm_pre_clip': 1.6003963947296143, 'timing/data': 0.000289292074739933, 'timing/model': 0.4164196141064167, 'learning_rate/qwen_vl_interface': 6.3309410777803025e-06, 'learning_rate/action_model': 6.330941077780302e-05, 'epoch': 0.32})
+[2026-07-28 05:19:02,935][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.029524371027946472, 'train/grad_norm_pre_clip': 0.6907339096069336, 'timing/data': 0.00025030504912137985, 'timing/model': 0.4127709558233619, 'learning_rate/qwen_vl_interface': 6.327214821124063e-06, 'learning_rate/action_model': 6.327214821124062e-05, 'epoch': 0.32})
+[2026-07-28 05:19:08,830][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.033159833401441574, 'train/grad_norm_pre_clip': 1.046799898147583, 'timing/data': 0.00028861407190561295, 'timing/model': 0.4178852392360568, 'learning_rate/qwen_vl_interface': 6.323487865474818e-06, 'learning_rate/action_model': 6.323487865474817e-05, 'epoch': 0.32})
+[2026-07-28 05:19:14,940][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.016134537756443024, 'train/grad_norm_pre_clip': 0.7365968227386475, 'timing/data': 0.0002830228768289089, 'timing/model': 0.6538947820663452, 'learning_rate/qwen_vl_interface': 6.319760213250952e-06, 'learning_rate/action_model': 6.319760213250951e-05, 'epoch': 0.32})
+[2026-07-28 05:19:21,060][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.07735781371593475, 'train/grad_norm_pre_clip': 1.616307258605957, 'timing/data': 0.00026726024225354195, 'timing/model': 0.4153796173632145, 'learning_rate/qwen_vl_interface': 6.316031866871297e-06, 'learning_rate/action_model': 6.316031866871297e-05, 'epoch': 0.32})
+[2026-07-28 05:19:26,718][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.04019388556480408, 'train/grad_norm_pre_clip': 1.0563430786132812, 'timing/data': 0.0002770940773189068, 'timing/model': 0.41498310305178165, 'learning_rate/qwen_vl_interface': 6.31230282875514e-06, 'learning_rate/action_model': 6.31230282875514e-05, 'epoch': 0.33})
+[2026-07-28 05:19:32,375][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.04779065400362015, 'train/grad_norm_pre_clip': 1.4601454734802246, 'timing/data': 0.00027149496600031853, 'timing/model': 0.41568972123786807, 'learning_rate/qwen_vl_interface': 6.308573101322209e-06, 'learning_rate/action_model': 6.308573101322209e-05, 'epoch': 0.33})
+[2026-07-28 05:19:38,035][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.06708677113056183, 'train/grad_norm_pre_clip': 1.3048226833343506, 'timing/data': 0.00027785589918494225, 'timing/model': 0.41621779184788465, 'learning_rate/qwen_vl_interface': 6.304842686992685e-06, 'learning_rate/action_model': 6.304842686992684e-05, 'epoch': 0.33})
+[2026-07-28 05:19:44,154][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.07289968430995941, 'train/grad_norm_pre_clip': 1.4821808338165283, 'timing/data': 0.00024589477106928825, 'timing/model': 0.6434819023124874, 'learning_rate/qwen_vl_interface': 6.301111588187194e-06, 'learning_rate/action_model': 6.301111588187194e-05, 'epoch': 0.33})
+[2026-07-28 05:19:50,075][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.039351340383291245, 'train/grad_norm_pre_clip': 1.0005306005477905, 'timing/data': 0.0002565290778875351, 'timing/model': 0.4148705736733973, 'learning_rate/qwen_vl_interface': 6.297379807326806e-06, 'learning_rate/action_model': 6.297379807326806e-05, 'epoch': 0.33})
+[2026-07-28 05:19:55,961][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.02843157947063446, 'train/grad_norm_pre_clip': 0.9710574150085449, 'timing/data': 0.00027819816023111343, 'timing/model': 0.4173619309440255, 'learning_rate/qwen_vl_interface': 6.293647346833033e-06, 'learning_rate/action_model': 6.293647346833032e-05, 'epoch': 0.33})
+[2026-07-28 05:20:01,855][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.06882929801940918, 'train/grad_norm_pre_clip': 1.5405081510543823, 'timing/data': 0.000279115978628397, 'timing/model': 0.4171478748321533, 'learning_rate/qwen_vl_interface': 6.289914209127826e-06, 'learning_rate/action_model': 6.289914209127825e-05, 'epoch': 0.33})
+[2026-07-28 05:20:08,065][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.057584285736083984, 'train/grad_norm_pre_clip': 1.8917006254196167, 'timing/data': 0.0002513732761144638, 'timing/model': 0.41538691287860274, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.33})
+[2026-07-28 05:20:14,003][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.02909621223807335, 'train/grad_norm_pre_clip': 1.1075011491775513, 'timing/data': 0.00027952296659350395, 'timing/model': 0.41697001177817583, 'learning_rate/qwen_vl_interface': 6.282445911773122e-06, 'learning_rate/action_model': 6.282445911773121e-05, 'epoch': 0.33})
+[2026-07-28 05:20:20,097][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.031038764864206314, 'train/grad_norm_pre_clip': 1.0072146654129028, 'timing/data': 0.00024842796847224236, 'timing/model': 0.4151854286901653, 'learning_rate/qwen_vl_interface': 6.27871075696972e-06, 'learning_rate/action_model': 6.27871075696972e-05, 'epoch': 0.33})
+[2026-07-28 05:20:25,977][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.05527012422680855, 'train/grad_norm_pre_clip': 1.6381052732467651, 'timing/data': 0.00027585122734308243, 'timing/model': 0.4165670401416719, 'learning_rate/qwen_vl_interface': 6.27497493464708e-06, 'learning_rate/action_model': 6.27497493464708e-05, 'epoch': 0.33})
+[2026-07-28 05:20:31,892][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.04121387004852295, 'train/grad_norm_pre_clip': 1.1025692224502563, 'timing/data': 0.0002783173695206642, 'timing/model': 0.4184370073489845, 'learning_rate/qwen_vl_interface': 6.271238447229332e-06, 'learning_rate/action_model': 6.271238447229332e-05, 'epoch': 0.33})
+[2026-07-28 05:20:37,998][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.06033347547054291, 'train/grad_norm_pre_clip': 1.1036279201507568, 'timing/data': 0.000279436819255352, 'timing/model': 0.653379085008055, 'learning_rate/qwen_vl_interface': 6.267501297141042e-06, 'learning_rate/action_model': 6.267501297141042e-05, 'epoch': 0.33})
+[2026-07-28 05:20:44,079][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.08789650350809097, 'train/grad_norm_pre_clip': 1.4973413944244385, 'timing/data': 0.00026685697957873344, 'timing/model': 0.4141849032603204, 'learning_rate/qwen_vl_interface': 6.263763486807208e-06, 'learning_rate/action_model': 6.263763486807208e-05, 'epoch': 0.33})
+[2026-07-28 05:20:49,965][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.05060272663831711, 'train/grad_norm_pre_clip': 1.7112897634506226, 'timing/data': 0.0002896399237215519, 'timing/model': 0.4163405359722674, 'learning_rate/qwen_vl_interface': 6.260025018653253e-06, 'learning_rate/action_model': 6.260025018653253e-05, 'epoch': 0.33})
+[2026-07-28 05:20:55,862][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.07663674652576447, 'train/grad_norm_pre_clip': 1.994891881942749, 'timing/data': 0.0002891821786761284, 'timing/model': 0.4167030989192426, 'learning_rate/qwen_vl_interface': 6.2562858951050294e-06, 'learning_rate/action_model': 6.256285895105029e-05, 'epoch': 0.33})
+[2026-07-28 05:21:02,015][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.04242970049381256, 'train/grad_norm_pre_clip': 0.8559948205947876, 'timing/data': 0.0002666981890797615, 'timing/model': 0.4153063599951565, 'learning_rate/qwen_vl_interface': 6.252546118588814e-06, 'learning_rate/action_model': 6.252546118588813e-05, 'epoch': 0.33})
+[2026-07-28 05:21:07,894][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.056578539311885834, 'train/grad_norm_pre_clip': 1.091997742652893, 'timing/data': 0.00026975106447935104, 'timing/model': 0.41625756211578846, 'learning_rate/qwen_vl_interface': 6.248805691531306e-06, 'learning_rate/action_model': 6.248805691531307e-05, 'epoch': 0.33})
+[2026-07-28 05:21:13,781][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.08324332535266876, 'train/grad_norm_pre_clip': 1.1747251749038696, 'timing/data': 0.00027984101325273514, 'timing/model': 0.4165469091385603, 'learning_rate/qwen_vl_interface': 6.245064616359628e-06, 'learning_rate/action_model': 6.245064616359628e-05, 'epoch': 0.33})
+[2026-07-28 05:21:20,122][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.05286310985684395, 'train/grad_norm_pre_clip': 2.0083279609680176, 'timing/data': 0.0002514519728720188, 'timing/model': 0.6044901879504323, 'learning_rate/qwen_vl_interface': 6.241322895501325e-06, 'learning_rate/action_model': 6.241322895501325e-05, 'epoch': 0.33})
+[2026-07-28 05:21:25,992][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.05446229875087738, 'train/grad_norm_pre_clip': 1.917012333869934, 'timing/data': 0.00024293502792716026, 'timing/model': 0.4230188960209489, 'learning_rate/qwen_vl_interface': 6.2375805313843564e-06, 'learning_rate/action_model': 6.237580531384356e-05, 'epoch': 0.33})
+[2026-07-28 05:21:31,674][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.03681080788373947, 'train/grad_norm_pre_clip': 0.7899231910705566, 'timing/data': 0.00030235666781663895, 'timing/model': 0.41674593510106206, 'learning_rate/qwen_vl_interface': 6.233837526437106e-06, 'learning_rate/action_model': 6.233837526437105e-05, 'epoch': 0.33})
+[2026-07-28 05:21:37,361][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.06537508964538574, 'train/grad_norm_pre_clip': 2.4156246185302734, 'timing/data': 0.00028010783717036247, 'timing/model': 0.41704654507339, 'learning_rate/qwen_vl_interface': 6.230093883088364e-06, 'learning_rate/action_model': 6.230093883088364e-05, 'epoch': 0.33})
+[2026-07-28 05:21:43,500][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.07087553292512894, 'train/grad_norm_pre_clip': 1.6313213109970093, 'timing/data': 0.00028353556990623474, 'timing/model': 0.6579883708618581, 'learning_rate/qwen_vl_interface': 6.226349603767341e-06, 'learning_rate/action_model': 6.22634960376734e-05, 'epoch': 0.33})
+[2026-07-28 05:21:49,416][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.04694437235593796, 'train/grad_norm_pre_clip': 1.1212360858917236, 'timing/data': 0.0002961689606308937, 'timing/model': 0.4168715919367969, 'learning_rate/qwen_vl_interface': 6.2226046909036616e-06, 'learning_rate/action_model': 6.222604690903662e-05, 'epoch': 0.33})
+[2026-07-28 05:21:55,307][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.04235482215881348, 'train/grad_norm_pre_clip': 1.2719627618789673, 'timing/data': 0.000277198851108551, 'timing/model': 0.41609305795282125, 'learning_rate/qwen_vl_interface': 6.218859146927358e-06, 'learning_rate/action_model': 6.218859146927358e-05, 'epoch': 0.33})
+[2026-07-28 05:22:00,999][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.04363138973712921, 'train/grad_norm_pre_clip': 1.7250381708145142, 'timing/data': 0.0002820719964802265, 'timing/model': 0.41749721486121416, 'learning_rate/qwen_vl_interface': 6.215112974268871e-06, 'learning_rate/action_model': 6.21511297426887e-05, 'epoch': 0.33})
+[2026-07-28 05:22:07,393][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.04518207162618637, 'train/grad_norm_pre_clip': 0.922322690486908, 'timing/data': 0.00032168813049793243, 'timing/model': 0.41467891866341233, 'learning_rate/qwen_vl_interface': 6.211366175359055e-06, 'learning_rate/action_model': 6.211366175359054e-05, 'epoch': 0.33})
+[2026-07-28 05:22:13,289][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.044666729867458344, 'train/grad_norm_pre_clip': 1.15691077709198, 'timing/data': 0.00029650330543518066, 'timing/model': 0.41595121705904603, 'learning_rate/qwen_vl_interface': 6.207618752629164e-06, 'learning_rate/action_model': 6.207618752629163e-05, 'epoch': 0.33})
+[2026-07-28 05:22:19,182][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.0451972596347332, 'train/grad_norm_pre_clip': 1.961356520652771, 'timing/data': 0.00026825396344065666, 'timing/model': 0.4153435309417546, 'learning_rate/qwen_vl_interface': 6.203870708510862e-06, 'learning_rate/action_model': 6.203870708510862e-05, 'epoch': 0.33})
+[2026-07-28 05:22:25,336][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.04346664249897003, 'train/grad_norm_pre_clip': 1.9547545909881592, 'timing/data': 0.0002538268454372883, 'timing/model': 0.415213939268142, 'learning_rate/qwen_vl_interface': 6.200122045436212e-06, 'learning_rate/action_model': 6.200122045436211e-05, 'epoch': 0.33})
+[2026-07-28 05:22:31,458][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.03972788527607918, 'train/grad_norm_pre_clip': 1.3655507564544678, 'timing/data': 0.0002814093604683876, 'timing/model': 0.6527824038639665, 'learning_rate/qwen_vl_interface': 6.196372765837682e-06, 'learning_rate/action_model': 6.196372765837682e-05, 'epoch': 0.33})
+[2026-07-28 05:22:37,376][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.04216909781098366, 'train/grad_norm_pre_clip': 2.6722419261932373, 'timing/data': 0.00026360293850302696, 'timing/model': 0.4682143861427903, 'learning_rate/qwen_vl_interface': 6.19262287214814e-06, 'learning_rate/action_model': 6.19262287214814e-05, 'epoch': 0.33})
+[2026-07-28 05:22:43,507][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.03154360502958298, 'train/grad_norm_pre_clip': 1.5314841270446777, 'timing/data': 0.0002785143442451954, 'timing/model': 0.6450174520723522, 'learning_rate/qwen_vl_interface': 6.18887236680085e-06, 'learning_rate/action_model': 6.18887236680085e-05, 'epoch': 0.33})
+[2026-07-28 05:22:49,441][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.043645065277814865, 'train/grad_norm_pre_clip': 1.9854665994644165, 'timing/data': 0.0002824380062520504, 'timing/model': 0.42508536111563444, 'learning_rate/qwen_vl_interface': 6.185121252229475e-06, 'learning_rate/action_model': 6.185121252229474e-05, 'epoch': 0.33})
+[2026-07-28 05:22:55,607][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.041719306260347366, 'train/grad_norm_pre_clip': 1.8687565326690674, 'timing/data': 0.0002596466802060604, 'timing/model': 0.4137479714117944, 'learning_rate/qwen_vl_interface': 6.181369530868071e-06, 'learning_rate/action_model': 6.181369530868071e-05, 'epoch': 0.33})
+[2026-07-28 05:23:01,500][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.08735385537147522, 'train/grad_norm_pre_clip': 2.6908483505249023, 'timing/data': 0.0002852790057659149, 'timing/model': 0.41542751202359796, 'learning_rate/qwen_vl_interface': 6.17761720515109e-06, 'learning_rate/action_model': 6.17761720515109e-05, 'epoch': 0.33})
+[2026-07-28 05:23:07,647][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.11770493537187576, 'train/grad_norm_pre_clip': 4.851280689239502, 'timing/data': 0.00024717627093195915, 'timing/model': 0.41263154661282897, 'learning_rate/qwen_vl_interface': 6.173864277513376e-06, 'learning_rate/action_model': 6.173864277513376e-05, 'epoch': 0.33})
+[2026-07-28 05:23:13,559][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.017793983221054077, 'train/grad_norm_pre_clip': 0.5694218277931213, 'timing/data': 0.00029222527518868446, 'timing/model': 0.4178988509811461, 'learning_rate/qwen_vl_interface': 6.170110750390163e-06, 'learning_rate/action_model': 6.170110750390163e-05, 'epoch': 0.33})
+[2026-07-28 05:23:19,676][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.034658774733543396, 'train/grad_norm_pre_clip': 1.3110681772232056, 'timing/data': 0.0003286912105977535, 'timing/model': 0.423790096770972, 'learning_rate/qwen_vl_interface': 6.1663566262170725e-06, 'learning_rate/action_model': 6.166356626217072e-05, 'epoch': 0.33})
+[2026-07-28 05:23:25,777][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.056921035051345825, 'train/grad_norm_pre_clip': 1.8856638669967651, 'timing/data': 0.0002561379224061966, 'timing/model': 0.4158172910101712, 'learning_rate/qwen_vl_interface': 6.162601907430116e-06, 'learning_rate/action_model': 6.162601907430116e-05, 'epoch': 0.33})
+[2026-07-28 05:23:31,700][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.03868017718195915, 'train/grad_norm_pre_clip': 0.9882553815841675, 'timing/data': 0.0003731600008904934, 'timing/model': 0.4173442479223013, 'learning_rate/qwen_vl_interface': 6.15884659646569e-06, 'learning_rate/action_model': 6.15884659646569e-05, 'epoch': 0.33})
+[2026-07-28 05:23:37,834][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.03564930334687233, 'train/grad_norm_pre_clip': 1.13981294631958, 'timing/data': 0.00025425106287002563, 'timing/model': 0.4137247707694769, 'learning_rate/qwen_vl_interface': 6.155090695760574e-06, 'learning_rate/action_model': 6.155090695760573e-05, 'epoch': 0.33})
+[2026-07-28 05:23:43,881][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.03783043101429939, 'train/grad_norm_pre_clip': 1.0627565383911133, 'timing/data': 0.00042983563616871834, 'timing/model': 0.584761243313551, 'learning_rate/qwen_vl_interface': 6.151334207751933e-06, 'learning_rate/action_model': 6.151334207751932e-05, 'epoch': 0.33})
+[2026-07-28 05:23:49,753][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.032845981419086456, 'train/grad_norm_pre_clip': 0.9746567606925964, 'timing/data': 0.0003422647714614868, 'timing/model': 0.41640376299619675, 'learning_rate/qwen_vl_interface': 6.1475771348773115e-06, 'learning_rate/action_model': 6.147577134877312e-05, 'epoch': 0.33})
+[2026-07-28 05:23:55,618][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.033706121146678925, 'train/grad_norm_pre_clip': 1.4940160512924194, 'timing/data': 0.0003633219748735428, 'timing/model': 0.4148172033019364, 'learning_rate/qwen_vl_interface': 6.14381947957463e-06, 'learning_rate/action_model': 6.14381947957463e-05, 'epoch': 0.33})
+[2026-07-28 05:24:01,731][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.06412675231695175, 'train/grad_norm_pre_clip': 2.4557132720947266, 'timing/data': 0.00028392206877470016, 'timing/model': 0.6530872890725732, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.33})
+[2026-07-28 05:24:07,824][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.029362667351961136, 'train/grad_norm_pre_clip': 0.9983536601066589, 'timing/data': 0.00029148906469345093, 'timing/model': 0.4166345470584929, 'learning_rate/qwen_vl_interface': 6.136302431438677e-06, 'learning_rate/action_model': 6.136302431438677e-05, 'epoch': 0.33})
+[2026-07-28 05:24:13,761][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.0548999086022377, 'train/grad_norm_pre_clip': 1.8267982006072998, 'timing/data': 0.0003717211075127125, 'timing/model': 0.41713567078113556, 'learning_rate/qwen_vl_interface': 6.132543043483137e-06, 'learning_rate/action_model': 6.132543043483136e-05, 'epoch': 0.33})
+[2026-07-28 05:24:19,672][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.05182075873017311, 'train/grad_norm_pre_clip': 1.3181264400482178, 'timing/data': 0.0002908390015363693, 'timing/model': 0.41488363593816757, 'learning_rate/qwen_vl_interface': 6.1287830828549975e-06, 'learning_rate/action_model': 6.128783082854997e-05, 'epoch': 0.33})
+[2026-07-28 05:24:25,784][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.0446392297744751, 'train/grad_norm_pre_clip': 1.1664365530014038, 'timing/data': 0.0002676672302186489, 'timing/model': 0.41508144699037075, 'learning_rate/qwen_vl_interface': 6.125022551994055e-06, 'learning_rate/action_model': 6.125022551994056e-05, 'epoch': 0.33})
+[2026-07-28 05:24:31,885][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.06174783036112785, 'train/grad_norm_pre_clip': 2.550452470779419, 'timing/data': 0.0003557349555194378, 'timing/model': 0.6444703196175396, 'learning_rate/qwen_vl_interface': 6.121261453340481e-06, 'learning_rate/action_model': 6.12126145334048e-05, 'epoch': 0.33})
+[2026-07-28 05:24:37,869][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.05076908692717552, 'train/grad_norm_pre_clip': 2.7336297035217285, 'timing/data': 0.0002604993060231209, 'timing/model': 0.41360717080533504, 'learning_rate/qwen_vl_interface': 6.117499789334806e-06, 'learning_rate/action_model': 6.117499789334807e-05, 'epoch': 0.33})
+[2026-07-28 05:24:44,003][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.07069669663906097, 'train/grad_norm_pre_clip': 1.9453415870666504, 'timing/data': 0.0002688998356461525, 'timing/model': 0.6564430221915245, 'learning_rate/qwen_vl_interface': 6.113737562417939e-06, 'learning_rate/action_model': 6.113737562417938e-05, 'epoch': 0.33})
+[2026-07-28 05:24:49,897][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.050710346549749374, 'train/grad_norm_pre_clip': 1.5778735876083374, 'timing/data': 0.00028582802042365074, 'timing/model': 0.4169647670350969, 'learning_rate/qwen_vl_interface': 6.109974775031144e-06, 'learning_rate/action_model': 6.109974775031144e-05, 'epoch': 0.34})
+[2026-07-28 05:24:55,814][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.051637496799230576, 'train/grad_norm_pre_clip': 1.946366548538208, 'timing/data': 0.0003560190089046955, 'timing/model': 0.4159438000060618, 'learning_rate/qwen_vl_interface': 6.106211429616054e-06, 'learning_rate/action_model': 6.106211429616053e-05, 'epoch': 0.34})
+[2026-07-28 05:25:01,709][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.02926282212138176, 'train/grad_norm_pre_clip': 0.5917530059814453, 'timing/data': 0.0002857716754078865, 'timing/model': 0.4162759860046208, 'learning_rate/qwen_vl_interface': 6.102447528614662e-06, 'learning_rate/action_model': 6.102447528614662e-05, 'epoch': 0.34})
+[2026-07-28 05:25:07,953][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.03946446627378464, 'train/grad_norm_pre_clip': 1.5694540739059448, 'timing/data': 0.0002558119595050812, 'timing/model': 0.4125841250643134, 'learning_rate/qwen_vl_interface': 6.098683074469325e-06, 'learning_rate/action_model': 6.0986830744693246e-05, 'epoch': 0.34})
+[2026-07-28 05:25:13,870][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.09014861285686493, 'train/grad_norm_pre_clip': 2.0124351978302, 'timing/data': 0.0002794191241264343, 'timing/model': 0.41439936868846416, 'learning_rate/qwen_vl_interface': 6.094918069622754e-06, 'learning_rate/action_model': 6.0949180696227534e-05, 'epoch': 0.34})
+[2026-07-28 05:25:19,559][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.04769125208258629, 'train/grad_norm_pre_clip': 0.8237180113792419, 'timing/data': 0.0002803611569106579, 'timing/model': 0.4156579151749611, 'learning_rate/qwen_vl_interface': 6.091152516518022e-06, 'learning_rate/action_model': 6.0911525165180216e-05, 'epoch': 0.34})
+[2026-07-28 05:25:25,221][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.034342311322689056, 'train/grad_norm_pre_clip': 1.2693973779678345, 'timing/data': 0.0002898736856877804, 'timing/model': 0.4164037946611643, 'learning_rate/qwen_vl_interface': 6.087386417598554e-06, 'learning_rate/action_model': 6.0873864175985537e-05, 'epoch': 0.34})
+[2026-07-28 05:25:31,369][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.03340204432606697, 'train/grad_norm_pre_clip': 1.0231916904449463, 'timing/data': 0.000256264116615057, 'timing/model': 0.4128913707099855, 'learning_rate/qwen_vl_interface': 6.083619775308129e-06, 'learning_rate/action_model': 6.083619775308129e-05, 'epoch': 0.34})
+[2026-07-28 05:25:37,268][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.06457001715898514, 'train/grad_norm_pre_clip': 1.6179487705230713, 'timing/data': 0.0003567398525774479, 'timing/model': 0.41490444308146834, 'learning_rate/qwen_vl_interface': 6.079852592090885e-06, 'learning_rate/action_model': 6.079852592090884e-05, 'epoch': 0.34})
+[2026-07-28 05:25:43,217][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.04538683220744133, 'train/grad_norm_pre_clip': 1.1394386291503906, 'timing/data': 0.00028795376420021057, 'timing/model': 0.4154119580052793, 'learning_rate/qwen_vl_interface': 6.076084870391303e-06, 'learning_rate/action_model': 6.076084870391302e-05, 'epoch': 0.34})
+[2026-07-28 05:25:49,153][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.07732074707746506, 'train/grad_norm_pre_clip': 2.540817975997925, 'timing/data': 0.0002745618112385273, 'timing/model': 0.662849276792258, 'learning_rate/qwen_vl_interface': 6.072316612654215e-06, 'learning_rate/action_model': 6.072316612654215e-05, 'epoch': 0.34})
+[2026-07-28 05:25:55,063][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.04147162288427353, 'train/grad_norm_pre_clip': 1.4246695041656494, 'timing/data': 0.00029096100479364395, 'timing/model': 0.4178553931415081, 'learning_rate/qwen_vl_interface': 6.068547821324808e-06, 'learning_rate/action_model': 6.068547821324807e-05, 'epoch': 0.34})
+[2026-07-28 05:26:00,762][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.02222958765923977, 'train/grad_norm_pre_clip': 1.1130962371826172, 'timing/data': 0.0002864869311451912, 'timing/model': 0.41708604292944074, 'learning_rate/qwen_vl_interface': 6.064778498848605e-06, 'learning_rate/action_model': 6.064778498848604e-05, 'epoch': 0.34})
+[2026-07-28 05:26:06,749][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.03296051546931267, 'train/grad_norm_pre_clip': 0.8765733242034912, 'timing/data': 0.00028043100610375404, 'timing/model': 0.4169740919023752, 'learning_rate/qwen_vl_interface': 6.061008647671479e-06, 'learning_rate/action_model': 6.061008647671479e-05, 'epoch': 0.34})
+[2026-07-28 05:26:12,943][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.04331820830702782, 'train/grad_norm_pre_clip': 1.049068570137024, 'timing/data': 0.0002584969624876976, 'timing/model': 0.4149372219108045, 'learning_rate/qwen_vl_interface': 6.05723827023965e-06, 'learning_rate/action_model': 6.057238270239649e-05, 'epoch': 0.34})
+[2026-07-28 05:26:19,072][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.0635499656200409, 'train/grad_norm_pre_clip': 2.012460708618164, 'timing/data': 0.00030247680842876434, 'timing/model': 0.6578285410068929, 'learning_rate/qwen_vl_interface': 6.053467368999667e-06, 'learning_rate/action_model': 6.0534673689996665e-05, 'epoch': 0.34})
+[2026-07-28 05:26:24,973][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.04362957179546356, 'train/grad_norm_pre_clip': 1.4311237335205078, 'timing/data': 0.0006011752411723137, 'timing/model': 0.41483811708167195, 'learning_rate/qwen_vl_interface': 6.049695946398433e-06, 'learning_rate/action_model': 6.049695946398433e-05, 'epoch': 0.34})
+[2026-07-28 05:26:31,133][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.05955793708562851, 'train/grad_norm_pre_clip': 1.7859691381454468, 'timing/data': 0.00026880204677581787, 'timing/model': 0.4183500148355961, 'learning_rate/qwen_vl_interface': 6.045924004883184e-06, 'learning_rate/action_model': 6.045924004883183e-05, 'epoch': 0.34})
+[2026-07-28 05:26:37,069][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.029709670692682266, 'train/grad_norm_pre_clip': 1.0990506410598755, 'timing/data': 0.00027748895809054375, 'timing/model': 0.4181379619985819, 'learning_rate/qwen_vl_interface': 6.042151546901487e-06, 'learning_rate/action_model': 6.042151546901487e-05, 'epoch': 0.34})
+[2026-07-28 05:26:43,019][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.038810767233371735, 'train/grad_norm_pre_clip': 1.0261116027832031, 'timing/data': 0.00027775485068559647, 'timing/model': 0.4178957138210535, 'learning_rate/qwen_vl_interface': 6.038378574901252e-06, 'learning_rate/action_model': 6.0383785749012514e-05, 'epoch': 0.34})
+[2026-07-28 05:26:48,914][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.07220356166362762, 'train/grad_norm_pre_clip': 1.7609889507293701, 'timing/data': 0.00028529996052384377, 'timing/model': 0.41665015975013375, 'learning_rate/qwen_vl_interface': 6.0346050913307195e-06, 'learning_rate/action_model': 6.0346050913307195e-05, 'epoch': 0.34})
+[2026-07-28 05:26:54,827][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.03689955919981003, 'train/grad_norm_pre_clip': 1.636317253112793, 'timing/data': 0.0002816971391439438, 'timing/model': 0.6610602280125022, 'learning_rate/qwen_vl_interface': 6.03083109863846e-06, 'learning_rate/action_model': 6.03083109863846e-05, 'epoch': 0.34})
+[2026-07-28 05:27:00,516][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.050685107707977295, 'train/grad_norm_pre_clip': 2.027817726135254, 'timing/data': 0.00032568490132689476, 'timing/model': 0.414339370559901, 'learning_rate/qwen_vl_interface': 6.02705659927338e-06, 'learning_rate/action_model': 6.02705659927338e-05, 'epoch': 0.34})
+[2026-07-28 05:27:06,412][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.023548398166894913, 'train/grad_norm_pre_clip': 1.1814316511154175, 'timing/data': 0.00039148423820734024, 'timing/model': 0.4162844279780984, 'learning_rate/qwen_vl_interface': 6.023281595684709e-06, 'learning_rate/action_model': 6.023281595684709e-05, 'epoch': 0.34})
+[2026-07-28 05:27:12,372][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.08880158513784409, 'train/grad_norm_pre_clip': 2.9705777168273926, 'timing/data': 0.0003401101566851139, 'timing/model': 0.4818582907319069, 'learning_rate/qwen_vl_interface': 6.019506090322003e-06, 'learning_rate/action_model': 6.019506090322004e-05, 'epoch': 0.34})
+[2026-07-28 05:27:18,262][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.013971572741866112, 'train/grad_norm_pre_clip': 1.7013274431228638, 'timing/data': 0.00047241803258657455, 'timing/model': 0.4186001028865576, 'learning_rate/qwen_vl_interface': 6.015730085635153e-06, 'learning_rate/action_model': 6.015730085635153e-05, 'epoch': 0.34})
+[2026-07-28 05:27:24,367][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.05042124167084694, 'train/grad_norm_pre_clip': 1.8534715175628662, 'timing/data': 0.00025004008784890175, 'timing/model': 0.4144035088829696, 'learning_rate/qwen_vl_interface': 6.011953584074361e-06, 'learning_rate/action_model': 6.0119535840743614e-05, 'epoch': 0.34})
+[2026-07-28 05:27:30,479][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.01311077456921339, 'train/grad_norm_pre_clip': 1.2751134634017944, 'timing/data': 0.0002538459375500679, 'timing/model': 0.4153451998718083, 'learning_rate/qwen_vl_interface': 6.008176588090162e-06, 'learning_rate/action_model': 6.0081765880901616e-05, 'epoch': 0.34})
+[2026-07-28 05:27:36,379][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.05655650421977043, 'train/grad_norm_pre_clip': 1.900075912475586, 'timing/data': 0.00027544423937797546, 'timing/model': 0.41638250509276986, 'learning_rate/qwen_vl_interface': 6.004399100133409e-06, 'learning_rate/action_model': 6.004399100133409e-05, 'epoch': 0.34})
+[2026-07-28 05:27:42,248][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.054301679134368896, 'train/grad_norm_pre_clip': 1.9372353553771973, 'timing/data': 0.00026563601568341255, 'timing/model': 0.4162979982793331, 'learning_rate/qwen_vl_interface': 6.0006211226552655e-06, 'learning_rate/action_model': 6.0006211226552655e-05, 'epoch': 0.34})
+[2026-07-28 05:27:48,383][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.04511227458715439, 'train/grad_norm_pre_clip': 1.7875899076461792, 'timing/data': 0.00025464314967393875, 'timing/model': 0.41575344838202, 'learning_rate/qwen_vl_interface': 5.996842658107228e-06, 'learning_rate/action_model': 5.996842658107228e-05, 'epoch': 0.34})
+[2026-07-28 05:27:54,510][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.04017435759305954, 'train/grad_norm_pre_clip': 2.1483683586120605, 'timing/data': 0.0002546580508351326, 'timing/model': 0.41562777012586594, 'learning_rate/qwen_vl_interface': 5.993063708941098e-06, 'learning_rate/action_model': 5.993063708941098e-05, 'epoch': 0.34})
+[2026-07-28 05:28:00,427][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.051423102617263794, 'train/grad_norm_pre_clip': 1.47663152217865, 'timing/data': 0.0003392440266907215, 'timing/model': 0.42392264772206545, 'learning_rate/qwen_vl_interface': 5.989284277608993e-06, 'learning_rate/action_model': 5.989284277608993e-05, 'epoch': 0.34})
+[2026-07-28 05:28:06,329][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.061663784086704254, 'train/grad_norm_pre_clip': 2.340501546859741, 'timing/data': 0.0003539728932082653, 'timing/model': 0.4166868692263961, 'learning_rate/qwen_vl_interface': 5.985504366563347e-06, 'learning_rate/action_model': 5.985504366563347e-05, 'epoch': 0.34})
+[2026-07-28 05:28:12,465][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.031190218403935432, 'train/grad_norm_pre_clip': 1.708270788192749, 'timing/data': 0.0002784659154713154, 'timing/model': 0.6457625520415604, 'learning_rate/qwen_vl_interface': 5.9817239782569045e-06, 'learning_rate/action_model': 5.9817239782569045e-05, 'epoch': 0.34})
+[2026-07-28 05:28:18,637][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.04114905744791031, 'train/grad_norm_pre_clip': 2.0937018394470215, 'timing/data': 0.00029274309054017067, 'timing/model': 0.6563191991299391, 'learning_rate/qwen_vl_interface': 5.977943115142715e-06, 'learning_rate/action_model': 5.977943115142715e-05, 'epoch': 0.34})
+[2026-07-28 05:28:24,641][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.07515639066696167, 'train/grad_norm_pre_clip': 3.2825310230255127, 'timing/data': 0.0002551330253481865, 'timing/model': 0.4396857828833163, 'learning_rate/qwen_vl_interface': 5.9741617796741415e-06, 'learning_rate/action_model': 5.9741617796741414e-05, 'epoch': 0.34})
+[2026-07-28 05:28:30,566][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.08061820268630981, 'train/grad_norm_pre_clip': 2.9034652709960938, 'timing/data': 0.0002858126536011696, 'timing/model': 0.41501789214089513, 'learning_rate/qwen_vl_interface': 5.970379974304851e-06, 'learning_rate/action_model': 5.9703799743048506e-05, 'epoch': 0.34})
+[2026-07-28 05:28:36,482][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.056685928255319595, 'train/grad_norm_pre_clip': 2.283315896987915, 'timing/data': 0.0002760309725999832, 'timing/model': 0.4159434079192579, 'learning_rate/qwen_vl_interface': 5.966597701488816e-06, 'learning_rate/action_model': 5.9665977014888154e-05, 'epoch': 0.34})
+[2026-07-28 05:28:42,646][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.06859256327152252, 'train/grad_norm_pre_clip': 2.3515138626098633, 'timing/data': 0.00033057108521461487, 'timing/model': 0.41698527801781893, 'learning_rate/qwen_vl_interface': 5.962814963680315e-06, 'learning_rate/action_model': 5.9628149636803144e-05, 'epoch': 0.34})
+[2026-07-28 05:28:48,523][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.0605614110827446, 'train/grad_norm_pre_clip': 2.2524068355560303, 'timing/data': 0.00027963100001215935, 'timing/model': 0.4156188350170851, 'learning_rate/qwen_vl_interface': 5.95903176333392e-06, 'learning_rate/action_model': 5.95903176333392e-05, 'epoch': 0.34})
+[2026-07-28 05:28:54,408][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.017092520371079445, 'train/grad_norm_pre_clip': 0.8782986402511597, 'timing/data': 0.0002695871517062187, 'timing/model': 0.41503722919151187, 'learning_rate/qwen_vl_interface': 5.955248102904513e-06, 'learning_rate/action_model': 5.955248102904512e-05, 'epoch': 0.34})
+[2026-07-28 05:29:00,536][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.033492133021354675, 'train/grad_norm_pre_clip': 0.9773305058479309, 'timing/data': 0.0002483590506017208, 'timing/model': 0.41334909200668335, 'learning_rate/qwen_vl_interface': 5.951463984847269e-06, 'learning_rate/action_model': 5.951463984847269e-05, 'epoch': 0.34})
+[2026-07-28 05:29:06,414][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.062257297337055206, 'train/grad_norm_pre_clip': 1.544370174407959, 'timing/data': 0.0005469457246363163, 'timing/model': 0.4183045644313097, 'learning_rate/qwen_vl_interface': 5.94767941161766e-06, 'learning_rate/action_model': 5.9476794116176596e-05, 'epoch': 0.34})
+[2026-07-28 05:29:12,539][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.06387919187545776, 'train/grad_norm_pre_clip': 2.421229600906372, 'timing/data': 0.00027409568428993225, 'timing/model': 0.6481544943526387, 'learning_rate/qwen_vl_interface': 5.943894385671456e-06, 'learning_rate/action_model': 5.9438943856714556e-05, 'epoch': 0.34})
+[2026-07-28 05:29:18,447][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.028199510648846626, 'train/grad_norm_pre_clip': 1.8521835803985596, 'timing/data': 0.0002791821025311947, 'timing/model': 0.4156817817129195, 'learning_rate/qwen_vl_interface': 5.940108909464719e-06, 'learning_rate/action_model': 5.940108909464719e-05, 'epoch': 0.34})
+[2026-07-28 05:29:24,616][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.03272709622979164, 'train/grad_norm_pre_clip': 1.9088242053985596, 'timing/data': 0.000347173772752285, 'timing/model': 0.6585737410932779, 'learning_rate/qwen_vl_interface': 5.936322985453802e-06, 'learning_rate/action_model': 5.936322985453802e-05, 'epoch': 0.34})
+[2026-07-28 05:29:30,509][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.06125984340906143, 'train/grad_norm_pre_clip': 1.8998862504959106, 'timing/data': 0.00033570220693945885, 'timing/model': 0.41794174863025546, 'learning_rate/qwen_vl_interface': 5.9325366160953516e-06, 'learning_rate/action_model': 5.932536616095351e-05, 'epoch': 0.34})
+[2026-07-28 05:29:36,425][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.05427902564406395, 'train/grad_norm_pre_clip': 1.1900532245635986, 'timing/data': 0.00034555792808532715, 'timing/model': 0.4164778967387974, 'learning_rate/qwen_vl_interface': 5.928749803846301e-06, 'learning_rate/action_model': 5.9287498038463006e-05, 'epoch': 0.34})
+[2026-07-28 05:29:42,641][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.0645904392004013, 'train/grad_norm_pre_clip': 1.8656545877456665, 'timing/data': 0.00027687475085258484, 'timing/model': 0.4176144772209227, 'learning_rate/qwen_vl_interface': 5.924962551163872e-06, 'learning_rate/action_model': 5.9249625511638716e-05, 'epoch': 0.34})
+[2026-07-28 05:29:48,758][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.042541272938251495, 'train/grad_norm_pre_clip': 10.386305809020996, 'timing/data': 0.00025263894349336624, 'timing/model': 0.4129035850055516, 'learning_rate/qwen_vl_interface': 5.921174860505573e-06, 'learning_rate/action_model': 5.9211748605055725e-05, 'epoch': 0.34})
+[2026-07-28 05:29:54,679][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.06094347685575485, 'train/grad_norm_pre_clip': 1.4413647651672363, 'timing/data': 0.0002668853849172592, 'timing/model': 0.41624662885442376, 'learning_rate/qwen_vl_interface': 5.9173867343291916e-06, 'learning_rate/action_model': 5.917386734329191e-05, 'epoch': 0.34})
+[2026-07-28 05:30:00,799][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.041259292513132095, 'train/grad_norm_pre_clip': 1.9412384033203125, 'timing/data': 0.00025325408205389977, 'timing/model': 0.4145923787727952, 'learning_rate/qwen_vl_interface': 5.913598175092804e-06, 'learning_rate/action_model': 5.913598175092804e-05, 'epoch': 0.34})
+[2026-07-28 05:30:06,694][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.04245137423276901, 'train/grad_norm_pre_clip': 1.7045060396194458, 'timing/data': 0.0002869642339646816, 'timing/model': 0.4159262152388692, 'learning_rate/qwen_vl_interface': 5.909809185254767e-06, 'learning_rate/action_model': 5.9098091852547665e-05, 'epoch': 0.34})
+[2026-07-28 05:30:12,839][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.09685137122869492, 'train/grad_norm_pre_clip': 2.6051769256591797, 'timing/data': 0.00025000888854265213, 'timing/model': 0.41349942702800035, 'learning_rate/qwen_vl_interface': 5.90601976727371e-06, 'learning_rate/action_model': 5.906019767273711e-05, 'epoch': 0.34})
+[2026-07-28 05:30:18,943][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.034997183829545975, 'train/grad_norm_pre_clip': 1.366417407989502, 'timing/data': 0.00034513603895902634, 'timing/model': 0.5917912018485367, 'learning_rate/qwen_vl_interface': 5.902229923608552e-06, 'learning_rate/action_model': 5.902229923608552e-05, 'epoch': 0.35})
+[2026-07-28 05:30:24,840][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.042589109390974045, 'train/grad_norm_pre_clip': 0.9240462183952332, 'timing/data': 0.00032888026908040047, 'timing/model': 0.41572464583441615, 'learning_rate/qwen_vl_interface': 5.898439656718476e-06, 'learning_rate/action_model': 5.898439656718476e-05, 'epoch': 0.35})
+[2026-07-28 05:30:30,976][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.035877618938684464, 'train/grad_norm_pre_clip': 1.1852096319198608, 'timing/data': 0.0003340989351272583, 'timing/model': 0.6597421108745039, 'learning_rate/qwen_vl_interface': 5.894648969062947e-06, 'learning_rate/action_model': 5.894648969062947e-05, 'epoch': 0.35})
+[2026-07-28 05:30:36,870][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.04046938568353653, 'train/grad_norm_pre_clip': 2.6782066822052, 'timing/data': 0.00032455800101161003, 'timing/model': 0.4158509145490825, 'learning_rate/qwen_vl_interface': 5.890857863101701e-06, 'learning_rate/action_model': 5.890857863101701e-05, 'epoch': 0.35})
+[2026-07-28 05:30:43,005][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.0810650885105133, 'train/grad_norm_pre_clip': 3.3511009216308594, 'timing/data': 0.0003437497653067112, 'timing/model': 0.41906622564420104, 'learning_rate/qwen_vl_interface': 5.887066341294745e-06, 'learning_rate/action_model': 5.8870663412947445e-05, 'epoch': 0.35})
+[2026-07-28 05:30:48,903][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.062045320868492126, 'train/grad_norm_pre_clip': 2.4383838176727295, 'timing/data': 0.000345572829246521, 'timing/model': 0.4177102819085121, 'learning_rate/qwen_vl_interface': 5.883274406102356e-06, 'learning_rate/action_model': 5.883274406102356e-05, 'epoch': 0.35})
+[2026-07-28 05:30:54,793][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.07632935792207718, 'train/grad_norm_pre_clip': 4.391557216644287, 'timing/data': 0.00033834343776106834, 'timing/model': 0.4172600726597011, 'learning_rate/qwen_vl_interface': 5.879482059985082e-06, 'learning_rate/action_model': 5.879482059985082e-05, 'epoch': 0.35})
+[2026-07-28 05:31:00,942][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.05485691875219345, 'train/grad_norm_pre_clip': 3.5258021354675293, 'timing/data': 0.00034448085352778435, 'timing/model': 0.6576315090060234, 'learning_rate/qwen_vl_interface': 5.875689305403732e-06, 'learning_rate/action_model': 5.875689305403732e-05, 'epoch': 0.35})
+[2026-07-28 05:31:07,089][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.04080784320831299, 'train/grad_norm_pre_clip': 1.0919910669326782, 'timing/data': 0.00030808616429567337, 'timing/model': 0.4162516128271818, 'learning_rate/qwen_vl_interface': 5.871896144819384e-06, 'learning_rate/action_model': 5.871896144819384e-05, 'epoch': 0.35})
+[2026-07-28 05:31:12,992][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.04435483366250992, 'train/grad_norm_pre_clip': 1.1477690935134888, 'timing/data': 0.00034995097666978836, 'timing/model': 0.41869483795017004, 'learning_rate/qwen_vl_interface': 5.868102580693381e-06, 'learning_rate/action_model': 5.868102580693381e-05, 'epoch': 0.35})
+[2026-07-28 05:31:18,897][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.04109703376889229, 'train/grad_norm_pre_clip': 1.0888394117355347, 'timing/data': 0.00034347688779234886, 'timing/model': 0.4166620369069278, 'learning_rate/qwen_vl_interface': 5.864308615487323e-06, 'learning_rate/action_model': 5.864308615487323e-05, 'epoch': 0.35})
+[2026-07-28 05:31:24,576][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.04033864289522171, 'train/grad_norm_pre_clip': 2.7034571170806885, 'timing/data': 0.0003448701463639736, 'timing/model': 0.41777264419943094, 'learning_rate/qwen_vl_interface': 5.860514251663075e-06, 'learning_rate/action_model': 5.860514251663075e-05, 'epoch': 0.35})
+[2026-07-28 05:31:30,264][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.06539924442768097, 'train/grad_norm_pre_clip': 1.1450822353363037, 'timing/data': 0.00032625580206513405, 'timing/model': 0.41608741134405136, 'learning_rate/qwen_vl_interface': 5.8567194916827575e-06, 'learning_rate/action_model': 5.8567194916827575e-05, 'epoch': 0.35})
+[2026-07-28 05:31:36,218][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.04075680673122406, 'train/grad_norm_pre_clip': 1.4846984148025513, 'timing/data': 0.00035083573311567307, 'timing/model': 0.6824280712753534, 'learning_rate/qwen_vl_interface': 5.852924338008748e-06, 'learning_rate/action_model': 5.8529243380087474e-05, 'epoch': 0.35})
+[2026-07-28 05:31:41,894][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.05022703856229782, 'train/grad_norm_pre_clip': 1.8962912559509277, 'timing/data': 0.0002996711991727352, 'timing/model': 0.4155278541147709, 'learning_rate/qwen_vl_interface': 5.849128793103682e-06, 'learning_rate/action_model': 5.849128793103682e-05, 'epoch': 0.35})
+[2026-07-28 05:31:47,619][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.010429618880152702, 'train/grad_norm_pre_clip': 0.508236289024353, 'timing/data': 0.0007326300255954266, 'timing/model': 0.4157948140054941, 'learning_rate/qwen_vl_interface': 5.845332859430446e-06, 'learning_rate/action_model': 5.845332859430446e-05, 'epoch': 0.35})
+[2026-07-28 05:31:53,362][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.08411725610494614, 'train/grad_norm_pre_clip': 2.8536596298217773, 'timing/data': 0.0003460049629211426, 'timing/model': 0.41714801732450724, 'learning_rate/qwen_vl_interface': 5.841536539452181e-06, 'learning_rate/action_model': 5.841536539452181e-05, 'epoch': 0.35})
+[2026-07-28 05:31:59,331][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.02660057693719864, 'train/grad_norm_pre_clip': 0.9336971640586853, 'timing/data': 0.0003391830250620842, 'timing/model': 0.41783915972337127, 'learning_rate/qwen_vl_interface': 5.83773983563228e-06, 'learning_rate/action_model': 5.8377398356322796e-05, 'epoch': 0.35})
+[2026-07-28 05:32:05,757][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.036197494715452194, 'train/grad_norm_pre_clip': 1.4496630430221558, 'timing/data': 0.0002585030160844326, 'timing/model': 0.41755607398226857, 'learning_rate/qwen_vl_interface': 5.833942750434377e-06, 'learning_rate/action_model': 5.8339427504343766e-05, 'epoch': 0.35})
+[2026-07-28 05:32:11,721][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.030535055324435234, 'train/grad_norm_pre_clip': 0.8302445411682129, 'timing/data': 0.00032520992681384087, 'timing/model': 0.4166266489773989, 'learning_rate/qwen_vl_interface': 5.830145286322365e-06, 'learning_rate/action_model': 5.830145286322365e-05, 'epoch': 0.35})
+[2026-07-28 05:32:17,680][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.04633522778749466, 'train/grad_norm_pre_clip': 1.4692715406417847, 'timing/data': 0.00035047903656959534, 'timing/model': 0.4180159168317914, 'learning_rate/qwen_vl_interface': 5.826347445760375e-06, 'learning_rate/action_model': 5.8263474457603754e-05, 'epoch': 0.35})
+[2026-07-28 05:32:23,869][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.05283958464860916, 'train/grad_norm_pre_clip': 1.3451244831085205, 'timing/data': 0.00030806008726358414, 'timing/model': 0.4176180218346417, 'learning_rate/qwen_vl_interface': 5.822549231212785e-06, 'learning_rate/action_model': 5.822549231212785e-05, 'epoch': 0.35})
+[2026-07-28 05:32:30,087][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.04610280692577362, 'train/grad_norm_pre_clip': 1.5700881481170654, 'timing/data': 0.0003238585777580738, 'timing/model': 0.41750325122848153, 'learning_rate/qwen_vl_interface': 5.8187506451442155e-06, 'learning_rate/action_model': 5.8187506451442155e-05, 'epoch': 0.35})
+[2026-07-28 05:32:36,293][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.05602634698152542, 'train/grad_norm_pre_clip': 1.6618682146072388, 'timing/data': 0.0002628215588629246, 'timing/model': 0.4498554337769747, 'learning_rate/qwen_vl_interface': 5.814951690019527e-06, 'learning_rate/action_model': 5.814951690019527e-05, 'epoch': 0.35})
+[2026-07-28 05:32:42,489][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.03393116965889931, 'train/grad_norm_pre_clip': 0.8856695294380188, 'timing/data': 0.0003255559131503105, 'timing/model': 0.45292131369933486, 'learning_rate/qwen_vl_interface': 5.811152368303819e-06, 'learning_rate/action_model': 5.811152368303819e-05, 'epoch': 0.35})
+[2026-07-28 05:32:48,722][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.04148496314883232, 'train/grad_norm_pre_clip': 1.3841361999511719, 'timing/data': 0.0002618841826915741, 'timing/model': 0.6487424187362194, 'learning_rate/qwen_vl_interface': 5.807352682462432e-06, 'learning_rate/action_model': 5.807352682462432e-05, 'epoch': 0.35})
+[2026-07-28 05:32:54,627][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.0596480593085289, 'train/grad_norm_pre_clip': 2.5861170291900635, 'timing/data': 0.00024463701993227005, 'timing/model': 0.4136871509253979, 'learning_rate/qwen_vl_interface': 5.803552634960942e-06, 'learning_rate/action_model': 5.8035526349609415e-05, 'epoch': 0.35})
+[2026-07-28 05:33:00,818][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.09328951686620712, 'train/grad_norm_pre_clip': 3.2547755241394043, 'timing/data': 0.0002653137780725956, 'timing/model': 0.6519964681938291, 'learning_rate/qwen_vl_interface': 5.799752228265157e-06, 'learning_rate/action_model': 5.799752228265156e-05, 'epoch': 0.35})
+[2026-07-28 05:33:06,700][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.051502328366041183, 'train/grad_norm_pre_clip': 1.3692498207092285, 'timing/data': 0.0002445932477712631, 'timing/model': 0.4137308979406953, 'learning_rate/qwen_vl_interface': 5.7959514648411175e-06, 'learning_rate/action_model': 5.795951464841117e-05, 'epoch': 0.35})
+[2026-07-28 05:33:12,638][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.04245515540242195, 'train/grad_norm_pre_clip': 0.9216362833976746, 'timing/data': 0.0002758488990366459, 'timing/model': 0.4159270008094609, 'learning_rate/qwen_vl_interface': 5.792150347155101e-06, 'learning_rate/action_model': 5.7921503471551007e-05, 'epoch': 0.35})
+[2026-07-28 05:33:18,551][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.03509548306465149, 'train/grad_norm_pre_clip': 1.271093726158142, 'timing/data': 0.0002815490588545799, 'timing/model': 0.415787145961076, 'learning_rate/qwen_vl_interface': 5.78834887767361e-06, 'learning_rate/action_model': 5.7883488776736093e-05, 'epoch': 0.35})
+[2026-07-28 05:33:24,691][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.03408075124025345, 'train/grad_norm_pre_clip': 0.779247522354126, 'timing/data': 0.00024564284831285477, 'timing/model': 0.4140718663111329, 'learning_rate/qwen_vl_interface': 5.784547058863377e-06, 'learning_rate/action_model': 5.784547058863377e-05, 'epoch': 0.35})
+[2026-07-28 05:33:30,595][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.03372735157608986, 'train/grad_norm_pre_clip': 0.8821839690208435, 'timing/data': 0.0002716630697250366, 'timing/model': 0.41377231013029814, 'learning_rate/qwen_vl_interface': 5.780744893191361e-06, 'learning_rate/action_model': 5.7807448931913614e-05, 'epoch': 0.35})
+[2026-07-28 05:33:36,482][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.06357064843177795, 'train/grad_norm_pre_clip': 1.1020268201828003, 'timing/data': 0.00026880204677581787, 'timing/model': 0.41713979095220566, 'learning_rate/qwen_vl_interface': 5.7769423831247474e-06, 'learning_rate/action_model': 5.776942383124747e-05, 'epoch': 0.35})
+[2026-07-28 05:33:42,646][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.027475081384181976, 'train/grad_norm_pre_clip': 0.6528543829917908, 'timing/data': 0.0002483581192791462, 'timing/model': 0.4156684768386185, 'learning_rate/qwen_vl_interface': 5.7731395311309405e-06, 'learning_rate/action_model': 5.77313953113094e-05, 'epoch': 0.35})
+[2026-07-28 05:33:48,811][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.039974138140678406, 'train/grad_norm_pre_clip': 1.8035105466842651, 'timing/data': 0.0002806512638926506, 'timing/model': 0.41887162206694484, 'learning_rate/qwen_vl_interface': 5.769336339677573e-06, 'learning_rate/action_model': 5.769336339677572e-05, 'epoch': 0.35})
+[2026-07-28 05:33:54,794][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.07732035219669342, 'train/grad_norm_pre_clip': 1.901508092880249, 'timing/data': 0.0002698339521884918, 'timing/model': 0.4163442300632596, 'learning_rate/qwen_vl_interface': 5.76553281123249e-06, 'learning_rate/action_model': 5.7655328112324904e-05, 'epoch': 0.35})
+[2026-07-28 05:34:00,488][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.0281389020383358, 'train/grad_norm_pre_clip': 0.7566134929656982, 'timing/data': 0.00035542482510209084, 'timing/model': 0.41679126024246216, 'learning_rate/qwen_vl_interface': 5.761728948263766e-06, 'learning_rate/action_model': 5.761728948263766e-05, 'epoch': 0.35})
+[2026-07-28 05:34:06,671][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.06160631775856018, 'train/grad_norm_pre_clip': 2.9357151985168457, 'timing/data': 0.00024251220747828484, 'timing/model': 0.6448746281675994, 'learning_rate/qwen_vl_interface': 5.757924753239683e-06, 'learning_rate/action_model': 5.757924753239683e-05, 'epoch': 0.35})
+[2026-07-28 05:34:12,563][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.05422036349773407, 'train/grad_norm_pre_clip': 1.7196881771087646, 'timing/data': 0.00025249412283301353, 'timing/model': 0.41545510198920965, 'learning_rate/qwen_vl_interface': 5.7541202286287405e-06, 'learning_rate/action_model': 5.75412022862874e-05, 'epoch': 0.35})
+[2026-07-28 05:34:18,234][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.06054965406656265, 'train/grad_norm_pre_clip': 1.216012954711914, 'timing/data': 0.0002643400803208351, 'timing/model': 0.41508339904248714, 'learning_rate/qwen_vl_interface': 5.750315376899655e-06, 'learning_rate/action_model': 5.7503153768996554e-05, 'epoch': 0.35})
+[2026-07-28 05:34:24,303][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.062280986458063126, 'train/grad_norm_pre_clip': 2.4508819580078125, 'timing/data': 0.00025781895965337753, 'timing/model': 0.5849803378805518, 'learning_rate/qwen_vl_interface': 5.746510200521354e-06, 'learning_rate/action_model': 5.746510200521355e-05, 'epoch': 0.35})
+[2026-07-28 05:34:30,469][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.03346160426735878, 'train/grad_norm_pre_clip': 0.8811169862747192, 'timing/data': 0.0003172038123011589, 'timing/model': 0.4141680793836713, 'learning_rate/qwen_vl_interface': 5.742704701962974e-06, 'learning_rate/action_model': 5.7427047019629744e-05, 'epoch': 0.35})
+[2026-07-28 05:34:36,372][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.04856666177511215, 'train/grad_norm_pre_clip': 1.3725767135620117, 'timing/data': 0.0002717869356274605, 'timing/model': 0.41629618033766747, 'learning_rate/qwen_vl_interface': 5.7388988836938645e-06, 'learning_rate/action_model': 5.738898883693864e-05, 'epoch': 0.35})
+[2026-07-28 05:34:42,290][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.028602145612239838, 'train/grad_norm_pre_clip': 1.5894126892089844, 'timing/data': 0.00026848725974559784, 'timing/model': 0.41627585608512163, 'learning_rate/qwen_vl_interface': 5.735092748183574e-06, 'learning_rate/action_model': 5.7350927481835736e-05, 'epoch': 0.35})
+[2026-07-28 05:34:48,438][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.05495559051632881, 'train/grad_norm_pre_clip': 1.5278115272521973, 'timing/data': 0.00027633877471089363, 'timing/model': 0.4168302621692419, 'learning_rate/qwen_vl_interface': 5.7312862979018676e-06, 'learning_rate/action_model': 5.731286297901868e-05, 'epoch': 0.35})
+[2026-07-28 05:34:54,575][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.05423945188522339, 'train/grad_norm_pre_clip': 1.55107581615448, 'timing/data': 0.00024597207084298134, 'timing/model': 0.4133778512477875, 'learning_rate/qwen_vl_interface': 5.727479535318708e-06, 'learning_rate/action_model': 5.727479535318707e-05, 'epoch': 0.35})
+[2026-07-28 05:35:00,520][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.04002991318702698, 'train/grad_norm_pre_clip': 1.243948221206665, 'timing/data': 0.00025422917678952217, 'timing/model': 0.41318627214059234, 'learning_rate/qwen_vl_interface': 5.7236724629042595e-06, 'learning_rate/action_model': 5.7236724629042594e-05, 'epoch': 0.35})
+[2026-07-28 05:35:06,454][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.0479302853345871, 'train/grad_norm_pre_clip': 1.3885536193847656, 'timing/data': 0.00026821019127964973, 'timing/model': 0.4160894942469895, 'learning_rate/qwen_vl_interface': 5.719865083128895e-06, 'learning_rate/action_model': 5.7198650831288944e-05, 'epoch': 0.35})
+[2026-07-28 05:35:12,844][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.03698495030403137, 'train/grad_norm_pre_clip': 0.8844303488731384, 'timing/data': 0.00024631619453430176, 'timing/model': 0.6471797022968531, 'learning_rate/qwen_vl_interface': 5.716057398463179e-06, 'learning_rate/action_model': 5.7160573984631785e-05, 'epoch': 0.35})
+[2026-07-28 05:35:18,557][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.044326022267341614, 'train/grad_norm_pre_clip': 0.8072965145111084, 'timing/data': 0.0002762060612440109, 'timing/model': 0.4162012101151049, 'learning_rate/qwen_vl_interface': 5.712249411377875e-06, 'learning_rate/action_model': 5.712249411377875e-05, 'epoch': 0.35})
+[2026-07-28 05:35:24,486][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.07353946566581726, 'train/grad_norm_pre_clip': 1.444699764251709, 'timing/data': 0.0002638823352754116, 'timing/model': 0.415797860827297, 'learning_rate/qwen_vl_interface': 5.708441124343948e-06, 'learning_rate/action_model': 5.708441124343947e-05, 'epoch': 0.35})
+[2026-07-28 05:35:30,408][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.06451631337404251, 'train/grad_norm_pre_clip': 1.3384686708450317, 'timing/data': 0.00026814406737685204, 'timing/model': 0.4166230522096157, 'learning_rate/qwen_vl_interface': 5.704632539832551e-06, 'learning_rate/action_model': 5.704632539832551e-05, 'epoch': 0.35})
+[2026-07-28 05:35:36,339][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.023046553134918213, 'train/grad_norm_pre_clip': 0.6097251772880554, 'timing/data': 0.00026627304032444954, 'timing/model': 0.41592316795140505, 'learning_rate/qwen_vl_interface': 5.700823660315038e-06, 'learning_rate/action_model': 5.7008236603150375e-05, 'epoch': 0.35})
+[2026-07-28 05:35:42,493][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.04169195890426636, 'train/grad_norm_pre_clip': 1.4398283958435059, 'timing/data': 0.00027233827859163284, 'timing/model': 0.6534148273058236, 'learning_rate/qwen_vl_interface': 5.697014488262945e-06, 'learning_rate/action_model': 5.697014488262945e-05, 'epoch': 0.36})
+[2026-07-28 05:35:48,180][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.04557107016444206, 'train/grad_norm_pre_clip': 1.0545014142990112, 'timing/data': 0.00024580396711826324, 'timing/model': 0.4125398271717131, 'learning_rate/qwen_vl_interface': 5.693205026148002e-06, 'learning_rate/action_model': 5.693205026148002e-05, 'epoch': 0.36})
+[2026-07-28 05:35:54,113][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.03510427847504616, 'train/grad_norm_pre_clip': 0.9999141097068787, 'timing/data': 0.0002771662548184395, 'timing/model': 0.41538709308952093, 'learning_rate/qwen_vl_interface': 5.689395276442131e-06, 'learning_rate/action_model': 5.689395276442131e-05, 'epoch': 0.36})
+[2026-07-28 05:36:00,057][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.0357956700026989, 'train/grad_norm_pre_clip': 1.6422016620635986, 'timing/data': 0.00027736276388168335, 'timing/model': 0.4177612140774727, 'learning_rate/qwen_vl_interface': 5.685585241617435e-06, 'learning_rate/action_model': 5.6855852416174346e-05, 'epoch': 0.36})
+[2026-07-28 05:36:06,020][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.05329524725675583, 'train/grad_norm_pre_clip': 1.7105313539505005, 'timing/data': 0.0002697119489312172, 'timing/model': 0.4176685339771211, 'learning_rate/qwen_vl_interface': 5.681774924146205e-06, 'learning_rate/action_model': 5.6817749241462046e-05, 'epoch': 0.36})
+[2026-07-28 05:36:12,166][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.053686223924160004, 'train/grad_norm_pre_clip': 2.4489808082580566, 'timing/data': 0.00024327542632818222, 'timing/model': 0.41388151375576854, 'learning_rate/qwen_vl_interface': 5.677964326500914e-06, 'learning_rate/action_model': 5.6779643265009144e-05, 'epoch': 0.36})
+[2026-07-28 05:36:18,344][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.0392640084028244, 'train/grad_norm_pre_clip': 1.2079708576202393, 'timing/data': 0.0002784538082778454, 'timing/model': 0.6479235510341823, 'learning_rate/qwen_vl_interface': 5.674153451154218e-06, 'learning_rate/action_model': 5.674153451154218e-05, 'epoch': 0.36})
+[2026-07-28 05:36:24,265][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.02151169255375862, 'train/grad_norm_pre_clip': 0.8436502814292908, 'timing/data': 0.0002460610121488571, 'timing/model': 0.41421074885874987, 'learning_rate/qwen_vl_interface': 5.670342300578954e-06, 'learning_rate/action_model': 5.670342300578954e-05, 'epoch': 0.36})
+[2026-07-28 05:36:30,442][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.04303457960486412, 'train/grad_norm_pre_clip': 1.7114059925079346, 'timing/data': 0.00023875292390584946, 'timing/model': 0.4159944551065564, 'learning_rate/qwen_vl_interface': 5.666530877248134e-06, 'learning_rate/action_model': 5.666530877248134e-05, 'epoch': 0.36})
+[2026-07-28 05:36:36,390][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.03152145445346832, 'train/grad_norm_pre_clip': 1.604415774345398, 'timing/data': 0.0003570783883333206, 'timing/model': 0.4151998762972653, 'learning_rate/qwen_vl_interface': 5.6627191836349484e-06, 'learning_rate/action_model': 5.662719183634948e-05, 'epoch': 0.36})
+[2026-07-28 05:36:42,322][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.06368108093738556, 'train/grad_norm_pre_clip': 1.2037397623062134, 'timing/data': 0.0002740449272096157, 'timing/model': 0.4161970540881157, 'learning_rate/qwen_vl_interface': 5.6589072222127665e-06, 'learning_rate/action_model': 5.658907222212766e-05, 'epoch': 0.36})
+[2026-07-28 05:36:48,496][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.05472417175769806, 'train/grad_norm_pre_clip': 1.132591724395752, 'timing/data': 0.00024818582460284233, 'timing/model': 0.4146103709936142, 'learning_rate/qwen_vl_interface': 5.655094995455128e-06, 'learning_rate/action_model': 5.655094995455128e-05, 'epoch': 0.36})
+[2026-07-28 05:36:54,659][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.03915072977542877, 'train/grad_norm_pre_clip': 1.4288625717163086, 'timing/data': 0.000256554689258337, 'timing/model': 0.6513268239796162, 'learning_rate/qwen_vl_interface': 5.6512825058357445e-06, 'learning_rate/action_model': 5.651282505835745e-05, 'epoch': 0.36})
+[2026-07-28 05:37:00,545][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.03841690719127655, 'train/grad_norm_pre_clip': 1.362505555152893, 'timing/data': 0.00023687677457928658, 'timing/model': 0.4138354351744056, 'learning_rate/qwen_vl_interface': 5.647469755828501e-06, 'learning_rate/action_model': 5.647469755828501e-05, 'epoch': 0.36})
+[2026-07-28 05:37:06,482][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.09310717135667801, 'train/grad_norm_pre_clip': 2.593432664871216, 'timing/data': 0.0002643582411110401, 'timing/model': 0.42415753472596407, 'learning_rate/qwen_vl_interface': 5.643656747907446e-06, 'learning_rate/action_model': 5.643656747907445e-05, 'epoch': 0.36})
+[2026-07-28 05:37:12,626][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.05406288057565689, 'train/grad_norm_pre_clip': 1.1538268327713013, 'timing/data': 0.0003741886466741562, 'timing/model': 0.41392081789672375, 'learning_rate/qwen_vl_interface': 5.639843484546801e-06, 'learning_rate/action_model': 5.639843484546801e-05, 'epoch': 0.36})
+[2026-07-28 05:37:18,529][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.04631929099559784, 'train/grad_norm_pre_clip': 1.5844802856445312, 'timing/data': 0.0002632932737469673, 'timing/model': 0.41675102338194847, 'learning_rate/qwen_vl_interface': 5.636029968220953e-06, 'learning_rate/action_model': 5.6360299682209525e-05, 'epoch': 0.36})
+[2026-07-28 05:37:24,416][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.022547585889697075, 'train/grad_norm_pre_clip': 1.141371726989746, 'timing/data': 0.0002520228736102581, 'timing/model': 0.4160950109362602, 'learning_rate/qwen_vl_interface': 5.632216201404449e-06, 'learning_rate/action_model': 5.632216201404449e-05, 'epoch': 0.36})
+[2026-07-28 05:37:30,562][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.05182008445262909, 'train/grad_norm_pre_clip': 1.7230079174041748, 'timing/data': 0.00023469561710953712, 'timing/model': 0.4132942189462483, 'learning_rate/qwen_vl_interface': 5.628402186572002e-06, 'learning_rate/action_model': 5.6284021865720015e-05, 'epoch': 0.36})
+[2026-07-28 05:37:36,453][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.04489622265100479, 'train/grad_norm_pre_clip': 2.0738139152526855, 'timing/data': 0.00027990108355879784, 'timing/model': 0.41510647209361196, 'learning_rate/qwen_vl_interface': 5.6245879261984835e-06, 'learning_rate/action_model': 5.6245879261984836e-05, 'epoch': 0.36})
+[2026-07-28 05:37:42,565][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.038008470088243484, 'train/grad_norm_pre_clip': 2.110508918762207, 'timing/data': 0.00025428272783756256, 'timing/model': 0.6456924830563366, 'learning_rate/qwen_vl_interface': 5.620773422758927e-06, 'learning_rate/action_model': 5.620773422758927e-05, 'epoch': 0.36})
+[2026-07-28 05:37:48,445][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.05507750064134598, 'train/grad_norm_pre_clip': 0.933048665523529, 'timing/data': 0.00024392688646912575, 'timing/model': 0.4134211749769747, 'learning_rate/qwen_vl_interface': 5.616958678728522e-06, 'learning_rate/action_model': 5.616958678728522e-05, 'epoch': 0.36})
+[2026-07-28 05:37:54,577][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.05115986242890358, 'train/grad_norm_pre_clip': 1.9850174188613892, 'timing/data': 0.00026291096583008766, 'timing/model': 0.4178710822016001, 'learning_rate/qwen_vl_interface': 5.6131436965826145e-06, 'learning_rate/action_model': 5.613143696582615e-05, 'epoch': 0.36})
+[2026-07-28 05:38:00,263][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.020845256745815277, 'train/grad_norm_pre_clip': 0.7152483463287354, 'timing/data': 0.00026426883414387703, 'timing/model': 0.4208732726983726, 'learning_rate/qwen_vl_interface': 5.609328478796705e-06, 'learning_rate/action_model': 5.6093284787967046e-05, 'epoch': 0.36})
+[2026-07-28 05:38:06,207][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.05430919677019119, 'train/grad_norm_pre_clip': 1.028738021850586, 'timing/data': 0.00026691704988479614, 'timing/model': 0.4219889431260526, 'learning_rate/qwen_vl_interface': 5.605513027846447e-06, 'learning_rate/action_model': 5.605513027846446e-05, 'epoch': 0.36})
+[2026-07-28 05:38:12,363][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.03237207606434822, 'train/grad_norm_pre_clip': 1.0235910415649414, 'timing/data': 0.00023541506379842758, 'timing/model': 0.4144574347883463, 'learning_rate/qwen_vl_interface': 5.601697346207644e-06, 'learning_rate/action_model': 5.601697346207644e-05, 'epoch': 0.36})
+[2026-07-28 05:38:18,284][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.028024498373270035, 'train/grad_norm_pre_clip': 0.8426833152770996, 'timing/data': 0.0002597789280116558, 'timing/model': 0.417723401915282, 'learning_rate/qwen_vl_interface': 5.597881436356252e-06, 'learning_rate/action_model': 5.597881436356252e-05, 'epoch': 0.36})
+[2026-07-28 05:38:24,503][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.04295791685581207, 'train/grad_norm_pre_clip': 1.5000686645507812, 'timing/data': 0.00024634599685668945, 'timing/model': 0.4141566976904869, 'learning_rate/qwen_vl_interface': 5.594065300768372e-06, 'learning_rate/action_model': 5.5940653007683716e-05, 'epoch': 0.36})
+[2026-07-28 05:38:30,606][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.08345334231853485, 'train/grad_norm_pre_clip': 2.8780200481414795, 'timing/data': 0.0002598278224468231, 'timing/model': 0.5921064401045442, 'learning_rate/qwen_vl_interface': 5.590248941920255e-06, 'learning_rate/action_model': 5.5902489419202554e-05, 'epoch': 0.36})
+[2026-07-28 05:38:36,556][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.05551169440150261, 'train/grad_norm_pre_clip': 2.5685176849365234, 'timing/data': 0.0002517816610634327, 'timing/model': 0.4152070046402514, 'learning_rate/qwen_vl_interface': 5.586432362288292e-06, 'learning_rate/action_model': 5.586432362288292e-05, 'epoch': 0.36})
+[2026-07-28 05:38:42,463][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.049637652933597565, 'train/grad_norm_pre_clip': 1.6198517084121704, 'timing/data': 0.00025261472910642624, 'timing/model': 0.4173852549865842, 'learning_rate/qwen_vl_interface': 5.582615564349024e-06, 'learning_rate/action_model': 5.582615564349024e-05, 'epoch': 0.36})
+[2026-07-28 05:38:48,598][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.059757575392723083, 'train/grad_norm_pre_clip': 2.58711314201355, 'timing/data': 0.0002628602087497711, 'timing/model': 0.6424811012111604, 'learning_rate/qwen_vl_interface': 5.578798550579126e-06, 'learning_rate/action_model': 5.578798550579125e-05, 'epoch': 0.36})
+[2026-07-28 05:38:54,562][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.03210850805044174, 'train/grad_norm_pre_clip': 1.0881798267364502, 'timing/data': 0.00026555638760328293, 'timing/model': 0.4183975108899176, 'learning_rate/qwen_vl_interface': 5.5749813234554175e-06, 'learning_rate/action_model': 5.574981323455417e-05, 'epoch': 0.36})
+[2026-07-28 05:39:00,473][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.045320380479097366, 'train/grad_norm_pre_clip': 2.057377338409424, 'timing/data': 0.0002559279091656208, 'timing/model': 0.42396712535992265, 'learning_rate/qwen_vl_interface': 5.571163885454859e-06, 'learning_rate/action_model': 5.5711638854548583e-05, 'epoch': 0.36})
+[2026-07-28 05:39:06,450][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.037851400673389435, 'train/grad_norm_pre_clip': 1.129962682723999, 'timing/data': 0.00026113586500287056, 'timing/model': 0.41444537695497274, 'learning_rate/qwen_vl_interface': 5.567346239054541e-06, 'learning_rate/action_model': 5.567346239054541e-05, 'epoch': 0.36})
+[2026-07-28 05:39:12,568][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.044891469180583954, 'train/grad_norm_pre_clip': 1.4331270456314087, 'timing/data': 0.00023691169917583466, 'timing/model': 0.41421737521886826, 'learning_rate/qwen_vl_interface': 5.563528386731692e-06, 'learning_rate/action_model': 5.5635283867316926e-05, 'epoch': 0.36})
+[2026-07-28 05:39:18,702][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.037625689059495926, 'train/grad_norm_pre_clip': 1.4124071598052979, 'timing/data': 0.0002369871363043785, 'timing/model': 0.41515490924939513, 'learning_rate/qwen_vl_interface': 5.559710330963678e-06, 'learning_rate/action_model': 5.5597103309636775e-05, 'epoch': 0.36})
+[2026-07-28 05:39:24,590][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.08104847371578217, 'train/grad_norm_pre_clip': 3.644353151321411, 'timing/data': 0.0002625873312354088, 'timing/model': 0.4170426339842379, 'learning_rate/qwen_vl_interface': 5.555892074227993e-06, 'learning_rate/action_model': 5.555892074227993e-05, 'epoch': 0.36})
+[2026-07-28 05:39:30,491][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.06403972208499908, 'train/grad_norm_pre_clip': 1.338432788848877, 'timing/data': 0.00034206686541438103, 'timing/model': 0.41609376994892955, 'learning_rate/qwen_vl_interface': 5.552073619002263e-06, 'learning_rate/action_model': 5.552073619002263e-05, 'epoch': 0.36})
+[2026-07-28 05:39:36,623][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.04388337582349777, 'train/grad_norm_pre_clip': 1.154505729675293, 'timing/data': 0.00024008098989725113, 'timing/model': 0.41454653488472104, 'learning_rate/qwen_vl_interface': 5.548254967764239e-06, 'learning_rate/action_model': 5.5482549677642394e-05, 'epoch': 0.36})
+[2026-07-28 05:39:42,513][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.04097471013665199, 'train/grad_norm_pre_clip': 1.7388793230056763, 'timing/data': 0.00025333603844046593, 'timing/model': 0.41533171897754073, 'learning_rate/qwen_vl_interface': 5.544436122991806e-06, 'learning_rate/action_model': 5.544436122991806e-05, 'epoch': 0.36})
+[2026-07-28 05:39:48,640][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.014502198435366154, 'train/grad_norm_pre_clip': 0.45779770612716675, 'timing/data': 0.00026006903499364853, 'timing/model': 0.6536124241538346, 'learning_rate/qwen_vl_interface': 5.54061708716297e-06, 'learning_rate/action_model': 5.540617087162969e-05, 'epoch': 0.36})
+[2026-07-28 05:39:54,751][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.05608392506837845, 'train/grad_norm_pre_clip': 1.4798409938812256, 'timing/data': 0.0005533029325306416, 'timing/model': 0.641512684058398, 'learning_rate/qwen_vl_interface': 5.536797862755859e-06, 'learning_rate/action_model': 5.5367978627558595e-05, 'epoch': 0.36})
+[2026-07-28 05:40:00,631][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.0629369467496872, 'train/grad_norm_pre_clip': 1.5163843631744385, 'timing/data': 0.00025131506845355034, 'timing/model': 0.41522922180593014, 'learning_rate/qwen_vl_interface': 5.53297845224873e-06, 'learning_rate/action_model': 5.5329784522487296e-05, 'epoch': 0.36})
+[2026-07-28 05:40:06,533][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.05604102835059166, 'train/grad_norm_pre_clip': 2.318504571914673, 'timing/data': 0.0003420822322368622, 'timing/model': 0.42362719820812345, 'learning_rate/qwen_vl_interface': 5.529158858119958e-06, 'learning_rate/action_model': 5.529158858119957e-05, 'epoch': 0.36})
+[2026-07-28 05:40:12,418][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.04948064312338829, 'train/grad_norm_pre_clip': 1.3852217197418213, 'timing/data': 0.0002559409476816654, 'timing/model': 0.4159051110036671, 'learning_rate/qwen_vl_interface': 5.52533908284803e-06, 'learning_rate/action_model': 5.52533908284803e-05, 'epoch': 0.36})
+[2026-07-28 05:40:18,559][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.03554227948188782, 'train/grad_norm_pre_clip': 1.818318247795105, 'timing/data': 0.00023561902344226837, 'timing/model': 0.41380306892096996, 'learning_rate/qwen_vl_interface': 5.521519128911561e-06, 'learning_rate/action_model': 5.521519128911561e-05, 'epoch': 0.36})
+[2026-07-28 05:40:24,723][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.029951784759759903, 'train/grad_norm_pre_clip': 0.8558621406555176, 'timing/data': 0.000245463103055954, 'timing/model': 0.6433245921507478, 'learning_rate/qwen_vl_interface': 5.5176989987892785e-06, 'learning_rate/action_model': 5.517698998789278e-05, 'epoch': 0.36})
+[2026-07-28 05:40:30,625][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.07578708231449127, 'train/grad_norm_pre_clip': 1.94835364818573, 'timing/data': 0.00023570796474814415, 'timing/model': 0.41145090106874704, 'learning_rate/qwen_vl_interface': 5.513878694960021e-06, 'learning_rate/action_model': 5.513878694960021e-05, 'epoch': 0.36})
+[2026-07-28 05:40:36,749][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.05222129076719284, 'train/grad_norm_pre_clip': 1.3553739786148071, 'timing/data': 0.0003082309849560261, 'timing/model': 0.4158468237146735, 'learning_rate/qwen_vl_interface': 5.510058219902744e-06, 'learning_rate/action_model': 5.5100582199027436e-05, 'epoch': 0.36})
+[2026-07-28 05:40:42,641][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.05475145950913429, 'train/grad_norm_pre_clip': 1.383234977722168, 'timing/data': 0.0002777702175080776, 'timing/model': 0.4163124840706587, 'learning_rate/qwen_vl_interface': 5.506237576096512e-06, 'learning_rate/action_model': 5.506237576096511e-05, 'epoch': 0.36})
+[2026-07-28 05:40:48,301][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.046418897807598114, 'train/grad_norm_pre_clip': 1.5108076333999634, 'timing/data': 0.00026785582304000854, 'timing/model': 0.41622235905379057, 'learning_rate/qwen_vl_interface': 5.502416766020497e-06, 'learning_rate/action_model': 5.502416766020496e-05, 'epoch': 0.36})
+[2026-07-28 05:40:54,189][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.022316355258226395, 'train/grad_norm_pre_clip': 0.7749425768852234, 'timing/data': 0.00024937698617577553, 'timing/model': 0.4177721180021763, 'learning_rate/qwen_vl_interface': 5.498595792153985e-06, 'learning_rate/action_model': 5.498595792153984e-05, 'epoch': 0.36})
+[2026-07-28 05:41:00,440][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.046842094510793686, 'train/grad_norm_pre_clip': 1.613499402999878, 'timing/data': 0.0003899969160556793, 'timing/model': 0.650132087059319, 'learning_rate/qwen_vl_interface': 5.4947746569763595e-06, 'learning_rate/action_model': 5.494774656976359e-05, 'epoch': 0.36})
+[2026-07-28 05:41:06,376][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.06327036023139954, 'train/grad_norm_pre_clip': 3.1013944149017334, 'timing/data': 0.00030037015676498413, 'timing/model': 0.41358912317082286, 'learning_rate/qwen_vl_interface': 5.490953362967119e-06, 'learning_rate/action_model': 5.4909533629671185e-05, 'epoch': 0.36})
+[2026-07-28 05:41:12,366][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.04272422567009926, 'train/grad_norm_pre_clip': 1.0848941802978516, 'timing/data': 0.00025758426636457443, 'timing/model': 0.41602019034326077, 'learning_rate/qwen_vl_interface': 5.487131912605859e-06, 'learning_rate/action_model': 5.4871319126058584e-05, 'epoch': 0.37})
+[2026-07-28 05:41:18,255][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.027620255947113037, 'train/grad_norm_pre_clip': 0.7928735017776489, 'timing/data': 0.00032784603536129, 'timing/model': 0.41542675718665123, 'learning_rate/qwen_vl_interface': 5.48331030837227e-06, 'learning_rate/action_model': 5.4833103083722705e-05, 'epoch': 0.37})
+[2026-07-28 05:41:24,148][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.060140226036310196, 'train/grad_norm_pre_clip': 2.076589822769165, 'timing/data': 0.00025935098528862, 'timing/model': 0.4169420348480344, 'learning_rate/qwen_vl_interface': 5.4794885527461596e-06, 'learning_rate/action_model': 5.479488552746159e-05, 'epoch': 0.37})
+[2026-07-28 05:41:30,263][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.017296165227890015, 'train/grad_norm_pre_clip': 0.49390503764152527, 'timing/data': 0.0002316078171133995, 'timing/model': 0.41429995791986585, 'learning_rate/qwen_vl_interface': 5.4756666482074164e-06, 'learning_rate/action_model': 5.475666648207416e-05, 'epoch': 0.37})
+[2026-07-28 05:41:36,396][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.04744388908147812, 'train/grad_norm_pre_clip': 1.0855028629302979, 'timing/data': 0.0003051501698791981, 'timing/model': 0.4136097370646894, 'learning_rate/qwen_vl_interface': 5.471844597236036e-06, 'learning_rate/action_model': 5.471844597236035e-05, 'epoch': 0.37})
+[2026-07-28 05:41:42,300][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.051018621772527695, 'train/grad_norm_pre_clip': 1.5375587940216064, 'timing/data': 0.00025579286739230156, 'timing/model': 0.41656936099752784, 'learning_rate/qwen_vl_interface': 5.468022402312105e-06, 'learning_rate/action_model': 5.468022402312105e-05, 'epoch': 0.37})
+[2026-07-28 05:41:48,208][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.060039110481739044, 'train/grad_norm_pre_clip': 1.2356841564178467, 'timing/data': 0.0003448640927672386, 'timing/model': 0.4189929789863527, 'learning_rate/qwen_vl_interface': 5.464200065915806e-06, 'learning_rate/action_model': 5.4642000659158054e-05, 'epoch': 0.37})
+[2026-07-28 05:41:54,368][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.048017654567956924, 'train/grad_norm_pre_clip': 1.348197102546692, 'timing/data': 0.00024355389177799225, 'timing/model': 0.41714124009013176, 'learning_rate/qwen_vl_interface': 5.460377590527407e-06, 'learning_rate/action_model': 5.460377590527407e-05, 'epoch': 0.37})
+[2026-07-28 05:42:00,520][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.014128782786428928, 'train/grad_norm_pre_clip': 0.6323524117469788, 'timing/data': 0.000273323617875576, 'timing/model': 0.42268216889351606, 'learning_rate/qwen_vl_interface': 5.456554978627276e-06, 'learning_rate/action_model': 5.456554978627275e-05, 'epoch': 0.37})
+[2026-07-28 05:42:06,442][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.040349848568439484, 'train/grad_norm_pre_clip': 1.6487623453140259, 'timing/data': 0.00025806622579693794, 'timing/model': 0.4187529138289392, 'learning_rate/qwen_vl_interface': 5.452732232695859e-06, 'learning_rate/action_model': 5.452732232695858e-05, 'epoch': 0.37})
+[2026-07-28 05:42:12,124][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.035957276821136475, 'train/grad_norm_pre_clip': 1.5920689105987549, 'timing/data': 0.0002509388141334057, 'timing/model': 0.41766993375495076, 'learning_rate/qwen_vl_interface': 5.4489093552136986e-06, 'learning_rate/action_model': 5.4489093552136984e-05, 'epoch': 0.37})
+[2026-07-28 05:42:18,039][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.02333907037973404, 'train/grad_norm_pre_clip': 0.7435986995697021, 'timing/data': 0.00030746031552553177, 'timing/model': 0.4283294351771474, 'learning_rate/qwen_vl_interface': 5.445086348661418e-06, 'learning_rate/action_model': 5.4450863486614177e-05, 'epoch': 0.37})
+[2026-07-28 05:42:24,183][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000
+[2026-07-28 05:42:29,016][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000
+[2026-07-28 09:47:03,050][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21'
+[2026-07-28 09:47:03,050][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 09:47:03,051][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 09:47:03,271][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 09:47:03,271][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 09:47:03,326][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 09:47:03,326][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 09:47:03,887][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 09:47:03,902][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 09:47:04,040][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 09:47:04,055][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 09:47:06,482][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 09:47:06,482][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 09:47:07,470][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 09:47:07,471][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 09:47:07,471][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 09:47:07,471][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 09:47:07,471][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 09:47:07,472][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 09:47:07,472][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 09:47:07,472][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 09:47:07,472][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 09:47:07,472][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 09:47:07,473][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 09:47:07,473][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 09:47:07,473][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 09:47:07,475][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 09:47:07,475][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 09:47:07,475][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 09:47:07,475][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 09:47:07,476][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 09:47:07,476][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 09:47:07,476][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 09:47:07,476][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 09:47:07,476][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 09:47:07,477][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 09:47:07,477][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 09:47:07,477][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 09:47:07,477][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 09:47:07,477][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 09:47:07,477][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 09:47:07,477][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 09:47:07,478][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 09:47:07,478][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 09:47:07,478][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 09:47:07,478][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 09:47:07,478][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 09:47:07,478][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 09:47:07,479][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 09:47:07,479][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 09:47:07,479][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 09:47:07,479][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 09:47:07,479][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 09:47:07,479][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 09:47:07,480][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 09:47:07,480][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 09:47:07,480][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 09:47:07,481][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 09:47:07,481][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 09:47:07,481][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 09:47:07,482][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 09:47:07,482][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 09:47:07,482][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 09:47:07,482][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 09:47:07,483][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 09:47:07,483][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 09:47:07,483][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 09:47:07,483][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 09:47:07,483][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 09:47:07,484][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 09:47:07,484][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 09:47:07,484][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 09:47:07,485][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 09:47:07,485][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 09:47:07,485][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 09:47:08,930][starVLA.dataloader][INFO] - Created train VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 09:47:08,933][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 09:47:09,034][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 09:47:09,052][starVLA.dataloader][INFO] - Created eval VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 09:47:09,059][starVLA.training.train_starvla][INFO] - LR Group qwen_vl_interface: lr=1e-05, num_params=713
+[2026-07-28 09:47:09,059][starVLA.training.train_starvla][INFO] - LR Group action_model: lr=0.0001, num_params=16
+[2026-07-28 09:47:10,364][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1500_state, steps: 1500
+[2026-07-28 09:47:10,375][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/config.full.yaml
+[2026-07-28 09:47:10,388][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/config.yaml
+[2026-07-28 09:47:10,389][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1500_state, steps: 1500
+[2026-07-28 09:47:10,394][accelerate.accelerator][INFO] - Since you passed both train and evaluation dataloader, `is_train_batch_min` (here True will decide the `train_batch_size` (8).
+[2026-07-28 09:47:32,171][accelerate.accelerator][INFO] - Loading states from /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1500_state
+[2026-07-28 09:47:32,171][accelerate.accelerator][INFO] - Loading DeepSpeed Model and Optimizer
+[2026-07-28 09:48:08,537][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer loaded from input dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_1500_state/pytorch_model
+[2026-07-28 09:48:08,537][accelerate.checkpointing][INFO] - All model weights loaded successfully
+[2026-07-28 09:48:08,537][accelerate.checkpointing][INFO] - All optimizer states loaded successfully
+[2026-07-28 09:48:08,537][accelerate.checkpointing][INFO] - All scheduler states loaded successfully
+[2026-07-28 09:48:08,537][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully
+[2026-07-28 09:48:08,546][accelerate.checkpointing][INFO] - All random states loaded successfully
+[2026-07-28 09:48:08,548][accelerate.accelerator][INFO] - Loading in 0 custom states
+[2026-07-28 09:48:11,838][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-07-28 09:48:11,839][starVLA.training.train_starvla][INFO] -   Total optimization steps = 4000
+[2026-07-28 09:48:11,839][starVLA.training.train_starvla][INFO] -   Per device batch size = 8
+[2026-07-28 09:48:11,839][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-07-28 09:48:11,839][starVLA.training.train_starvla][INFO] -   Total batch size = 256
+[2026-07-28 09:49:28,701][starVLA.training.train_starvla][INFO] - Step 1501, Loss: {'train/loss': 0.06581839919090271, 'train/grad_norm_pre_clip': 3.0759332180023193, 'timing/data': 0.0003394479863345623, 'timing/model': 0.6211512349545956, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.28})
+[2026-07-28 09:49:32,932][starVLA.training.train_starvla][INFO] - Step 1502, Loss: {'train/loss': 0.12610770761966705, 'train/grad_norm_pre_clip': 4.718153953552246, 'timing/data': 0.0004136413335800171, 'timing/model': 0.4062236831523478, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.28})
+[2026-07-28 09:49:37,070][starVLA.training.train_starvla][INFO] - Step 1503, Loss: {'train/loss': 0.09604956209659576, 'train/grad_norm_pre_clip': 3.8976972103118896, 'timing/data': 0.00031810393556952477, 'timing/model': 0.3361722696572542, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.28})
+[2026-07-28 09:49:41,192][starVLA.training.train_starvla][INFO] - Step 1504, Loss: {'train/loss': 0.04527005925774574, 'train/grad_norm_pre_clip': 1.3976354598999023, 'timing/data': 0.0005949740298092365, 'timing/model': 0.33644790668040514, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.28})
+[2026-07-28 09:49:45,313][starVLA.training.train_starvla][INFO] - Step 1505, Loss: {'train/loss': 0.06375034898519516, 'train/grad_norm_pre_clip': 4.82782506942749, 'timing/data': 0.0006339834071695805, 'timing/model': 0.3383883167989552, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.28})
+[2026-07-28 09:49:49,472][starVLA.training.train_starvla][INFO] - Step 1506, Loss: {'train/loss': 0.0661240890622139, 'train/grad_norm_pre_clip': 3.1549298763275146, 'timing/data': 0.00033055664971470833, 'timing/model': 0.3395016468130052, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.28})
+[2026-07-28 09:49:53,607][starVLA.training.train_starvla][INFO] - Step 1507, Loss: {'train/loss': 0.07352152466773987, 'train/grad_norm_pre_clip': 4.008698463439941, 'timing/data': 0.00030609918758273125, 'timing/model': 0.3354838867671788, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.28})
+[2026-07-28 09:49:57,730][starVLA.training.train_starvla][INFO] - Step 1508, Loss: {'train/loss': 0.03699466586112976, 'train/grad_norm_pre_clip': 0.8003612160682678, 'timing/data': 0.00038119684904813766, 'timing/model': 0.3376787598244846, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.28})
+[2026-07-28 09:50:01,878][starVLA.training.train_starvla][INFO] - Step 1509, Loss: {'train/loss': 0.05844513699412346, 'train/grad_norm_pre_clip': 2.4067788124084473, 'timing/data': 0.0004208143800497055, 'timing/model': 0.3380044843070209, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.28})
+[2026-07-28 09:50:06,013][starVLA.training.train_starvla][INFO] - Step 1510, Loss: {'train/loss': 0.08981245011091232, 'train/grad_norm_pre_clip': 2.307891368865967, 'timing/data': 0.0003117639571428299, 'timing/model': 0.3416771930642426, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.28})
+[2026-07-28 09:50:10,219][starVLA.training.train_starvla][INFO] - Step 1511, Loss: {'train/loss': 0.06977848708629608, 'train/grad_norm_pre_clip': 1.465651273727417, 'timing/data': 0.0003205430693924427, 'timing/model': 0.34062711615115404, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.28})
+[2026-07-28 09:50:14,398][starVLA.training.train_starvla][INFO] - Step 1512, Loss: {'train/loss': 0.08846008032560349, 'train/grad_norm_pre_clip': 3.437096118927002, 'timing/data': 0.0004986696876585484, 'timing/model': 0.3426820929162204, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.28})
+[2026-07-28 09:50:18,582][starVLA.training.train_starvla][INFO] - Step 1513, Loss: {'train/loss': 0.04838192090392113, 'train/grad_norm_pre_clip': 1.5520071983337402, 'timing/data': 0.00035549793392419815, 'timing/model': 0.33795344457030296, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.28})
+[2026-07-28 09:50:22,714][starVLA.training.train_starvla][INFO] - Step 1514, Loss: {'train/loss': 0.08870628476142883, 'train/grad_norm_pre_clip': 2.220034599304199, 'timing/data': 0.0003826860338449478, 'timing/model': 0.3350639659911394, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.28})
+[2026-07-28 09:50:26,895][starVLA.training.train_starvla][INFO] - Step 1515, Loss: {'train/loss': 0.06726101040840149, 'train/grad_norm_pre_clip': 1.8390017747879028, 'timing/data': 0.00035683391615748405, 'timing/model': 0.33801872562617064, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.28})
+[2026-07-28 09:50:31,075][starVLA.training.train_starvla][INFO] - Step 1516, Loss: {'train/loss': 0.07292984426021576, 'train/grad_norm_pre_clip': 3.5007810592651367, 'timing/data': 0.0003639138303697109, 'timing/model': 0.33567997021600604, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.28})
+[2026-07-28 09:50:35,205][starVLA.training.train_starvla][INFO] - Step 1517, Loss: {'train/loss': 0.0701533779501915, 'train/grad_norm_pre_clip': 1.2677946090698242, 'timing/data': 0.00043629109859466553, 'timing/model': 0.33658961532637477, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.28})
+[2026-07-28 09:50:39,306][starVLA.training.train_starvla][INFO] - Step 1518, Loss: {'train/loss': 0.06979811936616898, 'train/grad_norm_pre_clip': 1.5128767490386963, 'timing/data': 0.00031410902738571167, 'timing/model': 0.3369456841610372, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.28})
+[2026-07-28 09:50:43,452][starVLA.training.train_starvla][INFO] - Step 1519, Loss: {'train/loss': 0.04696724936366081, 'train/grad_norm_pre_clip': 1.073233723640442, 'timing/data': 0.00038168998435139656, 'timing/model': 0.3383759153075516, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.28})
+[2026-07-28 09:50:47,645][starVLA.training.train_starvla][INFO] - Step 1520, Loss: {'train/loss': 0.0666566714644432, 'train/grad_norm_pre_clip': 2.26951265335083, 'timing/data': 0.00033726124092936516, 'timing/model': 0.33811436500400305, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.28})
+[2026-07-28 09:50:51,849][starVLA.training.train_starvla][INFO] - Step 1521, Loss: {'train/loss': 0.06956689804792404, 'train/grad_norm_pre_clip': 2.3978381156921387, 'timing/data': 0.0003035827539861202, 'timing/model': 0.3415666692890227, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.28})
+[2026-07-28 09:50:55,974][starVLA.training.train_starvla][INFO] - Step 1522, Loss: {'train/loss': 0.03997790440917015, 'train/grad_norm_pre_clip': 0.8740379214286804, 'timing/data': 0.000451086089015007, 'timing/model': 0.33704084204509854, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.28})
+[2026-07-28 09:51:00,116][starVLA.training.train_starvla][INFO] - Step 1523, Loss: {'train/loss': 0.03571376949548721, 'train/grad_norm_pre_clip': 0.8863252997398376, 'timing/data': 0.000328725203871727, 'timing/model': 0.33587412210181355, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.28})
+[2026-07-28 09:51:04,220][starVLA.training.train_starvla][INFO] - Step 1524, Loss: {'train/loss': 0.03738900646567345, 'train/grad_norm_pre_clip': 0.7412291765213013, 'timing/data': 0.0003210371360182762, 'timing/model': 0.33658650889992714, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.28})
+[2026-07-28 09:51:08,336][starVLA.training.train_starvla][INFO] - Step 1525, Loss: {'train/loss': 0.06052695959806442, 'train/grad_norm_pre_clip': 1.1644635200500488, 'timing/data': 0.0004731849767267704, 'timing/model': 0.3354836283251643, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.28})
+[2026-07-28 09:51:12,439][starVLA.training.train_starvla][INFO] - Step 1526, Loss: {'train/loss': 0.052580878138542175, 'train/grad_norm_pre_clip': 1.0311943292617798, 'timing/data': 0.00030092382803559303, 'timing/model': 0.3352732649073005, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.28})
+[2026-07-28 09:51:16,577][starVLA.training.train_starvla][INFO] - Step 1527, Loss: {'train/loss': 0.05973692983388901, 'train/grad_norm_pre_clip': 2.687774658203125, 'timing/data': 0.00035178009420633316, 'timing/model': 0.33518000692129135, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.28})
+[2026-07-28 09:51:20,677][starVLA.training.train_starvla][INFO] - Step 1528, Loss: {'train/loss': 0.0633615255355835, 'train/grad_norm_pre_clip': 1.8703056573867798, 'timing/data': 0.00029273703694343567, 'timing/model': 0.33499817084521055, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.28})
+[2026-07-28 09:51:24,789][starVLA.training.train_starvla][INFO] - Step 1529, Loss: {'train/loss': 0.04946477711200714, 'train/grad_norm_pre_clip': 0.9827624559402466, 'timing/data': 0.0003358260728418827, 'timing/model': 0.3347297008149326, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.28})
+[2026-07-28 09:51:28,932][starVLA.training.train_starvla][INFO] - Step 1530, Loss: {'train/loss': 0.07717248052358627, 'train/grad_norm_pre_clip': 2.3511948585510254, 'timing/data': 0.00029762182384729385, 'timing/model': 0.3356285560876131, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.28})
+[2026-07-28 09:51:33,048][starVLA.training.train_starvla][INFO] - Step 1531, Loss: {'train/loss': 0.023677324876189232, 'train/grad_norm_pre_clip': 0.6521878838539124, 'timing/data': 0.0008850009180605412, 'timing/model': 0.33421371784061193, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.28})
+[2026-07-28 09:51:37,146][starVLA.training.train_starvla][INFO] - Step 1532, Loss: {'train/loss': 0.05699867010116577, 'train/grad_norm_pre_clip': 1.4326454401016235, 'timing/data': 0.0003701988607645035, 'timing/model': 0.3361285738646984, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.28})
+[2026-07-28 09:51:41,603][starVLA.training.train_starvla][INFO] - Step 1533, Loss: {'train/loss': 0.06131661683320999, 'train/grad_norm_pre_clip': 2.001371145248413, 'timing/data': 0.0003253878094255924, 'timing/model': 0.33726928709074855, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.28})
+[2026-07-28 09:51:46,151][starVLA.training.train_starvla][INFO] - Step 1534, Loss: {'train/loss': 0.04855230450630188, 'train/grad_norm_pre_clip': 2.4718942642211914, 'timing/data': 0.0003183279186487198, 'timing/model': 0.33480245200917125, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.28})
+[2026-07-28 09:51:50,271][starVLA.training.train_starvla][INFO] - Step 1535, Loss: {'train/loss': 0.03706004098057747, 'train/grad_norm_pre_clip': 0.5007351636886597, 'timing/data': 0.0003221677616238594, 'timing/model': 0.33692410215735435, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.28})
+[2026-07-28 09:51:54,402][starVLA.training.train_starvla][INFO] - Step 1536, Loss: {'train/loss': 0.04641889035701752, 'train/grad_norm_pre_clip': 0.9314345717430115, 'timing/data': 0.00034405989572405815, 'timing/model': 0.33727246802300215, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.28})
+[2026-07-28 09:51:58,577][starVLA.training.train_starvla][INFO] - Step 1537, Loss: {'train/loss': 0.02983001060783863, 'train/grad_norm_pre_clip': 0.9302554726600647, 'timing/data': 0.0003353608772158623, 'timing/model': 0.34208967676386237, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.28})
+[2026-07-28 09:52:02,722][starVLA.training.train_starvla][INFO] - Step 1538, Loss: {'train/loss': 0.03231426700949669, 'train/grad_norm_pre_clip': 1.1211482286453247, 'timing/data': 0.0003871796652674675, 'timing/model': 0.3478367826901376, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.28})
+[2026-07-28 09:52:06,859][starVLA.training.train_starvla][INFO] - Step 1539, Loss: {'train/loss': 0.08921751379966736, 'train/grad_norm_pre_clip': 3.6101841926574707, 'timing/data': 0.0003377450630068779, 'timing/model': 0.33834572695195675, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.28})
+[2026-07-28 09:52:11,023][starVLA.training.train_starvla][INFO] - Step 1540, Loss: {'train/loss': 0.036657433956861496, 'train/grad_norm_pre_clip': 1.1368643045425415, 'timing/data': 0.00032360898330807686, 'timing/model': 0.3417033995501697, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.28})
+[2026-07-28 09:52:15,196][starVLA.training.train_starvla][INFO] - Step 1541, Loss: {'train/loss': 0.034542158246040344, 'train/grad_norm_pre_clip': 1.3895379304885864, 'timing/data': 0.0006927149370312691, 'timing/model': 0.3426183219999075, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.28})
+[2026-07-28 09:52:19,356][starVLA.training.train_starvla][INFO] - Step 1542, Loss: {'train/loss': 0.07887222617864609, 'train/grad_norm_pre_clip': 2.038285732269287, 'timing/data': 0.0003257910721004009, 'timing/model': 0.34342149505391717, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.28})
+[2026-07-28 09:52:23,484][starVLA.training.train_starvla][INFO] - Step 1543, Loss: {'train/loss': 0.049187928438186646, 'train/grad_norm_pre_clip': 2.0081751346588135, 'timing/data': 0.00039793504402041435, 'timing/model': 0.34988473216071725, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.28})
+[2026-07-28 09:52:27,652][starVLA.training.train_starvla][INFO] - Step 1544, Loss: {'train/loss': 0.027955859899520874, 'train/grad_norm_pre_clip': 1.1150667667388916, 'timing/data': 0.00041485903784632683, 'timing/model': 0.3369434238411486, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.28})
+[2026-07-28 09:52:31,773][starVLA.training.train_starvla][INFO] - Step 1545, Loss: {'train/loss': 0.03526880964636803, 'train/grad_norm_pre_clip': 0.9686033725738525, 'timing/data': 0.00032620271667838097, 'timing/model': 0.3356015426106751, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.28})
+[2026-07-28 09:52:35,908][starVLA.training.train_starvla][INFO] - Step 1546, Loss: {'train/loss': 0.037549421191215515, 'train/grad_norm_pre_clip': 1.3555303812026978, 'timing/data': 0.0004239198751747608, 'timing/model': 0.3339578532613814, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.28})
+[2026-07-28 09:52:40,017][starVLA.training.train_starvla][INFO] - Step 1547, Loss: {'train/loss': 0.10740066319704056, 'train/grad_norm_pre_clip': 2.424168348312378, 'timing/data': 0.0004193936474621296, 'timing/model': 0.3357481458224356, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.28})
+[2026-07-28 09:52:44,129][starVLA.training.train_starvla][INFO] - Step 1548, Loss: {'train/loss': 0.05696217715740204, 'train/grad_norm_pre_clip': 2.639691114425659, 'timing/data': 0.00035615498200058937, 'timing/model': 0.3421643590554595, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.28})
+[2026-07-28 09:52:48,293][starVLA.training.train_starvla][INFO] - Step 1549, Loss: {'train/loss': 0.059478726238012314, 'train/grad_norm_pre_clip': 2.3533451557159424, 'timing/data': 0.00044941529631614685, 'timing/model': 0.33587705064564943, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.28})
+[2026-07-28 09:52:52,453][starVLA.training.train_starvla][INFO] - Step 1550, Loss: {'train/loss': 0.03402240201830864, 'train/grad_norm_pre_clip': 0.9785030484199524, 'timing/data': 0.0004055988974869251, 'timing/model': 0.34957247925922275, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.28})
+[2026-07-28 09:52:56,710][starVLA.training.train_starvla][INFO] - Step 1551, Loss: {'train/loss': 0.09937360882759094, 'train/grad_norm_pre_clip': 1.8658454418182373, 'timing/data': 0.0005138539709150791, 'timing/model': 0.3513728207908571, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.28})
+[2026-07-28 09:53:00,907][starVLA.training.train_starvla][INFO] - Step 1552, Loss: {'train/loss': 0.03383452445268631, 'train/grad_norm_pre_clip': 1.065634846687317, 'timing/data': 0.00032293377444148064, 'timing/model': 0.3357049976475537, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.29})
+[2026-07-28 09:53:05,037][starVLA.training.train_starvla][INFO] - Step 1553, Loss: {'train/loss': 0.028812622651457787, 'train/grad_norm_pre_clip': 1.7189840078353882, 'timing/data': 0.0003471840173006058, 'timing/model': 0.33517503878101707, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:53:09,204][starVLA.training.train_starvla][INFO] - Step 1554, Loss: {'train/loss': 0.08831611275672913, 'train/grad_norm_pre_clip': 2.708435535430908, 'timing/data': 0.0003403369337320328, 'timing/model': 0.3422982031479478, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.29})
+[2026-07-28 09:53:13,349][starVLA.training.train_starvla][INFO] - Step 1555, Loss: {'train/loss': 0.042614806443452835, 'train/grad_norm_pre_clip': 1.0207722187042236, 'timing/data': 0.00036125490441918373, 'timing/model': 0.334933347068727, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:53:17,550][starVLA.training.train_starvla][INFO] - Step 1556, Loss: {'train/loss': 0.03494100645184517, 'train/grad_norm_pre_clip': 0.9392683506011963, 'timing/data': 0.00033767567947506905, 'timing/model': 0.35067385202273726, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.29})
+[2026-07-28 09:53:21,706][starVLA.training.train_starvla][INFO] - Step 1557, Loss: {'train/loss': 0.03704797104001045, 'train/grad_norm_pre_clip': 1.3116159439086914, 'timing/data': 0.0005782400257885456, 'timing/model': 0.33880959590896964, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.29})
+[2026-07-28 09:53:25,901][starVLA.training.train_starvla][INFO] - Step 1558, Loss: {'train/loss': 0.07212573289871216, 'train/grad_norm_pre_clip': 1.8920491933822632, 'timing/data': 0.0008618929423391819, 'timing/model': 0.3442916879430413, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.29})
+[2026-07-28 09:53:30,011][starVLA.training.train_starvla][INFO] - Step 1559, Loss: {'train/loss': 0.06275202333927155, 'train/grad_norm_pre_clip': 1.7393252849578857, 'timing/data': 0.0004998561926186085, 'timing/model': 0.3356623249128461, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.29})
+[2026-07-28 09:53:34,122][starVLA.training.train_starvla][INFO] - Step 1560, Loss: {'train/loss': 0.029533274471759796, 'train/grad_norm_pre_clip': 1.3428418636322021, 'timing/data': 0.00037809694185853004, 'timing/model': 0.3344403519295156, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.29})
+[2026-07-28 09:53:38,254][starVLA.training.train_starvla][INFO] - Step 1561, Loss: {'train/loss': 0.0428466722369194, 'train/grad_norm_pre_clip': 1.1358305215835571, 'timing/data': 0.0003157821483910084, 'timing/model': 0.33661878714337945, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.29})
+[2026-07-28 09:53:42,389][starVLA.training.train_starvla][INFO] - Step 1562, Loss: {'train/loss': 0.027024973183870316, 'train/grad_norm_pre_clip': 0.9943414926528931, 'timing/data': 0.0003272928297519684, 'timing/model': 0.33603802090510726, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.29})
+[2026-07-28 09:53:46,501][starVLA.training.train_starvla][INFO] - Step 1563, Loss: {'train/loss': 0.019362006336450577, 'train/grad_norm_pre_clip': 1.2016314268112183, 'timing/data': 0.0003353869542479515, 'timing/model': 0.3361124908551574, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.29})
+[2026-07-28 09:53:50,621][starVLA.training.train_starvla][INFO] - Step 1564, Loss: {'train/loss': 0.04752257838845253, 'train/grad_norm_pre_clip': 1.2044360637664795, 'timing/data': 0.0003062239848077297, 'timing/model': 0.3540836749598384, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:53:54,725][starVLA.training.train_starvla][INFO] - Step 1565, Loss: {'train/loss': 0.04187851771712303, 'train/grad_norm_pre_clip': 3.176820993423462, 'timing/data': 0.0003241291269659996, 'timing/model': 0.33480508998036385, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:53:58,833][starVLA.training.train_starvla][INFO] - Step 1566, Loss: {'train/loss': 0.0499550998210907, 'train/grad_norm_pre_clip': 1.6294844150543213, 'timing/data': 0.00035419920459389687, 'timing/model': 0.3414230393245816, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.29})
+[2026-07-28 09:54:02,930][starVLA.training.train_starvla][INFO] - Step 1567, Loss: {'train/loss': 0.030004149302840233, 'train/grad_norm_pre_clip': 1.342077374458313, 'timing/data': 0.0003060596063733101, 'timing/model': 0.3343038922175765, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.29})
+[2026-07-28 09:54:07,036][starVLA.training.train_starvla][INFO] - Step 1568, Loss: {'train/loss': 0.043481145054101944, 'train/grad_norm_pre_clip': 2.1348185539245605, 'timing/data': 0.0003121723420917988, 'timing/model': 0.33618179708719254, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:54:11,153][starVLA.training.train_starvla][INFO] - Step 1569, Loss: {'train/loss': 0.04039379209280014, 'train/grad_norm_pre_clip': 1.960395336151123, 'timing/data': 0.0003412468358874321, 'timing/model': 0.3352003856562078, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.29})
+[2026-07-28 09:54:15,264][starVLA.training.train_starvla][INFO] - Step 1570, Loss: {'train/loss': 0.05438990890979767, 'train/grad_norm_pre_clip': 1.1944881677627563, 'timing/data': 0.0004151808097958565, 'timing/model': 0.3391783209517598, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.29})
+[2026-07-28 09:54:19,399][starVLA.training.train_starvla][INFO] - Step 1571, Loss: {'train/loss': 0.058786541223526, 'train/grad_norm_pre_clip': 2.1277713775634766, 'timing/data': 0.00033782003447413445, 'timing/model': 0.33706239098683, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.29})
+[2026-07-28 09:54:23,533][starVLA.training.train_starvla][INFO] - Step 1572, Loss: {'train/loss': 0.09307167679071426, 'train/grad_norm_pre_clip': 1.853750467300415, 'timing/data': 0.00041587697342038155, 'timing/model': 0.33530894899740815, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.29})
+[2026-07-28 09:54:27,644][starVLA.training.train_starvla][INFO] - Step 1573, Loss: {'train/loss': 0.07981596142053604, 'train/grad_norm_pre_clip': 2.753106117248535, 'timing/data': 0.00031113578006625175, 'timing/model': 0.334975210018456, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.29})
+[2026-07-28 09:54:31,753][starVLA.training.train_starvla][INFO] - Step 1574, Loss: {'train/loss': 0.05465936288237572, 'train/grad_norm_pre_clip': 0.8353221416473389, 'timing/data': 0.00033042626455426216, 'timing/model': 0.33785998495295644, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.29})
+[2026-07-28 09:54:35,888][starVLA.training.train_starvla][INFO] - Step 1575, Loss: {'train/loss': 0.05050530284643173, 'train/grad_norm_pre_clip': 1.9035533666610718, 'timing/data': 0.0002864100970327854, 'timing/model': 0.3348737535998225, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:54:40,300][starVLA.training.train_starvla][INFO] - Step 1576, Loss: {'train/loss': 0.06414572894573212, 'train/grad_norm_pre_clip': 2.1362521648406982, 'timing/data': 0.00037058815360069275, 'timing/model': 0.3361071618273854, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.29})
+[2026-07-28 09:54:44,402][starVLA.training.train_starvla][INFO] - Step 1577, Loss: {'train/loss': 0.060840904712677, 'train/grad_norm_pre_clip': 1.5967577695846558, 'timing/data': 0.0004340512678027153, 'timing/model': 0.33614528831094503, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.29})
+[2026-07-28 09:54:48,510][starVLA.training.train_starvla][INFO] - Step 1578, Loss: {'train/loss': 0.06680525839328766, 'train/grad_norm_pre_clip': 2.436948537826538, 'timing/data': 0.00033124303445219994, 'timing/model': 0.33469644095748663, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:54:52,615][starVLA.training.train_starvla][INFO] - Step 1579, Loss: {'train/loss': 0.0484321191906929, 'train/grad_norm_pre_clip': 1.8357902765274048, 'timing/data': 0.00029915524646639824, 'timing/model': 0.3363490430638194, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:54:56,717][starVLA.training.train_starvla][INFO] - Step 1580, Loss: {'train/loss': 0.08454199135303497, 'train/grad_norm_pre_clip': 1.5930719375610352, 'timing/data': 0.00035220710560679436, 'timing/model': 0.33623525919392705, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.29})
+[2026-07-28 09:55:00,815][starVLA.training.train_starvla][INFO] - Step 1581, Loss: {'train/loss': 0.05956430733203888, 'train/grad_norm_pre_clip': 1.7206850051879883, 'timing/data': 0.0003449888899922371, 'timing/model': 0.3334662262350321, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.29})
+[2026-07-28 09:55:05,207][starVLA.training.train_starvla][INFO] - Step 1582, Loss: {'train/loss': 0.05005134642124176, 'train/grad_norm_pre_clip': 1.5985727310180664, 'timing/data': 0.00046653300523757935, 'timing/model': 0.33548207534477115, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.29})
+[2026-07-28 09:55:09,299][starVLA.training.train_starvla][INFO] - Step 1583, Loss: {'train/loss': 0.033313654363155365, 'train/grad_norm_pre_clip': 1.0708520412445068, 'timing/data': 0.0004182159900665283, 'timing/model': 0.3340191999450326, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.29})
+[2026-07-28 09:55:13,393][starVLA.training.train_starvla][INFO] - Step 1584, Loss: {'train/loss': 0.03846076503396034, 'train/grad_norm_pre_clip': 1.2364866733551025, 'timing/data': 0.0003574620932340622, 'timing/model': 0.3368410300463438, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.29})
+[2026-07-28 09:55:17,630][starVLA.training.train_starvla][INFO] - Step 1585, Loss: {'train/loss': 0.03299878537654877, 'train/grad_norm_pre_clip': 1.0976098775863647, 'timing/data': 0.0009063943289220333, 'timing/model': 0.3601177781820297, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.29})
+[2026-07-28 09:55:21,791][starVLA.training.train_starvla][INFO] - Step 1586, Loss: {'train/loss': 0.05866820365190506, 'train/grad_norm_pre_clip': 1.6407580375671387, 'timing/data': 0.00035979971289634705, 'timing/model': 0.34572299290448427, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.29})
+[2026-07-28 09:55:25,914][starVLA.training.train_starvla][INFO] - Step 1587, Loss: {'train/loss': 0.06156563758850098, 'train/grad_norm_pre_clip': 1.296809434890747, 'timing/data': 0.0003568851388990879, 'timing/model': 0.3360122977755964, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.29})
+[2026-07-28 09:55:30,105][starVLA.training.train_starvla][INFO] - Step 1588, Loss: {'train/loss': 0.04520469903945923, 'train/grad_norm_pre_clip': 1.5186913013458252, 'timing/data': 0.00031236279755830765, 'timing/model': 0.33479355508461595, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:55:34,243][starVLA.training.train_starvla][INFO] - Step 1589, Loss: {'train/loss': 0.03488900512456894, 'train/grad_norm_pre_clip': 0.8958585858345032, 'timing/data': 0.00029610609635710716, 'timing/model': 0.33446358423680067, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:55:38,413][starVLA.training.train_starvla][INFO] - Step 1590, Loss: {'train/loss': 0.05048779398202896, 'train/grad_norm_pre_clip': 1.409719467163086, 'timing/data': 0.0003273370675742626, 'timing/model': 0.3347364068031311, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.29})
+[2026-07-28 09:55:42,540][starVLA.training.train_starvla][INFO] - Step 1591, Loss: {'train/loss': 0.06781542301177979, 'train/grad_norm_pre_clip': 1.5109331607818604, 'timing/data': 0.0003303680568933487, 'timing/model': 0.34087059600278735, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.29})
+[2026-07-28 09:55:46,648][starVLA.training.train_starvla][INFO] - Step 1592, Loss: {'train/loss': 0.047776706516742706, 'train/grad_norm_pre_clip': 1.9419419765472412, 'timing/data': 0.00030939094722270966, 'timing/model': 0.3370218020863831, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:55:50,849][starVLA.training.train_starvla][INFO] - Step 1593, Loss: {'train/loss': 0.10078003257513046, 'train/grad_norm_pre_clip': 3.2586374282836914, 'timing/data': 0.0003626421093940735, 'timing/model': 0.39968616981059313, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:55:55,075][starVLA.training.train_starvla][INFO] - Step 1594, Loss: {'train/loss': 0.07031261175870895, 'train/grad_norm_pre_clip': 1.2391583919525146, 'timing/data': 0.0003207651898264885, 'timing/model': 0.3414034456945956, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.29})
+[2026-07-28 09:55:59,206][starVLA.training.train_starvla][INFO] - Step 1595, Loss: {'train/loss': 0.07853007316589355, 'train/grad_norm_pre_clip': 1.9656040668487549, 'timing/data': 0.00034264521673321724, 'timing/model': 0.33472482999786735, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.29})
+[2026-07-28 09:56:03,354][starVLA.training.train_starvla][INFO] - Step 1596, Loss: {'train/loss': 0.054921552538871765, 'train/grad_norm_pre_clip': 2.7829151153564453, 'timing/data': 0.00032007088884711266, 'timing/model': 0.33476644894108176, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.29})
+[2026-07-28 09:56:07,491][starVLA.training.train_starvla][INFO] - Step 1597, Loss: {'train/loss': 0.05460534989833832, 'train/grad_norm_pre_clip': 2.5023980140686035, 'timing/data': 0.0003859270364046097, 'timing/model': 0.3385229827836156, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.29})
+[2026-07-28 09:56:11,625][starVLA.training.train_starvla][INFO] - Step 1598, Loss: {'train/loss': 0.0599864237010479, 'train/grad_norm_pre_clip': 2.3753504753112793, 'timing/data': 0.0003822711296379566, 'timing/model': 0.34135229885578156, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.29})
+[2026-07-28 09:56:15,846][starVLA.training.train_starvla][INFO] - Step 1599, Loss: {'train/loss': 0.058329299092292786, 'train/grad_norm_pre_clip': 2.2959632873535156, 'timing/data': 0.00039719603955745697, 'timing/model': 0.3459716970100999, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.29})
+[2026-07-28 09:56:19,989][starVLA.training.train_starvla][INFO] - Step 1600, Loss: {'train/loss': 0.024612147361040115, 'train/grad_norm_pre_clip': 0.662036120891571, 'timing/data': 0.00038343388587236404, 'timing/model': 0.3344673439860344, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.29})
+[2026-07-28 09:56:24,094][starVLA.training.train_starvla][INFO] - Step 1601, Loss: {'train/loss': 0.03889203816652298, 'train/grad_norm_pre_clip': 1.6550277471542358, 'timing/data': 0.00031847693026065826, 'timing/model': 0.3364651841111481, 'learning_rate/qwen_vl_interface': 9.999998458888285e-06, 'learning_rate/action_model': 9.999998458888285e-05, 'epoch': 0.29})
+[2026-07-28 09:56:28,197][starVLA.training.train_starvla][INFO] - Step 1602, Loss: {'train/loss': 0.07330909371376038, 'train/grad_norm_pre_clip': 1.6101064682006836, 'timing/data': 0.0003555379807949066, 'timing/model': 0.33530471893027425, 'learning_rate/qwen_vl_interface': 9.999993835554141e-06, 'learning_rate/action_model': 9.999993835554141e-05, 'epoch': 0.29})
+[2026-07-28 09:56:32,305][starVLA.training.train_starvla][INFO] - Step 1603, Loss: {'train/loss': 0.07031182944774628, 'train/grad_norm_pre_clip': 1.679451823234558, 'timing/data': 0.0003590090200304985, 'timing/model': 0.33486407715827227, 'learning_rate/qwen_vl_interface': 9.999986130000566e-06, 'learning_rate/action_model': 9.999986130000566e-05, 'epoch': 0.29})
+[2026-07-28 09:56:36,410][starVLA.training.train_starvla][INFO] - Step 1604, Loss: {'train/loss': 0.050094522535800934, 'train/grad_norm_pre_clip': 0.9888457655906677, 'timing/data': 0.0003330460749566555, 'timing/model': 0.33460339810699224, 'learning_rate/qwen_vl_interface': 9.99997534223256e-06, 'learning_rate/action_model': 9.99997534223256e-05, 'epoch': 0.29})
+[2026-07-28 09:56:40,513][starVLA.training.train_starvla][INFO] - Step 1605, Loss: {'train/loss': 0.049270953983068466, 'train/grad_norm_pre_clip': 0.7612764239311218, 'timing/data': 0.0003562602214515209, 'timing/model': 0.3338806270621717, 'learning_rate/qwen_vl_interface': 9.999961472257127e-06, 'learning_rate/action_model': 9.999961472257127e-05, 'epoch': 0.29})
+[2026-07-28 09:56:44,616][starVLA.training.train_starvla][INFO] - Step 1606, Loss: {'train/loss': 0.07561241835355759, 'train/grad_norm_pre_clip': 1.4292665719985962, 'timing/data': 0.00037156790494918823, 'timing/model': 0.3350667837075889, 'learning_rate/qwen_vl_interface': 9.999944520083262e-06, 'learning_rate/action_model': 9.999944520083261e-05, 'epoch': 0.29})
+[2026-07-28 09:56:48,737][starVLA.training.train_starvla][INFO] - Step 1607, Loss: {'train/loss': 0.0292527936398983, 'train/grad_norm_pre_clip': 0.7017472386360168, 'timing/data': 0.0004543946124613285, 'timing/model': 0.3358420100994408, 'learning_rate/qwen_vl_interface': 9.999924485721968e-06, 'learning_rate/action_model': 9.999924485721968e-05, 'epoch': 0.3})
+[2026-07-28 09:56:52,844][starVLA.training.train_starvla][INFO] - Step 1608, Loss: {'train/loss': 0.05398308113217354, 'train/grad_norm_pre_clip': 2.5665414333343506, 'timing/data': 0.0003145388327538967, 'timing/model': 0.33542371029034257, 'learning_rate/qwen_vl_interface': 9.999901369186244e-06, 'learning_rate/action_model': 9.999901369186244e-05, 'epoch': 0.3})
+[2026-07-28 09:56:56,955][starVLA.training.train_starvla][INFO] - Step 1609, Loss: {'train/loss': 0.051274579018354416, 'train/grad_norm_pre_clip': 2.0103001594543457, 'timing/data': 0.00030623702332377434, 'timing/model': 0.33616177598014474, 'learning_rate/qwen_vl_interface': 9.999875170491092e-06, 'learning_rate/action_model': 9.999875170491091e-05, 'epoch': 0.3})
+[2026-07-28 09:57:01,076][starVLA.training.train_starvla][INFO] - Step 1610, Loss: {'train/loss': 0.07529168576002121, 'train/grad_norm_pre_clip': 2.27913236618042, 'timing/data': 0.0003649592399597168, 'timing/model': 0.3350838446058333, 'learning_rate/qwen_vl_interface': 9.999845889653509e-06, 'learning_rate/action_model': 9.999845889653508e-05, 'epoch': 0.3})
+[2026-07-28 09:57:05,182][starVLA.training.train_starvla][INFO] - Step 1611, Loss: {'train/loss': 0.06789974123239517, 'train/grad_norm_pre_clip': 1.88502037525177, 'timing/data': 0.0002949167974293232, 'timing/model': 0.33603284833952785, 'learning_rate/qwen_vl_interface': 9.999813526692496e-06, 'learning_rate/action_model': 9.999813526692495e-05, 'epoch': 0.3})
+[2026-07-28 09:57:09,290][starVLA.training.train_starvla][INFO] - Step 1612, Loss: {'train/loss': 0.025170784443616867, 'train/grad_norm_pre_clip': 0.8835853934288025, 'timing/data': 0.0007570646703243256, 'timing/model': 0.33599611185491085, 'learning_rate/qwen_vl_interface': 9.999778081629055e-06, 'learning_rate/action_model': 9.999778081629054e-05, 'epoch': 0.3})
+[2026-07-28 09:57:13,466][starVLA.training.train_starvla][INFO] - Step 1613, Loss: {'train/loss': 0.047875843942165375, 'train/grad_norm_pre_clip': 1.361890196800232, 'timing/data': 0.0003097299486398697, 'timing/model': 0.357068357989192, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.3})
+[2026-07-28 09:57:17,571][starVLA.training.train_starvla][INFO] - Step 1614, Loss: {'train/loss': 0.04930644854903221, 'train/grad_norm_pre_clip': 2.0713672637939453, 'timing/data': 0.0004484322853386402, 'timing/model': 0.33740465296432376, 'learning_rate/qwen_vl_interface': 9.999697945288881e-06, 'learning_rate/action_model': 9.999697945288881e-05, 'epoch': 0.3})
+[2026-07-28 09:57:21,676][starVLA.training.train_starvla][INFO] - Step 1615, Loss: {'train/loss': 0.06742535531520844, 'train/grad_norm_pre_clip': 1.7210192680358887, 'timing/data': 0.0002970942296087742, 'timing/model': 0.3350455188192427, 'learning_rate/qwen_vl_interface': 9.999653254064153e-06, 'learning_rate/action_model': 9.999653254064152e-05, 'epoch': 0.3})
+[2026-07-28 09:57:25,786][starVLA.training.train_starvla][INFO] - Step 1616, Loss: {'train/loss': 0.04581073299050331, 'train/grad_norm_pre_clip': 1.280961036682129, 'timing/data': 0.0002888939343392849, 'timing/model': 0.33635821705684066, 'learning_rate/qwen_vl_interface': 9.99960548084099e-06, 'learning_rate/action_model': 9.99960548084099e-05, 'epoch': 0.3})
+[2026-07-28 09:57:29,909][starVLA.training.train_starvla][INFO] - Step 1617, Loss: {'train/loss': 0.07544480264186859, 'train/grad_norm_pre_clip': 1.937631607055664, 'timing/data': 0.00028894515708088875, 'timing/model': 0.3386780866421759, 'learning_rate/qwen_vl_interface': 9.999554625650398e-06, 'learning_rate/action_model': 9.999554625650399e-05, 'epoch': 0.3})
+[2026-07-28 09:57:34,019][starVLA.training.train_starvla][INFO] - Step 1618, Loss: {'train/loss': 0.09096494317054749, 'train/grad_norm_pre_clip': 2.358228921890259, 'timing/data': 0.00032983021810650826, 'timing/model': 0.33679543621838093, 'learning_rate/qwen_vl_interface': 9.999500688525376e-06, 'learning_rate/action_model': 9.999500688525375e-05, 'epoch': 0.3})
+[2026-07-28 09:57:38,129][starVLA.training.train_starvla][INFO] - Step 1619, Loss: {'train/loss': 0.0657673180103302, 'train/grad_norm_pre_clip': 2.3863348960876465, 'timing/data': 0.0003596390597522259, 'timing/model': 0.33694958267733455, 'learning_rate/qwen_vl_interface': 9.99944366950092e-06, 'learning_rate/action_model': 9.999443669500919e-05, 'epoch': 0.3})
+[2026-07-28 09:57:42,535][starVLA.training.train_starvla][INFO] - Step 1620, Loss: {'train/loss': 0.08605805784463882, 'train/grad_norm_pre_clip': 2.5430240631103516, 'timing/data': 0.000324849970638752, 'timing/model': 0.3357345671392977, 'learning_rate/qwen_vl_interface': 9.99938356861403e-06, 'learning_rate/action_model': 9.99938356861403e-05, 'epoch': 0.3})
+[2026-07-28 09:57:46,652][starVLA.training.train_starvla][INFO] - Step 1621, Loss: {'train/loss': 0.06961195170879364, 'train/grad_norm_pre_clip': 3.114264726638794, 'timing/data': 0.0003099837340414524, 'timing/model': 0.33571419585496187, 'learning_rate/qwen_vl_interface': 9.999320385903709e-06, 'learning_rate/action_model': 9.999320385903708e-05, 'epoch': 0.3})
+[2026-07-28 09:57:50,761][starVLA.training.train_starvla][INFO] - Step 1622, Loss: {'train/loss': 0.0453706756234169, 'train/grad_norm_pre_clip': 2.549321413040161, 'timing/data': 0.00031226081773638725, 'timing/model': 0.334965443238616, 'learning_rate/qwen_vl_interface': 9.999254121410948e-06, 'learning_rate/action_model': 9.999254121410949e-05, 'epoch': 0.3})
+[2026-07-28 09:57:54,880][starVLA.training.train_starvla][INFO] - Step 1623, Loss: {'train/loss': 0.07162166386842728, 'train/grad_norm_pre_clip': 2.665257692337036, 'timing/data': 0.0003072740510106087, 'timing/model': 0.33674873411655426, 'learning_rate/qwen_vl_interface': 9.999184775178751e-06, 'learning_rate/action_model': 9.999184775178751e-05, 'epoch': 0.3})
+[2026-07-28 09:57:59,032][starVLA.training.train_starvla][INFO] - Step 1624, Loss: {'train/loss': 0.08978430926799774, 'train/grad_norm_pre_clip': 3.3124780654907227, 'timing/data': 0.0003771660849452019, 'timing/model': 0.339536190032959, 'learning_rate/qwen_vl_interface': 9.999112347252117e-06, 'learning_rate/action_model': 9.999112347252116e-05, 'epoch': 0.3})
+[2026-07-28 09:58:03,145][starVLA.training.train_starvla][INFO] - Step 1625, Loss: {'train/loss': 0.057241059839725494, 'train/grad_norm_pre_clip': 2.9571754932403564, 'timing/data': 0.00034705689176917076, 'timing/model': 0.3337096720933914, 'learning_rate/qwen_vl_interface': 9.99903683767804e-06, 'learning_rate/action_model': 9.99903683767804e-05, 'epoch': 0.3})
+[2026-07-28 09:58:07,256][starVLA.training.train_starvla][INFO] - Step 1626, Loss: {'train/loss': 0.06276944279670715, 'train/grad_norm_pre_clip': 2.3409109115600586, 'timing/data': 0.0002993280068039894, 'timing/model': 0.3340204656124115, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.3})
+[2026-07-28 09:58:11,374][starVLA.training.train_starvla][INFO] - Step 1627, Loss: {'train/loss': 0.031017214059829712, 'train/grad_norm_pre_clip': 1.1808621883392334, 'timing/data': 0.0002938471734523773, 'timing/model': 0.33745081117376685, 'learning_rate/qwen_vl_interface': 9.998876573785547e-06, 'learning_rate/action_model': 9.998876573785547e-05, 'epoch': 0.3})
+[2026-07-28 09:58:15,517][starVLA.training.train_starvla][INFO] - Step 1628, Loss: {'train/loss': 0.09141260385513306, 'train/grad_norm_pre_clip': 1.7560744285583496, 'timing/data': 0.00030222581699490547, 'timing/model': 0.33619226794689894, 'learning_rate/qwen_vl_interface': 9.998791819571126e-06, 'learning_rate/action_model': 9.998791819571127e-05, 'epoch': 0.3})
+[2026-07-28 09:58:19,680][starVLA.training.train_starvla][INFO] - Step 1629, Loss: {'train/loss': 0.07222826033830643, 'train/grad_norm_pre_clip': 1.2951055765151978, 'timing/data': 0.0003438941203057766, 'timing/model': 0.334001153241843, 'learning_rate/qwen_vl_interface': 9.998703983917252e-06, 'learning_rate/action_model': 9.998703983917252e-05, 'epoch': 0.3})
+[2026-07-28 09:58:24,093][starVLA.training.train_starvla][INFO] - Step 1630, Loss: {'train/loss': 0.05684700608253479, 'train/grad_norm_pre_clip': 1.03084397315979, 'timing/data': 0.00034007197245955467, 'timing/model': 0.33849084889516234, 'learning_rate/qwen_vl_interface': 9.998613066880919e-06, 'learning_rate/action_model': 9.998613066880918e-05, 'epoch': 0.3})
+[2026-07-28 09:58:28,284][starVLA.training.train_starvla][INFO] - Step 1631, Loss: {'train/loss': 0.0410095751285553, 'train/grad_norm_pre_clip': 0.7319217324256897, 'timing/data': 0.00034661218523979187, 'timing/model': 0.33933736477047205, 'learning_rate/qwen_vl_interface': 9.99851906852112e-06, 'learning_rate/action_model': 9.998519068521119e-05, 'epoch': 0.3})
+[2026-07-28 09:58:32,457][starVLA.training.train_starvla][INFO] - Step 1632, Loss: {'train/loss': 0.09257718920707703, 'train/grad_norm_pre_clip': 1.7924202680587769, 'timing/data': 0.0003852918744087219, 'timing/model': 0.33990618493407965, 'learning_rate/qwen_vl_interface': 9.998421988898851e-06, 'learning_rate/action_model': 9.998421988898851e-05, 'epoch': 0.3})
+[2026-07-28 09:58:36,635][starVLA.training.train_starvla][INFO] - Step 1633, Loss: {'train/loss': 0.043307673186063766, 'train/grad_norm_pre_clip': 0.7648749947547913, 'timing/data': 0.000402962788939476, 'timing/model': 0.33651250693947077, 'learning_rate/qwen_vl_interface': 9.998321828077107e-06, 'learning_rate/action_model': 9.998321828077106e-05, 'epoch': 0.3})
+[2026-07-28 09:58:40,762][starVLA.training.train_starvla][INFO] - Step 1634, Loss: {'train/loss': 0.04891210421919823, 'train/grad_norm_pre_clip': 1.5277905464172363, 'timing/data': 0.0002961978316307068, 'timing/model': 0.3362742830067873, 'learning_rate/qwen_vl_interface': 9.99821858612088e-06, 'learning_rate/action_model': 9.99821858612088e-05, 'epoch': 0.3})
+[2026-07-28 09:58:44,876][starVLA.training.train_starvla][INFO] - Step 1635, Loss: {'train/loss': 0.04380190372467041, 'train/grad_norm_pre_clip': 0.7372319102287292, 'timing/data': 0.0003666607663035393, 'timing/model': 0.3378550233319402, 'learning_rate/qwen_vl_interface': 9.998112263097164e-06, 'learning_rate/action_model': 9.998112263097165e-05, 'epoch': 0.3})
+[2026-07-28 09:58:49,009][starVLA.training.train_starvla][INFO] - Step 1636, Loss: {'train/loss': 0.059555720537900925, 'train/grad_norm_pre_clip': 2.7968475818634033, 'timing/data': 0.0004506930708885193, 'timing/model': 0.3445232338272035, 'learning_rate/qwen_vl_interface': 9.99800285907495e-06, 'learning_rate/action_model': 9.99800285907495e-05, 'epoch': 0.3})
+[2026-07-28 09:58:53,191][starVLA.training.train_starvla][INFO] - Step 1637, Loss: {'train/loss': 0.06583516299724579, 'train/grad_norm_pre_clip': 4.13937520980835, 'timing/data': 0.00038626277819275856, 'timing/model': 0.3349886923097074, 'learning_rate/qwen_vl_interface': 9.997890374125228e-06, 'learning_rate/action_model': 9.997890374125227e-05, 'epoch': 0.3})
+[2026-07-28 09:58:57,313][starVLA.training.train_starvla][INFO] - Step 1638, Loss: {'train/loss': 0.07638618350028992, 'train/grad_norm_pre_clip': 1.9767218828201294, 'timing/data': 0.00038318801671266556, 'timing/model': 0.33491847664117813, 'learning_rate/qwen_vl_interface': 9.997774808320992e-06, 'learning_rate/action_model': 9.997774808320992e-05, 'epoch': 0.3})
+[2026-07-28 09:59:01,434][starVLA.training.train_starvla][INFO] - Step 1639, Loss: {'train/loss': 0.045918822288513184, 'train/grad_norm_pre_clip': 2.1141231060028076, 'timing/data': 0.0009114798158407211, 'timing/model': 0.33624024922028184, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.3})
+[2026-07-28 09:59:05,585][starVLA.training.train_starvla][INFO] - Step 1640, Loss: {'train/loss': 0.05467348173260689, 'train/grad_norm_pre_clip': 2.3788869380950928, 'timing/data': 0.00038227997720241547, 'timing/model': 0.33685029298067093, 'learning_rate/qwen_vl_interface': 9.997534434450922e-06, 'learning_rate/action_model': 9.997534434450921e-05, 'epoch': 0.3})
+[2026-07-28 09:59:09,705][starVLA.training.train_starvla][INFO] - Step 1641, Loss: {'train/loss': 0.038491588085889816, 'train/grad_norm_pre_clip': 0.7184939980506897, 'timing/data': 0.0003348630852997303, 'timing/model': 0.3375128782354295, 'learning_rate/qwen_vl_interface': 9.997409626541065e-06, 'learning_rate/action_model': 9.997409626541066e-05, 'epoch': 0.3})
+[2026-07-28 09:59:13,836][starVLA.training.train_starvla][INFO] - Step 1642, Loss: {'train/loss': 0.08187804371118546, 'train/grad_norm_pre_clip': 2.1641042232513428, 'timing/data': 0.00030872784554958344, 'timing/model': 0.3379180859774351, 'learning_rate/qwen_vl_interface': 9.997281738088647e-06, 'learning_rate/action_model': 9.997281738088646e-05, 'epoch': 0.3})
+[2026-07-28 09:59:17,965][starVLA.training.train_starvla][INFO] - Step 1643, Loss: {'train/loss': 0.02476317062973976, 'train/grad_norm_pre_clip': 1.3446776866912842, 'timing/data': 0.0003524851053953171, 'timing/model': 0.33774203388020396, 'learning_rate/qwen_vl_interface': 9.997150769176648e-06, 'learning_rate/action_model': 9.997150769176647e-05, 'epoch': 0.3})
+[2026-07-28 09:59:22,107][starVLA.training.train_starvla][INFO] - Step 1644, Loss: {'train/loss': 0.05803656578063965, 'train/grad_norm_pre_clip': 1.5924060344696045, 'timing/data': 0.0003048256039619446, 'timing/model': 0.3414252861402929, 'learning_rate/qwen_vl_interface': 9.997016719890052e-06, 'learning_rate/action_model': 9.997016719890052e-05, 'epoch': 0.3})
+[2026-07-28 09:59:26,260][starVLA.training.train_starvla][INFO] - Step 1645, Loss: {'train/loss': 0.05756494402885437, 'train/grad_norm_pre_clip': 1.7762959003448486, 'timing/data': 0.000323242973536253, 'timing/model': 0.3384877759963274, 'learning_rate/qwen_vl_interface': 9.996879590315847e-06, 'learning_rate/action_model': 9.996879590315846e-05, 'epoch': 0.3})
+[2026-07-28 09:59:30,457][starVLA.training.train_starvla][INFO] - Step 1646, Loss: {'train/loss': 0.11986281722784042, 'train/grad_norm_pre_clip': 2.233644723892212, 'timing/data': 0.0003700382076203823, 'timing/model': 0.33924014307558537, 'learning_rate/qwen_vl_interface': 9.99673938054301e-06, 'learning_rate/action_model': 9.996739380543009e-05, 'epoch': 0.3})
+[2026-07-28 09:59:34,590][starVLA.training.train_starvla][INFO] - Step 1647, Loss: {'train/loss': 0.09089671075344086, 'train/grad_norm_pre_clip': 1.9338068962097168, 'timing/data': 0.0003706919960677624, 'timing/model': 0.3386564003303647, 'learning_rate/qwen_vl_interface': 9.996596090662525e-06, 'learning_rate/action_model': 9.996596090662525e-05, 'epoch': 0.3})
+[2026-07-28 09:59:38,766][starVLA.training.train_starvla][INFO] - Step 1648, Loss: {'train/loss': 0.053379666060209274, 'train/grad_norm_pre_clip': 2.990910053253174, 'timing/data': 0.00039966311305761337, 'timing/model': 0.34606192680075765, 'learning_rate/qwen_vl_interface': 9.996449720767369e-06, 'learning_rate/action_model': 9.996449720767368e-05, 'epoch': 0.3})
+[2026-07-28 09:59:42,921][starVLA.training.train_starvla][INFO] - Step 1649, Loss: {'train/loss': 0.06925924122333527, 'train/grad_norm_pre_clip': 2.120889186859131, 'timing/data': 0.00029070302844047546, 'timing/model': 0.34913891600444913, 'learning_rate/qwen_vl_interface': 9.99630027095252e-06, 'learning_rate/action_model': 9.99630027095252e-05, 'epoch': 0.3})
+[2026-07-28 09:59:47,063][starVLA.training.train_starvla][INFO] - Step 1650, Loss: {'train/loss': 0.06255578994750977, 'train/grad_norm_pre_clip': 1.0736215114593506, 'timing/data': 0.0003075031563639641, 'timing/model': 0.33752399403601885, 'learning_rate/qwen_vl_interface': 9.996147741314956e-06, 'learning_rate/action_model': 9.996147741314957e-05, 'epoch': 0.3})
+[2026-07-28 09:59:51,203][starVLA.training.train_starvla][INFO] - Step 1651, Loss: {'train/loss': 0.12305378913879395, 'train/grad_norm_pre_clip': 4.057115077972412, 'timing/data': 0.0003701448440551758, 'timing/model': 0.34129602927714586, 'learning_rate/qwen_vl_interface': 9.99599213195365e-06, 'learning_rate/action_model': 9.99599213195365e-05, 'epoch': 0.3})
+[2026-07-28 09:59:55,385][starVLA.training.train_starvla][INFO] - Step 1652, Loss: {'train/loss': 0.0765734389424324, 'train/grad_norm_pre_clip': 3.323817014694214, 'timing/data': 0.0003771628253161907, 'timing/model': 0.34087134012952447, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.3})
+[2026-07-28 09:59:59,570][starVLA.training.train_starvla][INFO] - Step 1653, Loss: {'train/loss': 0.08099714666604996, 'train/grad_norm_pre_clip': 3.7287094593048096, 'timing/data': 0.00030289124697446823, 'timing/model': 0.33634680788964033, 'learning_rate/qwen_vl_interface': 9.995671674465709e-06, 'learning_rate/action_model': 9.995671674465709e-05, 'epoch': 0.3})
+[2026-07-28 10:00:03,748][starVLA.training.train_starvla][INFO] - Step 1654, Loss: {'train/loss': 0.06835240125656128, 'train/grad_norm_pre_clip': 3.181820869445801, 'timing/data': 0.0002994849346578121, 'timing/model': 0.3350484077818692, 'learning_rate/qwen_vl_interface': 9.99550682654701e-06, 'learning_rate/action_model': 9.995506826547011e-05, 'epoch': 0.3})
+[2026-07-28 10:00:07,929][starVLA.training.train_starvla][INFO] - Step 1655, Loss: {'train/loss': 0.09326690435409546, 'train/grad_norm_pre_clip': 2.7060387134552, 'timing/data': 0.0003034672699868679, 'timing/model': 0.3369145169854164, 'learning_rate/qwen_vl_interface': 9.995338899320457e-06, 'learning_rate/action_model': 9.995338899320456e-05, 'epoch': 0.3})
+[2026-07-28 10:00:12,077][starVLA.training.train_starvla][INFO] - Step 1656, Loss: {'train/loss': 0.054734937846660614, 'train/grad_norm_pre_clip': 1.0370995998382568, 'timing/data': 0.00034405477344989777, 'timing/model': 0.33596139727160335, 'learning_rate/qwen_vl_interface': 9.995167892895011e-06, 'learning_rate/action_model': 9.99516789289501e-05, 'epoch': 0.3})
+[2026-07-28 10:00:16,199][starVLA.training.train_starvla][INFO] - Step 1657, Loss: {'train/loss': 0.05336446315050125, 'train/grad_norm_pre_clip': 0.8450992107391357, 'timing/data': 0.00029962416738271713, 'timing/model': 0.33583149733021855, 'learning_rate/qwen_vl_interface': 9.994993807381635e-06, 'learning_rate/action_model': 9.994993807381635e-05, 'epoch': 0.3})
+[2026-07-28 10:00:20,320][starVLA.training.train_starvla][INFO] - Step 1658, Loss: {'train/loss': 0.06999008357524872, 'train/grad_norm_pre_clip': 1.1156227588653564, 'timing/data': 0.0004174690693616867, 'timing/model': 0.33559077931568027, 'learning_rate/qwen_vl_interface': 9.994816642893294e-06, 'learning_rate/action_model': 9.994816642893293e-05, 'epoch': 0.3})
+[2026-07-28 10:00:24,439][starVLA.training.train_starvla][INFO] - Step 1659, Loss: {'train/loss': 0.05527045205235481, 'train/grad_norm_pre_clip': 2.0051968097686768, 'timing/data': 0.0003123250789940357, 'timing/model': 0.3377537983469665, 'learning_rate/qwen_vl_interface': 9.994636399544948e-06, 'learning_rate/action_model': 9.994636399544947e-05, 'epoch': 0.3})
+[2026-07-28 10:00:28,558][starVLA.training.train_starvla][INFO] - Step 1660, Loss: {'train/loss': 0.08001835644245148, 'train/grad_norm_pre_clip': 2.5658555030822754, 'timing/data': 0.00030462583526968956, 'timing/model': 0.33732091821730137, 'learning_rate/qwen_vl_interface': 9.994453077453552e-06, 'learning_rate/action_model': 9.994453077453552e-05, 'epoch': 0.3})
+[2026-07-28 10:00:32,674][starVLA.training.train_starvla][INFO] - Step 1661, Loss: {'train/loss': 0.04379314184188843, 'train/grad_norm_pre_clip': 1.611595869064331, 'timing/data': 0.0004325821064412594, 'timing/model': 0.33796549681574106, 'learning_rate/qwen_vl_interface': 9.994266676738065e-06, 'learning_rate/action_model': 9.994266676738065e-05, 'epoch': 0.31})
+[2026-07-28 10:00:36,826][starVLA.training.train_starvla][INFO] - Step 1662, Loss: {'train/loss': 0.10381022095680237, 'train/grad_norm_pre_clip': 3.554889440536499, 'timing/data': 0.0003055357374250889, 'timing/model': 0.3353488161228597, 'learning_rate/qwen_vl_interface': 9.99407719751944e-06, 'learning_rate/action_model': 9.994077197519439e-05, 'epoch': 0.31})
+[2026-07-28 10:00:40,988][starVLA.training.train_starvla][INFO] - Step 1663, Loss: {'train/loss': 0.04956642538309097, 'train/grad_norm_pre_clip': 1.04496431350708, 'timing/data': 0.0003863861784338951, 'timing/model': 0.34264949802309275, 'learning_rate/qwen_vl_interface': 9.993884639920626e-06, 'learning_rate/action_model': 9.993884639920626e-05, 'epoch': 0.31})
+[2026-07-28 10:00:45,488][starVLA.training.train_starvla][INFO] - Step 1664, Loss: {'train/loss': 0.06430314481258392, 'train/grad_norm_pre_clip': 1.7118337154388428, 'timing/data': 0.00029472168534994125, 'timing/model': 0.3439680910669267, 'learning_rate/qwen_vl_interface': 9.993689004066573e-06, 'learning_rate/action_model': 9.993689004066573e-05, 'epoch': 0.31})
+[2026-07-28 10:00:49,954][starVLA.training.train_starvla][INFO] - Step 1665, Loss: {'train/loss': 0.045584142208099365, 'train/grad_norm_pre_clip': 2.006495714187622, 'timing/data': 0.0006677829660475254, 'timing/model': 0.366616218816489, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.31})
+[2026-07-28 10:00:54,198][starVLA.training.train_starvla][INFO] - Step 1666, Loss: {'train/loss': 0.05396241694688797, 'train/grad_norm_pre_clip': 1.952510952949524, 'timing/data': 0.0009292163886129856, 'timing/model': 0.33703396283090115, 'learning_rate/qwen_vl_interface': 9.993288498102531e-06, 'learning_rate/action_model': 9.993288498102532e-05, 'epoch': 0.31})
+[2026-07-28 10:00:58,372][starVLA.training.train_starvla][INFO] - Step 1667, Loss: {'train/loss': 0.08889704942703247, 'train/grad_norm_pre_clip': 3.553349018096924, 'timing/data': 0.0003427509218454361, 'timing/model': 0.33800815511494875, 'learning_rate/qwen_vl_interface': 9.993083628252424e-06, 'learning_rate/action_model': 9.993083628252424e-05, 'epoch': 0.31})
+[2026-07-28 10:01:02,501][starVLA.training.train_starvla][INFO] - Step 1668, Loss: {'train/loss': 0.06256633996963501, 'train/grad_norm_pre_clip': 1.0456067323684692, 'timing/data': 0.0003594108857214451, 'timing/model': 0.34085883293300867, 'learning_rate/qwen_vl_interface': 9.992875680666849e-06, 'learning_rate/action_model': 9.992875680666849e-05, 'epoch': 0.31})
+[2026-07-28 10:01:06,657][starVLA.training.train_starvla][INFO] - Step 1669, Loss: {'train/loss': 0.06548723578453064, 'train/grad_norm_pre_clip': 2.3466367721557617, 'timing/data': 0.0005084960721433163, 'timing/model': 0.3437354671768844, 'learning_rate/qwen_vl_interface': 9.992664655480736e-06, 'learning_rate/action_model': 9.992664655480735e-05, 'epoch': 0.31})
+[2026-07-28 10:01:10,964][starVLA.training.train_starvla][INFO] - Step 1670, Loss: {'train/loss': 0.08327339589595795, 'train/grad_norm_pre_clip': 1.4506880044937134, 'timing/data': 0.0004109330475330353, 'timing/model': 0.34819949278607965, 'learning_rate/qwen_vl_interface': 9.992450552831018e-06, 'learning_rate/action_model': 9.992450552831017e-05, 'epoch': 0.31})
+[2026-07-28 10:01:15,132][starVLA.training.train_starvla][INFO] - Step 1671, Loss: {'train/loss': 0.03323439881205559, 'train/grad_norm_pre_clip': 0.6962308287620544, 'timing/data': 0.00030544912442564964, 'timing/model': 0.35209764959290624, 'learning_rate/qwen_vl_interface': 9.992233372856623e-06, 'learning_rate/action_model': 9.992233372856624e-05, 'epoch': 0.31})
+[2026-07-28 10:01:19,403][starVLA.training.train_starvla][INFO] - Step 1672, Loss: {'train/loss': 0.06986777484416962, 'train/grad_norm_pre_clip': 0.9737241864204407, 'timing/data': 0.0003189770504832268, 'timing/model': 0.354597759898752, 'learning_rate/qwen_vl_interface': 9.99201311569848e-06, 'learning_rate/action_model': 9.99201311569848e-05, 'epoch': 0.31})
+[2026-07-28 10:01:23,731][starVLA.training.train_starvla][INFO] - Step 1673, Loss: {'train/loss': 0.06440970301628113, 'train/grad_norm_pre_clip': 2.070359706878662, 'timing/data': 0.00029520224779844284, 'timing/model': 0.3650719509460032, 'learning_rate/qwen_vl_interface': 9.99178978149951e-06, 'learning_rate/action_model': 9.99178978149951e-05, 'epoch': 0.31})
+[2026-07-28 10:01:28,023][starVLA.training.train_starvla][INFO] - Step 1674, Loss: {'train/loss': 0.05371471866965294, 'train/grad_norm_pre_clip': 2.6366097927093506, 'timing/data': 0.00036463281139731407, 'timing/model': 0.33689654199406505, 'learning_rate/qwen_vl_interface': 9.99156337040463e-06, 'learning_rate/action_model': 9.991563370404629e-05, 'epoch': 0.31})
+[2026-07-28 10:01:32,197][starVLA.training.train_starvla][INFO] - Step 1675, Loss: {'train/loss': 0.04610336944460869, 'train/grad_norm_pre_clip': 1.0459685325622559, 'timing/data': 0.0003628842532634735, 'timing/model': 0.3368757856078446, 'learning_rate/qwen_vl_interface': 9.991333882560759e-06, 'learning_rate/action_model': 9.991333882560759e-05, 'epoch': 0.31})
+[2026-07-28 10:01:36,351][starVLA.training.train_starvla][INFO] - Step 1676, Loss: {'train/loss': 0.07836584746837616, 'train/grad_norm_pre_clip': 1.8829894065856934, 'timing/data': 0.00035306625068187714, 'timing/model': 0.35098861204460263, 'learning_rate/qwen_vl_interface': 9.991101318116806e-06, 'learning_rate/action_model': 9.991101318116806e-05, 'epoch': 0.31})
+[2026-07-28 10:01:40,907][starVLA.training.train_starvla][INFO] - Step 1677, Loss: {'train/loss': 0.10192970931529999, 'train/grad_norm_pre_clip': 1.636066198348999, 'timing/data': 0.00029724882915616035, 'timing/model': 0.3372604981996119, 'learning_rate/qwen_vl_interface': 9.990865677223684e-06, 'learning_rate/action_model': 9.990865677223684e-05, 'epoch': 0.31})
+[2026-07-28 10:01:45,042][starVLA.training.train_starvla][INFO] - Step 1678, Loss: {'train/loss': 0.07643017172813416, 'train/grad_norm_pre_clip': 1.494033694267273, 'timing/data': 0.00036366796121001244, 'timing/model': 0.337752154096961, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.31})
+[2026-07-28 10:01:49,305][starVLA.training.train_starvla][INFO] - Step 1679, Loss: {'train/loss': 0.061754584312438965, 'train/grad_norm_pre_clip': 2.024421215057373, 'timing/data': 0.00030509103089571, 'timing/model': 0.34192338585853577, 'learning_rate/qwen_vl_interface': 9.990385166703534e-06, 'learning_rate/action_model': 9.990385166703533e-05, 'epoch': 0.31})
+[2026-07-28 10:01:53,517][starVLA.training.train_starvla][INFO] - Step 1680, Loss: {'train/loss': 0.043669622391462326, 'train/grad_norm_pre_clip': 1.3935813903808594, 'timing/data': 0.00030901096761226654, 'timing/model': 0.33737690187990665, 'learning_rate/qwen_vl_interface': 9.990140297388306e-06, 'learning_rate/action_model': 9.990140297388307e-05, 'epoch': 0.31})
+[2026-07-28 10:01:57,649][starVLA.training.train_starvla][INFO] - Step 1681, Loss: {'train/loss': 0.07243099063634872, 'train/grad_norm_pre_clip': 1.1765532493591309, 'timing/data': 0.000352335162460804, 'timing/model': 0.33530229004099965, 'learning_rate/qwen_vl_interface': 9.989892352247505e-06, 'learning_rate/action_model': 9.989892352247504e-05, 'epoch': 0.31})
+[2026-07-28 10:02:01,757][starVLA.training.train_starvla][INFO] - Step 1682, Loss: {'train/loss': 0.0634726732969284, 'train/grad_norm_pre_clip': 1.6432197093963623, 'timing/data': 0.0003657666966319084, 'timing/model': 0.33630257565528154, 'learning_rate/qwen_vl_interface': 9.989641331442014e-06, 'learning_rate/action_model': 9.989641331442013e-05, 'epoch': 0.31})
+[2026-07-28 10:02:05,869][starVLA.training.train_starvla][INFO] - Step 1683, Loss: {'train/loss': 0.07017091661691666, 'train/grad_norm_pre_clip': 2.2892003059387207, 'timing/data': 0.0003446717746555805, 'timing/model': 0.3352907416410744, 'learning_rate/qwen_vl_interface': 9.989387235134722e-06, 'learning_rate/action_model': 9.989387235134722e-05, 'epoch': 0.31})
+[2026-07-28 10:02:09,978][starVLA.training.train_starvla][INFO] - Step 1684, Loss: {'train/loss': 0.047057293355464935, 'train/grad_norm_pre_clip': 2.842724561691284, 'timing/data': 0.00031416164711117744, 'timing/model': 0.33858707593753934, 'learning_rate/qwen_vl_interface': 9.98913006349051e-06, 'learning_rate/action_model': 9.989130063490508e-05, 'epoch': 0.31})
+[2026-07-28 10:02:14,084][starVLA.training.train_starvla][INFO] - Step 1685, Loss: {'train/loss': 0.07192166894674301, 'train/grad_norm_pre_clip': 2.123903751373291, 'timing/data': 0.00031693466007709503, 'timing/model': 0.33866608003154397, 'learning_rate/qwen_vl_interface': 9.988869816676247e-06, 'learning_rate/action_model': 9.988869816676246e-05, 'epoch': 0.31})
+[2026-07-28 10:02:18,196][starVLA.training.train_starvla][INFO] - Step 1686, Loss: {'train/loss': 0.034185782074928284, 'train/grad_norm_pre_clip': 0.9387152194976807, 'timing/data': 0.00029639527201652527, 'timing/model': 0.334049120079726, 'learning_rate/qwen_vl_interface': 9.988606494860811e-06, 'learning_rate/action_model': 9.988606494860812e-05, 'epoch': 0.31})
+[2026-07-28 10:02:22,299][starVLA.training.train_starvla][INFO] - Step 1687, Loss: {'train/loss': 0.09226630628108978, 'train/grad_norm_pre_clip': 2.024562358856201, 'timing/data': 0.0002904837019741535, 'timing/model': 0.3357326122932136, 'learning_rate/qwen_vl_interface': 9.98834009821507e-06, 'learning_rate/action_model': 9.988340098215069e-05, 'epoch': 0.31})
+[2026-07-28 10:02:26,423][starVLA.training.train_starvla][INFO] - Step 1688, Loss: {'train/loss': 0.04731649160385132, 'train/grad_norm_pre_clip': 1.7377541065216064, 'timing/data': 0.0003626132383942604, 'timing/model': 0.35358072677627206, 'learning_rate/qwen_vl_interface': 9.98807062691188e-06, 'learning_rate/action_model': 9.988070626911879e-05, 'epoch': 0.31})
+[2026-07-28 10:02:30,667][starVLA.training.train_starvla][INFO] - Step 1689, Loss: {'train/loss': 0.07113977521657944, 'train/grad_norm_pre_clip': 1.6914528608322144, 'timing/data': 0.0003450480289757252, 'timing/model': 0.3889488657005131, 'learning_rate/qwen_vl_interface': 9.987798081126099e-06, 'learning_rate/action_model': 9.987798081126099e-05, 'epoch': 0.31})
+[2026-07-28 10:02:34,785][starVLA.training.train_starvla][INFO] - Step 1690, Loss: {'train/loss': 0.05410672724246979, 'train/grad_norm_pre_clip': 1.9720298051834106, 'timing/data': 0.00034561287611722946, 'timing/model': 0.3363738809712231, 'learning_rate/qwen_vl_interface': 9.987522461034584e-06, 'learning_rate/action_model': 9.987522461034584e-05, 'epoch': 0.31})
+[2026-07-28 10:02:38,920][starVLA.training.train_starvla][INFO] - Step 1691, Loss: {'train/loss': 0.03256400674581528, 'train/grad_norm_pre_clip': 1.8742165565490723, 'timing/data': 0.0002902313135564327, 'timing/model': 0.33620795141905546, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.31})
+[2026-07-28 10:02:43,050][starVLA.training.train_starvla][INFO] - Step 1692, Loss: {'train/loss': 0.04089091718196869, 'train/grad_norm_pre_clip': 1.2010133266448975, 'timing/data': 0.0003564287908375263, 'timing/model': 0.33658877201378345, 'learning_rate/qwen_vl_interface': 9.98696199865172e-06, 'learning_rate/action_model': 9.98696199865172e-05, 'epoch': 0.31})
+[2026-07-28 10:02:47,171][starVLA.training.train_starvla][INFO] - Step 1693, Loss: {'train/loss': 0.06998386979103088, 'train/grad_norm_pre_clip': 1.598662257194519, 'timing/data': 0.00045956438407301903, 'timing/model': 0.33667599176988006, 'learning_rate/qwen_vl_interface': 9.986677156724054e-06, 'learning_rate/action_model': 9.986677156724054e-05, 'epoch': 0.31})
+[2026-07-28 10:02:51,299][starVLA.training.train_starvla][INFO] - Step 1694, Loss: {'train/loss': 0.03612695634365082, 'train/grad_norm_pre_clip': 2.1305410861968994, 'timing/data': 0.0003502489998936653, 'timing/model': 0.3341222479939461, 'learning_rate/qwen_vl_interface': 9.986389241218003e-06, 'learning_rate/action_model': 9.986389241218003e-05, 'epoch': 0.31})
+[2026-07-28 10:02:55,419][starVLA.training.train_starvla][INFO] - Step 1695, Loss: {'train/loss': 0.08866100013256073, 'train/grad_norm_pre_clip': 2.023851156234741, 'timing/data': 0.0003600572235882282, 'timing/model': 0.3369920193217695, 'learning_rate/qwen_vl_interface': 9.986098252320397e-06, 'learning_rate/action_model': 9.986098252320397e-05, 'epoch': 0.31})
+[2026-07-28 10:02:59,601][starVLA.training.train_starvla][INFO] - Step 1696, Loss: {'train/loss': 0.05268940329551697, 'train/grad_norm_pre_clip': 1.1782335042953491, 'timing/data': 0.00031820498406887054, 'timing/model': 0.3373138108290732, 'learning_rate/qwen_vl_interface': 9.985804190220054e-06, 'learning_rate/action_model': 9.985804190220053e-05, 'epoch': 0.31})
+[2026-07-28 10:03:03,793][starVLA.training.train_starvla][INFO] - Step 1697, Loss: {'train/loss': 0.04186953976750374, 'train/grad_norm_pre_clip': 1.3923295736312866, 'timing/data': 0.0003747721202671528, 'timing/model': 0.33666370529681444, 'learning_rate/qwen_vl_interface': 9.985507055107788e-06, 'learning_rate/action_model': 9.985507055107788e-05, 'epoch': 0.31})
+[2026-07-28 10:03:07,957][starVLA.training.train_starvla][INFO] - Step 1698, Loss: {'train/loss': 0.05361676961183548, 'train/grad_norm_pre_clip': 1.780055284500122, 'timing/data': 0.00030206935480237007, 'timing/model': 0.33555034128949046, 'learning_rate/qwen_vl_interface': 9.985206847176406e-06, 'learning_rate/action_model': 9.985206847176406e-05, 'epoch': 0.31})
+[2026-07-28 10:03:12,122][starVLA.training.train_starvla][INFO] - Step 1699, Loss: {'train/loss': 0.048703186213970184, 'train/grad_norm_pre_clip': 1.093277096748352, 'timing/data': 0.0003034560941159725, 'timing/model': 0.3472199263051152, 'learning_rate/qwen_vl_interface': 9.984903566620711e-06, 'learning_rate/action_model': 9.984903566620712e-05, 'epoch': 0.31})
+[2026-07-28 10:03:16,316][starVLA.training.train_starvla][INFO] - Step 1700, Loss: {'train/loss': 0.04808103293180466, 'train/grad_norm_pre_clip': 1.19708251953125, 'timing/data': 0.0003712521865963936, 'timing/model': 0.34018756775185466, 'learning_rate/qwen_vl_interface': 9.984597213637497e-06, 'learning_rate/action_model': 9.984597213637497e-05, 'epoch': 0.31})
+[2026-07-28 10:03:20,503][starVLA.training.train_starvla][INFO] - Step 1701, Loss: {'train/loss': 0.0773022323846817, 'train/grad_norm_pre_clip': 2.434839963912964, 'timing/data': 0.0003658500500023365, 'timing/model': 0.3369916910305619, 'learning_rate/qwen_vl_interface': 9.984287788425555e-06, 'learning_rate/action_model': 9.984287788425555e-05, 'epoch': 0.31})
+[2026-07-28 10:03:24,695][starVLA.training.train_starvla][INFO] - Step 1702, Loss: {'train/loss': 0.03629457578063011, 'train/grad_norm_pre_clip': 1.6421682834625244, 'timing/data': 0.00036065420135855675, 'timing/model': 0.3394236699678004, 'learning_rate/qwen_vl_interface': 9.983975291185666e-06, 'learning_rate/action_model': 9.983975291185667e-05, 'epoch': 0.31})
+[2026-07-28 10:03:28,896][starVLA.training.train_starvla][INFO] - Step 1703, Loss: {'train/loss': 0.05173760652542114, 'train/grad_norm_pre_clip': 2.8547325134277344, 'timing/data': 0.0003127269446849823, 'timing/model': 0.34187421295791864, 'learning_rate/qwen_vl_interface': 9.983659722120607e-06, 'learning_rate/action_model': 9.983659722120608e-05, 'epoch': 0.31})
+[2026-07-28 10:03:33,046][starVLA.training.train_starvla][INFO] - Step 1704, Loss: {'train/loss': 0.05395574867725372, 'train/grad_norm_pre_clip': 2.2104063034057617, 'timing/data': 0.00035810796543955803, 'timing/model': 0.3361716140061617, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.31})
+[2026-07-28 10:03:37,190][starVLA.training.train_starvla][INFO] - Step 1705, Loss: {'train/loss': 0.0857321098446846, 'train/grad_norm_pre_clip': 3.4522647857666016, 'timing/data': 0.0007810653187334538, 'timing/model': 0.33780725533142686, 'learning_rate/qwen_vl_interface': 9.983019369336048e-06, 'learning_rate/action_model': 9.983019369336048e-05, 'epoch': 0.31})
+[2026-07-28 10:03:41,314][starVLA.training.train_starvla][INFO] - Step 1706, Loss: {'train/loss': 0.12113045901060104, 'train/grad_norm_pre_clip': 5.140448570251465, 'timing/data': 0.000359382014721632, 'timing/model': 0.33546361792832613, 'learning_rate/qwen_vl_interface': 9.982694586032066e-06, 'learning_rate/action_model': 9.982694586032065e-05, 'epoch': 0.31})
+[2026-07-28 10:03:45,452][starVLA.training.train_starvla][INFO] - Step 1707, Loss: {'train/loss': 0.0742703229188919, 'train/grad_norm_pre_clip': 1.2919766902923584, 'timing/data': 0.0003514993004500866, 'timing/model': 0.3363715847954154, 'learning_rate/qwen_vl_interface': 9.982366731733948e-06, 'learning_rate/action_model': 9.982366731733947e-05, 'epoch': 0.31})
+[2026-07-28 10:03:49,591][starVLA.training.train_starvla][INFO] - Step 1708, Loss: {'train/loss': 0.054145071655511856, 'train/grad_norm_pre_clip': 0.9911314249038696, 'timing/data': 0.00032007694244384766, 'timing/model': 0.3362332237884402, 'learning_rate/qwen_vl_interface': 9.982035806654437e-06, 'learning_rate/action_model': 9.982035806654437e-05, 'epoch': 0.31})
+[2026-07-28 10:03:54,091][starVLA.training.train_starvla][INFO] - Step 1709, Loss: {'train/loss': 0.06678108125925064, 'train/grad_norm_pre_clip': 1.93437659740448, 'timing/data': 0.00032292818650603294, 'timing/model': 0.33636857895180583, 'learning_rate/qwen_vl_interface': 9.981701811008267e-06, 'learning_rate/action_model': 9.981701811008267e-05, 'epoch': 0.31})
+[2026-07-28 10:03:58,308][starVLA.training.train_starvla][INFO] - Step 1710, Loss: {'train/loss': 0.03758993372321129, 'train/grad_norm_pre_clip': 1.2357593774795532, 'timing/data': 0.00030250195413827896, 'timing/model': 0.3368967971764505, 'learning_rate/qwen_vl_interface': 9.981364745012163e-06, 'learning_rate/action_model': 9.981364745012163e-05, 'epoch': 0.31})
+[2026-07-28 10:04:02,427][starVLA.training.train_starvla][INFO] - Step 1711, Loss: {'train/loss': 0.06728703528642654, 'train/grad_norm_pre_clip': 1.6827211380004883, 'timing/data': 0.00036387890577316284, 'timing/model': 0.33601280115544796, 'learning_rate/qwen_vl_interface': 9.981024608884842e-06, 'learning_rate/action_model': 9.981024608884843e-05, 'epoch': 0.31})
+[2026-07-28 10:04:06,561][starVLA.training.train_starvla][INFO] - Step 1712, Loss: {'train/loss': 0.0712370052933693, 'train/grad_norm_pre_clip': 2.685953378677368, 'timing/data': 0.00035638222470879555, 'timing/model': 0.33590859174728394, 'learning_rate/qwen_vl_interface': 9.980681402847017e-06, 'learning_rate/action_model': 9.980681402847016e-05, 'epoch': 0.31})
+[2026-07-28 10:04:10,705][starVLA.training.train_starvla][INFO] - Step 1713, Loss: {'train/loss': 0.06484635919332504, 'train/grad_norm_pre_clip': 3.4777934551239014, 'timing/data': 0.0003896700218319893, 'timing/model': 0.3356356625445187, 'learning_rate/qwen_vl_interface': 9.980335127121391e-06, 'learning_rate/action_model': 9.98033512712139e-05, 'epoch': 0.31})
+[2026-07-28 10:04:14,811][starVLA.training.train_starvla][INFO] - Step 1714, Loss: {'train/loss': 0.05236593633890152, 'train/grad_norm_pre_clip': 2.697120189666748, 'timing/data': 0.0003883461467921734, 'timing/model': 0.3357935920357704, 'learning_rate/qwen_vl_interface': 9.979985781932658e-06, 'learning_rate/action_model': 9.979985781932657e-05, 'epoch': 0.31})
+[2026-07-28 10:04:18,915][starVLA.training.train_starvla][INFO] - Step 1715, Loss: {'train/loss': 0.04068290442228317, 'train/grad_norm_pre_clip': 0.7827522158622742, 'timing/data': 0.00030469801276922226, 'timing/model': 0.3355022957548499, 'learning_rate/qwen_vl_interface': 9.979633367507501e-06, 'learning_rate/action_model': 9.979633367507502e-05, 'epoch': 0.32})
+[2026-07-28 10:04:23,080][starVLA.training.train_starvla][INFO] - Step 1716, Loss: {'train/loss': 0.06381303071975708, 'train/grad_norm_pre_clip': 1.3114620447158813, 'timing/data': 0.00031199678778648376, 'timing/model': 0.3365931026637554, 'learning_rate/qwen_vl_interface': 9.979277884074604e-06, 'learning_rate/action_model': 9.979277884074604e-05, 'epoch': 0.32})
+[2026-07-28 10:04:27,254][starVLA.training.train_starvla][INFO] - Step 1717, Loss: {'train/loss': 0.07474182546138763, 'train/grad_norm_pre_clip': 1.185916543006897, 'timing/data': 0.0005116420798003674, 'timing/model': 0.33563185716047883, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.32})
+[2026-07-28 10:04:31,369][starVLA.training.train_starvla][INFO] - Step 1718, Loss: {'train/loss': 0.05907324701547623, 'train/grad_norm_pre_clip': 1.0034074783325195, 'timing/data': 0.0003090193495154381, 'timing/model': 0.3359644110314548, 'learning_rate/qwen_vl_interface': 9.978557711110245e-06, 'learning_rate/action_model': 9.978557711110244e-05, 'epoch': 0.32})
+[2026-07-28 10:04:35,520][starVLA.training.train_starvla][INFO] - Step 1719, Loss: {'train/loss': 0.04144779592752457, 'train/grad_norm_pre_clip': 0.8460124731063843, 'timing/data': 0.00030975788831710815, 'timing/model': 0.33697052812203765, 'learning_rate/qwen_vl_interface': 9.978193022046097e-06, 'learning_rate/action_model': 9.978193022046097e-05, 'epoch': 0.32})
+[2026-07-28 10:04:39,685][starVLA.training.train_starvla][INFO] - Step 1720, Loss: {'train/loss': 0.0642806813120842, 'train/grad_norm_pre_clip': 0.9589282274246216, 'timing/data': 0.0003157062456011772, 'timing/model': 0.3351085912436247, 'learning_rate/qwen_vl_interface': 9.977825264908832e-06, 'learning_rate/action_model': 9.977825264908832e-05, 'epoch': 0.32})
+[2026-07-28 10:04:43,818][starVLA.training.train_starvla][INFO] - Step 1721, Loss: {'train/loss': 0.05775368586182594, 'train/grad_norm_pre_clip': 2.0188543796539307, 'timing/data': 0.0003139842301607132, 'timing/model': 0.33559068525210023, 'learning_rate/qwen_vl_interface': 9.97745443993708e-06, 'learning_rate/action_model': 9.97745443993708e-05, 'epoch': 0.32})
+[2026-07-28 10:04:47,948][starVLA.training.train_starvla][INFO] - Step 1722, Loss: {'train/loss': 0.028717514127492905, 'train/grad_norm_pre_clip': 1.2196063995361328, 'timing/data': 0.0003351522609591484, 'timing/model': 0.33522110478952527, 'learning_rate/qwen_vl_interface': 9.977080547371468e-06, 'learning_rate/action_model': 9.977080547371468e-05, 'epoch': 0.32})
+[2026-07-28 10:04:52,357][starVLA.training.train_starvla][INFO] - Step 1723, Loss: {'train/loss': 0.029126012697815895, 'train/grad_norm_pre_clip': 1.0761327743530273, 'timing/data': 0.0003112452104687691, 'timing/model': 0.33542059268802404, 'learning_rate/qwen_vl_interface': 9.97670358745461e-06, 'learning_rate/action_model': 9.97670358745461e-05, 'epoch': 0.32})
+[2026-07-28 10:04:56,510][starVLA.training.train_starvla][INFO] - Step 1724, Loss: {'train/loss': 0.07325991988182068, 'train/grad_norm_pre_clip': 3.54768967628479, 'timing/data': 0.0003884718753397465, 'timing/model': 0.33582467399537563, 'learning_rate/qwen_vl_interface': 9.976323560431112e-06, 'learning_rate/action_model': 9.976323560431111e-05, 'epoch': 0.32})
+[2026-07-28 10:05:00,709][starVLA.training.train_starvla][INFO] - Step 1725, Loss: {'train/loss': 0.04842741787433624, 'train/grad_norm_pre_clip': 1.354921579360962, 'timing/data': 0.00038131000474095345, 'timing/model': 0.3424707977101207, 'learning_rate/qwen_vl_interface': 9.975940466547567e-06, 'learning_rate/action_model': 9.975940466547567e-05, 'epoch': 0.32})
+[2026-07-28 10:05:04,841][starVLA.training.train_starvla][INFO] - Step 1726, Loss: {'train/loss': 0.09036078304052353, 'train/grad_norm_pre_clip': 2.1778500080108643, 'timing/data': 0.0003670398145914078, 'timing/model': 0.3366476180963218, 'learning_rate/qwen_vl_interface': 9.975554306052563e-06, 'learning_rate/action_model': 9.975554306052562e-05, 'epoch': 0.32})
+[2026-07-28 10:05:09,032][starVLA.training.train_starvla][INFO] - Step 1727, Loss: {'train/loss': 0.06206553056836128, 'train/grad_norm_pre_clip': 2.643996477127075, 'timing/data': 0.00030691083520650864, 'timing/model': 0.3873855769634247, 'learning_rate/qwen_vl_interface': 9.975165079196674e-06, 'learning_rate/action_model': 9.975165079196673e-05, 'epoch': 0.32})
+[2026-07-28 10:05:13,225][starVLA.training.train_starvla][INFO] - Step 1728, Loss: {'train/loss': 0.04255406931042671, 'train/grad_norm_pre_clip': 0.9648908376693726, 'timing/data': 0.0006011328659951687, 'timing/model': 0.33682867186143994, 'learning_rate/qwen_vl_interface': 9.974772786232466e-06, 'learning_rate/action_model': 9.974772786232466e-05, 'epoch': 0.32})
+[2026-07-28 10:05:17,402][starVLA.training.train_starvla][INFO] - Step 1729, Loss: {'train/loss': 0.04309976100921631, 'train/grad_norm_pre_clip': 1.091472864151001, 'timing/data': 0.0003068963997066021, 'timing/model': 0.33748139813542366, 'learning_rate/qwen_vl_interface': 9.974377427414493e-06, 'learning_rate/action_model': 9.974377427414493e-05, 'epoch': 0.32})
+[2026-07-28 10:05:21,533][starVLA.training.train_starvla][INFO] - Step 1730, Loss: {'train/loss': 0.03260205686092377, 'train/grad_norm_pre_clip': 0.9783661961555481, 'timing/data': 0.0002948208712041378, 'timing/model': 0.34993846993893385, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.32})
+[2026-07-28 10:05:25,717][starVLA.training.train_starvla][INFO] - Step 1731, Loss: {'train/loss': 0.07297856360673904, 'train/grad_norm_pre_clip': 1.7995543479919434, 'timing/data': 0.0002929489128291607, 'timing/model': 0.33718564407899976, 'learning_rate/qwen_vl_interface': 9.973577513245417e-06, 'learning_rate/action_model': 9.973577513245417e-05, 'epoch': 0.32})
+[2026-07-28 10:05:29,904][starVLA.training.train_starvla][INFO] - Step 1732, Loss: {'train/loss': 0.04114018380641937, 'train/grad_norm_pre_clip': 1.1787360906600952, 'timing/data': 0.0003133048303425312, 'timing/model': 0.3360658730380237, 'learning_rate/qwen_vl_interface': 9.973172958413372e-06, 'learning_rate/action_model': 9.97317295841337e-05, 'epoch': 0.32})
+[2026-07-28 10:05:34,046][starVLA.training.train_starvla][INFO] - Step 1733, Loss: {'train/loss': 0.039204247295856476, 'train/grad_norm_pre_clip': 0.777989387512207, 'timing/data': 0.000365247018635273, 'timing/model': 0.3412610883824527, 'learning_rate/qwen_vl_interface': 9.97276533876567e-06, 'learning_rate/action_model': 9.97276533876567e-05, 'epoch': 0.32})
+[2026-07-28 10:05:38,172][starVLA.training.train_starvla][INFO] - Step 1734, Loss: {'train/loss': 0.03507731109857559, 'train/grad_norm_pre_clip': 0.6840649247169495, 'timing/data': 0.00031741103157401085, 'timing/model': 0.33690202003344893, 'learning_rate/qwen_vl_interface': 9.972354654566816e-06, 'learning_rate/action_model': 9.972354654566816e-05, 'epoch': 0.32})
+[2026-07-28 10:05:42,369][starVLA.training.train_starvla][INFO] - Step 1735, Loss: {'train/loss': 0.041624732315540314, 'train/grad_norm_pre_clip': 1.5686453580856323, 'timing/data': 0.0003008968196809292, 'timing/model': 0.35598121769726276, 'learning_rate/qwen_vl_interface': 9.971940906083295e-06, 'learning_rate/action_model': 9.971940906083295e-05, 'epoch': 0.32})
+[2026-07-28 10:05:46,529][starVLA.training.train_starvla][INFO] - Step 1736, Loss: {'train/loss': 0.053888626396656036, 'train/grad_norm_pre_clip': 1.1242917776107788, 'timing/data': 0.0003661871887743473, 'timing/model': 0.33507928624749184, 'learning_rate/qwen_vl_interface': 9.971524093583586e-06, 'learning_rate/action_model': 9.971524093583586e-05, 'epoch': 0.32})
+[2026-07-28 10:05:50,728][starVLA.training.train_starvla][INFO] - Step 1737, Loss: {'train/loss': 0.09160314500331879, 'train/grad_norm_pre_clip': 1.7199947834014893, 'timing/data': 0.00036531686782836914, 'timing/model': 0.37497670901939273, 'learning_rate/qwen_vl_interface': 9.971104217338154e-06, 'learning_rate/action_model': 9.971104217338153e-05, 'epoch': 0.32})
+[2026-07-28 10:05:54,908][starVLA.training.train_starvla][INFO] - Step 1738, Loss: {'train/loss': 0.038751911371946335, 'train/grad_norm_pre_clip': 0.7869160175323486, 'timing/data': 0.00030887406319379807, 'timing/model': 0.3486534720286727, 'learning_rate/qwen_vl_interface': 9.970681277619452e-06, 'learning_rate/action_model': 9.970681277619452e-05, 'epoch': 0.32})
+[2026-07-28 10:05:59,123][starVLA.training.train_starvla][INFO] - Step 1739, Loss: {'train/loss': 0.04552634432911873, 'train/grad_norm_pre_clip': 1.2725434303283691, 'timing/data': 0.00033241789788007736, 'timing/model': 0.34417790500447154, 'learning_rate/qwen_vl_interface': 9.97025527470192e-06, 'learning_rate/action_model': 9.970255274701919e-05, 'epoch': 0.32})
+[2026-07-28 10:06:03,268][starVLA.training.train_starvla][INFO] - Step 1740, Loss: {'train/loss': 0.04912855848670006, 'train/grad_norm_pre_clip': 0.973983645439148, 'timing/data': 0.00036657415330410004, 'timing/model': 0.33562131179496646, 'learning_rate/qwen_vl_interface': 9.969826208861987e-06, 'learning_rate/action_model': 9.969826208861987e-05, 'epoch': 0.32})
+[2026-07-28 10:06:07,399][starVLA.training.train_starvla][INFO] - Step 1741, Loss: {'train/loss': 0.07424717396497726, 'train/grad_norm_pre_clip': 1.475188136100769, 'timing/data': 0.0003044581972062588, 'timing/model': 0.3405635701492429, 'learning_rate/qwen_vl_interface': 9.96939408037807e-06, 'learning_rate/action_model': 9.96939408037807e-05, 'epoch': 0.32})
+[2026-07-28 10:06:11,590][starVLA.training.train_starvla][INFO] - Step 1742, Loss: {'train/loss': 0.040027692914009094, 'train/grad_norm_pre_clip': 0.8072519898414612, 'timing/data': 0.00039765704423189163, 'timing/model': 0.35274801636114717, 'learning_rate/qwen_vl_interface': 9.968958889530573e-06, 'learning_rate/action_model': 9.968958889530573e-05, 'epoch': 0.32})
+[2026-07-28 10:06:15,782][starVLA.training.train_starvla][INFO] - Step 1743, Loss: {'train/loss': 0.06659458577632904, 'train/grad_norm_pre_clip': 0.812541663646698, 'timing/data': 0.00036708684638142586, 'timing/model': 0.3429088508710265, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.32})
+[2026-07-28 10:06:20,011][starVLA.training.train_starvla][INFO] - Step 1744, Loss: {'train/loss': 0.04907253384590149, 'train/grad_norm_pre_clip': 2.1830391883850098, 'timing/data': 0.0003226478584110737, 'timing/model': 0.34514510724693537, 'learning_rate/qwen_vl_interface': 9.968079321876382e-06, 'learning_rate/action_model': 9.968079321876382e-05, 'epoch': 0.32})
+[2026-07-28 10:06:24,201][starVLA.training.train_starvla][INFO] - Step 1745, Loss: {'train/loss': 0.04837874695658684, 'train/grad_norm_pre_clip': 1.3685287237167358, 'timing/data': 0.00043347617611289024, 'timing/model': 0.3480737800709903, 'learning_rate/qwen_vl_interface': 9.967634945640432e-06, 'learning_rate/action_model': 9.967634945640431e-05, 'epoch': 0.32})
+[2026-07-28 10:06:28,347][starVLA.training.train_starvla][INFO] - Step 1746, Loss: {'train/loss': 0.08556847274303436, 'train/grad_norm_pre_clip': 1.576562523841858, 'timing/data': 0.0004775538109242916, 'timing/model': 0.3365276660770178, 'learning_rate/qwen_vl_interface': 9.967187508182385e-06, 'learning_rate/action_model': 9.967187508182385e-05, 'epoch': 0.32})
+[2026-07-28 10:06:32,521][starVLA.training.train_starvla][INFO] - Step 1747, Loss: {'train/loss': 0.05569043010473251, 'train/grad_norm_pre_clip': 0.9335223436355591, 'timing/data': 0.0007939152419567108, 'timing/model': 0.33669650042429566, 'learning_rate/qwen_vl_interface': 9.966737009792577e-06, 'learning_rate/action_model': 9.966737009792577e-05, 'epoch': 0.32})
+[2026-07-28 10:06:36,706][starVLA.training.train_starvla][INFO] - Step 1748, Loss: {'train/loss': 0.05864661931991577, 'train/grad_norm_pre_clip': 1.351751446723938, 'timing/data': 0.00043168989941477776, 'timing/model': 0.3400787324644625, 'learning_rate/qwen_vl_interface': 9.966283450763333e-06, 'learning_rate/action_model': 9.966283450763333e-05, 'epoch': 0.32})
+[2026-07-28 10:06:40,835][starVLA.training.train_starvla][INFO] - Step 1749, Loss: {'train/loss': 0.0852801576256752, 'train/grad_norm_pre_clip': 2.0369625091552734, 'timing/data': 0.0004546958953142166, 'timing/model': 0.338448666036129, 'learning_rate/qwen_vl_interface': 9.965826831388962e-06, 'learning_rate/action_model': 9.965826831388961e-05, 'epoch': 0.32})
+[2026-07-28 10:06:44,971][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 1750
+[2026-07-28 10:07:59,190][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 1750
+[2026-07-28 10:12:44,233][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 1750
+[2026-07-28 10:12:44,234][starVLA.training.train_starvla][INFO] - Step 1750, Loss: {'train/loss': 0.040305521339178085, 'train/grad_norm_pre_clip': 1.2418779134750366, 'eval/action_loss/seconds': 74.2160446131602, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.05077151209115982, 'eval/latency_3/loss': 0.05077151209115982, 'eval/flappy/loss': 0.05077151209115982, 'eval/flappy/latency_3/loss': 0.05077151209115982, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 200.33000236973166, 'rl_games_eval/mid_train/mean_length': 1633.2, 'rl_games_eval/mid_train/std_reward': 167.7155352307171, 'rl_games_eval/mid_train/std_length': 1351.4404389391343, 'rl_games_eval/mid_train/macro_mean_reward': 200.33000236973166, 'rl_games_eval/mid_train/macro_mean_length': 1633.2, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 200.33000236973166, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 1633.2, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 167.7155352307171, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1351.4404389391343, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 200.33000236973166, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0002994160167872906, 'timing/model': 0.3351952042430639, 'learning_rate/qwen_vl_interface': 9.965367151965757e-06, 'learning_rate/action_model': 9.965367151965756e-05, 'epoch': 0.32})
+[2026-07-28 10:12:50,458][starVLA.training.train_starvla][INFO] - Step 1751, Loss: {'train/loss': 0.043165601789951324, 'train/grad_norm_pre_clip': 1.0202652215957642, 'timing/data': 0.0004190457984805107, 'timing/model': 0.4252889002673328, 'learning_rate/qwen_vl_interface': 9.964904412792002e-06, 'learning_rate/action_model': 9.964904412792002e-05, 'epoch': 0.32})
+[2026-07-28 10:12:56,835][starVLA.training.train_starvla][INFO] - Step 1752, Loss: {'train/loss': 0.0444432757794857, 'train/grad_norm_pre_clip': 1.2256313562393188, 'timing/data': 0.00033563002943992615, 'timing/model': 0.44055709708482027, 'learning_rate/qwen_vl_interface': 9.964438614167964e-06, 'learning_rate/action_model': 9.964438614167964e-05, 'epoch': 0.32})
+[2026-07-28 10:13:03,151][starVLA.training.train_starvla][INFO] - Step 1753, Loss: {'train/loss': 0.03255211189389229, 'train/grad_norm_pre_clip': 0.865839421749115, 'timing/data': 0.0004503820091485977, 'timing/model': 0.42904447531327605, 'learning_rate/qwen_vl_interface': 9.963969756395892e-06, 'learning_rate/action_model': 9.96396975639589e-05, 'epoch': 0.32})
+[2026-07-28 10:13:09,301][starVLA.training.train_starvla][INFO] - Step 1754, Loss: {'train/loss': 0.06787127256393433, 'train/grad_norm_pre_clip': 2.1868040561676025, 'timing/data': 0.0005080699920654297, 'timing/model': 0.4251368110999465, 'learning_rate/qwen_vl_interface': 9.963497839780025e-06, 'learning_rate/action_model': 9.963497839780024e-05, 'epoch': 0.32})
+[2026-07-28 10:13:15,096][starVLA.training.train_starvla][INFO] - Step 1755, Loss: {'train/loss': 0.04368508607149124, 'train/grad_norm_pre_clip': 1.1940701007843018, 'timing/data': 0.0003695432096719742, 'timing/model': 0.42187325609847903, 'learning_rate/qwen_vl_interface': 9.963022864626582e-06, 'learning_rate/action_model': 9.963022864626581e-05, 'epoch': 0.32})
+[2026-07-28 10:13:21,210][starVLA.training.train_starvla][INFO] - Step 1756, Loss: {'train/loss': 0.09265293926000595, 'train/grad_norm_pre_clip': 1.924194574356079, 'timing/data': 0.0003235023468732834, 'timing/model': 0.4244613228365779, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.32})
+[2026-07-28 10:13:27,794][starVLA.training.train_starvla][INFO] - Step 1757, Loss: {'train/loss': 0.06444346904754639, 'train/grad_norm_pre_clip': 1.1884901523590088, 'timing/data': 0.0003510559909045696, 'timing/model': 0.7308410033583641, 'learning_rate/qwen_vl_interface': 9.962063739941783e-06, 'learning_rate/action_model': 9.962063739941783e-05, 'epoch': 0.32})
+[2026-07-28 10:13:33,821][starVLA.training.train_starvla][INFO] - Step 1758, Loss: {'train/loss': 0.04132837429642677, 'train/grad_norm_pre_clip': 1.2615200281143188, 'timing/data': 0.0002953791990876198, 'timing/model': 0.4218598669394851, 'learning_rate/qwen_vl_interface': 9.961579591032791e-06, 'learning_rate/action_model': 9.96157959103279e-05, 'epoch': 0.32})
+[2026-07-28 10:13:39,905][starVLA.training.train_starvla][INFO] - Step 1759, Loss: {'train/loss': 0.03333934769034386, 'train/grad_norm_pre_clip': 1.0444419384002686, 'timing/data': 0.00034492090344429016, 'timing/model': 0.42353964410722256, 'learning_rate/qwen_vl_interface': 9.961092384830956e-06, 'learning_rate/action_model': 9.961092384830956e-05, 'epoch': 0.32})
+[2026-07-28 10:13:46,191][starVLA.training.train_starvla][INFO] - Step 1760, Loss: {'train/loss': 0.06764072924852371, 'train/grad_norm_pre_clip': 1.7677661180496216, 'timing/data': 0.00035289209336042404, 'timing/model': 0.421062754932791, 'learning_rate/qwen_vl_interface': 9.96060212165242e-06, 'learning_rate/action_model': 9.96060212165242e-05, 'epoch': 0.32})
+[2026-07-28 10:13:52,208][starVLA.training.train_starvla][INFO] - Step 1761, Loss: {'train/loss': 0.04033314436674118, 'train/grad_norm_pre_clip': 0.9212517738342285, 'timing/data': 0.0004102960228919983, 'timing/model': 0.4246390056796372, 'learning_rate/qwen_vl_interface': 9.96010880181531e-06, 'learning_rate/action_model': 9.960108801815309e-05, 'epoch': 0.32})
+[2026-07-28 10:13:58,519][starVLA.training.train_starvla][INFO] - Step 1762, Loss: {'train/loss': 0.06209105998277664, 'train/grad_norm_pre_clip': 0.8843457698822021, 'timing/data': 0.0004102718085050583, 'timing/model': 0.4210514766164124, 'learning_rate/qwen_vl_interface': 9.959612425639736e-06, 'learning_rate/action_model': 9.959612425639735e-05, 'epoch': 0.32})
+[2026-07-28 10:14:05,020][starVLA.training.train_starvla][INFO] - Step 1763, Loss: {'train/loss': 0.054275527596473694, 'train/grad_norm_pre_clip': 1.327142357826233, 'timing/data': 0.00030069099739193916, 'timing/model': 0.6824843850918114, 'learning_rate/qwen_vl_interface': 9.95911299344779e-06, 'learning_rate/action_model': 9.959112993447789e-05, 'epoch': 0.32})
+[2026-07-28 10:14:11,083][starVLA.training.train_starvla][INFO] - Step 1764, Loss: {'train/loss': 0.05203208327293396, 'train/grad_norm_pre_clip': 1.1478930711746216, 'timing/data': 0.0004607611335813999, 'timing/model': 0.42376350006088614, 'learning_rate/qwen_vl_interface': 9.958610505563546e-06, 'learning_rate/action_model': 9.958610505563546e-05, 'epoch': 0.32})
+[2026-07-28 10:14:17,188][starVLA.training.train_starvla][INFO] - Step 1765, Loss: {'train/loss': 0.056674882769584656, 'train/grad_norm_pre_clip': 0.9293444752693176, 'timing/data': 0.0003660861402750015, 'timing/model': 0.42304331669583917, 'learning_rate/qwen_vl_interface': 9.95810496231307e-06, 'learning_rate/action_model': 9.95810496231307e-05, 'epoch': 0.32})
+[2026-07-28 10:14:23,451][starVLA.training.train_starvla][INFO] - Step 1766, Loss: {'train/loss': 0.02955346554517746, 'train/grad_norm_pre_clip': 0.6331368088722229, 'timing/data': 0.0003146659582853317, 'timing/model': 0.4215379813686013, 'learning_rate/qwen_vl_interface': 9.957596364024396e-06, 'learning_rate/action_model': 9.957596364024396e-05, 'epoch': 0.32})
+[2026-07-28 10:14:29,696][starVLA.training.train_starvla][INFO] - Step 1767, Loss: {'train/loss': 0.04156632348895073, 'train/grad_norm_pre_clip': 0.8018500208854675, 'timing/data': 0.0003132098354399204, 'timing/model': 0.4224636177532375, 'learning_rate/qwen_vl_interface': 9.957084711027553e-06, 'learning_rate/action_model': 9.957084711027552e-05, 'epoch': 0.32})
+[2026-07-28 10:14:35,804][starVLA.training.train_starvla][INFO] - Step 1768, Loss: {'train/loss': 0.056521229445934296, 'train/grad_norm_pre_clip': 0.9453474879264832, 'timing/data': 0.00029959995299577713, 'timing/model': 0.42585381865501404, 'learning_rate/qwen_vl_interface': 9.956570003654544e-06, 'learning_rate/action_model': 9.956570003654544e-05, 'epoch': 0.32})
+[2026-07-28 10:14:41,646][starVLA.training.train_starvla][INFO] - Step 1769, Loss: {'train/loss': 0.05062481760978699, 'train/grad_norm_pre_clip': 1.2417948246002197, 'timing/data': 0.0002965722233057022, 'timing/model': 0.4269093810580671, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.32})
+[2026-07-28 10:14:47,686][starVLA.training.train_starvla][INFO] - Step 1770, Loss: {'train/loss': 0.10353738069534302, 'train/grad_norm_pre_clip': 3.2296736240386963, 'timing/data': 0.0003620600327849388, 'timing/model': 0.42853184789419174, 'learning_rate/qwen_vl_interface': 9.955531427117964e-06, 'learning_rate/action_model': 9.955531427117964e-05, 'epoch': 0.33})
+[2026-07-28 10:14:54,000][starVLA.training.train_starvla][INFO] - Step 1771, Loss: {'train/loss': 0.018780061975121498, 'train/grad_norm_pre_clip': 0.7434786558151245, 'timing/data': 0.0003215353935956955, 'timing/model': 0.6769254538230598, 'learning_rate/qwen_vl_interface': 9.955007558628316e-06, 'learning_rate/action_model': 9.955007558628316e-05, 'epoch': 0.33})
+[2026-07-28 10:15:00,001][starVLA.training.train_starvla][INFO] - Step 1772, Loss: {'train/loss': 0.0589161179959774, 'train/grad_norm_pre_clip': 1.5339173078536987, 'timing/data': 0.00031065288931131363, 'timing/model': 0.42167590791359544, 'learning_rate/qwen_vl_interface': 9.954480637110346e-06, 'learning_rate/action_model': 9.954480637110344e-05, 'epoch': 0.33})
+[2026-07-28 10:15:06,182][starVLA.training.train_starvla][INFO] - Step 1773, Loss: {'train/loss': 0.1104290783405304, 'train/grad_norm_pre_clip': 3.3388490676879883, 'timing/data': 0.0003272038884460926, 'timing/model': 0.5956505928188562, 'learning_rate/qwen_vl_interface': 9.953950662905963e-06, 'learning_rate/action_model': 9.953950662905964e-05, 'epoch': 0.33})
+[2026-07-28 10:15:12,180][starVLA.training.train_starvla][INFO] - Step 1774, Loss: {'train/loss': 0.057491790503263474, 'train/grad_norm_pre_clip': 1.6178194284439087, 'timing/data': 0.00046685291454195976, 'timing/model': 0.42453898675739765, 'learning_rate/qwen_vl_interface': 9.953417636359067e-06, 'learning_rate/action_model': 9.953417636359067e-05, 'epoch': 0.33})
+[2026-07-28 10:15:18,281][starVLA.training.train_starvla][INFO] - Step 1775, Loss: {'train/loss': 0.0467817597091198, 'train/grad_norm_pre_clip': 1.1194827556610107, 'timing/data': 0.0003353981301188469, 'timing/model': 0.44489728892222047, 'learning_rate/qwen_vl_interface': 9.95288155781553e-06, 'learning_rate/action_model': 9.95288155781553e-05, 'epoch': 0.33})
+[2026-07-28 10:15:24,055][starVLA.training.train_starvla][INFO] - Step 1776, Loss: {'train/loss': 0.07729928940534592, 'train/grad_norm_pre_clip': 3.4258179664611816, 'timing/data': 0.0003667897544801235, 'timing/model': 0.42380958423018456, 'learning_rate/qwen_vl_interface': 9.952342427623209e-06, 'learning_rate/action_model': 9.952342427623208e-05, 'epoch': 0.33})
+[2026-07-28 10:15:30,326][starVLA.training.train_starvla][INFO] - Step 1777, Loss: {'train/loss': 0.07279839366674423, 'train/grad_norm_pre_clip': 2.6241092681884766, 'timing/data': 0.0003340328112244606, 'timing/model': 0.4205695688724518, 'learning_rate/qwen_vl_interface': 9.95180024613194e-06, 'learning_rate/action_model': 9.951800246131939e-05, 'epoch': 0.33})
+[2026-07-28 10:15:36,668][starVLA.training.train_starvla][INFO] - Step 1778, Loss: {'train/loss': 0.04524762183427811, 'train/grad_norm_pre_clip': 0.9809848666191101, 'timing/data': 0.00036955298855900764, 'timing/model': 0.4226231570355594, 'learning_rate/qwen_vl_interface': 9.951255013693537e-06, 'learning_rate/action_model': 9.951255013693536e-05, 'epoch': 0.33})
+[2026-07-28 10:15:42,711][starVLA.training.train_starvla][INFO] - Step 1779, Loss: {'train/loss': 0.05450472608208656, 'train/grad_norm_pre_clip': 1.7811942100524902, 'timing/data': 0.00031654117628932, 'timing/model': 0.42361630499362946, 'learning_rate/qwen_vl_interface': 9.950706730661797e-06, 'learning_rate/action_model': 9.950706730661797e-05, 'epoch': 0.33})
+[2026-07-28 10:15:48,779][starVLA.training.train_starvla][INFO] - Step 1780, Loss: {'train/loss': 0.03476792946457863, 'train/grad_norm_pre_clip': 1.2240432500839233, 'timing/data': 0.0003033941611647606, 'timing/model': 0.4267566972412169, 'learning_rate/qwen_vl_interface': 9.950155397392494e-06, 'learning_rate/action_model': 9.950155397392493e-05, 'epoch': 0.33})
+[2026-07-28 10:15:55,032][starVLA.training.train_starvla][INFO] - Step 1781, Loss: {'train/loss': 0.053269218653440475, 'train/grad_norm_pre_clip': 1.651104211807251, 'timing/data': 0.0003117350861430168, 'timing/model': 0.6579114380292594, 'learning_rate/qwen_vl_interface': 9.949601014243382e-06, 'learning_rate/action_model': 9.949601014243382e-05, 'epoch': 0.33})
+[2026-07-28 10:16:01,402][starVLA.training.train_starvla][INFO] - Step 1782, Loss: {'train/loss': 0.05639917403459549, 'train/grad_norm_pre_clip': 2.641908884048462, 'timing/data': 0.0003295009955763817, 'timing/model': 0.7174296001903713, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.33})
+[2026-07-28 10:16:07,522][starVLA.training.train_starvla][INFO] - Step 1783, Loss: {'train/loss': 0.05935819447040558, 'train/grad_norm_pre_clip': 1.8262689113616943, 'timing/data': 0.0003046151250600815, 'timing/model': 0.43248269986361265, 'learning_rate/qwen_vl_interface': 9.948483099746645e-06, 'learning_rate/action_model': 9.948483099746645e-05, 'epoch': 0.33})
+[2026-07-28 10:16:13,421][starVLA.training.train_starvla][INFO] - Step 1784, Loss: {'train/loss': 0.10471805930137634, 'train/grad_norm_pre_clip': 2.3883254528045654, 'timing/data': 0.00046381400898098946, 'timing/model': 0.4247637651860714, 'learning_rate/qwen_vl_interface': 9.947919569124423e-06, 'learning_rate/action_model': 9.947919569124422e-05, 'epoch': 0.33})
+[2026-07-28 10:16:19,722][starVLA.training.train_starvla][INFO] - Step 1785, Loss: {'train/loss': 0.07972250878810883, 'train/grad_norm_pre_clip': 2.175853729248047, 'timing/data': 0.00032791867852211, 'timing/model': 0.42476249719038606, 'learning_rate/qwen_vl_interface': 9.947352990073196e-06, 'learning_rate/action_model': 9.947352990073195e-05, 'epoch': 0.33})
+[2026-07-28 10:16:26,201][starVLA.training.train_starvla][INFO] - Step 1786, Loss: {'train/loss': 0.07955534756183624, 'train/grad_norm_pre_clip': 1.949668526649475, 'timing/data': 0.0003782310523092747, 'timing/model': 0.42441536160185933, 'learning_rate/qwen_vl_interface': 9.946783362960612e-06, 'learning_rate/action_model': 9.946783362960612e-05, 'epoch': 0.33})
+[2026-07-28 10:16:32,514][starVLA.training.train_starvla][INFO] - Step 1787, Loss: {'train/loss': 0.04602938890457153, 'train/grad_norm_pre_clip': 4.768019676208496, 'timing/data': 0.00036675482988357544, 'timing/model': 0.42353585362434387, 'learning_rate/qwen_vl_interface': 9.946210688156297e-06, 'learning_rate/action_model': 9.946210688156297e-05, 'epoch': 0.33})
+[2026-07-28 10:16:38,584][starVLA.training.train_starvla][INFO] - Step 1788, Loss: {'train/loss': 0.09166228771209717, 'train/grad_norm_pre_clip': 2.041015625, 'timing/data': 0.0003390866331756115, 'timing/model': 0.423657467123121, 'learning_rate/qwen_vl_interface': 9.945634966031852e-06, 'learning_rate/action_model': 9.945634966031851e-05, 'epoch': 0.33})
+[2026-07-28 10:16:44,641][starVLA.training.train_starvla][INFO] - Step 1789, Loss: {'train/loss': 0.05506183207035065, 'train/grad_norm_pre_clip': 1.8948140144348145, 'timing/data': 0.0005435040220618248, 'timing/model': 0.42653678404167295, 'learning_rate/qwen_vl_interface': 9.945056196960858e-06, 'learning_rate/action_model': 9.945056196960858e-05, 'epoch': 0.33})
+[2026-07-28 10:16:50,999][starVLA.training.train_starvla][INFO] - Step 1790, Loss: {'train/loss': 0.0594056099653244, 'train/grad_norm_pre_clip': 1.193959355354309, 'timing/data': 0.0005023232661187649, 'timing/model': 0.42409910215064883, 'learning_rate/qwen_vl_interface': 9.944474381318871e-06, 'learning_rate/action_model': 9.94447438131887e-05, 'epoch': 0.33})
+[2026-07-28 10:16:57,313][starVLA.training.train_starvla][INFO] - Step 1791, Loss: {'train/loss': 0.027738220989704132, 'train/grad_norm_pre_clip': 1.087654948234558, 'timing/data': 0.00036154408007860184, 'timing/model': 0.4229789259843528, 'learning_rate/qwen_vl_interface': 9.943889519483424e-06, 'learning_rate/action_model': 9.943889519483424e-05, 'epoch': 0.33})
+[2026-07-28 10:17:03,356][starVLA.training.train_starvla][INFO] - Step 1792, Loss: {'train/loss': 0.05586288869380951, 'train/grad_norm_pre_clip': 2.2686426639556885, 'timing/data': 0.0004189140163362026, 'timing/model': 0.4270295328460634, 'learning_rate/qwen_vl_interface': 9.94330161183403e-06, 'learning_rate/action_model': 9.94330161183403e-05, 'epoch': 0.33})
+[2026-07-28 10:17:09,748][starVLA.training.train_starvla][INFO] - Step 1793, Loss: {'train/loss': 0.08627858012914658, 'train/grad_norm_pre_clip': 1.4195730686187744, 'timing/data': 0.0005057263188064098, 'timing/model': 0.6743026808835566, 'learning_rate/qwen_vl_interface': 9.942710658752175e-06, 'learning_rate/action_model': 9.942710658752174e-05, 'epoch': 0.33})
+[2026-07-28 10:17:16,028][starVLA.training.train_starvla][INFO] - Step 1794, Loss: {'train/loss': 0.0622701495885849, 'train/grad_norm_pre_clip': 1.8740339279174805, 'timing/data': 0.0005417629145085812, 'timing/model': 0.6597256301902235, 'learning_rate/qwen_vl_interface': 9.942116660621319e-06, 'learning_rate/action_model': 9.94211666062132e-05, 'epoch': 0.33})
+[2026-07-28 10:17:22,061][starVLA.training.train_starvla][INFO] - Step 1795, Loss: {'train/loss': 0.06189720705151558, 'train/grad_norm_pre_clip': 2.978851795196533, 'timing/data': 0.0005368138663470745, 'timing/model': 0.42666127206757665, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.33})
+[2026-07-28 10:17:28,121][starVLA.training.train_starvla][INFO] - Step 1796, Loss: {'train/loss': 0.052430346608161926, 'train/grad_norm_pre_clip': 2.114985466003418, 'timing/data': 0.00029608188197016716, 'timing/model': 0.4238804061897099, 'learning_rate/qwen_vl_interface': 9.940919530756347e-06, 'learning_rate/action_model': 9.940919530756347e-05, 'epoch': 0.33})
+[2026-07-28 10:17:34,165][starVLA.training.train_starvla][INFO] - Step 1797, Loss: {'train/loss': 0.06659907847642899, 'train/grad_norm_pre_clip': 0.9828266501426697, 'timing/data': 0.0003149043768644333, 'timing/model': 0.42622282402589917, 'learning_rate/qwen_vl_interface': 9.940316399799032e-06, 'learning_rate/action_model': 9.940316399799033e-05, 'epoch': 0.33})
+[2026-07-28 10:17:40,268][starVLA.training.train_starvla][INFO] - Step 1798, Loss: {'train/loss': 0.044170014560222626, 'train/grad_norm_pre_clip': 0.8148400187492371, 'timing/data': 0.0003487127833068371, 'timing/model': 0.6924758777022362, 'learning_rate/qwen_vl_interface': 9.939710225346329e-06, 'learning_rate/action_model': 9.939710225346329e-05, 'epoch': 0.33})
+[2026-07-28 10:17:46,068][starVLA.training.train_starvla][INFO] - Step 1799, Loss: {'train/loss': 0.059702854603528976, 'train/grad_norm_pre_clip': 1.375456690788269, 'timing/data': 0.00034295208752155304, 'timing/model': 0.4249004297889769, 'learning_rate/qwen_vl_interface': 9.939101007791575e-06, 'learning_rate/action_model': 9.939101007791575e-05, 'epoch': 0.33})
+[2026-07-28 10:17:51,899][starVLA.training.train_starvla][INFO] - Step 1800, Loss: {'train/loss': 0.0828956812620163, 'train/grad_norm_pre_clip': 2.990295648574829, 'timing/data': 0.0004697060212492943, 'timing/model': 0.4252127967774868, 'learning_rate/qwen_vl_interface': 9.938488747530087e-06, 'learning_rate/action_model': 9.938488747530086e-05, 'epoch': 0.33})
+[2026-07-28 10:17:57,995][starVLA.training.train_starvla][INFO] - Step 1801, Loss: {'train/loss': 0.032971128821372986, 'train/grad_norm_pre_clip': 1.1791127920150757, 'timing/data': 0.00040526827797293663, 'timing/model': 0.42889728071168065, 'learning_rate/qwen_vl_interface': 9.937873444959151e-06, 'learning_rate/action_model': 9.937873444959151e-05, 'epoch': 0.33})
+[2026-07-28 10:18:04,121][starVLA.training.train_starvla][INFO] - Step 1802, Loss: {'train/loss': 0.07865412533283234, 'train/grad_norm_pre_clip': 1.7290927171707153, 'timing/data': 0.0003275289200246334, 'timing/model': 0.4240163532085717, 'learning_rate/qwen_vl_interface': 9.937255100478036e-06, 'learning_rate/action_model': 9.937255100478036e-05, 'epoch': 0.33})
+[2026-07-28 10:18:10,179][starVLA.training.train_starvla][INFO] - Step 1803, Loss: {'train/loss': 0.09033013880252838, 'train/grad_norm_pre_clip': 1.9751958847045898, 'timing/data': 0.00031916098669171333, 'timing/model': 0.4261747729033232, 'learning_rate/qwen_vl_interface': 9.93663371448797e-06, 'learning_rate/action_model': 9.93663371448797e-05, 'epoch': 0.33})
+[2026-07-28 10:18:16,751][starVLA.training.train_starvla][INFO] - Step 1804, Loss: {'train/loss': 0.062476709485054016, 'train/grad_norm_pre_clip': 1.288468837738037, 'timing/data': 0.0003137192688882351, 'timing/model': 0.6714312289841473, 'learning_rate/qwen_vl_interface': 9.936009287392173e-06, 'learning_rate/action_model': 9.936009287392171e-05, 'epoch': 0.33})
+[2026-07-28 10:18:22,786][starVLA.training.train_starvla][INFO] - Step 1805, Loss: {'train/loss': 0.032481834292411804, 'train/grad_norm_pre_clip': 1.2517536878585815, 'timing/data': 0.00035385508090257645, 'timing/model': 0.43233284028247, 'learning_rate/qwen_vl_interface': 9.935381819595822e-06, 'learning_rate/action_model': 9.935381819595822e-05, 'epoch': 0.33})
+[2026-07-28 10:18:28,831][starVLA.training.train_starvla][INFO] - Step 1806, Loss: {'train/loss': 0.048565540462732315, 'train/grad_norm_pre_clip': 1.192064642906189, 'timing/data': 0.0003519519232213497, 'timing/model': 0.4367275536060333, 'learning_rate/qwen_vl_interface': 9.934751311506078e-06, 'learning_rate/action_model': 9.934751311506077e-05, 'epoch': 0.33})
+[2026-07-28 10:18:34,946][starVLA.training.train_starvla][INFO] - Step 1807, Loss: {'train/loss': 0.08591201156377792, 'train/grad_norm_pre_clip': 1.9969197511672974, 'timing/data': 0.00032238196581602097, 'timing/model': 0.4247289700433612, 'learning_rate/qwen_vl_interface': 9.93411776353207e-06, 'learning_rate/action_model': 9.934117763532069e-05, 'epoch': 0.33})
+[2026-07-28 10:18:41,254][starVLA.training.train_starvla][INFO] - Step 1808, Loss: {'train/loss': 0.04362190514802933, 'train/grad_norm_pre_clip': 1.4956510066986084, 'timing/data': 0.0006106458604335785, 'timing/model': 0.431389847304672, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.33})
+[2026-07-28 10:18:47,602][starVLA.training.train_starvla][INFO] - Step 1809, Loss: {'train/loss': 0.032431334257125854, 'train/grad_norm_pre_clip': 1.2935583591461182, 'timing/data': 0.00044568581506609917, 'timing/model': 0.4217440979555249, 'learning_rate/qwen_vl_interface': 9.932841549577642e-06, 'learning_rate/action_model': 9.932841549577641e-05, 'epoch': 0.33})
+[2026-07-28 10:18:53,739][starVLA.training.train_starvla][INFO] - Step 1810, Loss: {'train/loss': 0.07547641545534134, 'train/grad_norm_pre_clip': 2.363976240158081, 'timing/data': 0.00028521986678242683, 'timing/model': 0.42982553876936436, 'learning_rate/qwen_vl_interface': 9.932198884425343e-06, 'learning_rate/action_model': 9.932198884425342e-05, 'epoch': 0.33})
+[2026-07-28 10:19:00,144][starVLA.training.train_starvla][INFO] - Step 1811, Loss: {'train/loss': 0.08983497321605682, 'train/grad_norm_pre_clip': 3.838650703430176, 'timing/data': 0.00043177278712391853, 'timing/model': 0.427312140353024, 'learning_rate/qwen_vl_interface': 9.931553181045022e-06, 'learning_rate/action_model': 9.931553181045022e-05, 'epoch': 0.33})
+[2026-07-28 10:19:06,268][starVLA.training.train_starvla][INFO] - Step 1812, Loss: {'train/loss': 0.06867733597755432, 'train/grad_norm_pre_clip': 3.480620861053467, 'timing/data': 0.0003208122216165066, 'timing/model': 0.4269361370243132, 'learning_rate/qwen_vl_interface': 9.930904439855667e-06, 'learning_rate/action_model': 9.930904439855667e-05, 'epoch': 0.33})
+[2026-07-28 10:19:12,655][starVLA.training.train_starvla][INFO] - Step 1813, Loss: {'train/loss': 0.04616333916783333, 'train/grad_norm_pre_clip': 1.765473484992981, 'timing/data': 0.00039977673441171646, 'timing/model': 0.42283060820773244, 'learning_rate/qwen_vl_interface': 9.930252661278245e-06, 'learning_rate/action_model': 9.930252661278244e-05, 'epoch': 0.33})
+[2026-07-28 10:19:18,955][starVLA.training.train_starvla][INFO] - Step 1814, Loss: {'train/loss': 0.03202205151319504, 'train/grad_norm_pre_clip': 1.5079761743545532, 'timing/data': 0.0004427400417625904, 'timing/model': 0.6006334698759019, 'learning_rate/qwen_vl_interface': 9.92959784573568e-06, 'learning_rate/action_model': 9.92959784573568e-05, 'epoch': 0.33})
+[2026-07-28 10:19:25,056][starVLA.training.train_starvla][INFO] - Step 1815, Loss: {'train/loss': 0.05928872898221016, 'train/grad_norm_pre_clip': 1.5766295194625854, 'timing/data': 0.0004427810199558735, 'timing/model': 0.4258097978308797, 'learning_rate/qwen_vl_interface': 9.92893999365288e-06, 'learning_rate/action_model': 9.92893999365288e-05, 'epoch': 0.33})
+[2026-07-28 10:19:31,104][starVLA.training.train_starvla][INFO] - Step 1816, Loss: {'train/loss': 0.03581566363573074, 'train/grad_norm_pre_clip': 1.669999122619629, 'timing/data': 0.00033653294667601585, 'timing/model': 0.42617935920134187, 'learning_rate/qwen_vl_interface': 9.928279105456717e-06, 'learning_rate/action_model': 9.928279105456718e-05, 'epoch': 0.33})
+[2026-07-28 10:19:37,289][starVLA.training.train_starvla][INFO] - Step 1817, Loss: {'train/loss': 0.09867598116397858, 'train/grad_norm_pre_clip': 2.8343725204467773, 'timing/data': 0.00037073297426104546, 'timing/model': 0.4263898693025112, 'learning_rate/qwen_vl_interface': 9.927615181576034e-06, 'learning_rate/action_model': 9.927615181576034e-05, 'epoch': 0.33})
+[2026-07-28 10:19:43,633][starVLA.training.train_starvla][INFO] - Step 1818, Loss: {'train/loss': 0.09068372845649719, 'train/grad_norm_pre_clip': 4.27050256729126, 'timing/data': 0.0004910030402243137, 'timing/model': 0.7296752543188632, 'learning_rate/qwen_vl_interface': 9.926948222441644e-06, 'learning_rate/action_model': 9.926948222441643e-05, 'epoch': 0.33})
+[2026-07-28 10:19:49,736][starVLA.training.train_starvla][INFO] - Step 1819, Loss: {'train/loss': 0.026206104084849358, 'train/grad_norm_pre_clip': 1.8860574960708618, 'timing/data': 0.00038235681131482124, 'timing/model': 0.4227672880515456, 'learning_rate/qwen_vl_interface': 9.926278228486328e-06, 'learning_rate/action_model': 9.926278228486327e-05, 'epoch': 0.33})
+[2026-07-28 10:19:55,831][starVLA.training.train_starvla][INFO] - Step 1820, Loss: {'train/loss': 0.10951017588376999, 'train/grad_norm_pre_clip': 4.19132661819458, 'timing/data': 0.0003803796134889126, 'timing/model': 0.4244516082108021, 'learning_rate/qwen_vl_interface': 9.92560520014484e-06, 'learning_rate/action_model': 9.92560520014484e-05, 'epoch': 0.33})
+[2026-07-28 10:20:01,886][starVLA.training.train_starvla][INFO] - Step 1821, Loss: {'train/loss': 0.043549828231334686, 'train/grad_norm_pre_clip': 1.781731128692627, 'timing/data': 0.00044763507321476936, 'timing/model': 0.4275367436930537, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.33})
+[2026-07-28 10:20:08,311][starVLA.training.train_starvla][INFO] - Step 1822, Loss: {'train/loss': 0.07160685211420059, 'train/grad_norm_pre_clip': 1.700278878211975, 'timing/data': 0.0003930996172130108, 'timing/model': 0.4228481990285218, 'learning_rate/qwen_vl_interface': 9.924250042052197e-06, 'learning_rate/action_model': 9.924250042052197e-05, 'epoch': 0.33})
+[2026-07-28 10:20:14,374][starVLA.training.train_starvla][INFO] - Step 1823, Loss: {'train/loss': 0.08147724717855453, 'train/grad_norm_pre_clip': 2.377627372741699, 'timing/data': 0.00038745487108826637, 'timing/model': 0.4272818360477686, 'learning_rate/qwen_vl_interface': 9.92356791318039e-06, 'learning_rate/action_model': 9.92356791318039e-05, 'epoch': 0.33})
+[2026-07-28 10:20:20,454][starVLA.training.train_starvla][INFO] - Step 1824, Loss: {'train/loss': 0.10768604278564453, 'train/grad_norm_pre_clip': 3.381596088409424, 'timing/data': 0.0003845030441880226, 'timing/model': 0.42511688405647874, 'learning_rate/qwen_vl_interface': 9.922882751681104e-06, 'learning_rate/action_model': 9.922882751681104e-05, 'epoch': 0.34})
+[2026-07-28 10:20:26,701][starVLA.training.train_starvla][INFO] - Step 1825, Loss: {'train/loss': 0.0535607747733593, 'train/grad_norm_pre_clip': 2.0908148288726807, 'timing/data': 0.00033953506499528885, 'timing/model': 0.4259693161584437, 'learning_rate/qwen_vl_interface': 9.922194557998932e-06, 'learning_rate/action_model': 9.922194557998933e-05, 'epoch': 0.34})
+[2026-07-28 10:20:32,765][starVLA.training.train_starvla][INFO] - Step 1826, Loss: {'train/loss': 0.055263496935367584, 'train/grad_norm_pre_clip': 2.6937501430511475, 'timing/data': 0.00034340983256697655, 'timing/model': 0.42612553713843226, 'learning_rate/qwen_vl_interface': 9.921503332580439e-06, 'learning_rate/action_model': 9.921503332580439e-05, 'epoch': 0.34})
+[2026-07-28 10:20:39,060][starVLA.training.train_starvla][INFO] - Step 1827, Loss: {'train/loss': 0.07256806641817093, 'train/grad_norm_pre_clip': 1.778212308883667, 'timing/data': 0.00030946871265769005, 'timing/model': 0.6624164236709476, 'learning_rate/qwen_vl_interface': 9.92080907587415e-06, 'learning_rate/action_model': 9.92080907587415e-05, 'epoch': 0.34})
+[2026-07-28 10:20:45,136][starVLA.training.train_starvla][INFO] - Step 1828, Loss: {'train/loss': 0.06381634622812271, 'train/grad_norm_pre_clip': 1.4697840213775635, 'timing/data': 0.0004836348816752434, 'timing/model': 0.42254535388201475, 'learning_rate/qwen_vl_interface': 9.92011178833056e-06, 'learning_rate/action_model': 9.92011178833056e-05, 'epoch': 0.34})
+[2026-07-28 10:20:51,411][starVLA.training.train_starvla][INFO] - Step 1829, Loss: {'train/loss': 0.05921435356140137, 'train/grad_norm_pre_clip': 0.8527711033821106, 'timing/data': 0.0003427439369261265, 'timing/model': 0.6652009440585971, 'learning_rate/qwen_vl_interface': 9.919411470402136e-06, 'learning_rate/action_model': 9.919411470402136e-05, 'epoch': 0.34})
+[2026-07-28 10:20:57,436][starVLA.training.train_starvla][INFO] - Step 1830, Loss: {'train/loss': 0.04296012967824936, 'train/grad_norm_pre_clip': 1.2641692161560059, 'timing/data': 0.00037945667281746864, 'timing/model': 0.42373683396726847, 'learning_rate/qwen_vl_interface': 9.918708122543305e-06, 'learning_rate/action_model': 9.918708122543305e-05, 'epoch': 0.34})
+[2026-07-28 10:21:03,490][starVLA.training.train_starvla][INFO] - Step 1831, Loss: {'train/loss': 0.062456753104925156, 'train/grad_norm_pre_clip': 3.138751268386841, 'timing/data': 0.00035127531737089157, 'timing/model': 0.42740875389426947, 'learning_rate/qwen_vl_interface': 9.918001745210458e-06, 'learning_rate/action_model': 9.918001745210456e-05, 'epoch': 0.34})
+[2026-07-28 10:21:09,754][starVLA.training.train_starvla][INFO] - Step 1832, Loss: {'train/loss': 0.12641561031341553, 'train/grad_norm_pre_clip': 3.8114402294158936, 'timing/data': 0.0003718719817698002, 'timing/model': 0.4231051942333579, 'learning_rate/qwen_vl_interface': 9.917292338861958e-06, 'learning_rate/action_model': 9.917292338861957e-05, 'epoch': 0.34})
+[2026-07-28 10:21:16,008][starVLA.training.train_starvla][INFO] - Step 1833, Loss: {'train/loss': 0.04257989674806595, 'train/grad_norm_pre_clip': 1.687503457069397, 'timing/data': 0.0004589138552546501, 'timing/model': 0.42355516692623496, 'learning_rate/qwen_vl_interface': 9.916579903958134e-06, 'learning_rate/action_model': 9.916579903958133e-05, 'epoch': 0.34})
+[2026-07-28 10:21:22,035][starVLA.training.train_starvla][INFO] - Step 1834, Loss: {'train/loss': 0.042653199285268784, 'train/grad_norm_pre_clip': 0.8320173621177673, 'timing/data': 0.00035994313657283783, 'timing/model': 0.42836570739746094, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.34})
+[2026-07-28 10:21:28,428][starVLA.training.train_starvla][INFO] - Step 1835, Loss: {'train/loss': 0.07910460978746414, 'train/grad_norm_pre_clip': 1.9745514392852783, 'timing/data': 0.000346112996339798, 'timing/model': 0.42937607876956463, 'learning_rate/qwen_vl_interface': 9.915145950335631e-06, 'learning_rate/action_model': 9.915145950335632e-05, 'epoch': 0.34})
+[2026-07-28 10:21:34,546][starVLA.training.train_starvla][INFO] - Step 1836, Loss: {'train/loss': 0.0561302974820137, 'train/grad_norm_pre_clip': 2.367196798324585, 'timing/data': 0.0003086607903242111, 'timing/model': 0.42409286741167307, 'learning_rate/qwen_vl_interface': 9.914424432547434e-06, 'learning_rate/action_model': 9.914424432547433e-05, 'epoch': 0.34})
+[2026-07-28 10:21:40,796][starVLA.training.train_starvla][INFO] - Step 1837, Loss: {'train/loss': 0.07267864793539047, 'train/grad_norm_pre_clip': 1.919770359992981, 'timing/data': 0.0003452827222645283, 'timing/model': 0.42400024086236954, 'learning_rate/qwen_vl_interface': 9.913699888064862e-06, 'learning_rate/action_model': 9.913699888064861e-05, 'epoch': 0.34})
+[2026-07-28 10:21:46,857][starVLA.training.train_starvla][INFO] - Step 1838, Loss: {'train/loss': 0.09360707551240921, 'train/grad_norm_pre_clip': 1.6807708740234375, 'timing/data': 0.0003806650638580322, 'timing/model': 0.42330888379365206, 'learning_rate/qwen_vl_interface': 9.912972317358066e-06, 'learning_rate/action_model': 9.912972317358066e-05, 'epoch': 0.34})
+[2026-07-28 10:21:53,169][starVLA.training.train_starvla][INFO] - Step 1839, Loss: {'train/loss': 0.09978073835372925, 'train/grad_norm_pre_clip': 3.5687310695648193, 'timing/data': 0.00031991535797715187, 'timing/model': 0.6620852877385914, 'learning_rate/qwen_vl_interface': 9.912241720899157e-06, 'learning_rate/action_model': 9.912241720899157e-05, 'epoch': 0.34})
+[2026-07-28 10:21:59,453][starVLA.training.train_starvla][INFO] - Step 1840, Loss: {'train/loss': 0.05012580379843712, 'train/grad_norm_pre_clip': 1.7838373184204102, 'timing/data': 0.00030612386763095856, 'timing/model': 0.6645263191312551, 'learning_rate/qwen_vl_interface': 9.911508099162215e-06, 'learning_rate/action_model': 9.911508099162214e-05, 'epoch': 0.34})
+[2026-07-28 10:22:05,488][starVLA.training.train_starvla][INFO] - Step 1841, Loss: {'train/loss': 0.04686836153268814, 'train/grad_norm_pre_clip': 0.9207471609115601, 'timing/data': 0.0002988409250974655, 'timing/model': 0.42445188062265515, 'learning_rate/qwen_vl_interface': 9.910771452623276e-06, 'learning_rate/action_model': 9.910771452623275e-05, 'epoch': 0.34})
+[2026-07-28 10:22:11,274][starVLA.training.train_starvla][INFO] - Step 1842, Loss: {'train/loss': 0.06839116662740707, 'train/grad_norm_pre_clip': 1.5422166585922241, 'timing/data': 0.0004691239446401596, 'timing/model': 0.42311950819566846, 'learning_rate/qwen_vl_interface': 9.910031781760342e-06, 'learning_rate/action_model': 9.910031781760342e-05, 'epoch': 0.34})
+[2026-07-28 10:22:17,350][starVLA.training.train_starvla][INFO] - Step 1843, Loss: {'train/loss': 0.033985015004873276, 'train/grad_norm_pre_clip': 0.6443592309951782, 'timing/data': 0.0003027748316526413, 'timing/model': 0.4248057808727026, 'learning_rate/qwen_vl_interface': 9.909289087053378e-06, 'learning_rate/action_model': 9.909289087053378e-05, 'epoch': 0.34})
+[2026-07-28 10:22:23,387][starVLA.training.train_starvla][INFO] - Step 1844, Loss: {'train/loss': 0.05847448855638504, 'train/grad_norm_pre_clip': 1.8301607370376587, 'timing/data': 0.0003542369231581688, 'timing/model': 0.4349620300345123, 'learning_rate/qwen_vl_interface': 9.90854336898431e-06, 'learning_rate/action_model': 9.90854336898431e-05, 'epoch': 0.34})
+[2026-07-28 10:22:29,874][starVLA.training.train_starvla][INFO] - Step 1845, Loss: {'train/loss': 0.0525696836411953, 'train/grad_norm_pre_clip': 1.247010588645935, 'timing/data': 0.00043777283281087875, 'timing/model': 0.6713714939542115, 'learning_rate/qwen_vl_interface': 9.907794628037028e-06, 'learning_rate/action_model': 9.907794628037028e-05, 'epoch': 0.34})
+[2026-07-28 10:22:35,872][starVLA.training.train_starvla][INFO] - Step 1846, Loss: {'train/loss': 0.038429901003837585, 'train/grad_norm_pre_clip': 0.9393904805183411, 'timing/data': 0.0004374431446194649, 'timing/model': 0.42329706717282534, 'learning_rate/qwen_vl_interface': 9.90704286469738e-06, 'learning_rate/action_model': 9.90704286469738e-05, 'epoch': 0.34})
+[2026-07-28 10:22:41,959][starVLA.training.train_starvla][INFO] - Step 1847, Loss: {'train/loss': 0.09018057584762573, 'train/grad_norm_pre_clip': 2.32307767868042, 'timing/data': 0.00037803594022989273, 'timing/model': 0.42522220220416784, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.34})
+[2026-07-28 10:22:47,983][starVLA.training.train_starvla][INFO] - Step 1848, Loss: {'train/loss': 0.05415170639753342, 'train/grad_norm_pre_clip': 2.9413743019104004, 'timing/data': 0.00030283303931355476, 'timing/model': 0.4250147929415107, 'learning_rate/qwen_vl_interface': 9.905530272794195e-06, 'learning_rate/action_model': 9.905530272794195e-05, 'epoch': 0.34})
+[2026-07-28 10:22:54,306][starVLA.training.train_starvla][INFO] - Step 1849, Loss: {'train/loss': 0.06903346627950668, 'train/grad_norm_pre_clip': 1.819405436515808, 'timing/data': 0.00030073896050453186, 'timing/model': 0.4271893408149481, 'learning_rate/qwen_vl_interface': 9.904769445212163e-06, 'learning_rate/action_model': 9.904769445212162e-05, 'epoch': 0.34})
+[2026-07-28 10:23:00,356][starVLA.training.train_starvla][INFO] - Step 1850, Loss: {'train/loss': 0.040270570665597916, 'train/grad_norm_pre_clip': 1.258049726486206, 'timing/data': 0.0003224089741706848, 'timing/model': 0.43007550574839115, 'learning_rate/qwen_vl_interface': 9.904005597200772e-06, 'learning_rate/action_model': 9.904005597200772e-05, 'epoch': 0.34})
+[2026-07-28 10:23:06,703][starVLA.training.train_starvla][INFO] - Step 1851, Loss: {'train/loss': 0.03895862028002739, 'train/grad_norm_pre_clip': 1.367684245109558, 'timing/data': 0.0003069210797548294, 'timing/model': 0.6816493500955403, 'learning_rate/qwen_vl_interface': 9.903238729255678e-06, 'learning_rate/action_model': 9.903238729255677e-05, 'epoch': 0.34})
+[2026-07-28 10:23:13,005][starVLA.training.train_starvla][INFO] - Step 1852, Loss: {'train/loss': 0.05879060924053192, 'train/grad_norm_pre_clip': 1.377644419670105, 'timing/data': 0.00036203116178512573, 'timing/model': 0.4244163231924176, 'learning_rate/qwen_vl_interface': 9.902468841874492e-06, 'learning_rate/action_model': 9.902468841874492e-05, 'epoch': 0.34})
+[2026-07-28 10:23:18,868][starVLA.training.train_starvla][INFO] - Step 1853, Loss: {'train/loss': 0.06561672687530518, 'train/grad_norm_pre_clip': 1.2074486017227173, 'timing/data': 0.00038312654942274094, 'timing/model': 0.42414939403533936, 'learning_rate/qwen_vl_interface': 9.901695935556784e-06, 'learning_rate/action_model': 9.901695935556783e-05, 'epoch': 0.34})
+[2026-07-28 10:23:24,938][starVLA.training.train_starvla][INFO] - Step 1854, Loss: {'train/loss': 0.043251026421785355, 'train/grad_norm_pre_clip': 1.5125728845596313, 'timing/data': 0.0003214692696928978, 'timing/model': 0.427354933694005, 'learning_rate/qwen_vl_interface': 9.900920010804088e-06, 'learning_rate/action_model': 9.900920010804088e-05, 'epoch': 0.34})
+[2026-07-28 10:23:30,995][starVLA.training.train_starvla][INFO] - Step 1855, Loss: {'train/loss': 0.047026921063661575, 'train/grad_norm_pre_clip': 1.009953260421753, 'timing/data': 0.0005837441422045231, 'timing/model': 0.4265364711172879, 'learning_rate/qwen_vl_interface': 9.900141068119891e-06, 'learning_rate/action_model': 9.900141068119891e-05, 'epoch': 0.34})
+[2026-07-28 10:23:37,508][starVLA.training.train_starvla][INFO] - Step 1856, Loss: {'train/loss': 0.021350625902414322, 'train/grad_norm_pre_clip': 0.9583354592323303, 'timing/data': 0.00030166469514369965, 'timing/model': 0.421641465742141, 'learning_rate/qwen_vl_interface': 9.89935910800964e-06, 'learning_rate/action_model': 9.89935910800964e-05, 'epoch': 0.34})
+[2026-07-28 10:23:43,595][starVLA.training.train_starvla][INFO] - Step 1857, Loss: {'train/loss': 0.050751909613609314, 'train/grad_norm_pre_clip': 1.1808953285217285, 'timing/data': 0.00032022781670093536, 'timing/model': 0.4270564168691635, 'learning_rate/qwen_vl_interface': 9.898574130980743e-06, 'learning_rate/action_model': 9.898574130980742e-05, 'epoch': 0.34})
+[2026-07-28 10:23:49,645][starVLA.training.train_starvla][INFO] - Step 1858, Loss: {'train/loss': 0.039925072342157364, 'train/grad_norm_pre_clip': 1.0642036199569702, 'timing/data': 0.00033015618100762367, 'timing/model': 0.42795853689312935, 'learning_rate/qwen_vl_interface': 9.89778613754256e-06, 'learning_rate/action_model': 9.89778613754256e-05, 'epoch': 0.34})
+[2026-07-28 10:23:55,703][starVLA.training.train_starvla][INFO] - Step 1859, Loss: {'train/loss': 0.04217544198036194, 'train/grad_norm_pre_clip': 1.9247254133224487, 'timing/data': 0.00029753614217042923, 'timing/model': 0.42559715593233705, 'learning_rate/qwen_vl_interface': 9.896995128206412e-06, 'learning_rate/action_model': 9.896995128206411e-05, 'epoch': 0.34})
+[2026-07-28 10:24:01,768][starVLA.training.train_starvla][INFO] - Step 1860, Loss: {'train/loss': 0.09022928774356842, 'train/grad_norm_pre_clip': 2.883169174194336, 'timing/data': 0.00038488488644361496, 'timing/model': 0.4216691763140261, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.34})
+[2026-07-28 10:24:07,852][starVLA.training.train_starvla][INFO] - Step 1861, Loss: {'train/loss': 0.08112961798906326, 'train/grad_norm_pre_clip': 3.2805652618408203, 'timing/data': 0.00032597919926047325, 'timing/model': 0.42774401837959886, 'learning_rate/qwen_vl_interface': 9.89540406389529e-06, 'learning_rate/action_model': 9.89540406389529e-05, 'epoch': 0.34})
+[2026-07-28 10:24:13,902][starVLA.training.train_starvla][INFO] - Step 1862, Loss: {'train/loss': 0.07180286943912506, 'train/grad_norm_pre_clip': 1.6035128831863403, 'timing/data': 0.00037151994183659554, 'timing/model': 0.4225463792681694, 'learning_rate/qwen_vl_interface': 9.894604009952742e-06, 'learning_rate/action_model': 9.89460400995274e-05, 'epoch': 0.34})
+[2026-07-28 10:24:19,888][starVLA.training.train_starvla][INFO] - Step 1863, Loss: {'train/loss': 0.02392156794667244, 'train/grad_norm_pre_clip': 1.0196858644485474, 'timing/data': 0.00030114268884062767, 'timing/model': 0.42243445105850697, 'learning_rate/qwen_vl_interface': 9.893800942177073e-06, 'learning_rate/action_model': 9.893800942177073e-05, 'epoch': 0.34})
+[2026-07-28 10:24:25,883][starVLA.training.train_starvla][INFO] - Step 1864, Loss: {'train/loss': 0.0633845329284668, 'train/grad_norm_pre_clip': 1.3080865144729614, 'timing/data': 0.00038289185613393784, 'timing/model': 0.4234371446073055, 'learning_rate/qwen_vl_interface': 9.892994861089393e-06, 'learning_rate/action_model': 9.892994861089393e-05, 'epoch': 0.34})
+[2026-07-28 10:24:32,111][starVLA.training.train_starvla][INFO] - Step 1865, Loss: {'train/loss': 0.05031196027994156, 'train/grad_norm_pre_clip': 2.017808437347412, 'timing/data': 0.00037696631625294685, 'timing/model': 0.6668298640288413, 'learning_rate/qwen_vl_interface': 9.892185767212756e-06, 'learning_rate/action_model': 9.892185767212756e-05, 'epoch': 0.34})
+[2026-07-28 10:24:37,880][starVLA.training.train_starvla][INFO] - Step 1866, Loss: {'train/loss': 0.024944836273789406, 'train/grad_norm_pre_clip': 1.047509789466858, 'timing/data': 0.0003084661439061165, 'timing/model': 0.42267440166324377, 'learning_rate/qwen_vl_interface': 9.891373661072172e-06, 'learning_rate/action_model': 9.891373661072173e-05, 'epoch': 0.34})
+[2026-07-28 10:24:43,892][starVLA.training.train_starvla][INFO] - Step 1867, Loss: {'train/loss': 0.027923062443733215, 'train/grad_norm_pre_clip': 1.5076169967651367, 'timing/data': 0.0003613377921283245, 'timing/model': 0.4241412109695375, 'learning_rate/qwen_vl_interface': 9.89055854319461e-06, 'learning_rate/action_model': 9.890558543194611e-05, 'epoch': 0.34})
+[2026-07-28 10:24:49,903][starVLA.training.train_starvla][INFO] - Step 1868, Loss: {'train/loss': 0.06940431892871857, 'train/grad_norm_pre_clip': 2.410780906677246, 'timing/data': 0.00039608636870980263, 'timing/model': 0.4212213479913771, 'learning_rate/qwen_vl_interface': 9.889740414108994e-06, 'learning_rate/action_model': 9.889740414108995e-05, 'epoch': 0.34})
+[2026-07-28 10:24:55,901][starVLA.training.train_starvla][INFO] - Step 1869, Loss: {'train/loss': 0.050077375024557114, 'train/grad_norm_pre_clip': 1.9548120498657227, 'timing/data': 0.00030886009335517883, 'timing/model': 0.4231392308138311, 'learning_rate/qwen_vl_interface': 9.888919274346196e-06, 'learning_rate/action_model': 9.888919274346195e-05, 'epoch': 0.34})
+[2026-07-28 10:25:02,185][starVLA.training.train_starvla][INFO] - Step 1870, Loss: {'train/loss': 0.049951180815696716, 'train/grad_norm_pre_clip': 1.4102321863174438, 'timing/data': 0.00030195293948054314, 'timing/model': 0.44094624975696206, 'learning_rate/qwen_vl_interface': 9.888095124439043e-06, 'learning_rate/action_model': 9.888095124439043e-05, 'epoch': 0.34})
+[2026-07-28 10:25:08,235][starVLA.training.train_starvla][INFO] - Step 1871, Loss: {'train/loss': 0.06020103022456169, 'train/grad_norm_pre_clip': 3.0741446018218994, 'timing/data': 0.00036852527409791946, 'timing/model': 0.424024302046746, 'learning_rate/qwen_vl_interface': 9.88726796492232e-06, 'learning_rate/action_model': 9.88726796492232e-05, 'epoch': 0.34})
+[2026-07-28 10:25:14,499][starVLA.training.train_starvla][INFO] - Step 1872, Loss: {'train/loss': 0.045983634889125824, 'train/grad_norm_pre_clip': 1.1685757637023926, 'timing/data': 0.0003602178767323494, 'timing/model': 0.42115516727790236, 'learning_rate/qwen_vl_interface': 9.886437796332759e-06, 'learning_rate/action_model': 9.88643779633276e-05, 'epoch': 0.34})
+[2026-07-28 10:25:20,971][starVLA.training.train_starvla][INFO] - Step 1873, Loss: {'train/loss': 0.053254105150699615, 'train/grad_norm_pre_clip': 1.2247940301895142, 'timing/data': 0.0003007729537785053, 'timing/model': 0.42086326517164707, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.34})
+[2026-07-28 10:25:27,024][starVLA.training.train_starvla][INFO] - Step 1874, Loss: {'train/loss': 0.041065558791160583, 'train/grad_norm_pre_clip': 1.1478710174560547, 'timing/data': 0.0003547007218003273, 'timing/model': 0.44169955933466554, 'learning_rate/qwen_vl_interface': 9.884768434091832e-06, 'learning_rate/action_model': 9.884768434091832e-05, 'epoch': 0.34})
+[2026-07-28 10:25:33,045][starVLA.training.train_starvla][INFO] - Step 1875, Loss: {'train/loss': 0.03183490037918091, 'train/grad_norm_pre_clip': 0.8856268525123596, 'timing/data': 0.0003671380691230297, 'timing/model': 0.4314907039515674, 'learning_rate/qwen_vl_interface': 9.883929241523698e-06, 'learning_rate/action_model': 9.883929241523697e-05, 'epoch': 0.34})
+[2026-07-28 10:25:39,295][starVLA.training.train_starvla][INFO] - Step 1876, Loss: {'train/loss': 0.06812892109155655, 'train/grad_norm_pre_clip': 1.528984785079956, 'timing/data': 0.00036436691880226135, 'timing/model': 0.6705265450291336, 'learning_rate/qwen_vl_interface': 9.883087042049185e-06, 'learning_rate/action_model': 9.883087042049186e-05, 'epoch': 0.34})
+[2026-07-28 10:25:45,531][starVLA.training.train_starvla][INFO] - Step 1877, Loss: {'train/loss': 0.06981505453586578, 'train/grad_norm_pre_clip': 1.170212745666504, 'timing/data': 0.0003774566575884819, 'timing/model': 0.6595100490376353, 'learning_rate/qwen_vl_interface': 9.882241836214794e-06, 'learning_rate/action_model': 9.882241836214794e-05, 'epoch': 0.34})
+[2026-07-28 10:25:51,536][starVLA.training.train_starvla][INFO] - Step 1878, Loss: {'train/loss': 0.07726119458675385, 'train/grad_norm_pre_clip': 1.4200444221496582, 'timing/data': 0.00042078597471117973, 'timing/model': 0.4305093348957598, 'learning_rate/qwen_vl_interface': 9.881393624568965e-06, 'learning_rate/action_model': 9.881393624568965e-05, 'epoch': 0.34})
+[2026-07-28 10:25:57,542][starVLA.training.train_starvla][INFO] - Step 1879, Loss: {'train/loss': 0.0813072994351387, 'train/grad_norm_pre_clip': 1.477566123008728, 'timing/data': 0.00039284396916627884, 'timing/model': 0.42387002101168036, 'learning_rate/qwen_vl_interface': 9.880542407662098e-06, 'learning_rate/action_model': 9.880542407662097e-05, 'epoch': 0.35})
+[2026-07-28 10:26:03,817][starVLA.training.train_starvla][INFO] - Step 1880, Loss: {'train/loss': 0.04914022237062454, 'train/grad_norm_pre_clip': 1.9389041662216187, 'timing/data': 0.00036748917773365974, 'timing/model': 0.429820965975523, 'learning_rate/qwen_vl_interface': 9.879688186046535e-06, 'learning_rate/action_model': 9.879688186046534e-05, 'epoch': 0.35})
+[2026-07-28 10:26:09,871][starVLA.training.train_starvla][INFO] - Step 1881, Loss: {'train/loss': 0.1168263703584671, 'train/grad_norm_pre_clip': 2.1068742275238037, 'timing/data': 0.0003057289868593216, 'timing/model': 0.43227655021473765, 'learning_rate/qwen_vl_interface': 9.87883096027657e-06, 'learning_rate/action_model': 9.87883096027657e-05, 'epoch': 0.35})
+[2026-07-28 10:26:15,898][starVLA.training.train_starvla][INFO] - Step 1882, Loss: {'train/loss': 0.08126122504472733, 'train/grad_norm_pre_clip': 1.6924595832824707, 'timing/data': 0.00039585307240486145, 'timing/model': 0.4274014299735427, 'learning_rate/qwen_vl_interface': 9.87797073090845e-06, 'learning_rate/action_model': 9.87797073090845e-05, 'epoch': 0.35})
+[2026-07-28 10:26:22,239][starVLA.training.train_starvla][INFO] - Step 1883, Loss: {'train/loss': 0.06561350077390671, 'train/grad_norm_pre_clip': 0.9749537706375122, 'timing/data': 0.0003476548008620739, 'timing/model': 0.4309710212983191, 'learning_rate/qwen_vl_interface': 9.877107498500368e-06, 'learning_rate/action_model': 9.877107498500367e-05, 'epoch': 0.35})
+[2026-07-28 10:26:28,560][starVLA.training.train_starvla][INFO] - Step 1884, Loss: {'train/loss': 0.06438907235860825, 'train/grad_norm_pre_clip': 1.4442802667617798, 'timing/data': 0.0002884543500840664, 'timing/model': 0.4282724279910326, 'learning_rate/qwen_vl_interface': 9.876241263612467e-06, 'learning_rate/action_model': 9.876241263612465e-05, 'epoch': 0.35})
+[2026-07-28 10:26:34,579][starVLA.training.train_starvla][INFO] - Step 1885, Loss: {'train/loss': 0.06498932838439941, 'train/grad_norm_pre_clip': 0.8980137705802917, 'timing/data': 0.00034839892759919167, 'timing/model': 0.4394202115945518, 'learning_rate/qwen_vl_interface': 9.875372026806835e-06, 'learning_rate/action_model': 9.875372026806835e-05, 'epoch': 0.35})
+[2026-07-28 10:26:40,640][starVLA.training.train_starvla][INFO] - Step 1886, Loss: {'train/loss': 0.04763544350862503, 'train/grad_norm_pre_clip': 0.8567733764648438, 'timing/data': 0.0005036750808358192, 'timing/model': 0.4464946808293462, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.35})
+[2026-07-28 10:26:47,123][starVLA.training.train_starvla][INFO] - Step 1887, Loss: {'train/loss': 0.035024240612983704, 'train/grad_norm_pre_clip': 0.7412005066871643, 'timing/data': 0.0003758068196475506, 'timing/model': 0.6912762899883091, 'learning_rate/qwen_vl_interface': 9.873624549700485e-06, 'learning_rate/action_model': 9.873624549700484e-05, 'epoch': 0.35})
+[2026-07-28 10:26:53,110][starVLA.training.train_starvla][INFO] - Step 1888, Loss: {'train/loss': 0.0896722823381424, 'train/grad_norm_pre_clip': 2.347837209701538, 'timing/data': 0.0002978709526360035, 'timing/model': 0.42269841162487864, 'learning_rate/qwen_vl_interface': 9.872746310533683e-06, 'learning_rate/action_model': 9.872746310533684e-05, 'epoch': 0.35})
+[2026-07-28 10:26:59,119][starVLA.training.train_starvla][INFO] - Step 1889, Loss: {'train/loss': 0.049946919083595276, 'train/grad_norm_pre_clip': 1.0583125352859497, 'timing/data': 0.0002959221601486206, 'timing/model': 0.4225154658779502, 'learning_rate/qwen_vl_interface': 9.871865071716991e-06, 'learning_rate/action_model': 9.87186507171699e-05, 'epoch': 0.35})
+[2026-07-28 10:27:05,332][starVLA.training.train_starvla][INFO] - Step 1890, Loss: {'train/loss': 0.038169976323843, 'train/grad_norm_pre_clip': 1.143448829650879, 'timing/data': 0.0003364807926118374, 'timing/model': 0.4216531617566943, 'learning_rate/qwen_vl_interface': 9.870980833822231e-06, 'learning_rate/action_model': 9.87098083382223e-05, 'epoch': 0.35})
+[2026-07-28 10:27:11,357][starVLA.training.train_starvla][INFO] - Step 1891, Loss: {'train/loss': 0.05042879655957222, 'train/grad_norm_pre_clip': 1.1130459308624268, 'timing/data': 0.00034905271604657173, 'timing/model': 0.43051032768562436, 'learning_rate/qwen_vl_interface': 9.87009359742318e-06, 'learning_rate/action_model': 9.87009359742318e-05, 'epoch': 0.35})
+[2026-07-28 10:27:17,654][starVLA.training.train_starvla][INFO] - Step 1892, Loss: {'train/loss': 0.0953768938779831, 'train/grad_norm_pre_clip': 2.3021020889282227, 'timing/data': 0.00043771136552095413, 'timing/model': 0.6826491048559546, 'learning_rate/qwen_vl_interface': 9.86920336309555e-06, 'learning_rate/action_model': 9.86920336309555e-05, 'epoch': 0.35})
+[2026-07-28 10:27:23,680][starVLA.training.train_starvla][INFO] - Step 1893, Loss: {'train/loss': 0.054412778466939926, 'train/grad_norm_pre_clip': 2.0127999782562256, 'timing/data': 0.0003306041471660137, 'timing/model': 0.4226014260202646, 'learning_rate/qwen_vl_interface': 9.86831013141701e-06, 'learning_rate/action_model': 9.86831013141701e-05, 'epoch': 0.35})
+[2026-07-28 10:27:29,909][starVLA.training.train_starvla][INFO] - Step 1894, Loss: {'train/loss': 0.05142296105623245, 'train/grad_norm_pre_clip': 2.1977155208587646, 'timing/data': 0.0003261566162109375, 'timing/model': 0.6606427296064794, 'learning_rate/qwen_vl_interface': 9.867413902967166e-06, 'learning_rate/action_model': 9.867413902967166e-05, 'epoch': 0.35})
+[2026-07-28 10:27:35,897][starVLA.training.train_starvla][INFO] - Step 1895, Loss: {'train/loss': 0.07134270668029785, 'train/grad_norm_pre_clip': 1.8906298875808716, 'timing/data': 0.00042736437171697617, 'timing/model': 0.42111037392169237, 'learning_rate/qwen_vl_interface': 9.86651467832757e-06, 'learning_rate/action_model': 9.86651467832757e-05, 'epoch': 0.35})
+[2026-07-28 10:27:41,906][starVLA.training.train_starvla][INFO] - Step 1896, Loss: {'train/loss': 0.0499560609459877, 'train/grad_norm_pre_clip': 2.759500503540039, 'timing/data': 0.0002909889444708824, 'timing/model': 0.4235007488168776, 'learning_rate/qwen_vl_interface': 9.865612458081718e-06, 'learning_rate/action_model': 9.865612458081719e-05, 'epoch': 0.35})
+[2026-07-28 10:27:47,930][starVLA.training.train_starvla][INFO] - Step 1897, Loss: {'train/loss': 0.08031423389911652, 'train/grad_norm_pre_clip': 2.5800938606262207, 'timing/data': 0.00046057533472776413, 'timing/model': 0.4225523532368243, 'learning_rate/qwen_vl_interface': 9.864707242815057e-06, 'learning_rate/action_model': 9.864707242815056e-05, 'epoch': 0.35})
+[2026-07-28 10:27:54,247][starVLA.training.train_starvla][INFO] - Step 1898, Loss: {'train/loss': 0.06966028362512589, 'train/grad_norm_pre_clip': 1.807348608970642, 'timing/data': 0.000309123657643795, 'timing/model': 0.6674179080873728, 'learning_rate/qwen_vl_interface': 9.863799033114966e-06, 'learning_rate/action_model': 9.863799033114966e-05, 'epoch': 0.35})
+[2026-07-28 10:28:00,269][starVLA.training.train_starvla][INFO] - Step 1899, Loss: {'train/loss': 0.025751953944563866, 'train/grad_norm_pre_clip': 1.4501993656158447, 'timing/data': 0.00032703811302781105, 'timing/model': 0.4244265640154481, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.35})
+[2026-07-28 10:28:06,500][starVLA.training.train_starvla][INFO] - Step 1900, Loss: {'train/loss': 0.1393413096666336, 'train/grad_norm_pre_clip': 3.922588586807251, 'timing/data': 0.0002897060476243496, 'timing/model': 0.4209009171463549, 'learning_rate/qwen_vl_interface': 9.861973632773748e-06, 'learning_rate/action_model': 9.861973632773748e-05, 'epoch': 0.35})
+[2026-07-28 10:28:12,557][starVLA.training.train_starvla][INFO] - Step 1901, Loss: {'train/loss': 0.12605911493301392, 'train/grad_norm_pre_clip': 3.494983673095703, 'timing/data': 0.00034171994775533676, 'timing/model': 0.4373510661534965, 'learning_rate/qwen_vl_interface': 9.861056443317105e-06, 'learning_rate/action_model': 9.861056443317105e-05, 'epoch': 0.35})
+[2026-07-28 10:28:18,612][starVLA.training.train_starvla][INFO] - Step 1902, Loss: {'train/loss': 0.11381825804710388, 'train/grad_norm_pre_clip': 3.4040307998657227, 'timing/data': 0.0002873949706554413, 'timing/model': 0.42611935129389167, 'learning_rate/qwen_vl_interface': 9.860136261796e-06, 'learning_rate/action_model': 9.860136261795998e-05, 'epoch': 0.35})
+[2026-07-28 10:28:24,844][starVLA.training.train_starvla][INFO] - Step 1903, Loss: {'train/loss': 0.08590098470449448, 'train/grad_norm_pre_clip': 3.3782215118408203, 'timing/data': 0.00030802981927990913, 'timing/model': 0.4221298238262534, 'learning_rate/qwen_vl_interface': 9.859213088807521e-06, 'learning_rate/action_model': 9.85921308880752e-05, 'epoch': 0.35})
+[2026-07-28 10:28:31,129][starVLA.training.train_starvla][INFO] - Step 1904, Loss: {'train/loss': 0.11466778814792633, 'train/grad_norm_pre_clip': 3.4830243587493896, 'timing/data': 0.0004178406670689583, 'timing/model': 0.4488739678636193, 'learning_rate/qwen_vl_interface': 9.858286924950714e-06, 'learning_rate/action_model': 9.858286924950713e-05, 'epoch': 0.35})
+[2026-07-28 10:28:37,205][starVLA.training.train_starvla][INFO] - Step 1905, Loss: {'train/loss': 0.03587770462036133, 'train/grad_norm_pre_clip': 1.0173885822296143, 'timing/data': 0.0004895869642496109, 'timing/model': 0.4227664810605347, 'learning_rate/qwen_vl_interface': 9.85735777082655e-06, 'learning_rate/action_model': 9.857357770826551e-05, 'epoch': 0.35})
+[2026-07-28 10:28:43,288][starVLA.training.train_starvla][INFO] - Step 1906, Loss: {'train/loss': 0.04438866674900055, 'train/grad_norm_pre_clip': 1.2398920059204102, 'timing/data': 0.00034450413659214973, 'timing/model': 0.4241901901550591, 'learning_rate/qwen_vl_interface': 9.856425627037951e-06, 'learning_rate/action_model': 9.856425627037952e-05, 'epoch': 0.35})
+[2026-07-28 10:28:49,804][starVLA.training.train_starvla][INFO] - Step 1907, Loss: {'train/loss': 0.034022003412246704, 'train/grad_norm_pre_clip': 1.027748703956604, 'timing/data': 0.00033809617161750793, 'timing/model': 0.6526893805712461, 'learning_rate/qwen_vl_interface': 9.855490494189777e-06, 'learning_rate/action_model': 9.855490494189776e-05, 'epoch': 0.35})
+[2026-07-28 10:28:55,884][starVLA.training.train_starvla][INFO] - Step 1908, Loss: {'train/loss': 0.06703703105449677, 'train/grad_norm_pre_clip': 2.228106737136841, 'timing/data': 0.00028353510424494743, 'timing/model': 0.42352956905961037, 'learning_rate/qwen_vl_interface': 9.854552372888819e-06, 'learning_rate/action_model': 9.854552372888819e-05, 'epoch': 0.35})
+[2026-07-28 10:29:01,972][starVLA.training.train_starvla][INFO] - Step 1909, Loss: {'train/loss': 0.06387381255626678, 'train/grad_norm_pre_clip': 1.8008623123168945, 'timing/data': 0.0003020847216248512, 'timing/model': 0.4235995286144316, 'learning_rate/qwen_vl_interface': 9.853611263743817e-06, 'learning_rate/action_model': 9.853611263743817e-05, 'epoch': 0.35})
+[2026-07-28 10:29:07,800][starVLA.training.train_starvla][INFO] - Step 1910, Loss: {'train/loss': 0.06420987844467163, 'train/grad_norm_pre_clip': 1.8540775775909424, 'timing/data': 0.0003003017045557499, 'timing/model': 0.4272590922191739, 'learning_rate/qwen_vl_interface': 9.85266716736545e-06, 'learning_rate/action_model': 9.852667167365449e-05, 'epoch': 0.35})
+[2026-07-28 10:29:13,637][starVLA.training.train_starvla][INFO] - Step 1911, Loss: {'train/loss': 0.061819083988666534, 'train/grad_norm_pre_clip': 1.169915795326233, 'timing/data': 0.0003856169059872627, 'timing/model': 0.4243012070655823, 'learning_rate/qwen_vl_interface': 9.851720084366324e-06, 'learning_rate/action_model': 9.851720084366324e-05, 'epoch': 0.35})
+[2026-07-28 10:29:19,899][starVLA.training.train_starvla][INFO] - Step 1912, Loss: {'train/loss': 0.07232159376144409, 'train/grad_norm_pre_clip': 2.078110456466675, 'timing/data': 0.00030416715890169144, 'timing/model': 0.6604892890900373, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.35})
+[2026-07-28 10:29:25,896][starVLA.training.train_starvla][INFO] - Step 1913, Loss: {'train/loss': 0.13563652336597443, 'train/grad_norm_pre_clip': 4.07162618637085, 'timing/data': 0.00031666224822402, 'timing/model': 0.42133785923942924, 'learning_rate/qwen_vl_interface': 9.849816960965959e-06, 'learning_rate/action_model': 9.849816960965959e-05, 'epoch': 0.35})
+[2026-07-28 10:29:31,898][starVLA.training.train_starvla][INFO] - Step 1914, Loss: {'train/loss': 0.03446323424577713, 'train/grad_norm_pre_clip': 1.0383230447769165, 'timing/data': 0.0002910369075834751, 'timing/model': 0.42277912655845284, 'learning_rate/qwen_vl_interface': 9.848860921799633e-06, 'learning_rate/action_model': 9.848860921799632e-05, 'epoch': 0.35})
+[2026-07-28 10:29:37,882][starVLA.training.train_starvla][INFO] - Step 1915, Loss: {'train/loss': 0.042100414633750916, 'train/grad_norm_pre_clip': 1.12586510181427, 'timing/data': 0.00031534815207123756, 'timing/model': 0.4222583370283246, 'learning_rate/qwen_vl_interface': 9.847901898482385e-06, 'learning_rate/action_model': 9.847901898482385e-05, 'epoch': 0.35})
+[2026-07-28 10:29:43,858][starVLA.training.train_starvla][INFO] - Step 1916, Loss: {'train/loss': 0.08495578169822693, 'train/grad_norm_pre_clip': 1.6443244218826294, 'timing/data': 0.0002957712858915329, 'timing/model': 0.42231589322909713, 'learning_rate/qwen_vl_interface': 9.846939891636512e-06, 'learning_rate/action_model': 9.846939891636512e-05, 'epoch': 0.35})
+[2026-07-28 10:29:50,071][starVLA.training.train_starvla][INFO] - Step 1917, Loss: {'train/loss': 0.06097247451543808, 'train/grad_norm_pre_clip': 1.0896644592285156, 'timing/data': 0.0003237989731132984, 'timing/model': 0.6568237859755754, 'learning_rate/qwen_vl_interface': 9.845974901886253e-06, 'learning_rate/action_model': 9.845974901886252e-05, 'epoch': 0.35})
+[2026-07-28 10:29:56,084][starVLA.training.train_starvla][INFO] - Step 1918, Loss: {'train/loss': 0.08998134732246399, 'train/grad_norm_pre_clip': 2.391427993774414, 'timing/data': 0.00030536064878106117, 'timing/model': 0.4254570440389216, 'learning_rate/qwen_vl_interface': 9.845006929857777e-06, 'learning_rate/action_model': 9.845006929857776e-05, 'epoch': 0.35})
+[2026-07-28 10:30:02,331][starVLA.training.train_starvla][INFO] - Step 1919, Loss: {'train/loss': 0.0440327525138855, 'train/grad_norm_pre_clip': 1.0000841617584229, 'timing/data': 0.00026291096583008766, 'timing/model': 0.421827910002321, 'learning_rate/qwen_vl_interface': 9.844035976179193e-06, 'learning_rate/action_model': 9.844035976179192e-05, 'epoch': 0.35})
+[2026-07-28 10:30:08,330][starVLA.training.train_starvla][INFO] - Step 1920, Loss: {'train/loss': 0.023641828447580338, 'train/grad_norm_pre_clip': 0.5173819065093994, 'timing/data': 0.00033651478588581085, 'timing/model': 0.42130112275481224, 'learning_rate/qwen_vl_interface': 9.843062041480538e-06, 'learning_rate/action_model': 9.843062041480539e-05, 'epoch': 0.35})
+[2026-07-28 10:30:14,320][starVLA.training.train_starvla][INFO] - Step 1921, Loss: {'train/loss': 0.07718779146671295, 'train/grad_norm_pre_clip': 2.039980411529541, 'timing/data': 0.0003075390122830868, 'timing/model': 0.42284231074154377, 'learning_rate/qwen_vl_interface': 9.842085126393793e-06, 'learning_rate/action_model': 9.842085126393792e-05, 'epoch': 0.35})
+[2026-07-28 10:30:20,554][starVLA.training.train_starvla][INFO] - Step 1922, Loss: {'train/loss': 0.044937148690223694, 'train/grad_norm_pre_clip': 2.052750587463379, 'timing/data': 0.0003234436735510826, 'timing/model': 0.42274382803589106, 'learning_rate/qwen_vl_interface': 9.841105231552863e-06, 'learning_rate/action_model': 9.841105231552864e-05, 'epoch': 0.35})
+[2026-07-28 10:30:26,781][starVLA.training.train_starvla][INFO] - Step 1923, Loss: {'train/loss': 0.06132045015692711, 'train/grad_norm_pre_clip': 1.1697065830230713, 'timing/data': 0.000430254265666008, 'timing/model': 0.664764080196619, 'learning_rate/qwen_vl_interface': 9.840122357593594e-06, 'learning_rate/action_model': 9.840122357593594e-05, 'epoch': 0.35})
+[2026-07-28 10:30:32,764][starVLA.training.train_starvla][INFO] - Step 1924, Loss: {'train/loss': 0.05132053792476654, 'train/grad_norm_pre_clip': 1.015464186668396, 'timing/data': 0.0003663268871605396, 'timing/model': 0.42363322898745537, 'learning_rate/qwen_vl_interface': 9.839136505153762e-06, 'learning_rate/action_model': 9.839136505153762e-05, 'epoch': 0.35})
+[2026-07-28 10:30:39,039][starVLA.training.train_starvla][INFO] - Step 1925, Loss: {'train/loss': 0.03485562652349472, 'train/grad_norm_pre_clip': 1.3121700286865234, 'timing/data': 0.0004453626461327076, 'timing/model': 0.42214517714455724, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.35})
+[2026-07-28 10:30:45,049][starVLA.training.train_starvla][INFO] - Step 1926, Loss: {'train/loss': 0.05150119960308075, 'train/grad_norm_pre_clip': 0.6938843727111816, 'timing/data': 0.0003239554353058338, 'timing/model': 0.4228728599846363, 'learning_rate/qwen_vl_interface': 9.837155867393176e-06, 'learning_rate/action_model': 9.837155867393175e-05, 'epoch': 0.35})
+[2026-07-28 10:30:51,305][starVLA.training.train_starvla][INFO] - Step 1927, Loss: {'train/loss': 0.049798667430877686, 'train/grad_norm_pre_clip': 0.8370426893234253, 'timing/data': 0.0002598450519144535, 'timing/model': 0.4216095879673958, 'learning_rate/qwen_vl_interface': 9.836161083357635e-06, 'learning_rate/action_model': 9.836161083357635e-05, 'epoch': 0.35})
+[2026-07-28 10:30:57,496][starVLA.training.train_starvla][INFO] - Step 1928, Loss: {'train/loss': 0.0796351507306099, 'train/grad_norm_pre_clip': 2.0974185466766357, 'timing/data': 0.00033053336665034294, 'timing/model': 0.5969943949021399, 'learning_rate/qwen_vl_interface': 9.835163323411957e-06, 'learning_rate/action_model': 9.835163323411958e-05, 'epoch': 0.35})
+[2026-07-28 10:31:03,519][starVLA.training.train_starvla][INFO] - Step 1929, Loss: {'train/loss': 0.07276755571365356, 'train/grad_norm_pre_clip': 1.685477614402771, 'timing/data': 0.00032976502552628517, 'timing/model': 0.4240527977235615, 'learning_rate/qwen_vl_interface': 9.834162588203581e-06, 'learning_rate/action_model': 9.83416258820358e-05, 'epoch': 0.35})
+[2026-07-28 10:31:09,540][starVLA.training.train_starvla][INFO] - Step 1930, Loss: {'train/loss': 0.0401507243514061, 'train/grad_norm_pre_clip': 1.3695704936981201, 'timing/data': 0.00036143092438578606, 'timing/model': 0.4267112291418016, 'learning_rate/qwen_vl_interface': 9.833158878381872e-06, 'learning_rate/action_model': 9.83315887838187e-05, 'epoch': 0.35})
+[2026-07-28 10:31:15,806][starVLA.training.train_starvla][INFO] - Step 1931, Loss: {'train/loss': 0.03559482842683792, 'train/grad_norm_pre_clip': 1.4995534420013428, 'timing/data': 0.00031845271587371826, 'timing/model': 0.4231238141655922, 'learning_rate/qwen_vl_interface': 9.832152194598123e-06, 'learning_rate/action_model': 9.832152194598122e-05, 'epoch': 0.35})
+[2026-07-28 10:31:22,046][starVLA.training.train_starvla][INFO] - Step 1932, Loss: {'train/loss': 0.043215878307819366, 'train/grad_norm_pre_clip': 1.134735107421875, 'timing/data': 0.00031310971826314926, 'timing/model': 0.4235948738642037, 'learning_rate/qwen_vl_interface': 9.831142537505563e-06, 'learning_rate/action_model': 9.831142537505562e-05, 'epoch': 0.35})
+[2026-07-28 10:31:28,104][starVLA.training.train_starvla][INFO] - Step 1933, Loss: {'train/loss': 0.06656840443611145, 'train/grad_norm_pre_clip': 1.7262382507324219, 'timing/data': 0.0003166021779179573, 'timing/model': 0.42367698391899467, 'learning_rate/qwen_vl_interface': 9.830129907759347e-06, 'learning_rate/action_model': 9.830129907759347e-05, 'epoch': 0.36})
+[2026-07-28 10:31:34,388][starVLA.training.train_starvla][INFO] - Step 1934, Loss: {'train/loss': 0.06766834855079651, 'train/grad_norm_pre_clip': 2.221956729888916, 'timing/data': 0.0003624800592660904, 'timing/model': 0.6877329931594431, 'learning_rate/qwen_vl_interface': 9.82911430601656e-06, 'learning_rate/action_model': 9.82911430601656e-05, 'epoch': 0.36})
+[2026-07-28 10:31:40,608][starVLA.training.train_starvla][INFO] - Step 1935, Loss: {'train/loss': 0.042102716863155365, 'train/grad_norm_pre_clip': 1.0180411338806152, 'timing/data': 0.0003347671590745449, 'timing/model': 0.6551690096966922, 'learning_rate/qwen_vl_interface': 9.828095732936216e-06, 'learning_rate/action_model': 9.828095732936216e-05, 'epoch': 0.36})
+[2026-07-28 10:31:46,638][starVLA.training.train_starvla][INFO] - Step 1936, Loss: {'train/loss': 0.06771396845579147, 'train/grad_norm_pre_clip': 2.3881032466888428, 'timing/data': 0.0004718988202512264, 'timing/model': 0.4247603961266577, 'learning_rate/qwen_vl_interface': 9.827074189179254e-06, 'learning_rate/action_model': 9.827074189179253e-05, 'epoch': 0.36})
+[2026-07-28 10:31:52,648][starVLA.training.train_starvla][INFO] - Step 1937, Loss: {'train/loss': 0.0738733559846878, 'train/grad_norm_pre_clip': 2.6539461612701416, 'timing/data': 0.000303077045828104, 'timing/model': 0.4243115810677409, 'learning_rate/qwen_vl_interface': 9.82604967540854e-06, 'learning_rate/action_model': 9.82604967540854e-05, 'epoch': 0.36})
+[2026-07-28 10:31:58,679][starVLA.training.train_starvla][INFO] - Step 1938, Loss: {'train/loss': 0.04361243546009064, 'train/grad_norm_pre_clip': 1.257514238357544, 'timing/data': 0.0003503561019897461, 'timing/model': 0.42678374564275146, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.36})
+[2026-07-28 10:32:05,201][starVLA.training.train_starvla][INFO] - Step 1939, Loss: {'train/loss': 0.032324742525815964, 'train/grad_norm_pre_clip': 1.2251957654953003, 'timing/data': 0.00036049308255314827, 'timing/model': 0.677045791875571, 'learning_rate/qwen_vl_interface': 9.823991740486982e-06, 'learning_rate/action_model': 9.823991740486982e-05, 'epoch': 0.36})
+[2026-07-28 10:32:11,198][starVLA.training.train_starvla][INFO] - Step 1940, Loss: {'train/loss': 0.023254860192537308, 'train/grad_norm_pre_clip': 0.7439587712287903, 'timing/data': 0.0003168112598359585, 'timing/model': 0.42213903507217765, 'learning_rate/qwen_vl_interface': 9.822958320671506e-06, 'learning_rate/action_model': 9.822958320671506e-05, 'epoch': 0.36})
+[2026-07-28 10:32:17,201][starVLA.training.train_starvla][INFO] - Step 1941, Loss: {'train/loss': 0.06214692443609238, 'train/grad_norm_pre_clip': 1.6605263948440552, 'timing/data': 0.00037482473999261856, 'timing/model': 0.4235210162587464, 'learning_rate/qwen_vl_interface': 9.821921933513023e-06, 'learning_rate/action_model': 9.821921933513023e-05, 'epoch': 0.36})
+[2026-07-28 10:32:23,436][starVLA.training.train_starvla][INFO] - Step 1942, Loss: {'train/loss': 0.04797804728150368, 'train/grad_norm_pre_clip': 2.4863741397857666, 'timing/data': 0.00033466704189777374, 'timing/model': 0.4195420700125396, 'learning_rate/qwen_vl_interface': 9.820882579684037e-06, 'learning_rate/action_model': 9.820882579684037e-05, 'epoch': 0.36})
+[2026-07-28 10:32:29,421][starVLA.training.train_starvla][INFO] - Step 1943, Loss: {'train/loss': 0.06691302359104156, 'train/grad_norm_pre_clip': 2.1464996337890625, 'timing/data': 0.00041267508640885353, 'timing/model': 0.4243961158208549, 'learning_rate/qwen_vl_interface': 9.819840259858968e-06, 'learning_rate/action_model': 9.819840259858968e-05, 'epoch': 0.36})
+[2026-07-28 10:32:35,442][starVLA.training.train_starvla][INFO] - Step 1944, Loss: {'train/loss': 0.04866257309913635, 'train/grad_norm_pre_clip': 2.2767417430877686, 'timing/data': 0.00036961818113923073, 'timing/model': 0.4220902184024453, 'learning_rate/qwen_vl_interface': 9.818794974714167e-06, 'learning_rate/action_model': 9.818794974714168e-05, 'epoch': 0.36})
+[2026-07-28 10:32:41,696][starVLA.training.train_starvla][INFO] - Step 1945, Loss: {'train/loss': 0.05970919132232666, 'train/grad_norm_pre_clip': 2.011336088180542, 'timing/data': 0.0003259270451962948, 'timing/model': 0.6665166961029172, 'learning_rate/qwen_vl_interface': 9.817746724927912e-06, 'learning_rate/action_model': 9.817746724927912e-05, 'epoch': 0.36})
+[2026-07-28 10:32:47,698][starVLA.training.train_starvla][INFO] - Step 1946, Loss: {'train/loss': 0.10868577659130096, 'train/grad_norm_pre_clip': 3.386641502380371, 'timing/data': 0.00025879545137286186, 'timing/model': 0.4225277337245643, 'learning_rate/qwen_vl_interface': 9.8166955111804e-06, 'learning_rate/action_model': 9.816695511180399e-05, 'epoch': 0.36})
+[2026-07-28 10:32:53,730][starVLA.training.train_starvla][INFO] - Step 1947, Loss: {'train/loss': 0.08342249691486359, 'train/grad_norm_pre_clip': 3.3666884899139404, 'timing/data': 0.00031293323263525963, 'timing/model': 0.4604040221311152, 'learning_rate/qwen_vl_interface': 9.815641334153746e-06, 'learning_rate/action_model': 9.815641334153746e-05, 'epoch': 0.36})
+[2026-07-28 10:32:59,716][starVLA.training.train_starvla][INFO] - Step 1948, Loss: {'train/loss': 0.03153456747531891, 'train/grad_norm_pre_clip': 1.5214194059371948, 'timing/data': 0.00043191108852624893, 'timing/model': 0.42188036907464266, 'learning_rate/qwen_vl_interface': 9.814584194532001e-06, 'learning_rate/action_model': 9.814584194532001e-05, 'epoch': 0.36})
+[2026-07-28 10:33:05,963][starVLA.training.train_starvla][INFO] - Step 1949, Loss: {'train/loss': 0.061496585607528687, 'train/grad_norm_pre_clip': 2.8253018856048584, 'timing/data': 0.000309918075799942, 'timing/model': 0.42358680814504623, 'learning_rate/qwen_vl_interface': 9.813524093001128e-06, 'learning_rate/action_model': 9.813524093001129e-05, 'epoch': 0.36})
+[2026-07-28 10:33:12,231][starVLA.training.train_starvla][INFO] - Step 1950, Loss: {'train/loss': 0.07264362275600433, 'train/grad_norm_pre_clip': 3.054542064666748, 'timing/data': 0.00036113476380705833, 'timing/model': 0.421513550914824, 'learning_rate/qwen_vl_interface': 9.81246103024902e-06, 'learning_rate/action_model': 9.81246103024902e-05, 'epoch': 0.36})
+[2026-07-28 10:33:18,248][starVLA.training.train_starvla][INFO] - Step 1951, Loss: {'train/loss': 0.06289182603359222, 'train/grad_norm_pre_clip': 1.7759013175964355, 'timing/data': 0.0003867154009640217, 'timing/model': 0.42360073886811733, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.36})
+[2026-07-28 10:33:24,513][starVLA.training.train_starvla][INFO] - Step 1952, Loss: {'train/loss': 0.10153371840715408, 'train/grad_norm_pre_clip': 0.9731241464614868, 'timing/data': 0.0002750749699771404, 'timing/model': 0.4237422477453947, 'learning_rate/qwen_vl_interface': 9.810326023842245e-06, 'learning_rate/action_model': 9.810326023842245e-05, 'epoch': 0.36})
+[2026-07-28 10:33:30,531][starVLA.training.train_starvla][INFO] - Step 1953, Loss: {'train/loss': 0.05901821330189705, 'train/grad_norm_pre_clip': 2.140146017074585, 'timing/data': 0.00031264685094356537, 'timing/model': 0.42374640703201294, 'learning_rate/qwen_vl_interface': 9.809254081572963e-06, 'learning_rate/action_model': 9.809254081572961e-05, 'epoch': 0.36})
+[2026-07-28 10:33:36,777][starVLA.training.train_starvla][INFO] - Step 1954, Loss: {'train/loss': 0.06454679369926453, 'train/grad_norm_pre_clip': 0.993076741695404, 'timing/data': 0.0002996278926730156, 'timing/model': 0.4221926317550242, 'learning_rate/qwen_vl_interface': 9.808179180853204e-06, 'learning_rate/action_model': 9.808179180853204e-05, 'epoch': 0.36})
+[2026-07-28 10:33:42,991][starVLA.training.train_starvla][INFO] - Step 1955, Loss: {'train/loss': 0.07494053989648819, 'train/grad_norm_pre_clip': 1.487699031829834, 'timing/data': 0.00029336102306842804, 'timing/model': 0.595920829102397, 'learning_rate/qwen_vl_interface': 9.807101322380464e-06, 'learning_rate/action_model': 9.807101322380464e-05, 'epoch': 0.36})
+[2026-07-28 10:33:48,983][starVLA.training.train_starvla][INFO] - Step 1956, Loss: {'train/loss': 0.13550005853176117, 'train/grad_norm_pre_clip': 4.188057899475098, 'timing/data': 0.0003509819507598877, 'timing/model': 0.42384079983457923, 'learning_rate/qwen_vl_interface': 9.806020506854152e-06, 'learning_rate/action_model': 9.806020506854153e-05, 'epoch': 0.36})
+[2026-07-28 10:33:54,980][starVLA.training.train_starvla][INFO] - Step 1957, Loss: {'train/loss': 0.05059320107102394, 'train/grad_norm_pre_clip': 1.664959192276001, 'timing/data': 0.0004313201643526554, 'timing/model': 0.4227022482082248, 'learning_rate/qwen_vl_interface': 9.804936734975596e-06, 'learning_rate/action_model': 9.804936734975596e-05, 'epoch': 0.36})
+[2026-07-28 10:34:00,992][starVLA.training.train_starvla][INFO] - Step 1958, Loss: {'train/loss': 0.07513889670372009, 'train/grad_norm_pre_clip': 1.5098402500152588, 'timing/data': 0.00031322799623012543, 'timing/model': 0.42349872179329395, 'learning_rate/qwen_vl_interface': 9.803850007448046e-06, 'learning_rate/action_model': 9.803850007448045e-05, 'epoch': 0.36})
+[2026-07-28 10:34:07,278][starVLA.training.train_starvla][INFO] - Step 1959, Loss: {'train/loss': 0.11652691662311554, 'train/grad_norm_pre_clip': 2.203751802444458, 'timing/data': 0.0003086291253566742, 'timing/model': 0.6658267448656261, 'learning_rate/qwen_vl_interface': 9.802760324976666e-06, 'learning_rate/action_model': 9.802760324976667e-05, 'epoch': 0.36})
+[2026-07-28 10:34:13,280][starVLA.training.train_starvla][INFO] - Step 1960, Loss: {'train/loss': 0.03276609629392624, 'train/grad_norm_pre_clip': 0.8134897947311401, 'timing/data': 0.00030458392575383186, 'timing/model': 0.4228063249029219, 'learning_rate/qwen_vl_interface': 9.80166768826854e-06, 'learning_rate/action_model': 9.80166768826854e-05, 'epoch': 0.36})
+[2026-07-28 10:34:19,316][starVLA.training.train_starvla][INFO] - Step 1961, Loss: {'train/loss': 0.046754270792007446, 'train/grad_norm_pre_clip': 1.0778088569641113, 'timing/data': 0.00044173188507556915, 'timing/model': 0.42377769388258457, 'learning_rate/qwen_vl_interface': 9.800572098032668e-06, 'learning_rate/action_model': 9.800572098032668e-05, 'epoch': 0.36})
+[2026-07-28 10:34:25,329][starVLA.training.train_starvla][INFO] - Step 1962, Loss: {'train/loss': 0.052542805671691895, 'train/grad_norm_pre_clip': 1.8978233337402344, 'timing/data': 0.0003104149363934994, 'timing/model': 0.4233954921364784, 'learning_rate/qwen_vl_interface': 9.799473554979966e-06, 'learning_rate/action_model': 9.799473554979965e-05, 'epoch': 0.36})
+[2026-07-28 10:34:31,564][starVLA.training.train_starvla][INFO] - Step 1963, Loss: {'train/loss': 0.047475144267082214, 'train/grad_norm_pre_clip': 0.8516272902488708, 'timing/data': 0.00048289215192198753, 'timing/model': 0.42102537397295237, 'learning_rate/qwen_vl_interface': 9.798372059823268e-06, 'learning_rate/action_model': 9.798372059823268e-05, 'epoch': 0.36})
+[2026-07-28 10:34:37,587][starVLA.training.train_starvla][INFO] - Step 1964, Loss: {'train/loss': 0.048395100980997086, 'train/grad_norm_pre_clip': 1.2697350978851318, 'timing/data': 0.00036797719076275826, 'timing/model': 0.4248995198868215, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.36})
+[2026-07-28 10:34:43,620][starVLA.training.train_starvla][INFO] - Step 1965, Loss: {'train/loss': 0.07826592028141022, 'train/grad_norm_pre_clip': 1.4601346254348755, 'timing/data': 0.0003710179589688778, 'timing/model': 0.4284982308745384, 'learning_rate/qwen_vl_interface': 9.79616021605879e-06, 'learning_rate/action_model': 9.796160216058788e-05, 'epoch': 0.36})
+[2026-07-28 10:34:49,913][starVLA.training.train_starvla][INFO] - Step 1966, Loss: {'train/loss': 0.048451073467731476, 'train/grad_norm_pre_clip': 1.7721824645996094, 'timing/data': 0.00029375310987234116, 'timing/model': 0.42199350986629725, 'learning_rate/qwen_vl_interface': 9.795049868886247e-06, 'learning_rate/action_model': 9.795049868886247e-05, 'epoch': 0.36})
+[2026-07-28 10:34:55,939][starVLA.training.train_starvla][INFO] - Step 1967, Loss: {'train/loss': 0.07162511348724365, 'train/grad_norm_pre_clip': 1.519496202468872, 'timing/data': 0.0003760852850973606, 'timing/model': 0.42534417612478137, 'learning_rate/qwen_vl_interface': 9.793936572480192e-06, 'learning_rate/action_model': 9.793936572480192e-05, 'epoch': 0.36})
+[2026-07-28 10:35:02,390][starVLA.training.train_starvla][INFO] - Step 1968, Loss: {'train/loss': 0.04060672968626022, 'train/grad_norm_pre_clip': 1.2471719980239868, 'timing/data': 0.0003899489529430866, 'timing/model': 0.6588110299780965, 'learning_rate/qwen_vl_interface': 9.792820327563028e-06, 'learning_rate/action_model': 9.792820327563028e-05, 'epoch': 0.36})
+[2026-07-28 10:35:08,449][starVLA.training.train_starvla][INFO] - Step 1969, Loss: {'train/loss': 0.04582750052213669, 'train/grad_norm_pre_clip': 0.7633166313171387, 'timing/data': 0.0003094119019806385, 'timing/model': 0.42332832608371973, 'learning_rate/qwen_vl_interface': 9.791701134859076e-06, 'learning_rate/action_model': 9.791701134859075e-05, 'epoch': 0.36})
+[2026-07-28 10:35:14,749][starVLA.training.train_starvla][INFO] - Step 1970, Loss: {'train/loss': 0.07345352321863174, 'train/grad_norm_pre_clip': 2.0301358699798584, 'timing/data': 0.00042595481500029564, 'timing/model': 0.6653229109942913, 'learning_rate/qwen_vl_interface': 9.790578995094565e-06, 'learning_rate/action_model': 9.790578995094563e-05, 'epoch': 0.36})
+[2026-07-28 10:35:20,781][starVLA.training.train_starvla][INFO] - Step 1971, Loss: {'train/loss': 0.043954282999038696, 'train/grad_norm_pre_clip': 1.6235460042953491, 'timing/data': 0.0003352188505232334, 'timing/model': 0.42377541586756706, 'learning_rate/qwen_vl_interface': 9.789453908997639e-06, 'learning_rate/action_model': 9.789453908997638e-05, 'epoch': 0.36})
+[2026-07-28 10:35:26,940][starVLA.training.train_starvla][INFO] - Step 1972, Loss: {'train/loss': 0.056656576693058014, 'train/grad_norm_pre_clip': 1.497708797454834, 'timing/data': 0.00032032933086156845, 'timing/model': 0.42461195308715105, 'learning_rate/qwen_vl_interface': 9.788325877298355e-06, 'learning_rate/action_model': 9.788325877298355e-05, 'epoch': 0.36})
+[2026-07-28 10:35:33,172][starVLA.training.train_starvla][INFO] - Step 1973, Loss: {'train/loss': 0.03483922779560089, 'train/grad_norm_pre_clip': 1.2894575595855713, 'timing/data': 0.0003315610811114311, 'timing/model': 0.4225850240327418, 'learning_rate/qwen_vl_interface': 9.787194900728682e-06, 'learning_rate/action_model': 9.787194900728683e-05, 'epoch': 0.36})
+[2026-07-28 10:35:39,447][starVLA.training.train_starvla][INFO] - Step 1974, Loss: {'train/loss': 0.06787355989217758, 'train/grad_norm_pre_clip': 2.429076910018921, 'timing/data': 0.0003482750616967678, 'timing/model': 0.4254254908300936, 'learning_rate/qwen_vl_interface': 9.786060980022499e-06, 'learning_rate/action_model': 9.786060980022498e-05, 'epoch': 0.36})
+[2026-07-28 10:35:45,463][starVLA.training.train_starvla][INFO] - Step 1975, Loss: {'train/loss': 0.038325704634189606, 'train/grad_norm_pre_clip': 1.929294466972351, 'timing/data': 0.00039006536826491356, 'timing/model': 0.4227296030148864, 'learning_rate/qwen_vl_interface': 9.784924115915592e-06, 'learning_rate/action_model': 9.784924115915592e-05, 'epoch': 0.36})
+[2026-07-28 10:35:51,696][starVLA.training.train_starvla][INFO] - Step 1976, Loss: {'train/loss': 0.03651481121778488, 'train/grad_norm_pre_clip': 1.705062747001648, 'timing/data': 0.0003097681328654289, 'timing/model': 0.42024393705651164, 'learning_rate/qwen_vl_interface': 9.783784309145663e-06, 'learning_rate/action_model': 9.783784309145661e-05, 'epoch': 0.36})
+[2026-07-28 10:35:57,692][starVLA.training.train_starvla][INFO] - Step 1977, Loss: {'train/loss': 0.036213383078575134, 'train/grad_norm_pre_clip': 1.5216288566589355, 'timing/data': 0.0003448878414928913, 'timing/model': 0.42410475481301546, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.36})
+[2026-07-28 10:36:03,955][starVLA.training.train_starvla][INFO] - Step 1978, Loss: {'train/loss': 0.013572749681770802, 'train/grad_norm_pre_clip': 1.2789876461029053, 'timing/data': 0.00036679208278656006, 'timing/model': 0.4231243641115725, 'learning_rate/qwen_vl_interface': 9.78149587057707e-06, 'learning_rate/action_model': 9.78149587057707e-05, 'epoch': 0.36})
+[2026-07-28 10:36:10,011][starVLA.training.train_starvla][INFO] - Step 1979, Loss: {'train/loss': 0.114857017993927, 'train/grad_norm_pre_clip': 2.770341157913208, 'timing/data': 0.000470960047096014, 'timing/model': 0.424846975132823, 'learning_rate/qwen_vl_interface': 9.780347240263354e-06, 'learning_rate/action_model': 9.780347240263353e-05, 'epoch': 0.36})
+[2026-07-28 10:36:16,280][starVLA.training.train_starvla][INFO] - Step 1980, Loss: {'train/loss': 0.057382240891456604, 'train/grad_norm_pre_clip': 2.475023031234741, 'timing/data': 0.00047980668023228645, 'timing/model': 0.6674623591825366, 'learning_rate/qwen_vl_interface': 9.779195670256498e-06, 'learning_rate/action_model': 9.779195670256497e-05, 'epoch': 0.36})
+[2026-07-28 10:36:22,587][starVLA.training.train_starvla][INFO] - Step 1981, Loss: {'train/loss': 0.06344541907310486, 'train/grad_norm_pre_clip': 1.9451977014541626, 'timing/data': 0.0002969382330775261, 'timing/model': 0.6646924228407443, 'learning_rate/qwen_vl_interface': 9.778041161303746e-06, 'learning_rate/action_model': 9.778041161303745e-05, 'epoch': 0.36})
+[2026-07-28 10:36:28,615][starVLA.training.train_starvla][INFO] - Step 1982, Loss: {'train/loss': 0.0354173444211483, 'train/grad_norm_pre_clip': 1.23410165309906, 'timing/data': 0.0004555908963084221, 'timing/model': 0.42224333994090557, 'learning_rate/qwen_vl_interface': 9.776883714154242e-06, 'learning_rate/action_model': 9.776883714154241e-05, 'epoch': 0.36})
+[2026-07-28 10:36:34,429][starVLA.training.train_starvla][INFO] - Step 1983, Loss: {'train/loss': 0.05651375651359558, 'train/grad_norm_pre_clip': 2.030400514602661, 'timing/data': 0.00031852396205067635, 'timing/model': 0.42454298166558146, 'learning_rate/qwen_vl_interface': 9.775723329559044e-06, 'learning_rate/action_model': 9.775723329559044e-05, 'epoch': 0.36})
+[2026-07-28 10:36:40,441][starVLA.training.train_starvla][INFO] - Step 1984, Loss: {'train/loss': 0.06215587630867958, 'train/grad_norm_pre_clip': 2.3067524433135986, 'timing/data': 0.000430478248745203, 'timing/model': 0.4253596840426326, 'learning_rate/qwen_vl_interface': 9.774560008271114e-06, 'learning_rate/action_model': 9.774560008271114e-05, 'epoch': 0.36})
+[2026-07-28 10:36:46,476][starVLA.training.train_starvla][INFO] - Step 1985, Loss: {'train/loss': 0.05793596804141998, 'train/grad_norm_pre_clip': 1.9224404096603394, 'timing/data': 0.0003992603160440922, 'timing/model': 0.4232822717167437, 'learning_rate/qwen_vl_interface': 9.773393751045315e-06, 'learning_rate/action_model': 9.773393751045315e-05, 'epoch': 0.36})
+[2026-07-28 10:36:53,051][starVLA.training.train_starvla][INFO] - Step 1986, Loss: {'train/loss': 0.07525568455457687, 'train/grad_norm_pre_clip': 1.8370450735092163, 'timing/data': 0.00030732061713933945, 'timing/model': 0.6687662787735462, 'learning_rate/qwen_vl_interface': 9.772224558638422e-06, 'learning_rate/action_model': 9.772224558638422e-05, 'epoch': 0.36})
+[2026-07-28 10:36:59,151][starVLA.training.train_starvla][INFO] - Step 1987, Loss: {'train/loss': 0.09206290543079376, 'train/grad_norm_pre_clip': 2.735069990158081, 'timing/data': 0.0003206823021173477, 'timing/model': 0.42367308074608445, 'learning_rate/qwen_vl_interface': 9.77105243180911e-06, 'learning_rate/action_model': 9.771052431809109e-05, 'epoch': 0.36})
+[2026-07-28 10:37:05,228][starVLA.training.train_starvla][INFO] - Step 1988, Loss: {'train/loss': 0.12343715131282806, 'train/grad_norm_pre_clip': 2.4225971698760986, 'timing/data': 0.0003159348852932453, 'timing/model': 0.42482314677909017, 'learning_rate/qwen_vl_interface': 9.76987737131796e-06, 'learning_rate/action_model': 9.76987737131796e-05, 'epoch': 0.37})
+[2026-07-28 10:37:11,238][starVLA.training.train_starvla][INFO] - Step 1989, Loss: {'train/loss': 0.07317144423723221, 'train/grad_norm_pre_clip': 1.576475739479065, 'timing/data': 0.0003752117045223713, 'timing/model': 0.42207304015755653, 'learning_rate/qwen_vl_interface': 9.768699377927454e-06, 'learning_rate/action_model': 9.768699377927453e-05, 'epoch': 0.37})
+[2026-07-28 10:37:17,500][starVLA.training.train_starvla][INFO] - Step 1990, Loss: {'train/loss': 0.05859163776040077, 'train/grad_norm_pre_clip': 2.473376750946045, 'timing/data': 0.0003116363659501076, 'timing/model': 0.42434408282861114, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.37})
+[2026-07-28 10:37:23,603][starVLA.training.train_starvla][INFO] - Step 1991, Loss: {'train/loss': 0.043239038437604904, 'train/grad_norm_pre_clip': 0.5219770669937134, 'timing/data': 0.0003272444009780884, 'timing/model': 0.4245100920088589, 'learning_rate/qwen_vl_interface': 9.76633459550783e-06, 'learning_rate/action_model': 9.76633459550783e-05, 'epoch': 0.37})
+[2026-07-28 10:37:29,890][starVLA.training.train_starvla][INFO] - Step 1992, Loss: {'train/loss': 0.05851427838206291, 'train/grad_norm_pre_clip': 1.2436316013336182, 'timing/data': 0.00039749499410390854, 'timing/model': 0.6719622341915965, 'learning_rate/qwen_vl_interface': 9.765147808013193e-06, 'learning_rate/action_model': 9.765147808013192e-05, 'epoch': 0.37})
+[2026-07-28 10:37:36,140][starVLA.training.train_starvla][INFO] - Step 1993, Loss: {'train/loss': 0.07115773111581802, 'train/grad_norm_pre_clip': 2.1303529739379883, 'timing/data': 0.0003214632160961628, 'timing/model': 0.4207548601552844, 'learning_rate/qwen_vl_interface': 9.763958090688165e-06, 'learning_rate/action_model': 9.763958090688164e-05, 'epoch': 0.37})
+[2026-07-28 10:37:41,916][starVLA.training.train_starvla][INFO] - Step 1994, Loss: {'train/loss': 0.07363604009151459, 'train/grad_norm_pre_clip': 1.4220811128616333, 'timing/data': 0.0004868311807513237, 'timing/model': 0.4243919919244945, 'learning_rate/qwen_vl_interface': 9.762765444304737e-06, 'learning_rate/action_model': 9.762765444304737e-05, 'epoch': 0.37})
+[2026-07-28 10:37:47,974][starVLA.training.train_starvla][INFO] - Step 1995, Loss: {'train/loss': 0.07396949827671051, 'train/grad_norm_pre_clip': 1.1531354188919067, 'timing/data': 0.0004918919876217842, 'timing/model': 0.4242973658256233, 'learning_rate/qwen_vl_interface': 9.761569869636808e-06, 'learning_rate/action_model': 9.761569869636808e-05, 'epoch': 0.37})
+[2026-07-28 10:37:53,989][starVLA.training.train_starvla][INFO] - Step 1996, Loss: {'train/loss': 0.05595829337835312, 'train/grad_norm_pre_clip': 1.0047495365142822, 'timing/data': 0.00037933187559247017, 'timing/model': 0.4249537060968578, 'learning_rate/qwen_vl_interface': 9.760371367460172e-06, 'learning_rate/action_model': 9.760371367460171e-05, 'epoch': 0.37})
+[2026-07-28 10:38:00,527][starVLA.training.train_starvla][INFO] - Step 1997, Loss: {'train/loss': 0.04006126523017883, 'train/grad_norm_pre_clip': 0.8772849440574646, 'timing/data': 0.0004980708472430706, 'timing/model': 0.4211266487836838, 'learning_rate/qwen_vl_interface': 9.759169938552524e-06, 'learning_rate/action_model': 9.759169938552524e-05, 'epoch': 0.37})
+[2026-07-28 10:38:06,561][starVLA.training.train_starvla][INFO] - Step 1998, Loss: {'train/loss': 0.053701870143413544, 'train/grad_norm_pre_clip': 0.9022988080978394, 'timing/data': 0.00034226197749376297, 'timing/model': 0.4278406691737473, 'learning_rate/qwen_vl_interface': 9.757965583693458e-06, 'learning_rate/action_model': 9.757965583693457e-05, 'epoch': 0.37})
+[2026-07-28 10:38:12,608][starVLA.training.train_starvla][INFO] - Step 1999, Loss: {'train/loss': 0.05734195187687874, 'train/grad_norm_pre_clip': 2.086256742477417, 'timing/data': 0.0003230469301342964, 'timing/model': 0.42501289304345846, 'learning_rate/qwen_vl_interface': 9.756758303664467e-06, 'learning_rate/action_model': 9.756758303664468e-05, 'epoch': 0.37})
+[2026-07-28 10:38:18,633][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2000
+[2026-07-28 10:38:23,698][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2000
+[2026-07-28 10:46:22,544][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2000
+[2026-07-28 10:46:22,545][starVLA.training.train_starvla][INFO] - Step 2000, Loss: {'train/loss': 0.060104645788669586, 'train/grad_norm_pre_clip': 1.680938959121704, 'eval/action_loss/seconds': 5.061645636335015, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.04619395360350609, 'eval/latency_3/loss': 0.04619395360350609, 'eval/flappy/loss': 0.04619395360350609, 'eval/flappy/latency_3/loss': 0.04619395360350609, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 360.93000424727796, 'rl_games_eval/mid_train/mean_length': 2926.9, 'rl_games_eval/mid_train/std_reward': 141.95610790358714, 'rl_games_eval/mid_train/std_length': 1143.4599643188212, 'rl_games_eval/mid_train/macro_mean_reward': 360.93000424727796, 'rl_games_eval/mid_train/macro_mean_length': 2926.9, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 360.93000424727796, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2926.9, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 141.95610790358714, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1143.4599643188212, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 360.93000424727796, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.000317601952701807, 'timing/model': 0.4243993447162211, 'learning_rate/qwen_vl_interface': 9.755548099248942e-06, 'learning_rate/action_model': 9.75554809924894e-05, 'epoch': 0.37})
+[2026-07-28 10:46:22,548][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_2000_state
+[2026-07-28 10:46:22,548][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 10:47:00,360][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_2000_state/pytorch_model
+[2026-07-28 10:47:00,361][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_2000_state/sampler.bin
+[2026-07-28 10:47:00,361][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_2000_state/sampler_1.bin
+[2026-07-28 10:47:00,363][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_2000_state/random_states_0.pkl
+[2026-07-28 10:47:14,230][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2000
+[2026-07-28 10:47:14,231][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 10:47:14,254][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 10:47:20,511][starVLA.training.train_starvla][INFO] - Step 2001, Loss: {'train/loss': 0.041606370359659195, 'train/grad_norm_pre_clip': 1.0281982421875, 'timing/data': 0.000299219973385334, 'timing/model': 0.4242425970733166, 'learning_rate/qwen_vl_interface': 9.75433497123217e-06, 'learning_rate/action_model': 9.754334971232171e-05, 'epoch': 0.37})
+[2026-07-28 10:47:26,552][starVLA.training.train_starvla][INFO] - Step 2002, Loss: {'train/loss': 0.06030619889497757, 'train/grad_norm_pre_clip': 2.3399882316589355, 'timing/data': 0.000296842772513628, 'timing/model': 0.42383211199194193, 'learning_rate/qwen_vl_interface': 9.753118920401341e-06, 'learning_rate/action_model': 9.75311892040134e-05, 'epoch': 0.37})
+[2026-07-28 10:47:32,802][starVLA.training.train_starvla][INFO] - Step 2003, Loss: {'train/loss': 0.020989220589399338, 'train/grad_norm_pre_clip': 0.944314181804657, 'timing/data': 0.0002936641685664654, 'timing/model': 0.676565685775131, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.37})
+[2026-07-28 10:47:39,048][starVLA.training.train_starvla][INFO] - Step 2004, Loss: {'train/loss': 0.04174939543008804, 'train/grad_norm_pre_clip': 1.8326295614242554, 'timing/data': 0.0002858908846974373, 'timing/model': 0.6609168681316078, 'learning_rate/qwen_vl_interface': 9.750678053455725e-06, 'learning_rate/action_model': 9.750678053455725e-05, 'epoch': 0.37})
+[2026-07-28 10:47:45,083][starVLA.training.train_starvla][INFO] - Step 2005, Loss: {'train/loss': 0.036947570741176605, 'train/grad_norm_pre_clip': 0.8291400074958801, 'timing/data': 0.00031058210879564285, 'timing/model': 0.4211067329160869, 'learning_rate/qwen_vl_interface': 9.749453238924793e-06, 'learning_rate/action_model': 9.749453238924792e-05, 'epoch': 0.37})
+[2026-07-28 10:47:51,160][starVLA.training.train_starvla][INFO] - Step 2006, Loss: {'train/loss': 0.06130732595920563, 'train/grad_norm_pre_clip': 1.9582220315933228, 'timing/data': 0.0003972342237830162, 'timing/model': 0.422737886197865, 'learning_rate/qwen_vl_interface': 9.7482255047475e-06, 'learning_rate/action_model': 9.7482255047475e-05, 'epoch': 0.37})
+[2026-07-28 10:47:57,211][starVLA.training.train_starvla][INFO] - Step 2007, Loss: {'train/loss': 0.040328241884708405, 'train/grad_norm_pre_clip': 1.5080385208129883, 'timing/data': 0.00033565703779459, 'timing/model': 0.42343905474990606, 'learning_rate/qwen_vl_interface': 9.746994851720518e-06, 'learning_rate/action_model': 9.746994851720517e-05, 'epoch': 0.37})
+[2026-07-28 10:48:03,315][starVLA.training.train_starvla][INFO] - Step 2008, Loss: {'train/loss': 0.039042726159095764, 'train/grad_norm_pre_clip': 2.3821828365325928, 'timing/data': 0.00031194090843200684, 'timing/model': 0.4287366080097854, 'learning_rate/qwen_vl_interface': 9.745761280642397e-06, 'learning_rate/action_model': 9.745761280642397e-05, 'epoch': 0.37})
+[2026-07-28 10:48:09,717][starVLA.training.train_starvla][INFO] - Step 2009, Loss: {'train/loss': 0.024807652458548546, 'train/grad_norm_pre_clip': 1.3160057067871094, 'timing/data': 0.0002857060171663761, 'timing/model': 0.42390858195722103, 'learning_rate/qwen_vl_interface': 9.744524792313591e-06, 'learning_rate/action_model': 9.74452479231359e-05, 'epoch': 0.37})
+[2026-07-28 10:48:16,036][starVLA.training.train_starvla][INFO] - Step 2010, Loss: {'train/loss': 0.09229207783937454, 'train/grad_norm_pre_clip': 2.1650421619415283, 'timing/data': 0.00041054608300328255, 'timing/model': 0.42258632415905595, 'learning_rate/qwen_vl_interface': 9.743285387536445e-06, 'learning_rate/action_model': 9.743285387536444e-05, 'epoch': 0.37})
+[2026-07-28 10:48:21,936][starVLA.training.train_starvla][INFO] - Step 2011, Loss: {'train/loss': 0.08398561179637909, 'train/grad_norm_pre_clip': 2.6781318187713623, 'timing/data': 0.0003667287528514862, 'timing/model': 0.42452538711950183, 'learning_rate/qwen_vl_interface': 9.742043067115193e-06, 'learning_rate/action_model': 9.742043067115192e-05, 'epoch': 0.37})
+[2026-07-28 10:48:28,047][starVLA.training.train_starvla][INFO] - Step 2012, Loss: {'train/loss': 0.03966239467263222, 'train/grad_norm_pre_clip': 1.6021919250488281, 'timing/data': 0.0003030463121831417, 'timing/model': 0.4245655038394034, 'learning_rate/qwen_vl_interface': 9.740797831855965e-06, 'learning_rate/action_model': 9.740797831855965e-05, 'epoch': 0.37})
+[2026-07-28 10:48:34,136][starVLA.training.train_starvla][INFO] - Step 2013, Loss: {'train/loss': 0.08744439482688904, 'train/grad_norm_pre_clip': 1.7210261821746826, 'timing/data': 0.00043478840962052345, 'timing/model': 0.4264439307153225, 'learning_rate/qwen_vl_interface': 9.73954968256678e-06, 'learning_rate/action_model': 9.73954968256678e-05, 'epoch': 0.37})
+[2026-07-28 10:48:40,479][starVLA.training.train_starvla][INFO] - Step 2014, Loss: {'train/loss': 0.017125966027379036, 'train/grad_norm_pre_clip': 0.5432499051094055, 'timing/data': 0.00037376629188656807, 'timing/model': 0.42268336517736316, 'learning_rate/qwen_vl_interface': 9.738298620057547e-06, 'learning_rate/action_model': 9.738298620057546e-05, 'epoch': 0.37})
+[2026-07-28 10:48:46,855][starVLA.training.train_starvla][INFO] - Step 2015, Loss: {'train/loss': 0.06112655997276306, 'train/grad_norm_pre_clip': 0.8995862603187561, 'timing/data': 0.00027230987325310707, 'timing/model': 0.42070077219977975, 'learning_rate/qwen_vl_interface': 9.73704464514007e-06, 'learning_rate/action_model': 9.737044645140068e-05, 'epoch': 0.37})
+[2026-07-28 10:48:52,898][starVLA.training.train_starvla][INFO] - Step 2016, Loss: {'train/loss': 0.04774802550673485, 'train/grad_norm_pre_clip': 0.9992547035217285, 'timing/data': 0.0002912580966949463, 'timing/model': 0.4239384988322854, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.37})
+[2026-07-28 10:48:59,164][starVLA.training.train_starvla][INFO] - Step 2017, Loss: {'train/loss': 0.06373338401317596, 'train/grad_norm_pre_clip': 1.0291024446487427, 'timing/data': 0.0005704937502741814, 'timing/model': 0.4291201885789633, 'learning_rate/qwen_vl_interface': 9.734527961337028e-06, 'learning_rate/action_model': 9.734527961337028e-05, 'epoch': 0.37})
+[2026-07-28 10:49:05,246][starVLA.training.train_starvla][INFO] - Step 2018, Loss: {'train/loss': 0.041513197124004364, 'train/grad_norm_pre_clip': 1.115530014038086, 'timing/data': 0.00028592394664883614, 'timing/model': 0.4227513950318098, 'learning_rate/qwen_vl_interface': 9.733265254084514e-06, 'learning_rate/action_model': 9.733265254084513e-05, 'epoch': 0.37})
+[2026-07-28 10:49:11,313][starVLA.training.train_starvla][INFO] - Step 2019, Loss: {'train/loss': 0.06062420457601547, 'train/grad_norm_pre_clip': 1.3597341775894165, 'timing/data': 0.00029809633269906044, 'timing/model': 0.4396848031319678, 'learning_rate/qwen_vl_interface': 9.73199963768985e-06, 'learning_rate/action_model': 9.731999637689849e-05, 'epoch': 0.37})
+[2026-07-28 10:49:17,893][starVLA.training.train_starvla][INFO] - Step 2020, Loss: {'train/loss': 0.0696730762720108, 'train/grad_norm_pre_clip': 0.9155360460281372, 'timing/data': 0.0003630891442298889, 'timing/model': 0.6001894599758089, 'learning_rate/qwen_vl_interface': 9.73073111297428e-06, 'learning_rate/action_model': 9.73073111297428e-05, 'epoch': 0.37})
+[2026-07-28 10:49:24,009][starVLA.training.train_starvla][INFO] - Step 2021, Loss: {'train/loss': 0.05106540024280548, 'train/grad_norm_pre_clip': 0.9177026748657227, 'timing/data': 0.0002940976992249489, 'timing/model': 0.4243210698477924, 'learning_rate/qwen_vl_interface': 9.72945968076094e-06, 'learning_rate/action_model': 9.729459680760939e-05, 'epoch': 0.37})
+[2026-07-28 10:49:30,120][starVLA.training.train_starvla][INFO] - Step 2022, Loss: {'train/loss': 0.0291915163397789, 'train/grad_norm_pre_clip': 1.1436208486557007, 'timing/data': 0.0003332458436489105, 'timing/model': 0.46905774902552366, 'learning_rate/qwen_vl_interface': 9.728185341874844e-06, 'learning_rate/action_model': 9.728185341874844e-05, 'epoch': 0.37})
+[2026-07-28 10:49:36,268][starVLA.training.train_starvla][INFO] - Step 2023, Loss: {'train/loss': 0.035974517464637756, 'train/grad_norm_pre_clip': 1.5037842988967896, 'timing/data': 0.0003145006485283375, 'timing/model': 0.4415974817238748, 'learning_rate/qwen_vl_interface': 9.7269080971429e-06, 'learning_rate/action_model': 9.726908097142899e-05, 'epoch': 0.37})
+[2026-07-28 10:49:42,872][starVLA.training.train_starvla][INFO] - Step 2024, Loss: {'train/loss': 0.10838145017623901, 'train/grad_norm_pre_clip': 2.374577522277832, 'timing/data': 0.00029552681371569633, 'timing/model': 0.6512515149079263, 'learning_rate/qwen_vl_interface': 9.725627947393896e-06, 'learning_rate/action_model': 9.725627947393896e-05, 'epoch': 0.37})
+[2026-07-28 10:49:48,982][starVLA.training.train_starvla][INFO] - Step 2025, Loss: {'train/loss': 0.04886788874864578, 'train/grad_norm_pre_clip': 1.5789357423782349, 'timing/data': 0.0004066750407218933, 'timing/model': 0.4252605689689517, 'learning_rate/qwen_vl_interface': 9.72434489345851e-06, 'learning_rate/action_model': 9.72434489345851e-05, 'epoch': 0.37})
+[2026-07-28 10:49:55,142][starVLA.training.train_starvla][INFO] - Step 2026, Loss: {'train/loss': 0.03155875951051712, 'train/grad_norm_pre_clip': 0.7800263166427612, 'timing/data': 0.0004510427825152874, 'timing/model': 0.42738719284534454, 'learning_rate/qwen_vl_interface': 9.723058936169298e-06, 'learning_rate/action_model': 9.723058936169297e-05, 'epoch': 0.37})
+[2026-07-28 10:50:00,993][starVLA.training.train_starvla][INFO] - Step 2027, Loss: {'train/loss': 0.060759834945201874, 'train/grad_norm_pre_clip': 2.084522247314453, 'timing/data': 0.0002950201742351055, 'timing/model': 0.4228917662985623, 'learning_rate/qwen_vl_interface': 9.721770076360707e-06, 'learning_rate/action_model': 9.721770076360706e-05, 'epoch': 0.37})
+[2026-07-28 10:50:06,827][starVLA.training.train_starvla][INFO] - Step 2028, Loss: {'train/loss': 0.0417494997382164, 'train/grad_norm_pre_clip': 1.056802749633789, 'timing/data': 0.00032978691160678864, 'timing/model': 0.42983245477080345, 'learning_rate/qwen_vl_interface': 9.720478314869061e-06, 'learning_rate/action_model': 9.72047831486906e-05, 'epoch': 0.37})
+[2026-07-28 10:50:13,258][starVLA.training.train_starvla][INFO] - Step 2029, Loss: {'train/loss': 0.04225800558924675, 'train/grad_norm_pre_clip': 0.6600711345672607, 'timing/data': 0.00039403699338436127, 'timing/model': 0.4207433587871492, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.37})
+[2026-07-28 10:50:19,376][starVLA.training.train_starvla][INFO] - Step 2030, Loss: {'train/loss': 0.085569828748703, 'train/grad_norm_pre_clip': 1.6126322746276855, 'timing/data': 0.00030222581699490547, 'timing/model': 0.43323715403676033, 'learning_rate/qwen_vl_interface': 9.717886090191334e-06, 'learning_rate/action_model': 9.717886090191333e-05, 'epoch': 0.37})
+[2026-07-28 10:50:25,451][starVLA.training.train_starvla][INFO] - Step 2031, Loss: {'train/loss': 0.022465255111455917, 'train/grad_norm_pre_clip': 0.749127984046936, 'timing/data': 0.0003977809101343155, 'timing/model': 0.424819263163954, 'learning_rate/qwen_vl_interface': 9.716585628687317e-06, 'learning_rate/action_model': 9.716585628687316e-05, 'epoch': 0.37})
+[2026-07-28 10:50:31,599][starVLA.training.train_starvla][INFO] - Step 2032, Loss: {'train/loss': 0.08762189745903015, 'train/grad_norm_pre_clip': 1.9320368766784668, 'timing/data': 0.00034484732896089554, 'timing/model': 0.4348724661394954, 'learning_rate/qwen_vl_interface': 9.715282268864379e-06, 'learning_rate/action_model': 9.715282268864378e-05, 'epoch': 0.37})
+[2026-07-28 10:50:37,719][starVLA.training.train_starvla][INFO] - Step 2033, Loss: {'train/loss': 0.0724605917930603, 'train/grad_norm_pre_clip': 2.77630615234375, 'timing/data': 0.0003402777947485447, 'timing/model': 0.42637929879128933, 'learning_rate/qwen_vl_interface': 9.713976011568256e-06, 'learning_rate/action_model': 9.713976011568256e-05, 'epoch': 0.37})
+[2026-07-28 10:50:44,073][starVLA.training.train_starvla][INFO] - Step 2034, Loss: {'train/loss': 0.028793368488550186, 'train/grad_norm_pre_clip': 1.2901115417480469, 'timing/data': 0.0002824850380420685, 'timing/model': 0.42179801501333714, 'learning_rate/qwen_vl_interface': 9.712666857646565e-06, 'learning_rate/action_model': 9.712666857646564e-05, 'epoch': 0.37})
+[2026-07-28 10:50:50,405][starVLA.training.train_starvla][INFO] - Step 2035, Loss: {'train/loss': 0.060209475457668304, 'train/grad_norm_pre_clip': 2.1098575592041016, 'timing/data': 0.00032031256705522537, 'timing/model': 0.7150142341852188, 'learning_rate/qwen_vl_interface': 9.711354807948798e-06, 'learning_rate/action_model': 9.711354807948798e-05, 'epoch': 0.37})
+[2026-07-28 10:50:56,427][starVLA.training.train_starvla][INFO] - Step 2036, Loss: {'train/loss': 0.049836717545986176, 'train/grad_norm_pre_clip': 0.9432117938995361, 'timing/data': 0.00028223125264048576, 'timing/model': 0.4227878609672189, 'learning_rate/qwen_vl_interface': 9.710039863326334e-06, 'learning_rate/action_model': 9.710039863326334e-05, 'epoch': 0.37})
+[2026-07-28 10:51:02,718][starVLA.training.train_starvla][INFO] - Step 2037, Loss: {'train/loss': 0.03675242140889168, 'train/grad_norm_pre_clip': 0.9050787687301636, 'timing/data': 0.0004913290031254292, 'timing/model': 0.42481583496555686, 'learning_rate/qwen_vl_interface': 9.708722024632425e-06, 'learning_rate/action_model': 9.708722024632424e-05, 'epoch': 0.37})
+[2026-07-28 10:51:09,369][starVLA.training.train_starvla][INFO] - Step 2038, Loss: {'train/loss': 0.043616242706775665, 'train/grad_norm_pre_clip': 1.449141263961792, 'timing/data': 0.0002725217491388321, 'timing/model': 0.430331835988909, 'learning_rate/qwen_vl_interface': 9.7074012927222e-06, 'learning_rate/action_model': 9.7074012927222e-05, 'epoch': 0.37})
+[2026-07-28 10:51:15,654][starVLA.training.train_starvla][INFO] - Step 2039, Loss: {'train/loss': 0.012121817097067833, 'train/grad_norm_pre_clip': 0.41863125562667847, 'timing/data': 0.00027786195278167725, 'timing/model': 0.4349757479503751, 'learning_rate/qwen_vl_interface': 9.706077668452671e-06, 'learning_rate/action_model': 9.706077668452671e-05, 'epoch': 0.37})
+[2026-07-28 10:51:22,002][starVLA.training.train_starvla][INFO] - Step 2040, Loss: {'train/loss': 0.06715373694896698, 'train/grad_norm_pre_clip': 3.5160343647003174, 'timing/data': 0.0002715671434998512, 'timing/model': 0.4202167927287519, 'learning_rate/qwen_vl_interface': 9.704751152682722e-06, 'learning_rate/action_model': 9.704751152682721e-05, 'epoch': 0.37})
+[2026-07-28 10:51:28,544][starVLA.training.train_starvla][INFO] - Step 2041, Loss: {'train/loss': 0.06702175736427307, 'train/grad_norm_pre_clip': 3.2789554595947266, 'timing/data': 0.0002921121194958687, 'timing/model': 0.7317123021930456, 'learning_rate/qwen_vl_interface': 9.703421746273115e-06, 'learning_rate/action_model': 9.703421746273114e-05, 'epoch': 0.37})
+[2026-07-28 10:51:34,716][starVLA.training.train_starvla][INFO] - Step 2042, Loss: {'train/loss': 0.0807824581861496, 'train/grad_norm_pre_clip': 2.689427137374878, 'timing/data': 0.0003007822670042515, 'timing/model': 0.43150985101237893, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.38})
+[2026-07-28 10:51:40,790][starVLA.training.train_starvla][INFO] - Step 2043, Loss: {'train/loss': 0.04305156320333481, 'train/grad_norm_pre_clip': 1.6545387506484985, 'timing/data': 0.00031299516558647156, 'timing/model': 0.4249918879941106, 'learning_rate/qwen_vl_interface': 9.700754264987347e-06, 'learning_rate/action_model': 9.700754264987347e-05, 'epoch': 0.38})
+[2026-07-28 10:51:47,091][starVLA.training.train_starvla][INFO] - Step 2044, Loss: {'train/loss': 0.06615421175956726, 'train/grad_norm_pre_clip': 1.3259445428848267, 'timing/data': 0.0005478006787598133, 'timing/model': 0.4220976447686553, 'learning_rate/qwen_vl_interface': 9.69941619184209e-06, 'learning_rate/action_model': 9.699416191842089e-05, 'epoch': 0.38})
+[2026-07-28 10:51:53,433][starVLA.training.train_starvla][INFO] - Step 2045, Loss: {'train/loss': 0.10023912787437439, 'train/grad_norm_pre_clip': 2.2092013359069824, 'timing/data': 0.0003771660849452019, 'timing/model': 0.4211512850597501, 'learning_rate/qwen_vl_interface': 9.69807523151897e-06, 'learning_rate/action_model': 9.698075231518969e-05, 'epoch': 0.38})
+[2026-07-28 10:51:59,527][starVLA.training.train_starvla][INFO] - Step 2046, Loss: {'train/loss': 0.05072171986103058, 'train/grad_norm_pre_clip': 1.2095016241073608, 'timing/data': 0.00029634684324264526, 'timing/model': 0.4306645868346095, 'learning_rate/qwen_vl_interface': 9.696731384888125e-06, 'learning_rate/action_model': 9.696731384888125e-05, 'epoch': 0.38})
+[2026-07-28 10:52:05,932][starVLA.training.train_starvla][INFO] - Step 2047, Loss: {'train/loss': 0.033928364515304565, 'train/grad_norm_pre_clip': 1.228196620941162, 'timing/data': 0.00037679681554436684, 'timing/model': 0.6699136290699244, 'learning_rate/qwen_vl_interface': 9.695384652821562e-06, 'learning_rate/action_model': 9.695384652821561e-05, 'epoch': 0.38})
+[2026-07-28 10:52:12,234][starVLA.training.train_starvla][INFO] - Step 2048, Loss: {'train/loss': 0.06498537957668304, 'train/grad_norm_pre_clip': 1.2055542469024658, 'timing/data': 0.000379255972802639, 'timing/model': 0.4206301970407367, 'learning_rate/qwen_vl_interface': 9.694035036193162e-06, 'learning_rate/action_model': 9.694035036193161e-05, 'epoch': 0.38})
+[2026-07-28 10:52:18,299][starVLA.training.train_starvla][INFO] - Step 2049, Loss: {'train/loss': 0.04104822501540184, 'train/grad_norm_pre_clip': 0.7322491407394409, 'timing/data': 0.0005483110435307026, 'timing/model': 0.4232091177254915, 'learning_rate/qwen_vl_interface': 9.692682535878674e-06, 'learning_rate/action_model': 9.692682535878673e-05, 'epoch': 0.38})
+[2026-07-28 10:52:24,387][starVLA.training.train_starvla][INFO] - Step 2050, Loss: {'train/loss': 0.04125642776489258, 'train/grad_norm_pre_clip': 1.1866902112960815, 'timing/data': 0.00029182620346546173, 'timing/model': 0.42457327572628856, 'learning_rate/qwen_vl_interface': 9.691327152755721e-06, 'learning_rate/action_model': 9.691327152755722e-05, 'epoch': 0.38})
+[2026-07-28 10:52:30,463][starVLA.training.train_starvla][INFO] - Step 2051, Loss: {'train/loss': 0.05226786807179451, 'train/grad_norm_pre_clip': 1.3272862434387207, 'timing/data': 0.00039357319474220276, 'timing/model': 0.4271223396062851, 'learning_rate/qwen_vl_interface': 9.6899688877038e-06, 'learning_rate/action_model': 9.689968887703798e-05, 'epoch': 0.38})
+[2026-07-28 10:52:36,817][starVLA.training.train_starvla][INFO] - Step 2052, Loss: {'train/loss': 0.04191522300243378, 'train/grad_norm_pre_clip': 0.9543175101280212, 'timing/data': 0.0002837688662111759, 'timing/model': 0.7157403407618403, 'learning_rate/qwen_vl_interface': 9.688607741604269e-06, 'learning_rate/action_model': 9.688607741604269e-05, 'epoch': 0.38})
+[2026-07-28 10:52:42,691][starVLA.training.train_starvla][INFO] - Step 2053, Loss: {'train/loss': 0.07056906819343567, 'train/grad_norm_pre_clip': 2.2425060272216797, 'timing/data': 0.0002949703484773636, 'timing/model': 0.4241874651052058, 'learning_rate/qwen_vl_interface': 9.687243715340365e-06, 'learning_rate/action_model': 9.687243715340364e-05, 'epoch': 0.38})
+[2026-07-28 10:52:48,511][starVLA.training.train_starvla][INFO] - Step 2054, Loss: {'train/loss': 0.04911845922470093, 'train/grad_norm_pre_clip': 2.6172969341278076, 'timing/data': 0.0004095020703971386, 'timing/model': 0.4231055323034525, 'learning_rate/qwen_vl_interface': 9.685876809797188e-06, 'learning_rate/action_model': 9.685876809797188e-05, 'epoch': 0.38})
+[2026-07-28 10:52:54,561][starVLA.training.train_starvla][INFO] - Step 2055, Loss: {'train/loss': 0.04597293958067894, 'train/grad_norm_pre_clip': 1.1690784692764282, 'timing/data': 0.0003198999911546707, 'timing/model': 0.42388479318469763, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.38})
+[2026-07-28 10:53:00,696][starVLA.training.train_starvla][INFO] - Step 2056, Loss: {'train/loss': 0.054973676800727844, 'train/grad_norm_pre_clip': 1.205704689025879, 'timing/data': 0.00029812101274728775, 'timing/model': 0.4227503230795264, 'learning_rate/qwen_vl_interface': 9.683134364422766e-06, 'learning_rate/action_model': 9.683134364422765e-05, 'epoch': 0.38})
+[2026-07-28 10:53:06,702][starVLA.training.train_starvla][INFO] - Step 2057, Loss: {'train/loss': 0.06692831963300705, 'train/grad_norm_pre_clip': 1.8938623666763306, 'timing/data': 0.0003054649569094181, 'timing/model': 0.42522356985136867, 'learning_rate/qwen_vl_interface': 9.681758826371064e-06, 'learning_rate/action_model': 9.681758826371063e-05, 'epoch': 0.38})
+[2026-07-28 10:53:13,272][starVLA.training.train_starvla][INFO] - Step 2058, Loss: {'train/loss': 0.056920409202575684, 'train/grad_norm_pre_clip': 1.6430485248565674, 'timing/data': 0.00033246399834752083, 'timing/model': 0.6600811742246151, 'learning_rate/qwen_vl_interface': 9.680380412599174e-06, 'learning_rate/action_model': 9.680380412599173e-05, 'epoch': 0.38})
+[2026-07-28 10:53:19,325][starVLA.training.train_starvla][INFO] - Step 2059, Loss: {'train/loss': 0.055418066680431366, 'train/grad_norm_pre_clip': 1.9249910116195679, 'timing/data': 0.00027704378589987755, 'timing/model': 0.42019716603681445, 'learning_rate/qwen_vl_interface': 9.678999124001533e-06, 'learning_rate/action_model': 9.678999124001533e-05, 'epoch': 0.38})
+[2026-07-28 10:53:25,351][starVLA.training.train_starvla][INFO] - Step 2060, Loss: {'train/loss': 0.10142402350902557, 'train/grad_norm_pre_clip': 1.1072741746902466, 'timing/data': 0.0002859393134713173, 'timing/model': 0.4228174132294953, 'learning_rate/qwen_vl_interface': 9.67761496147445e-06, 'learning_rate/action_model': 9.677614961474449e-05, 'epoch': 0.38})
+[2026-07-28 10:53:31,386][starVLA.training.train_starvla][INFO] - Step 2061, Loss: {'train/loss': 0.10727684199810028, 'train/grad_norm_pre_clip': 4.347090244293213, 'timing/data': 0.00038238195702433586, 'timing/model': 0.42248116014525294, 'learning_rate/qwen_vl_interface': 9.676227925916082e-06, 'learning_rate/action_model': 9.676227925916082e-05, 'epoch': 0.38})
+[2026-07-28 10:53:37,668][starVLA.training.train_starvla][INFO] - Step 2062, Loss: {'train/loss': 0.06037522852420807, 'train/grad_norm_pre_clip': 1.7997188568115234, 'timing/data': 0.0002793031744658947, 'timing/model': 0.42861367482692003, 'learning_rate/qwen_vl_interface': 9.674838018226472e-06, 'learning_rate/action_model': 9.674838018226471e-05, 'epoch': 0.38})
+[2026-07-28 10:53:43,931][starVLA.training.train_starvla][INFO] - Step 2063, Loss: {'train/loss': 0.0864889919757843, 'train/grad_norm_pre_clip': 3.279006004333496, 'timing/data': 0.00028199469670653343, 'timing/model': 0.41997983073815703, 'learning_rate/qwen_vl_interface': 9.67344523930751e-06, 'learning_rate/action_model': 9.673445239307511e-05, 'epoch': 0.38})
+[2026-07-28 10:53:49,959][starVLA.training.train_starvla][INFO] - Step 2064, Loss: {'train/loss': 0.05598771572113037, 'train/grad_norm_pre_clip': 1.95383882522583, 'timing/data': 0.0003477446734905243, 'timing/model': 0.4283289280720055, 'learning_rate/qwen_vl_interface': 9.672049590062959e-06, 'learning_rate/action_model': 9.672049590062958e-05, 'epoch': 0.38})
+[2026-07-28 10:53:56,307][starVLA.training.train_starvla][INFO] - Step 2065, Loss: {'train/loss': 0.06267879903316498, 'train/grad_norm_pre_clip': 1.0730741024017334, 'timing/data': 0.0004840381443500519, 'timing/model': 0.6672493889927864, 'learning_rate/qwen_vl_interface': 9.670651071398436e-06, 'learning_rate/action_model': 9.670651071398436e-05, 'epoch': 0.38})
+[2026-07-28 10:54:02,631][starVLA.training.train_starvla][INFO] - Step 2066, Loss: {'train/loss': 0.09375971555709839, 'train/grad_norm_pre_clip': 1.294839859008789, 'timing/data': 0.000453081913292408, 'timing/model': 0.4410760523751378, 'learning_rate/qwen_vl_interface': 9.669249684221427e-06, 'learning_rate/action_model': 9.669249684221428e-05, 'epoch': 0.38})
+[2026-07-28 10:54:08,966][starVLA.training.train_starvla][INFO] - Step 2067, Loss: {'train/loss': 0.026620710268616676, 'train/grad_norm_pre_clip': 0.8190024495124817, 'timing/data': 0.0003207889385521412, 'timing/model': 0.4271611929871142, 'learning_rate/qwen_vl_interface': 9.66784542944128e-06, 'learning_rate/action_model': 9.66784542944128e-05, 'epoch': 0.38})
+[2026-07-28 10:54:15,074][starVLA.training.train_starvla][INFO] - Step 2068, Loss: {'train/loss': 0.0851113572716713, 'train/grad_norm_pre_clip': 2.546276807785034, 'timing/data': 0.00030354084447026253, 'timing/model': 0.42574746208265424, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.38})
+[2026-07-28 10:54:21,150][starVLA.training.train_starvla][INFO] - Step 2069, Loss: {'train/loss': 0.057648494839668274, 'train/grad_norm_pre_clip': 1.2955899238586426, 'timing/data': 0.0003022640012204647, 'timing/model': 0.42365287337452173, 'learning_rate/qwen_vl_interface': 9.66502832071824e-06, 'learning_rate/action_model': 9.66502832071824e-05, 'epoch': 0.38})
+[2026-07-28 10:54:27,438][starVLA.training.train_starvla][INFO] - Step 2070, Loss: {'train/loss': 0.10333699733018875, 'train/grad_norm_pre_clip': 2.891569137573242, 'timing/data': 0.00031101005151867867, 'timing/model': 0.4219639031216502, 'learning_rate/qwen_vl_interface': 9.663615468603344e-06, 'learning_rate/action_model': 9.663615468603343e-05, 'epoch': 0.38})
+[2026-07-28 10:54:33,527][starVLA.training.train_starvla][INFO] - Step 2071, Loss: {'train/loss': 0.06799700856208801, 'train/grad_norm_pre_clip': 1.4684535264968872, 'timing/data': 0.000402974896132946, 'timing/model': 0.4262290857732296, 'learning_rate/qwen_vl_interface': 9.662199752541287e-06, 'learning_rate/action_model': 9.662199752541287e-05, 'epoch': 0.38})
+[2026-07-28 10:54:39,916][starVLA.training.train_starvla][INFO] - Step 2072, Loss: {'train/loss': 0.046269528567790985, 'train/grad_norm_pre_clip': 1.0146090984344482, 'timing/data': 0.00029640505090355873, 'timing/model': 0.677133072167635, 'learning_rate/qwen_vl_interface': 9.660781173450711e-06, 'learning_rate/action_model': 9.660781173450712e-05, 'epoch': 0.38})
+[2026-07-28 10:54:45,950][starVLA.training.train_starvla][INFO] - Step 2073, Loss: {'train/loss': 0.048779670149087906, 'train/grad_norm_pre_clip': 2.0090103149414062, 'timing/data': 0.000284428708255291, 'timing/model': 0.42630500439554453, 'learning_rate/qwen_vl_interface': 9.659359732252121e-06, 'learning_rate/action_model': 9.659359732252121e-05, 'epoch': 0.38})
+[2026-07-28 10:54:52,039][starVLA.training.train_starvla][INFO] - Step 2074, Loss: {'train/loss': 0.046401914209127426, 'train/grad_norm_pre_clip': 1.6410424709320068, 'timing/data': 0.0005863397382199764, 'timing/model': 0.4216313208453357, 'learning_rate/qwen_vl_interface': 9.657935429867872e-06, 'learning_rate/action_model': 9.657935429867873e-05, 'epoch': 0.38})
+[2026-07-28 10:54:58,265][starVLA.training.train_starvla][INFO] - Step 2075, Loss: {'train/loss': 0.04101332277059555, 'train/grad_norm_pre_clip': 0.8285636901855469, 'timing/data': 0.00027380790561437607, 'timing/model': 0.6604082230478525, 'learning_rate/qwen_vl_interface': 9.65650826722218e-06, 'learning_rate/action_model': 9.656508267222179e-05, 'epoch': 0.38})
+[2026-07-28 10:55:04,338][starVLA.training.train_starvla][INFO] - Step 2076, Loss: {'train/loss': 0.053059861063957214, 'train/grad_norm_pre_clip': 0.9715957641601562, 'timing/data': 0.00031806714832782745, 'timing/model': 0.43476728489622474, 'learning_rate/qwen_vl_interface': 9.655078245241114e-06, 'learning_rate/action_model': 9.655078245241112e-05, 'epoch': 0.38})
+[2026-07-28 10:55:10,354][starVLA.training.train_starvla][INFO] - Step 2077, Loss: {'train/loss': 0.0742630660533905, 'train/grad_norm_pre_clip': 2.0628890991210938, 'timing/data': 0.0003055129200220108, 'timing/model': 0.4220740678720176, 'learning_rate/qwen_vl_interface': 9.653645364852598e-06, 'learning_rate/action_model': 9.653645364852597e-05, 'epoch': 0.38})
+[2026-07-28 10:55:16,458][starVLA.training.train_starvla][INFO] - Step 2078, Loss: {'train/loss': 0.08287608623504639, 'train/grad_norm_pre_clip': 1.5377929210662842, 'timing/data': 0.00032049510627985, 'timing/model': 0.4346497496590018, 'learning_rate/qwen_vl_interface': 9.652209626986414e-06, 'learning_rate/action_model': 9.652209626986413e-05, 'epoch': 0.38})
+[2026-07-28 10:55:22,756][starVLA.training.train_starvla][INFO] - Step 2079, Loss: {'train/loss': 0.012158703058958054, 'train/grad_norm_pre_clip': 0.9891056418418884, 'timing/data': 0.0003020702861249447, 'timing/model': 0.4219463490881026, 'learning_rate/qwen_vl_interface': 9.6507710325742e-06, 'learning_rate/action_model': 9.6507710325742e-05, 'epoch': 0.38})
+[2026-07-28 10:55:29,026][starVLA.training.train_starvla][INFO] - Step 2080, Loss: {'train/loss': 0.015623149462044239, 'train/grad_norm_pre_clip': 0.4923132359981537, 'timing/data': 0.0003935457207262516, 'timing/model': 0.4260088140144944, 'learning_rate/qwen_vl_interface': 9.649329582549439e-06, 'learning_rate/action_model': 9.649329582549439e-05, 'epoch': 0.38})
+[2026-07-28 10:55:35,076][starVLA.training.train_starvla][INFO] - Step 2081, Loss: {'train/loss': 0.0497162826359272, 'train/grad_norm_pre_clip': 0.8189122676849365, 'timing/data': 0.0002989852800965309, 'timing/model': 0.42445468017831445, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.38})
+[2026-07-28 10:55:40,858][starVLA.training.train_starvla][INFO] - Step 2082, Loss: {'train/loss': 0.042706310749053955, 'train/grad_norm_pre_clip': 2.1954121589660645, 'timing/data': 0.00035882415249943733, 'timing/model': 0.4229140062816441, 'learning_rate/qwen_vl_interface': 9.646438119405503e-06, 'learning_rate/action_model': 9.646438119405503e-05, 'epoch': 0.38})
+[2026-07-28 10:55:47,128][starVLA.training.train_starvla][INFO] - Step 2083, Loss: {'train/loss': 0.027415912598371506, 'train/grad_norm_pre_clip': 2.7364537715911865, 'timing/data': 0.00031276559457182884, 'timing/model': 0.6655701813288033, 'learning_rate/qwen_vl_interface': 9.644988108162566e-06, 'learning_rate/action_model': 9.644988108162567e-05, 'epoch': 0.38})
+[2026-07-28 10:55:53,199][starVLA.training.train_starvla][INFO] - Step 2084, Loss: {'train/loss': 0.07665751874446869, 'train/grad_norm_pre_clip': 2.2682995796203613, 'timing/data': 0.00030946824699640274, 'timing/model': 0.42418140079826117, 'learning_rate/qwen_vl_interface': 9.643535245059562e-06, 'learning_rate/action_model': 9.643535245059561e-05, 'epoch': 0.38})
+[2026-07-28 10:55:59,015][starVLA.training.train_starvla][INFO] - Step 2085, Loss: {'train/loss': 0.049496088176965714, 'train/grad_norm_pre_clip': 1.747698187828064, 'timing/data': 0.00031342124566435814, 'timing/model': 0.42512340284883976, 'learning_rate/qwen_vl_interface': 9.642079531039235e-06, 'learning_rate/action_model': 9.642079531039235e-05, 'epoch': 0.38})
+[2026-07-28 10:56:05,236][starVLA.training.train_starvla][INFO] - Step 2086, Loss: {'train/loss': 0.06273908168077469, 'train/grad_norm_pre_clip': 2.639726400375366, 'timing/data': 0.0002970048226416111, 'timing/model': 0.6166391726583242, 'learning_rate/qwen_vl_interface': 9.640620967046186e-06, 'learning_rate/action_model': 9.640620967046186e-05, 'epoch': 0.38})
+[2026-07-28 10:56:11,508][starVLA.training.train_starvla][INFO] - Step 2087, Loss: {'train/loss': 0.011221702210605145, 'train/grad_norm_pre_clip': 0.6629819869995117, 'timing/data': 0.0002704579383134842, 'timing/model': 0.42211179388687015, 'learning_rate/qwen_vl_interface': 9.63915955402686e-06, 'learning_rate/action_model': 9.63915955402686e-05, 'epoch': 0.38})
+[2026-07-28 10:56:17,576][starVLA.training.train_starvla][INFO] - Step 2088, Loss: {'train/loss': 0.04856827110052109, 'train/grad_norm_pre_clip': 3.206625461578369, 'timing/data': 0.0003240578807890415, 'timing/model': 0.4253702340647578, 'learning_rate/qwen_vl_interface': 9.637695292929552e-06, 'learning_rate/action_model': 9.637695292929552e-05, 'epoch': 0.38})
+[2026-07-28 10:56:23,630][starVLA.training.train_starvla][INFO] - Step 2089, Loss: {'train/loss': 0.05656049773097038, 'train/grad_norm_pre_clip': 2.18780255317688, 'timing/data': 0.000384259968996048, 'timing/model': 0.42409130092710257, 'learning_rate/qwen_vl_interface': 9.636228184704405e-06, 'learning_rate/action_model': 9.636228184704405e-05, 'epoch': 0.38})
+[2026-07-28 10:56:29,913][starVLA.training.train_starvla][INFO] - Step 2090, Loss: {'train/loss': 0.049768466502428055, 'train/grad_norm_pre_clip': 2.3611340522766113, 'timing/data': 0.00037928298115730286, 'timing/model': 0.4238403611816466, 'learning_rate/qwen_vl_interface': 9.634758230303408e-06, 'learning_rate/action_model': 9.634758230303408e-05, 'epoch': 0.38})
+[2026-07-28 10:56:36,166][starVLA.training.train_starvla][INFO] - Step 2091, Loss: {'train/loss': 0.06997015327215195, 'train/grad_norm_pre_clip': 3.0367777347564697, 'timing/data': 0.0002776351757347584, 'timing/model': 0.42172853788360953, 'learning_rate/qwen_vl_interface': 9.633285430680401e-06, 'learning_rate/action_model': 9.633285430680401e-05, 'epoch': 0.38})
+[2026-07-28 10:56:42,172][starVLA.training.train_starvla][INFO] - Step 2092, Loss: {'train/loss': 0.07601400464773178, 'train/grad_norm_pre_clip': 1.2207520008087158, 'timing/data': 0.00029243016615509987, 'timing/model': 0.4242621441371739, 'learning_rate/qwen_vl_interface': 9.631809786791066e-06, 'learning_rate/action_model': 9.631809786791066e-05, 'epoch': 0.38})
+[2026-07-28 10:56:48,209][starVLA.training.train_starvla][INFO] - Step 2093, Loss: {'train/loss': 0.031015688553452492, 'train/grad_norm_pre_clip': 1.4105579853057861, 'timing/data': 0.0004038452170789242, 'timing/model': 0.4237212878651917, 'learning_rate/qwen_vl_interface': 9.630331299592932e-06, 'learning_rate/action_model': 9.630331299592933e-05, 'epoch': 0.38})
+[2026-07-28 10:56:54,831][starVLA.training.train_starvla][INFO] - Step 2094, Loss: {'train/loss': 0.047146063297986984, 'train/grad_norm_pre_clip': 1.0414572954177856, 'timing/data': 0.00028842873871326447, 'timing/model': 0.7143306480720639, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.38})
+[2026-07-28 10:57:00,614][starVLA.training.train_starvla][INFO] - Step 2095, Loss: {'train/loss': 0.04712239280343056, 'train/grad_norm_pre_clip': 0.9799882769584656, 'timing/data': 0.00028593884781003, 'timing/model': 0.43104240763932467, 'learning_rate/qwen_vl_interface': 9.627365799109617e-06, 'learning_rate/action_model': 9.627365799109617e-05, 'epoch': 0.38})
+[2026-07-28 10:57:06,689][starVLA.training.train_starvla][INFO] - Step 2096, Loss: {'train/loss': 0.08165080845355988, 'train/grad_norm_pre_clip': 2.341472625732422, 'timing/data': 0.0003685462288558483, 'timing/model': 0.4242937038652599, 'learning_rate/qwen_vl_interface': 9.625878787748712e-06, 'learning_rate/action_model': 9.625878787748713e-05, 'epoch': 0.38})
+[2026-07-28 10:57:12,760][starVLA.training.train_starvla][INFO] - Step 2097, Loss: {'train/loss': 0.06867839395999908, 'train/grad_norm_pre_clip': 1.9699559211730957, 'timing/data': 0.00029878271743655205, 'timing/model': 0.43688893411308527, 'learning_rate/qwen_vl_interface': 9.624388936927572e-06, 'learning_rate/action_model': 9.624388936927571e-05, 'epoch': 0.39})
+[2026-07-28 10:57:18,923][starVLA.training.train_starvla][INFO] - Step 2098, Loss: {'train/loss': 0.04015945643186569, 'train/grad_norm_pre_clip': 1.175517201423645, 'timing/data': 0.0007579210214316845, 'timing/model': 0.44544860115274787, 'learning_rate/qwen_vl_interface': 9.622896247612943e-06, 'learning_rate/action_model': 9.622896247612942e-05, 'epoch': 0.39})
+[2026-07-28 10:57:25,257][starVLA.training.train_starvla][INFO] - Step 2099, Loss: {'train/loss': 0.08213905245065689, 'train/grad_norm_pre_clip': 1.3103982210159302, 'timing/data': 0.0003150659613311291, 'timing/model': 0.6929436321370304, 'learning_rate/qwen_vl_interface': 9.621400720773414e-06, 'learning_rate/action_model': 9.621400720773414e-05, 'epoch': 0.39})
+[2026-07-28 10:57:31,117][starVLA.training.train_starvla][INFO] - Step 2100, Loss: {'train/loss': 0.05594597011804581, 'train/grad_norm_pre_clip': 1.7307204008102417, 'timing/data': 0.0003445036709308624, 'timing/model': 0.42186562390998006, 'learning_rate/qwen_vl_interface': 9.619902357379416e-06, 'learning_rate/action_model': 9.619902357379415e-05, 'epoch': 0.39})
+[2026-07-28 10:57:37,189][starVLA.training.train_starvla][INFO] - Step 2101, Loss: {'train/loss': 0.0957169234752655, 'train/grad_norm_pre_clip': 2.3346188068389893, 'timing/data': 0.0003111870028078556, 'timing/model': 0.43355876486748457, 'learning_rate/qwen_vl_interface': 9.618401158403222e-06, 'learning_rate/action_model': 9.618401158403222e-05, 'epoch': 0.39})
+[2026-07-28 10:57:43,237][starVLA.training.train_starvla][INFO] - Step 2102, Loss: {'train/loss': 0.0601392537355423, 'train/grad_norm_pre_clip': 0.8074675798416138, 'timing/data': 0.0003697550855576992, 'timing/model': 0.42322047939524055, 'learning_rate/qwen_vl_interface': 9.616897124818944e-06, 'learning_rate/action_model': 9.616897124818945e-05, 'epoch': 0.39})
+[2026-07-28 10:57:49,313][starVLA.training.train_starvla][INFO] - Step 2103, Loss: {'train/loss': 0.08741831034421921, 'train/grad_norm_pre_clip': 3.9110231399536133, 'timing/data': 0.0002902350388467312, 'timing/model': 0.4352501006796956, 'learning_rate/qwen_vl_interface': 9.615390257602532e-06, 'learning_rate/action_model': 9.615390257602531e-05, 'epoch': 0.39})
+[2026-07-28 10:57:55,581][starVLA.training.train_starvla][INFO] - Step 2104, Loss: {'train/loss': 0.04189349338412285, 'train/grad_norm_pre_clip': 2.2660460472106934, 'timing/data': 0.00033519230782985687, 'timing/model': 0.4231616067700088, 'learning_rate/qwen_vl_interface': 9.613880557731776e-06, 'learning_rate/action_model': 9.613880557731776e-05, 'epoch': 0.39})
+[2026-07-28 10:58:01,967][starVLA.training.train_starvla][INFO] - Step 2105, Loss: {'train/loss': 0.05841698870062828, 'train/grad_norm_pre_clip': 2.413426637649536, 'timing/data': 0.00035743601620197296, 'timing/model': 0.6669281013309956, 'learning_rate/qwen_vl_interface': 9.612368026186304e-06, 'learning_rate/action_model': 9.612368026186304e-05, 'epoch': 0.39})
+[2026-07-28 10:58:08,031][starVLA.training.train_starvla][INFO] - Step 2106, Loss: {'train/loss': 0.07288722693920135, 'train/grad_norm_pre_clip': 2.954491138458252, 'timing/data': 0.00027729524299502373, 'timing/model': 0.4287263839505613, 'learning_rate/qwen_vl_interface': 9.610852663947582e-06, 'learning_rate/action_model': 9.610852663947583e-05, 'epoch': 0.39})
+[2026-07-28 10:58:14,367][starVLA.training.train_starvla][INFO] - Step 2107, Loss: {'train/loss': 0.04433602839708328, 'train/grad_norm_pre_clip': 1.247313380241394, 'timing/data': 0.0003129690885543823, 'timing/model': 0.4219831880182028, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.39})
+[2026-07-28 10:58:20,411][starVLA.training.train_starvla][INFO] - Step 2108, Loss: {'train/loss': 0.04910384863615036, 'train/grad_norm_pre_clip': 1.055033802986145, 'timing/data': 0.0002736542373895645, 'timing/model': 0.4242875981144607, 'learning_rate/qwen_vl_interface': 9.607813451325431e-06, 'learning_rate/action_model': 9.60781345132543e-05, 'epoch': 0.39})
+[2026-07-28 10:58:26,510][starVLA.training.train_starvla][INFO] - Step 2109, Loss: {'train/loss': 0.06920883804559708, 'train/grad_norm_pre_clip': 0.9828562140464783, 'timing/data': 0.0003605661913752556, 'timing/model': 0.4324820158071816, 'learning_rate/qwen_vl_interface': 9.606289602914113e-06, 'learning_rate/action_model': 9.606289602914112e-05, 'epoch': 0.39})
+[2026-07-28 10:58:32,860][starVLA.training.train_starvla][INFO] - Step 2110, Loss: {'train/loss': 0.08503228425979614, 'train/grad_norm_pre_clip': 7.07882833480835, 'timing/data': 0.0002954029478132725, 'timing/model': 0.4252782007679343, 'learning_rate/qwen_vl_interface': 9.604762927753768e-06, 'learning_rate/action_model': 9.604762927753768e-05, 'epoch': 0.39})
+[2026-07-28 10:58:39,168][starVLA.training.train_starvla][INFO] - Step 2111, Loss: {'train/loss': 0.07364009320735931, 'train/grad_norm_pre_clip': 1.766168475151062, 'timing/data': 0.0003770231269299984, 'timing/model': 0.6817503622733057, 'learning_rate/qwen_vl_interface': 9.603233426835037e-06, 'learning_rate/action_model': 9.603233426835036e-05, 'epoch': 0.39})
+[2026-07-28 10:58:45,224][starVLA.training.train_starvla][INFO] - Step 2112, Loss: {'train/loss': 0.04669336974620819, 'train/grad_norm_pre_clip': 0.8905275464057922, 'timing/data': 0.0002802540548145771, 'timing/model': 0.42200688226148486, 'learning_rate/qwen_vl_interface': 9.601701101150399e-06, 'learning_rate/action_model': 9.601701101150398e-05, 'epoch': 0.39})
+[2026-07-28 10:58:51,295][starVLA.training.train_starvla][INFO] - Step 2113, Loss: {'train/loss': 0.06561294198036194, 'train/grad_norm_pre_clip': 1.892349123954773, 'timing/data': 0.00038695894181728363, 'timing/model': 0.42271037492901087, 'learning_rate/qwen_vl_interface': 9.600165951694162e-06, 'learning_rate/action_model': 9.60016595169416e-05, 'epoch': 0.39})
+[2026-07-28 10:58:57,640][starVLA.training.train_starvla][INFO] - Step 2114, Loss: {'train/loss': 0.07959050685167313, 'train/grad_norm_pre_clip': 2.605844259262085, 'timing/data': 0.00030212895944714546, 'timing/model': 0.42086373176425695, 'learning_rate/qwen_vl_interface': 9.598627979462466e-06, 'learning_rate/action_model': 9.598627979462467e-05, 'epoch': 0.39})
+[2026-07-28 10:59:03,691][starVLA.training.train_starvla][INFO] - Step 2115, Loss: {'train/loss': 0.029015056788921356, 'train/grad_norm_pre_clip': 1.424593210220337, 'timing/data': 0.0003114459104835987, 'timing/model': 0.42130992002785206, 'learning_rate/qwen_vl_interface': 9.59708718545329e-06, 'learning_rate/action_model': 9.597087185453288e-05, 'epoch': 0.39})
+[2026-07-28 10:59:09,706][starVLA.training.train_starvla][INFO] - Step 2116, Loss: {'train/loss': 0.05655861645936966, 'train/grad_norm_pre_clip': 1.4763189554214478, 'timing/data': 0.0003515910357236862, 'timing/model': 0.42289916798472404, 'learning_rate/qwen_vl_interface': 9.595543570666433e-06, 'learning_rate/action_model': 9.595543570666432e-05, 'epoch': 0.39})
+[2026-07-28 10:59:15,968][starVLA.training.train_starvla][INFO] - Step 2117, Loss: {'train/loss': 0.05521547049283981, 'train/grad_norm_pre_clip': 0.7127615809440613, 'timing/data': 0.00027976976707577705, 'timing/model': 0.42412094213068485, 'learning_rate/qwen_vl_interface': 9.593997136103533e-06, 'learning_rate/action_model': 9.593997136103533e-05, 'epoch': 0.39})
+[2026-07-28 10:59:22,046][starVLA.training.train_starvla][INFO] - Step 2118, Loss: {'train/loss': 0.05924238637089729, 'train/grad_norm_pre_clip': 0.7534244656562805, 'timing/data': 0.0003828941844403744, 'timing/model': 0.423079386819154, 'learning_rate/qwen_vl_interface': 9.592447882768053e-06, 'learning_rate/action_model': 9.592447882768052e-05, 'epoch': 0.39})
+[2026-07-28 10:59:28,458][starVLA.training.train_starvla][INFO] - Step 2119, Loss: {'train/loss': 0.03906319662928581, 'train/grad_norm_pre_clip': 0.9607049226760864, 'timing/data': 0.00030470313504338264, 'timing/model': 0.6699026729911566, 'learning_rate/qwen_vl_interface': 9.590895811665284e-06, 'learning_rate/action_model': 9.590895811665285e-05, 'epoch': 0.39})
+[2026-07-28 10:59:34,477][starVLA.training.train_starvla][INFO] - Step 2120, Loss: {'train/loss': 0.03604475408792496, 'train/grad_norm_pre_clip': 1.0245189666748047, 'timing/data': 0.00035237008705735207, 'timing/model': 0.423122791107744, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.39})
+[2026-07-28 10:59:40,826][starVLA.training.train_starvla][INFO] - Step 2121, Loss: {'train/loss': 0.04800311475992203, 'train/grad_norm_pre_clip': 2.572251796722412, 'timing/data': 0.00029321806505322456, 'timing/model': 0.4421182903461158, 'learning_rate/qwen_vl_interface': 9.587783220188212e-06, 'learning_rate/action_model': 9.587783220188212e-05, 'epoch': 0.39})
+[2026-07-28 10:59:46,650][starVLA.training.train_starvla][INFO] - Step 2122, Loss: {'train/loss': 0.055100493133068085, 'train/grad_norm_pre_clip': 0.9228184223175049, 'timing/data': 0.00028922734782099724, 'timing/model': 0.42399724991992116, 'learning_rate/qwen_vl_interface': 9.586222701833631e-06, 'learning_rate/action_model': 9.58622270183363e-05, 'epoch': 0.39})
+[2026-07-28 10:59:52,744][starVLA.training.train_starvla][INFO] - Step 2123, Loss: {'train/loss': 0.08399499207735062, 'train/grad_norm_pre_clip': 2.9417221546173096, 'timing/data': 0.00030634505674242973, 'timing/model': 0.43498564092442393, 'learning_rate/qwen_vl_interface': 9.584659369751216e-06, 'learning_rate/action_model': 9.584659369751215e-05, 'epoch': 0.39})
+[2026-07-28 10:59:58,994][starVLA.training.train_starvla][INFO] - Step 2124, Loss: {'train/loss': 0.04630444943904877, 'train/grad_norm_pre_clip': 1.3839019536972046, 'timing/data': 0.0003365371376276016, 'timing/model': 0.42254106188192964, 'learning_rate/qwen_vl_interface': 9.583093224955397e-06, 'learning_rate/action_model': 9.583093224955397e-05, 'epoch': 0.39})
+[2026-07-28 11:00:05,084][starVLA.training.train_starvla][INFO] - Step 2125, Loss: {'train/loss': 0.06641820073127747, 'train/grad_norm_pre_clip': 1.4585567712783813, 'timing/data': 0.0005554207600653172, 'timing/model': 0.43240980012342334, 'learning_rate/qwen_vl_interface': 9.58152426846243e-06, 'learning_rate/action_model': 9.581524268462429e-05, 'epoch': 0.39})
+[2026-07-28 11:00:11,343][starVLA.training.train_starvla][INFO] - Step 2126, Loss: {'train/loss': 0.06259579956531525, 'train/grad_norm_pre_clip': 1.7368972301483154, 'timing/data': 0.00041523901745676994, 'timing/model': 0.4318315777927637, 'learning_rate/qwen_vl_interface': 9.579952501290388e-06, 'learning_rate/action_model': 9.579952501290387e-05, 'epoch': 0.39})
+[2026-07-28 11:00:17,432][starVLA.training.train_starvla][INFO] - Step 2127, Loss: {'train/loss': 0.06301005929708481, 'train/grad_norm_pre_clip': 2.078214645385742, 'timing/data': 0.00041606510058045387, 'timing/model': 0.43047873489558697, 'learning_rate/qwen_vl_interface': 9.57837792445918e-06, 'learning_rate/action_model': 9.578377924459181e-05, 'epoch': 0.39})
+[2026-07-28 11:00:23,787][starVLA.training.train_starvla][INFO] - Step 2128, Loss: {'train/loss': 0.06287544220685959, 'train/grad_norm_pre_clip': 0.7294681072235107, 'timing/data': 0.00030438927933573723, 'timing/model': 0.4320299453102052, 'learning_rate/qwen_vl_interface': 9.57680053899053e-06, 'learning_rate/action_model': 9.576800538990529e-05, 'epoch': 0.39})
+[2026-07-28 11:00:29,845][starVLA.training.train_starvla][INFO] - Step 2129, Loss: {'train/loss': 0.06113552302122116, 'train/grad_norm_pre_clip': 1.6199164390563965, 'timing/data': 0.0003724009729921818, 'timing/model': 0.42409402085468173, 'learning_rate/qwen_vl_interface': 9.575220345907984e-06, 'learning_rate/action_model': 9.575220345907984e-05, 'epoch': 0.39})
+[2026-07-28 11:00:36,242][starVLA.training.train_starvla][INFO] - Step 2130, Loss: {'train/loss': 0.09298794716596603, 'train/grad_norm_pre_clip': 2.0031449794769287, 'timing/data': 0.00045836856588721275, 'timing/model': 0.7037743651308119, 'learning_rate/qwen_vl_interface': 9.573637346236915e-06, 'learning_rate/action_model': 9.573637346236915e-05, 'epoch': 0.39})
+[2026-07-28 11:00:42,393][starVLA.training.train_starvla][INFO] - Step 2131, Loss: {'train/loss': 0.03703188896179199, 'train/grad_norm_pre_clip': 0.8234393000602722, 'timing/data': 0.0003870176151394844, 'timing/model': 0.42343999398872256, 'learning_rate/qwen_vl_interface': 9.572051541004513e-06, 'learning_rate/action_model': 9.572051541004513e-05, 'epoch': 0.39})
+[2026-07-28 11:00:48,744][starVLA.training.train_starvla][INFO] - Step 2132, Loss: {'train/loss': 0.03164491429924965, 'train/grad_norm_pre_clip': 1.1372888088226318, 'timing/data': 0.00029161758720874786, 'timing/model': 0.42204663902521133, 'learning_rate/qwen_vl_interface': 9.57046293123979e-06, 'learning_rate/action_model': 9.570462931239789e-05, 'epoch': 0.39})
+[2026-07-28 11:00:54,837][starVLA.training.train_starvla][INFO] - Step 2133, Loss: {'train/loss': 0.025397783145308495, 'train/grad_norm_pre_clip': 0.7138988971710205, 'timing/data': 0.0003045680932700634, 'timing/model': 0.425665030721575, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.39})
+[2026-07-28 11:01:01,179][starVLA.training.train_starvla][INFO] - Step 2134, Loss: {'train/loss': 0.030150586739182472, 'train/grad_norm_pre_clip': 0.7421267032623291, 'timing/data': 0.0003742659464478493, 'timing/model': 0.4224567278288305, 'learning_rate/qwen_vl_interface': 9.567277302238524e-06, 'learning_rate/action_model': 9.567277302238524e-05, 'epoch': 0.39})
+[2026-07-28 11:01:07,261][starVLA.training.train_starvla][INFO] - Step 2135, Loss: {'train/loss': 0.0578603520989418, 'train/grad_norm_pre_clip': 2.4080822467803955, 'timing/data': 0.000331126619130373, 'timing/model': 0.43420536909252405, 'learning_rate/qwen_vl_interface': 9.565680285069104e-06, 'learning_rate/action_model': 9.565680285069103e-05, 'epoch': 0.39})
+[2026-07-28 11:01:13,317][starVLA.training.train_starvla][INFO] - Step 2136, Loss: {'train/loss': 0.05190819874405861, 'train/grad_norm_pre_clip': 1.1910874843597412, 'timing/data': 0.00036856671795248985, 'timing/model': 0.4390260810032487, 'learning_rate/qwen_vl_interface': 9.564080467501601e-06, 'learning_rate/action_model': 9.5640804675016e-05, 'epoch': 0.39})
+[2026-07-28 11:01:19,945][starVLA.training.train_starvla][INFO] - Step 2137, Loss: {'train/loss': 0.030274417251348495, 'train/grad_norm_pre_clip': 0.9083030819892883, 'timing/data': 0.00039155595004558563, 'timing/model': 0.6770306332036853, 'learning_rate/qwen_vl_interface': 9.562477850574118e-06, 'learning_rate/action_model': 9.562477850574119e-05, 'epoch': 0.39})
+[2026-07-28 11:01:26,192][starVLA.training.train_starvla][INFO] - Step 2138, Loss: {'train/loss': 0.05938108265399933, 'train/grad_norm_pre_clip': 1.9605791568756104, 'timing/data': 0.00038655009120702744, 'timing/model': 0.4278361457400024, 'learning_rate/qwen_vl_interface': 9.56087243532658e-06, 'learning_rate/action_model': 9.56087243532658e-05, 'epoch': 0.39})
+[2026-07-28 11:01:32,239][starVLA.training.train_starvla][INFO] - Step 2139, Loss: {'train/loss': 0.0733686164021492, 'train/grad_norm_pre_clip': 4.098689079284668, 'timing/data': 0.0003667757846415043, 'timing/model': 0.42437367839738727, 'learning_rate/qwen_vl_interface': 9.55926422280072e-06, 'learning_rate/action_model': 9.55926422280072e-05, 'epoch': 0.39})
+[2026-07-28 11:01:38,130][starVLA.training.train_starvla][INFO] - Step 2140, Loss: {'train/loss': 0.1287289559841156, 'train/grad_norm_pre_clip': 2.141028642654419, 'timing/data': 0.00037062307819724083, 'timing/model': 0.4256087951362133, 'learning_rate/qwen_vl_interface': 9.55765321404009e-06, 'learning_rate/action_model': 9.557653214040091e-05, 'epoch': 0.39})
+[2026-07-28 11:01:44,625][starVLA.training.train_starvla][INFO] - Step 2141, Loss: {'train/loss': 0.030189435929059982, 'train/grad_norm_pre_clip': 1.4493740797042847, 'timing/data': 0.00033180322498083115, 'timing/model': 0.7228500777855515, 'learning_rate/qwen_vl_interface': 9.556039410090061e-06, 'learning_rate/action_model': 9.556039410090061e-05, 'epoch': 0.39})
+[2026-07-28 11:01:50,908][starVLA.training.train_starvla][INFO] - Step 2142, Loss: {'train/loss': 0.043639279901981354, 'train/grad_norm_pre_clip': 1.3963615894317627, 'timing/data': 0.0002738749608397484, 'timing/model': 0.42436380218714476, 'learning_rate/qwen_vl_interface': 9.554422811997804e-06, 'learning_rate/action_model': 9.554422811997805e-05, 'epoch': 0.39})
+[2026-07-28 11:01:57,044][starVLA.training.train_starvla][INFO] - Step 2143, Loss: {'train/loss': 0.07373501360416412, 'train/grad_norm_pre_clip': 1.6862722635269165, 'timing/data': 0.00037887273356318474, 'timing/model': 0.4265169259160757, 'learning_rate/qwen_vl_interface': 9.55280342081232e-06, 'learning_rate/action_model': 9.55280342081232e-05, 'epoch': 0.39})
+[2026-07-28 11:02:02,893][starVLA.training.train_starvla][INFO] - Step 2144, Loss: {'train/loss': 0.026108037680387497, 'train/grad_norm_pre_clip': 0.6071197986602783, 'timing/data': 0.0005336790345609188, 'timing/model': 0.4307757611386478, 'learning_rate/qwen_vl_interface': 9.55118123758441e-06, 'learning_rate/action_model': 9.551181237584409e-05, 'epoch': 0.39})
+[2026-07-28 11:02:09,233][starVLA.training.train_starvla][INFO] - Step 2145, Loss: {'train/loss': 0.06641186028718948, 'train/grad_norm_pre_clip': 1.9578701257705688, 'timing/data': 0.00026718294247984886, 'timing/model': 0.4233724339865148, 'learning_rate/qwen_vl_interface': 9.549556263366691e-06, 'learning_rate/action_model': 9.549556263366692e-05, 'epoch': 0.39})
+[2026-07-28 11:02:15,513][starVLA.training.train_starvla][INFO] - Step 2146, Loss: {'train/loss': 0.06621356308460236, 'train/grad_norm_pre_clip': 1.1927675008773804, 'timing/data': 0.0003049592487514019, 'timing/model': 0.6703782072290778, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.39})
+[2026-07-28 11:02:21,618][starVLA.training.train_starvla][INFO] - Step 2147, Loss: {'train/loss': 0.054019659757614136, 'train/grad_norm_pre_clip': 2.1961302757263184, 'timing/data': 0.0002687140367925167, 'timing/model': 0.42339878901839256, 'learning_rate/qwen_vl_interface': 9.546297946181354e-06, 'learning_rate/action_model': 9.546297946181354e-05, 'epoch': 0.39})
+[2026-07-28 11:02:27,673][starVLA.training.train_starvla][INFO] - Step 2148, Loss: {'train/loss': 0.051230285316705704, 'train/grad_norm_pre_clip': 1.6026873588562012, 'timing/data': 0.00036742212250828743, 'timing/model': 0.4243559241294861, 'learning_rate/qwen_vl_interface': 9.544664605328022e-06, 'learning_rate/action_model': 9.544664605328022e-05, 'epoch': 0.39})
+[2026-07-28 11:02:34,073][starVLA.training.train_starvla][INFO] - Step 2149, Loss: {'train/loss': 0.05015682429075241, 'train/grad_norm_pre_clip': 1.2743942737579346, 'timing/data': 0.0003374321386218071, 'timing/model': 0.6620814041234553, 'learning_rate/qwen_vl_interface': 9.543028477713455e-06, 'learning_rate/action_model': 9.543028477713453e-05, 'epoch': 0.39})
+[2026-07-28 11:02:40,138][starVLA.training.train_starvla][INFO] - Step 2150, Loss: {'train/loss': 0.052599869668483734, 'train/grad_norm_pre_clip': 2.683506488800049, 'timing/data': 0.00033510103821754456, 'timing/model': 0.4240989563986659, 'learning_rate/qwen_vl_interface': 9.541389564399316e-06, 'learning_rate/action_model': 9.541389564399315e-05, 'epoch': 0.39})
+[2026-07-28 11:02:46,215][starVLA.training.train_starvla][INFO] - Step 2151, Loss: {'train/loss': 0.03061084821820259, 'train/grad_norm_pre_clip': 2.076610803604126, 'timing/data': 0.0003227856941521168, 'timing/model': 0.4389857742935419, 'learning_rate/qwen_vl_interface': 9.53974786644908e-06, 'learning_rate/action_model': 9.53974786644908e-05, 'epoch': 0.4})
+[2026-07-28 11:02:52,584][starVLA.training.train_starvla][INFO] - Step 2152, Loss: {'train/loss': 0.035018280148506165, 'train/grad_norm_pre_clip': 0.7795265913009644, 'timing/data': 0.0004019318148493767, 'timing/model': 0.721430481877178, 'learning_rate/qwen_vl_interface': 9.538103384928029e-06, 'learning_rate/action_model': 9.538103384928027e-05, 'epoch': 0.4})
+[2026-07-28 11:02:58,663][starVLA.training.train_starvla][INFO] - Step 2153, Loss: {'train/loss': 0.06952960789203644, 'train/grad_norm_pre_clip': 2.0649402141571045, 'timing/data': 0.0002804230898618698, 'timing/model': 0.4231158113107085, 'learning_rate/qwen_vl_interface': 9.536456120903244e-06, 'learning_rate/action_model': 9.536456120903244e-05, 'epoch': 0.4})
+[2026-07-28 11:03:04,695][starVLA.training.train_starvla][INFO] - Step 2154, Loss: {'train/loss': 0.06621155887842178, 'train/grad_norm_pre_clip': 3.001119375228882, 'timing/data': 0.00030821701511740685, 'timing/model': 0.42492166301235557, 'learning_rate/qwen_vl_interface': 9.534806075443619e-06, 'learning_rate/action_model': 9.534806075443619e-05, 'epoch': 0.4})
+[2026-07-28 11:03:10,955][starVLA.training.train_starvla][INFO] - Step 2155, Loss: {'train/loss': 0.0529172345995903, 'train/grad_norm_pre_clip': 2.176849842071533, 'timing/data': 0.00034368690103292465, 'timing/model': 0.4206937663257122, 'learning_rate/qwen_vl_interface': 9.533153249619853e-06, 'learning_rate/action_model': 9.533153249619853e-05, 'epoch': 0.4})
+[2026-07-28 11:03:17,043][starVLA.training.train_starvla][INFO] - Step 2156, Loss: {'train/loss': 0.03364402800798416, 'train/grad_norm_pre_clip': 1.336543321609497, 'timing/data': 0.000308087095618248, 'timing/model': 0.4241550089791417, 'learning_rate/qwen_vl_interface': 9.531497644504444e-06, 'learning_rate/action_model': 9.531497644504445e-05, 'epoch': 0.4})
+[2026-07-28 11:03:23,363][starVLA.training.train_starvla][INFO] - Step 2157, Loss: {'train/loss': 0.027896521613001823, 'train/grad_norm_pre_clip': 1.196349024772644, 'timing/data': 0.0003209407441318035, 'timing/model': 0.4225611798465252, 'learning_rate/qwen_vl_interface': 9.529839261171698e-06, 'learning_rate/action_model': 9.529839261171696e-05, 'epoch': 0.4})
+[2026-07-28 11:03:29,449][starVLA.training.train_starvla][INFO] - Step 2158, Loss: {'train/loss': 0.0674569234251976, 'train/grad_norm_pre_clip': 2.625171661376953, 'timing/data': 0.0002992958761751652, 'timing/model': 0.4871452571824193, 'learning_rate/qwen_vl_interface': 9.52817810069772e-06, 'learning_rate/action_model': 9.52817810069772e-05, 'epoch': 0.4})
+[2026-07-28 11:03:35,783][starVLA.training.train_starvla][INFO] - Step 2159, Loss: {'train/loss': 0.09091310203075409, 'train/grad_norm_pre_clip': 1.7018619775772095, 'timing/data': 0.00029434915632009506, 'timing/model': 0.42511380603536963, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.4})
+[2026-07-28 11:03:41,826][starVLA.training.train_starvla][INFO] - Step 2160, Loss: {'train/loss': 0.04567672312259674, 'train/grad_norm_pre_clip': 1.6782180070877075, 'timing/data': 0.0003089657984673977, 'timing/model': 0.4237782680429518, 'learning_rate/qwen_vl_interface': 9.524847452639509e-06, 'learning_rate/action_model': 9.524847452639509e-05, 'epoch': 0.4})
+[2026-07-28 11:03:48,186][starVLA.training.train_starvla][INFO] - Step 2161, Loss: {'train/loss': 0.04482917860150337, 'train/grad_norm_pre_clip': 1.2804293632507324, 'timing/data': 0.00042270682752132416, 'timing/model': 0.4222487499937415, 'learning_rate/qwen_vl_interface': 9.523177967216496e-06, 'learning_rate/action_model': 9.523177967216496e-05, 'epoch': 0.4})
+[2026-07-28 11:03:54,440][starVLA.training.train_starvla][INFO] - Step 2162, Loss: {'train/loss': 0.08699008822441101, 'train/grad_norm_pre_clip': 1.4434295892715454, 'timing/data': 0.00029616616666316986, 'timing/model': 0.42205192195251584, 'learning_rate/qwen_vl_interface': 9.521505708974694e-06, 'learning_rate/action_model': 9.521505708974694e-05, 'epoch': 0.4})
+[2026-07-28 11:04:00,494][starVLA.training.train_starvla][INFO] - Step 2163, Loss: {'train/loss': 0.048453833907842636, 'train/grad_norm_pre_clip': 1.3557955026626587, 'timing/data': 0.0003100922331213951, 'timing/model': 0.42486163508147, 'learning_rate/qwen_vl_interface': 9.519830678999214e-06, 'learning_rate/action_model': 9.519830678999213e-05, 'epoch': 0.4})
+[2026-07-28 11:04:06,579][starVLA.training.train_starvla][INFO] - Step 2164, Loss: {'train/loss': 0.048894964158535004, 'train/grad_norm_pre_clip': 1.0965733528137207, 'timing/data': 0.0003017042763531208, 'timing/model': 0.4261921741999686, 'learning_rate/qwen_vl_interface': 9.518152878376961e-06, 'learning_rate/action_model': 9.51815287837696e-05, 'epoch': 0.4})
+[2026-07-28 11:04:12,845][starVLA.training.train_starvla][INFO] - Step 2165, Loss: {'train/loss': 0.03584641218185425, 'train/grad_norm_pre_clip': 1.3393354415893555, 'timing/data': 0.00037549715489149094, 'timing/model': 0.6285295318812132, 'learning_rate/qwen_vl_interface': 9.516472308196646e-06, 'learning_rate/action_model': 9.516472308196645e-05, 'epoch': 0.4})
+[2026-07-28 11:04:19,172][starVLA.training.train_starvla][INFO] - Step 2166, Loss: {'train/loss': 0.04205142334103584, 'train/grad_norm_pre_clip': 1.532244324684143, 'timing/data': 0.0003794371150434017, 'timing/model': 0.7263669418171048, 'learning_rate/qwen_vl_interface': 9.51478896954877e-06, 'learning_rate/action_model': 9.51478896954877e-05, 'epoch': 0.4})
+[2026-07-28 11:04:25,266][starVLA.training.train_starvla][INFO] - Step 2167, Loss: {'train/loss': 0.0351417250931263, 'train/grad_norm_pre_clip': 1.1333603858947754, 'timing/data': 0.00035812193527817726, 'timing/model': 0.42374155716970563, 'learning_rate/qwen_vl_interface': 9.51310286352563e-06, 'learning_rate/action_model': 9.51310286352563e-05, 'epoch': 0.4})
+[2026-07-28 11:04:31,351][starVLA.training.train_starvla][INFO] - Step 2168, Loss: {'train/loss': 0.06268846243619919, 'train/grad_norm_pre_clip': 1.2364240884780884, 'timing/data': 0.00031381379812955856, 'timing/model': 0.426084169652313, 'learning_rate/qwen_vl_interface': 9.51141399122133e-06, 'learning_rate/action_model': 9.511413991221331e-05, 'epoch': 0.4})
+[2026-07-28 11:04:37,641][starVLA.training.train_starvla][INFO] - Step 2169, Loss: {'train/loss': 0.0496315062046051, 'train/grad_norm_pre_clip': 0.8378323316574097, 'timing/data': 0.0002865418791770935, 'timing/model': 0.42609026189893484, 'learning_rate/qwen_vl_interface': 9.509722353731757e-06, 'learning_rate/action_model': 9.509722353731756e-05, 'epoch': 0.4})
+[2026-07-28 11:04:43,726][starVLA.training.train_starvla][INFO] - Step 2170, Loss: {'train/loss': 0.030879981815814972, 'train/grad_norm_pre_clip': 2.917450189590454, 'timing/data': 0.0003983280621469021, 'timing/model': 0.42732846178114414, 'learning_rate/qwen_vl_interface': 9.508027952154595e-06, 'learning_rate/action_model': 9.508027952154593e-05, 'epoch': 0.4})
+[2026-07-28 11:04:49,782][starVLA.training.train_starvla][INFO] - Step 2171, Loss: {'train/loss': 0.03805674612522125, 'train/grad_norm_pre_clip': 1.629502534866333, 'timing/data': 0.00029667187482118607, 'timing/model': 0.4259382248856127, 'learning_rate/qwen_vl_interface': 9.506330787589321e-06, 'learning_rate/action_model': 9.506330787589322e-05, 'epoch': 0.4})
+[2026-07-28 11:04:55,818][starVLA.training.train_starvla][INFO] - Step 2172, Loss: {'train/loss': 0.06778331100940704, 'train/grad_norm_pre_clip': 1.1935919523239136, 'timing/data': 0.0003141718916594982, 'timing/model': 0.42444201093167067, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.4})
+[2026-07-28 11:05:02,304][starVLA.training.train_starvla][INFO] - Step 2173, Loss: {'train/loss': 0.03542023152112961, 'train/grad_norm_pre_clip': 0.8236339688301086, 'timing/data': 0.00029151421040296555, 'timing/model': 0.655839669983834, 'learning_rate/qwen_vl_interface': 9.50292817390133e-06, 'learning_rate/action_model': 9.502928173901329e-05, 'epoch': 0.4})
+[2026-07-28 11:05:08,358][starVLA.training.train_starvla][INFO] - Step 2174, Loss: {'train/loss': 0.03149338811635971, 'train/grad_norm_pre_clip': 0.9044791460037231, 'timing/data': 0.00029627373442053795, 'timing/model': 0.4232516949996352, 'learning_rate/qwen_vl_interface': 9.501222726986525e-06, 'learning_rate/action_model': 9.501222726986524e-05, 'epoch': 0.4})
+[2026-07-28 11:05:14,696][starVLA.training.train_starvla][INFO] - Step 2175, Loss: {'train/loss': 0.022243082523345947, 'train/grad_norm_pre_clip': 0.7746553421020508, 'timing/data': 0.0005134539678692818, 'timing/model': 0.4870473858900368, 'learning_rate/qwen_vl_interface': 9.49951452149945e-06, 'learning_rate/action_model': 9.499514521499451e-05, 'epoch': 0.4})
+[2026-07-28 11:05:20,950][starVLA.training.train_starvla][INFO] - Step 2176, Loss: {'train/loss': 0.053420133888721466, 'train/grad_norm_pre_clip': 0.7318936586380005, 'timing/data': 0.0003229929134249687, 'timing/model': 0.4760327162221074, 'learning_rate/qwen_vl_interface': 9.497803558548539e-06, 'learning_rate/action_model': 9.497803558548539e-05, 'epoch': 0.4})
+[2026-07-28 11:05:27,287][starVLA.training.train_starvla][INFO] - Step 2177, Loss: {'train/loss': 0.05852484703063965, 'train/grad_norm_pre_clip': 1.9253849983215332, 'timing/data': 0.00032401876524090767, 'timing/model': 0.7214157222770154, 'learning_rate/qwen_vl_interface': 9.496089839244015e-06, 'learning_rate/action_model': 9.496089839244015e-05, 'epoch': 0.4})
+[2026-07-28 11:05:33,302][starVLA.training.train_starvla][INFO] - Step 2178, Loss: {'train/loss': 0.0883767157793045, 'train/grad_norm_pre_clip': 1.5640984773635864, 'timing/data': 0.0003553689457476139, 'timing/model': 0.4207152407616377, 'learning_rate/qwen_vl_interface': 9.494373364697895e-06, 'learning_rate/action_model': 9.494373364697894e-05, 'epoch': 0.4})
+[2026-07-28 11:05:39,609][starVLA.training.train_starvla][INFO] - Step 2179, Loss: {'train/loss': 0.03490424156188965, 'train/grad_norm_pre_clip': 2.455697774887085, 'timing/data': 0.0008600312285125256, 'timing/model': 0.4219625270925462, 'learning_rate/qwen_vl_interface': 9.492654136023978e-06, 'learning_rate/action_model': 9.492654136023979e-05, 'epoch': 0.4})
+[2026-07-28 11:05:45,616][starVLA.training.train_starvla][INFO] - Step 2180, Loss: {'train/loss': 0.03661373257637024, 'train/grad_norm_pre_clip': 1.3644864559173584, 'timing/data': 0.00030969688668847084, 'timing/model': 0.42361740535125136, 'learning_rate/qwen_vl_interface': 9.490932154337855e-06, 'learning_rate/action_model': 9.490932154337855e-05, 'epoch': 0.4})
+[2026-07-28 11:05:51,969][starVLA.training.train_starvla][INFO] - Step 2181, Loss: {'train/loss': 0.03442098945379257, 'train/grad_norm_pre_clip': 1.6084247827529907, 'timing/data': 0.00038651004433631897, 'timing/model': 0.4303201702423394, 'learning_rate/qwen_vl_interface': 9.4892074207569e-06, 'learning_rate/action_model': 9.489207420756899e-05, 'epoch': 0.4})
+[2026-07-28 11:05:58,171][starVLA.training.train_starvla][INFO] - Step 2182, Loss: {'train/loss': 0.06008297950029373, 'train/grad_norm_pre_clip': 1.9774514436721802, 'timing/data': 0.0003448519855737686, 'timing/model': 0.4227332677692175, 'learning_rate/qwen_vl_interface': 9.487479936400274e-06, 'learning_rate/action_model': 9.487479936400272e-05, 'epoch': 0.4})
+[2026-07-28 11:06:04,413][starVLA.training.train_starvla][INFO] - Step 2183, Loss: {'train/loss': 0.034575752913951874, 'train/grad_norm_pre_clip': 1.164471983909607, 'timing/data': 0.00027141720056533813, 'timing/model': 0.42055639112368226, 'learning_rate/qwen_vl_interface': 9.485749702388923e-06, 'learning_rate/action_model': 9.485749702388923e-05, 'epoch': 0.4})
+[2026-07-28 11:06:10,740][starVLA.training.train_starvla][INFO] - Step 2184, Loss: {'train/loss': 0.042457517236471176, 'train/grad_norm_pre_clip': 2.5839226245880127, 'timing/data': 0.00034517189487814903, 'timing/model': 0.42253714380785823, 'learning_rate/qwen_vl_interface': 9.484016719845576e-06, 'learning_rate/action_model': 9.484016719845576e-05, 'epoch': 0.4})
+[2026-07-28 11:06:17,282][starVLA.training.train_starvla][INFO] - Step 2185, Loss: {'train/loss': 0.06710788607597351, 'train/grad_norm_pre_clip': 1.787200927734375, 'timing/data': 0.0003548143431544304, 'timing/model': 0.4198776837438345, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.4})
+[2026-07-28 11:06:23,847][starVLA.training.train_starvla][INFO] - Step 2186, Loss: {'train/loss': 0.03546711802482605, 'train/grad_norm_pre_clip': 0.7672891020774841, 'timing/data': 0.000337876845151186, 'timing/model': 0.6564127900637686, 'learning_rate/qwen_vl_interface': 9.480542513662734e-06, 'learning_rate/action_model': 9.480542513662734e-05, 'epoch': 0.4})
+[2026-07-28 11:06:29,888][starVLA.training.train_starvla][INFO] - Step 2187, Loss: {'train/loss': 0.01989147812128067, 'train/grad_norm_pre_clip': 1.2266970872879028, 'timing/data': 0.00033430103212594986, 'timing/model': 0.4241455979645252, 'learning_rate/qwen_vl_interface': 9.478801292277617e-06, 'learning_rate/action_model': 9.478801292277616e-05, 'epoch': 0.4})
+[2026-07-28 11:06:36,227][starVLA.training.train_starvla][INFO] - Step 2188, Loss: {'train/loss': 0.04975932836532593, 'train/grad_norm_pre_clip': 0.9604687094688416, 'timing/data': 0.000300487969070673, 'timing/model': 0.6678967350162566, 'learning_rate/qwen_vl_interface': 9.477057326869249e-06, 'learning_rate/action_model': 9.477057326869249e-05, 'epoch': 0.4})
+[2026-07-28 11:06:42,233][starVLA.training.train_starvla][INFO] - Step 2189, Loss: {'train/loss': 0.05666409432888031, 'train/grad_norm_pre_clip': 1.7567647695541382, 'timing/data': 0.0002937549725174904, 'timing/model': 0.42384559009224176, 'learning_rate/qwen_vl_interface': 9.475310618569274e-06, 'learning_rate/action_model': 9.475310618569274e-05, 'epoch': 0.4})
+[2026-07-28 11:06:48,519][starVLA.training.train_starvla][INFO] - Step 2190, Loss: {'train/loss': 0.06458331644535065, 'train/grad_norm_pre_clip': 1.6952760219573975, 'timing/data': 0.00027375202625989914, 'timing/model': 0.4204259901307523, 'learning_rate/qwen_vl_interface': 9.473561168511115e-06, 'learning_rate/action_model': 9.473561168511115e-05, 'epoch': 0.4})
+[2026-07-28 11:06:54,827][starVLA.training.train_starvla][INFO] - Step 2191, Loss: {'train/loss': 0.0699322372674942, 'train/grad_norm_pre_clip': 1.7667396068572998, 'timing/data': 0.00032256776466965675, 'timing/model': 0.4216805943287909, 'learning_rate/qwen_vl_interface': 9.471808977829967e-06, 'learning_rate/action_model': 9.471808977829967e-05, 'epoch': 0.4})
+[2026-07-28 11:07:00,963][starVLA.training.train_starvla][INFO] - Step 2192, Loss: {'train/loss': 0.05576544255018234, 'train/grad_norm_pre_clip': 1.4436490535736084, 'timing/data': 0.0003106687217950821, 'timing/model': 0.4222051058895886, 'learning_rate/qwen_vl_interface': 9.470054047662808e-06, 'learning_rate/action_model': 9.470054047662808e-05, 'epoch': 0.4})
+[2026-07-28 11:07:07,275][starVLA.training.train_starvla][INFO] - Step 2193, Loss: {'train/loss': 0.08230791985988617, 'train/grad_norm_pre_clip': 1.767598032951355, 'timing/data': 0.0004202732816338539, 'timing/model': 0.42120001604780555, 'learning_rate/qwen_vl_interface': 9.468296379148394e-06, 'learning_rate/action_model': 9.468296379148393e-05, 'epoch': 0.4})
+[2026-07-28 11:07:13,545][starVLA.training.train_starvla][INFO] - Step 2194, Loss: {'train/loss': 0.04867113381624222, 'train/grad_norm_pre_clip': 1.4555623531341553, 'timing/data': 0.0003024186007678509, 'timing/model': 0.42439304292201996, 'learning_rate/qwen_vl_interface': 9.466535973427259e-06, 'learning_rate/action_model': 9.466535973427258e-05, 'epoch': 0.4})
+[2026-07-28 11:07:19,589][starVLA.training.train_starvla][INFO] - Step 2195, Loss: {'train/loss': 0.06936574727296829, 'train/grad_norm_pre_clip': 2.402397632598877, 'timing/data': 0.00038274889811873436, 'timing/model': 0.4252309179864824, 'learning_rate/qwen_vl_interface': 9.464772831641706e-06, 'learning_rate/action_model': 9.464772831641707e-05, 'epoch': 0.4})
+[2026-07-28 11:07:25,784][starVLA.training.train_starvla][INFO] - Step 2196, Loss: {'train/loss': 0.05530058965086937, 'train/grad_norm_pre_clip': 0.7163838744163513, 'timing/data': 0.00031992606818675995, 'timing/model': 0.5960136521607637, 'learning_rate/qwen_vl_interface': 9.463006954935823e-06, 'learning_rate/action_model': 9.463006954935822e-05, 'epoch': 0.4})
+[2026-07-28 11:07:31,851][starVLA.training.train_starvla][INFO] - Step 2197, Loss: {'train/loss': 0.0650404840707779, 'train/grad_norm_pre_clip': 2.4652135372161865, 'timing/data': 0.00038123270496726036, 'timing/model': 0.4236215427517891, 'learning_rate/qwen_vl_interface': 9.461238344455468e-06, 'learning_rate/action_model': 9.461238344455468e-05, 'epoch': 0.4})
+[2026-07-28 11:07:37,908][starVLA.training.train_starvla][INFO] - Step 2198, Loss: {'train/loss': 0.07408465445041656, 'train/grad_norm_pre_clip': 3.1789934635162354, 'timing/data': 0.00033475179225206375, 'timing/model': 0.42313650716096163, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.4})
+[2026-07-28 11:07:43,972][starVLA.training.train_starvla][INFO] - Step 2199, Loss: {'train/loss': 0.06841188669204712, 'train/grad_norm_pre_clip': 1.5292664766311646, 'timing/data': 0.0003042309544980526, 'timing/model': 0.42341786390170455, 'learning_rate/qwen_vl_interface': 9.457692926763638e-06, 'learning_rate/action_model': 9.457692926763638e-05, 'epoch': 0.4})
+[2026-07-28 11:07:50,326][starVLA.training.train_starvla][INFO] - Step 2200, Loss: {'train/loss': 0.0345076248049736, 'train/grad_norm_pre_clip': 1.013038992881775, 'timing/data': 0.0002820161171257496, 'timing/model': 0.4200339401140809, 'learning_rate/qwen_vl_interface': 9.455916121852748e-06, 'learning_rate/action_model': 9.455916121852747e-05, 'epoch': 0.4})
+[2026-07-28 11:07:56,391][starVLA.training.train_starvla][INFO] - Step 2201, Loss: {'train/loss': 0.06563928723335266, 'train/grad_norm_pre_clip': 1.6707721948623657, 'timing/data': 0.0004085199907422066, 'timing/model': 0.4223201097920537, 'learning_rate/qwen_vl_interface': 9.454136587768549e-06, 'learning_rate/action_model': 9.454136587768548e-05, 'epoch': 0.4})
+[2026-07-28 11:08:02,436][starVLA.training.train_starvla][INFO] - Step 2202, Loss: {'train/loss': 0.05532463639974594, 'train/grad_norm_pre_clip': 1.068651556968689, 'timing/data': 0.00035146716982126236, 'timing/model': 0.42058247001841664, 'learning_rate/qwen_vl_interface': 9.452354325665763e-06, 'learning_rate/action_model': 9.452354325665763e-05, 'epoch': 0.4})
+[2026-07-28 11:08:08,717][starVLA.training.train_starvla][INFO] - Step 2203, Loss: {'train/loss': 0.02233458310365677, 'train/grad_norm_pre_clip': 0.6927849054336548, 'timing/data': 0.00031181611120700836, 'timing/model': 0.6654942841269076, 'learning_rate/qwen_vl_interface': 9.450569336700878e-06, 'learning_rate/action_model': 9.450569336700878e-05, 'epoch': 0.4})
+[2026-07-28 11:08:14,788][starVLA.training.train_starvla][INFO] - Step 2204, Loss: {'train/loss': 0.08239362388849258, 'train/grad_norm_pre_clip': 1.2431598901748657, 'timing/data': 0.00030151521787047386, 'timing/model': 0.4390843762084842, 'learning_rate/qwen_vl_interface': 9.448781622032155e-06, 'learning_rate/action_model': 9.448781622032155e-05, 'epoch': 0.4})
+[2026-07-28 11:08:21,052][starVLA.training.train_starvla][INFO] - Step 2205, Loss: {'train/loss': 0.04283168166875839, 'train/grad_norm_pre_clip': 0.8570376634597778, 'timing/data': 0.00032268185168504715, 'timing/model': 0.42136831022799015, 'learning_rate/qwen_vl_interface': 9.446991182819624e-06, 'learning_rate/action_model': 9.446991182819623e-05, 'epoch': 0.41})
+[2026-07-28 11:08:27,112][starVLA.training.train_starvla][INFO] - Step 2206, Loss: {'train/loss': 0.04706985130906105, 'train/grad_norm_pre_clip': 1.4759521484375, 'timing/data': 0.000285422895103693, 'timing/model': 0.4214417291805148, 'learning_rate/qwen_vl_interface': 9.44519802022508e-06, 'learning_rate/action_model': 9.44519802022508e-05, 'epoch': 0.41})
+[2026-07-28 11:08:33,206][starVLA.training.train_starvla][INFO] - Step 2207, Loss: {'train/loss': 0.025234777480363846, 'train/grad_norm_pre_clip': 1.8846393823623657, 'timing/data': 0.0003078100271522999, 'timing/model': 0.45666180085390806, 'learning_rate/qwen_vl_interface': 9.443402135412088e-06, 'learning_rate/action_model': 9.443402135412088e-05, 'epoch': 0.41})
+[2026-07-28 11:08:39,507][starVLA.training.train_starvla][INFO] - Step 2208, Loss: {'train/loss': 0.03831872344017029, 'train/grad_norm_pre_clip': 1.3453032970428467, 'timing/data': 0.00036302441731095314, 'timing/model': 0.4213568330742419, 'learning_rate/qwen_vl_interface': 9.441603529545977e-06, 'learning_rate/action_model': 9.441603529545976e-05, 'epoch': 0.41})
+[2026-07-28 11:08:45,796][starVLA.training.train_starvla][INFO] - Step 2209, Loss: {'train/loss': 0.06148017942905426, 'train/grad_norm_pre_clip': 2.496961832046509, 'timing/data': 0.0002661007456481457, 'timing/model': 0.4204925517551601, 'learning_rate/qwen_vl_interface': 9.439802203793843e-06, 'learning_rate/action_model': 9.439802203793842e-05, 'epoch': 0.41})
+[2026-07-28 11:08:51,620][starVLA.training.train_starvla][INFO] - Step 2210, Loss: {'train/loss': 0.04709586501121521, 'train/grad_norm_pre_clip': 0.9563068151473999, 'timing/data': 0.0002865409478545189, 'timing/model': 0.4221332189626992, 'learning_rate/qwen_vl_interface': 9.437998159324547e-06, 'learning_rate/action_model': 9.437998159324547e-05, 'epoch': 0.41})
+[2026-07-28 11:08:57,442][starVLA.training.train_starvla][INFO] - Step 2211, Loss: {'train/loss': 0.049639828503131866, 'train/grad_norm_pre_clip': 1.5259003639221191, 'timing/data': 0.00035546906292438507, 'timing/model': 0.4231355511583388, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.41})
+[2026-07-28 11:09:03,280][starVLA.training.train_starvla][INFO] - Step 2212, Loss: {'train/loss': 0.03669131174683571, 'train/grad_norm_pre_clip': 1.4377377033233643, 'timing/data': 0.0003554662689566612, 'timing/model': 0.42295344499871135, 'learning_rate/qwen_vl_interface': 9.434381918918731e-06, 'learning_rate/action_model': 9.434381918918731e-05, 'epoch': 0.41})
+[2026-07-28 11:09:09,336][starVLA.training.train_starvla][INFO] - Step 2213, Loss: {'train/loss': 0.1100815013051033, 'train/grad_norm_pre_clip': 1.6029127836227417, 'timing/data': 0.0003484850749373436, 'timing/model': 0.65848758071661, 'learning_rate/qwen_vl_interface': 9.432569725328753e-06, 'learning_rate/action_model': 9.432569725328753e-05, 'epoch': 0.41})
+[2026-07-28 11:09:15,631][starVLA.training.train_starvla][INFO] - Step 2214, Loss: {'train/loss': 0.02827603742480278, 'train/grad_norm_pre_clip': 0.8327667713165283, 'timing/data': 0.0003583421930670738, 'timing/model': 0.41943383729085326, 'learning_rate/qwen_vl_interface': 9.430754817714686e-06, 'learning_rate/action_model': 9.430754817714687e-05, 'epoch': 0.41})
+[2026-07-28 11:09:21,766][starVLA.training.train_starvla][INFO] - Step 2215, Loss: {'train/loss': 0.10543090105056763, 'train/grad_norm_pre_clip': 2.831373929977417, 'timing/data': 0.000404581893235445, 'timing/model': 0.4548580008558929, 'learning_rate/qwen_vl_interface': 9.42893719725421e-06, 'learning_rate/action_model': 9.42893719725421e-05, 'epoch': 0.41})
+[2026-07-28 11:09:27,860][starVLA.training.train_starvla][INFO] - Step 2216, Loss: {'train/loss': 0.10216863453388214, 'train/grad_norm_pre_clip': 2.6187257766723633, 'timing/data': 0.0003951569087803364, 'timing/model': 0.4237100896425545, 'learning_rate/qwen_vl_interface': 9.427116865126757e-06, 'learning_rate/action_model': 9.427116865126756e-05, 'epoch': 0.41})
+[2026-07-28 11:09:33,926][starVLA.training.train_starvla][INFO] - Step 2217, Loss: {'train/loss': 0.03001604788005352, 'train/grad_norm_pre_clip': 1.3261884450912476, 'timing/data': 0.0003997143357992172, 'timing/model': 0.4239248209632933, 'learning_rate/qwen_vl_interface': 9.42529382251352e-06, 'learning_rate/action_model': 9.42529382251352e-05, 'epoch': 0.41})
+[2026-07-28 11:09:40,238][starVLA.training.train_starvla][INFO] - Step 2218, Loss: {'train/loss': 0.06483893096446991, 'train/grad_norm_pre_clip': 2.41196608543396, 'timing/data': 0.0002982751466333866, 'timing/model': 0.4221120299771428, 'learning_rate/qwen_vl_interface': 9.423468070597453e-06, 'learning_rate/action_model': 9.423468070597452e-05, 'epoch': 0.41})
+[2026-07-28 11:09:46,496][starVLA.training.train_starvla][INFO] - Step 2219, Loss: {'train/loss': 0.029456742107868195, 'train/grad_norm_pre_clip': 0.6504999995231628, 'timing/data': 0.0003950269892811775, 'timing/model': 0.4227399849332869, 'learning_rate/qwen_vl_interface': 9.421639610563264e-06, 'learning_rate/action_model': 9.421639610563264e-05, 'epoch': 0.41})
+[2026-07-28 11:09:52,846][starVLA.training.train_starvla][INFO] - Step 2220, Loss: {'train/loss': 0.09947444498538971, 'train/grad_norm_pre_clip': 2.870830774307251, 'timing/data': 0.00035079242661595345, 'timing/model': 0.7276840968988836, 'learning_rate/qwen_vl_interface': 9.419808443597422e-06, 'learning_rate/action_model': 9.419808443597423e-05, 'epoch': 0.41})
+[2026-07-28 11:09:59,275][starVLA.training.train_starvla][INFO] - Step 2221, Loss: {'train/loss': 0.07712310552597046, 'train/grad_norm_pre_clip': 1.8589985370635986, 'timing/data': 0.00037031201645731926, 'timing/model': 0.4242396969348192, 'learning_rate/qwen_vl_interface': 9.417974570888152e-06, 'learning_rate/action_model': 9.417974570888153e-05, 'epoch': 0.41})
+[2026-07-28 11:10:05,388][starVLA.training.train_starvla][INFO] - Step 2222, Loss: {'train/loss': 0.055777110159397125, 'train/grad_norm_pre_clip': 0.9154215455055237, 'timing/data': 0.00039485516026616096, 'timing/model': 0.4356814161874354, 'learning_rate/qwen_vl_interface': 9.416137993625437e-06, 'learning_rate/action_model': 9.416137993625437e-05, 'epoch': 0.41})
+[2026-07-28 11:10:11,417][starVLA.training.train_starvla][INFO] - Step 2223, Loss: {'train/loss': 0.06296302378177643, 'train/grad_norm_pre_clip': 2.372480869293213, 'timing/data': 0.00034905504435300827, 'timing/model': 0.423847293946892, 'learning_rate/qwen_vl_interface': 9.414298713001007e-06, 'learning_rate/action_model': 9.414298713001006e-05, 'epoch': 0.41})
+[2026-07-28 11:10:17,748][starVLA.training.train_starvla][INFO] - Step 2224, Loss: {'train/loss': 0.04931509494781494, 'train/grad_norm_pre_clip': 2.2156596183776855, 'timing/data': 0.00030242279171943665, 'timing/model': 0.6226798668503761, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.41})
+[2026-07-28 11:10:24,022][starVLA.training.train_starvla][INFO] - Step 2225, Loss: {'train/loss': 0.07163333892822266, 'train/grad_norm_pre_clip': 2.4360194206237793, 'timing/data': 0.0002794540487229824, 'timing/model': 0.41960917273536325, 'learning_rate/qwen_vl_interface': 9.410612046442719e-06, 'learning_rate/action_model': 9.410612046442718e-05, 'epoch': 0.41})
+[2026-07-28 11:10:30,079][starVLA.training.train_starvla][INFO] - Step 2226, Loss: {'train/loss': 0.06752948462963104, 'train/grad_norm_pre_clip': 1.6894242763519287, 'timing/data': 0.00039381859824061394, 'timing/model': 0.42177849588915706, 'learning_rate/qwen_vl_interface': 9.408764662901097e-06, 'learning_rate/action_model': 9.408764662901098e-05, 'epoch': 0.41})
+[2026-07-28 11:10:36,152][starVLA.training.train_starvla][INFO] - Step 2227, Loss: {'train/loss': 0.05698191374540329, 'train/grad_norm_pre_clip': 0.7300441265106201, 'timing/data': 0.00029262807220220566, 'timing/model': 0.4225512631237507, 'learning_rate/qwen_vl_interface': 9.406914580782238e-06, 'learning_rate/action_model': 9.406914580782238e-05, 'epoch': 0.41})
+[2026-07-28 11:10:42,401][starVLA.training.train_starvla][INFO] - Step 2228, Loss: {'train/loss': 0.05781546235084534, 'train/grad_norm_pre_clip': 1.5261921882629395, 'timing/data': 0.0002848873846232891, 'timing/model': 0.4214382655918598, 'learning_rate/qwen_vl_interface': 9.405061801286638e-06, 'learning_rate/action_model': 9.405061801286638e-05, 'epoch': 0.41})
+[2026-07-28 11:10:48,708][starVLA.training.train_starvla][INFO] - Step 2229, Loss: {'train/loss': 0.0701034888625145, 'train/grad_norm_pre_clip': 0.8811649680137634, 'timing/data': 0.00039754994213581085, 'timing/model': 0.42926658410578966, 'learning_rate/qwen_vl_interface': 9.403206325616546e-06, 'learning_rate/action_model': 9.403206325616545e-05, 'epoch': 0.41})
+[2026-07-28 11:10:54,766][starVLA.training.train_starvla][INFO] - Step 2230, Loss: {'train/loss': 0.02604435384273529, 'train/grad_norm_pre_clip': 0.6886489987373352, 'timing/data': 0.00037969695404171944, 'timing/model': 0.4498144178651273, 'learning_rate/qwen_vl_interface': 9.40134815497596e-06, 'learning_rate/action_model': 9.40134815497596e-05, 'epoch': 0.41})
+[2026-07-28 11:11:01,054][starVLA.training.train_starvla][INFO] - Step 2231, Loss: {'train/loss': 0.054250285029411316, 'train/grad_norm_pre_clip': 2.013237237930298, 'timing/data': 0.0003993515856564045, 'timing/model': 0.6785059371031821, 'learning_rate/qwen_vl_interface': 9.399487290570629e-06, 'learning_rate/action_model': 9.399487290570629e-05, 'epoch': 0.41})
+[2026-07-28 11:11:07,377][starVLA.training.train_starvla][INFO] - Step 2232, Loss: {'train/loss': 0.08480998128652573, 'train/grad_norm_pre_clip': 3.6654865741729736, 'timing/data': 0.0003036358393728733, 'timing/model': 0.6782673806883395, 'learning_rate/qwen_vl_interface': 9.397623733608042e-06, 'learning_rate/action_model': 9.397623733608042e-05, 'epoch': 0.41})
+[2026-07-28 11:11:13,414][starVLA.training.train_starvla][INFO] - Step 2233, Loss: {'train/loss': 0.06828334927558899, 'train/grad_norm_pre_clip': 1.3980263471603394, 'timing/data': 0.0008483650162816048, 'timing/model': 0.4288661233149469, 'learning_rate/qwen_vl_interface': 9.395757485297447e-06, 'learning_rate/action_model': 9.395757485297446e-05, 'epoch': 0.41})
+[2026-07-28 11:11:19,444][starVLA.training.train_starvla][INFO] - Step 2234, Loss: {'train/loss': 0.07789488881826401, 'train/grad_norm_pre_clip': 1.9912837743759155, 'timing/data': 0.0003232099115848541, 'timing/model': 0.43016600608825684, 'learning_rate/qwen_vl_interface': 9.393888546849828e-06, 'learning_rate/action_model': 9.393888546849828e-05, 'epoch': 0.41})
+[2026-07-28 11:11:25,770][starVLA.training.train_starvla][INFO] - Step 2235, Loss: {'train/loss': 0.05600300803780556, 'train/grad_norm_pre_clip': 1.1936379671096802, 'timing/data': 0.00030455924570560455, 'timing/model': 0.4300679978914559, 'learning_rate/qwen_vl_interface': 9.392016919477922e-06, 'learning_rate/action_model': 9.392016919477921e-05, 'epoch': 0.41})
+[2026-07-28 11:11:31,844][starVLA.training.train_starvla][INFO] - Step 2236, Loss: {'train/loss': 0.0678803101181984, 'train/grad_norm_pre_clip': 2.2366628646850586, 'timing/data': 0.0003130766563117504, 'timing/model': 0.42815185710787773, 'learning_rate/qwen_vl_interface': 9.390142604396206e-06, 'learning_rate/action_model': 9.390142604396205e-05, 'epoch': 0.41})
+[2026-07-28 11:11:37,882][starVLA.training.train_starvla][INFO] - Step 2237, Loss: {'train/loss': 0.05779797211289406, 'train/grad_norm_pre_clip': 2.624081611633301, 'timing/data': 0.0002832910977303982, 'timing/model': 0.4318829639814794, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.41})
+[2026-07-28 11:11:44,226][starVLA.training.train_starvla][INFO] - Step 2238, Loss: {'train/loss': 0.043136775493621826, 'train/grad_norm_pre_clip': 1.480860948562622, 'timing/data': 0.00031197117641568184, 'timing/model': 0.4247314389795065, 'learning_rate/qwen_vl_interface': 9.386385915969975e-06, 'learning_rate/action_model': 9.386385915969976e-05, 'epoch': 0.41})
+[2026-07-28 11:11:50,360][starVLA.training.train_starvla][INFO] - Step 2239, Loss: {'train/loss': 0.05476918816566467, 'train/grad_norm_pre_clip': 1.7076857089996338, 'timing/data': 0.0003887801431119442, 'timing/model': 0.42377658979967237, 'learning_rate/qwen_vl_interface': 9.384503545063138e-06, 'learning_rate/action_model': 9.384503545063137e-05, 'epoch': 0.41})
+[2026-07-28 11:11:56,385][starVLA.training.train_starvla][INFO] - Step 2240, Loss: {'train/loss': 0.026775160804390907, 'train/grad_norm_pre_clip': 1.4816820621490479, 'timing/data': 0.00031182728707790375, 'timing/model': 0.428459997754544, 'learning_rate/qwen_vl_interface': 9.382618491321837e-06, 'learning_rate/action_model': 9.382618491321837e-05, 'epoch': 0.41})
+[2026-07-28 11:12:02,898][starVLA.training.train_starvla][INFO] - Step 2241, Loss: {'train/loss': 0.035971660166978836, 'train/grad_norm_pre_clip': 1.3010536432266235, 'timing/data': 0.0003092321567237377, 'timing/model': 0.7628424749709666, 'learning_rate/qwen_vl_interface': 9.380730755969265e-06, 'learning_rate/action_model': 9.380730755969264e-05, 'epoch': 0.41})
+[2026-07-28 11:12:09,329][starVLA.training.train_starvla][INFO] - Step 2242, Loss: {'train/loss': 0.0654330849647522, 'train/grad_norm_pre_clip': 3.106506824493408, 'timing/data': 0.00037591997534036636, 'timing/model': 0.43395330756902695, 'learning_rate/qwen_vl_interface': 9.378840340230353e-06, 'learning_rate/action_model': 9.378840340230352e-05, 'epoch': 0.41})
+[2026-07-28 11:12:15,612][starVLA.training.train_starvla][INFO] - Step 2243, Loss: {'train/loss': 0.06651545315980911, 'train/grad_norm_pre_clip': 1.2474875450134277, 'timing/data': 0.0003019077703356743, 'timing/model': 0.44000162510201335, 'learning_rate/qwen_vl_interface': 9.376947245331769e-06, 'learning_rate/action_model': 9.376947245331767e-05, 'epoch': 0.41})
+[2026-07-28 11:12:22,262][starVLA.training.train_starvla][INFO] - Step 2244, Loss: {'train/loss': 0.047257985919713974, 'train/grad_norm_pre_clip': 1.8483744859695435, 'timing/data': 0.00027928268536925316, 'timing/model': 0.42683177907019854, 'learning_rate/qwen_vl_interface': 9.375051472501922e-06, 'learning_rate/action_model': 9.375051472501923e-05, 'epoch': 0.41})
+[2026-07-28 11:12:28,868][starVLA.training.train_starvla][INFO] - Step 2245, Loss: {'train/loss': 0.07724522799253464, 'train/grad_norm_pre_clip': 1.5777816772460938, 'timing/data': 0.0003574918955564499, 'timing/model': 0.4319662409834564, 'learning_rate/qwen_vl_interface': 9.373153022970961e-06, 'learning_rate/action_model': 9.373153022970961e-05, 'epoch': 0.41})
+[2026-07-28 11:12:34,814][starVLA.training.train_starvla][INFO] - Step 2246, Loss: {'train/loss': 0.04965739697217941, 'train/grad_norm_pre_clip': 1.0804710388183594, 'timing/data': 0.000302713830024004, 'timing/model': 0.43442743737250566, 'learning_rate/qwen_vl_interface': 9.371251897970768e-06, 'learning_rate/action_model': 9.371251897970768e-05, 'epoch': 0.41})
+[2026-07-28 11:12:41,166][starVLA.training.train_starvla][INFO] - Step 2247, Loss: {'train/loss': 0.059801045805215836, 'train/grad_norm_pre_clip': 4.270776271820068, 'timing/data': 0.0003630947321653366, 'timing/model': 0.46487003611400723, 'learning_rate/qwen_vl_interface': 9.369348098734961e-06, 'learning_rate/action_model': 9.369348098734961e-05, 'epoch': 0.41})
+[2026-07-28 11:12:47,784][starVLA.training.train_starvla][INFO] - Step 2248, Loss: {'train/loss': 0.02893778681755066, 'train/grad_norm_pre_clip': 1.3503144979476929, 'timing/data': 0.0003367471508681774, 'timing/model': 0.423236692789942, 'learning_rate/qwen_vl_interface': 9.367441626498897e-06, 'learning_rate/action_model': 9.367441626498896e-05, 'epoch': 0.41})
+[2026-07-28 11:12:54,419][starVLA.training.train_starvla][INFO] - Step 2249, Loss: {'train/loss': 0.052635304629802704, 'train/grad_norm_pre_clip': 1.511620044708252, 'timing/data': 0.0002752169966697693, 'timing/model': 0.4380165571346879, 'learning_rate/qwen_vl_interface': 9.365532482499663e-06, 'learning_rate/action_model': 9.365532482499662e-05, 'epoch': 0.41})
+[2026-07-28 11:13:00,744][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2250
+[2026-07-28 11:13:05,836][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2250
+[2026-07-28 11:18:04,853][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2250
+[2026-07-28 11:18:04,857][starVLA.training.train_starvla][INFO] - Step 2250, Loss: {'train/loss': 0.05582848936319351, 'train/grad_norm_pre_clip': 1.1502363681793213, 'eval/action_loss/seconds': 5.088708254974335, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.056802425533533096, 'eval/latency_3/loss': 0.056802425533533096, 'eval/flappy/loss': 0.056802425533533096, 'eval/flappy/latency_3/loss': 0.056802425533533096, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 214.26500253416597, 'rl_games_eval/mid_train/mean_length': 1746.45, 'rl_games_eval/mid_train/std_reward': 146.78073708385077, 'rl_games_eval/mid_train/std_length': 1182.8461639198902, 'rl_games_eval/mid_train/macro_mean_reward': 214.26500253416597, 'rl_games_eval/mid_train/macro_mean_length': 1746.45, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 214.26500253416597, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 1746.45, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 146.78073708385077, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1182.8461639198902, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 214.26500253416597, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0002797259949147701, 'timing/model': 0.428545817732811, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.41})
+[2026-07-28 11:18:11,252][starVLA.training.train_starvla][INFO] - Step 2251, Loss: {'train/loss': 0.0658213198184967, 'train/grad_norm_pre_clip': 1.1607669591903687, 'timing/data': 0.00034320075064897537, 'timing/model': 0.4386043297126889, 'learning_rate/qwen_vl_interface': 9.361706184168714e-06, 'learning_rate/action_model': 9.361706184168713e-05, 'epoch': 0.41})
+[2026-07-28 11:18:17,838][starVLA.training.train_starvla][INFO] - Step 2252, Loss: {'train/loss': 0.08210517466068268, 'train/grad_norm_pre_clip': 2.470724105834961, 'timing/data': 0.0003154119476675987, 'timing/model': 0.7474792059510946, 'learning_rate/qwen_vl_interface': 9.359789032319843e-06, 'learning_rate/action_model': 9.359789032319842e-05, 'epoch': 0.41})
+[2026-07-28 11:18:24,012][starVLA.training.train_starvla][INFO] - Step 2253, Loss: {'train/loss': 0.03054061345756054, 'train/grad_norm_pre_clip': 1.5571925640106201, 'timing/data': 0.000273166224360466, 'timing/model': 0.46915273973718286, 'learning_rate/qwen_vl_interface': 9.357869213673484e-06, 'learning_rate/action_model': 9.357869213673485e-05, 'epoch': 0.41})
+[2026-07-28 11:18:30,159][starVLA.training.train_starvla][INFO] - Step 2254, Loss: {'train/loss': 0.06281687319278717, 'train/grad_norm_pre_clip': 0.8481106162071228, 'timing/data': 0.000357787124812603, 'timing/model': 0.4285818091593683, 'learning_rate/qwen_vl_interface': 9.355946729475393e-06, 'learning_rate/action_model': 9.355946729475393e-05, 'epoch': 0.41})
+[2026-07-28 11:18:36,315][starVLA.training.train_starvla][INFO] - Step 2255, Loss: {'train/loss': 0.03204384073615074, 'train/grad_norm_pre_clip': 0.9853853583335876, 'timing/data': 0.0003071208484470844, 'timing/model': 0.4308107257820666, 'learning_rate/qwen_vl_interface': 9.354021580973049e-06, 'learning_rate/action_model': 9.354021580973048e-05, 'epoch': 0.41})
+[2026-07-28 11:18:42,725][starVLA.training.train_starvla][INFO] - Step 2256, Loss: {'train/loss': 0.029669148847460747, 'train/grad_norm_pre_clip': 0.6839432120323181, 'timing/data': 0.00037592509761452675, 'timing/model': 0.462249874137342, 'learning_rate/qwen_vl_interface': 9.352093769415655e-06, 'learning_rate/action_model': 9.352093769415655e-05, 'epoch': 0.41})
+[2026-07-28 11:18:49,248][starVLA.training.train_starvla][INFO] - Step 2257, Loss: {'train/loss': 0.038398705422878265, 'train/grad_norm_pre_clip': 0.9686126112937927, 'timing/data': 0.00030197296291589737, 'timing/model': 0.444063289090991, 'learning_rate/qwen_vl_interface': 9.35016329605415e-06, 'learning_rate/action_model': 9.350163296054151e-05, 'epoch': 0.41})
+[2026-07-28 11:18:55,539][starVLA.training.train_starvla][INFO] - Step 2258, Loss: {'train/loss': 0.01837291568517685, 'train/grad_norm_pre_clip': 0.6444904804229736, 'timing/data': 0.0003181588836014271, 'timing/model': 0.44128644466400146, 'learning_rate/qwen_vl_interface': 9.348230162141198e-06, 'learning_rate/action_model': 9.348230162141198e-05, 'epoch': 0.41})
+[2026-07-28 11:19:02,111][starVLA.training.train_starvla][INFO] - Step 2259, Loss: {'train/loss': 0.030123133212327957, 'train/grad_norm_pre_clip': 1.3234584331512451, 'timing/data': 0.00047726696357131004, 'timing/model': 0.4292865409515798, 'learning_rate/qwen_vl_interface': 9.346294368931188e-06, 'learning_rate/action_model': 9.346294368931188e-05, 'epoch': 0.41})
+[2026-07-28 11:19:08,623][starVLA.training.train_starvla][INFO] - Step 2260, Loss: {'train/loss': 0.042811762541532516, 'train/grad_norm_pre_clip': 1.2908663749694824, 'timing/data': 0.00035071326419711113, 'timing/model': 0.6691332100890577, 'learning_rate/qwen_vl_interface': 9.344355917680235e-06, 'learning_rate/action_model': 9.344355917680235e-05, 'epoch': 0.42})
+[2026-07-28 11:19:14,858][starVLA.training.train_starvla][INFO] - Step 2261, Loss: {'train/loss': 0.12190961837768555, 'train/grad_norm_pre_clip': 5.23472785949707, 'timing/data': 0.000360049307346344, 'timing/model': 0.43201745906844735, 'learning_rate/qwen_vl_interface': 9.342414809646179e-06, 'learning_rate/action_model': 9.342414809646179e-05, 'epoch': 0.42})
+[2026-07-28 11:19:21,589][starVLA.training.train_starvla][INFO] - Step 2262, Loss: {'train/loss': 0.0901508629322052, 'train/grad_norm_pre_clip': 3.056088924407959, 'timing/data': 0.0003086780197918415, 'timing/model': 0.764785089995712, 'learning_rate/qwen_vl_interface': 9.340471046088581e-06, 'learning_rate/action_model': 9.34047104608858e-05, 'epoch': 0.42})
+[2026-07-28 11:19:27,997][starVLA.training.train_starvla][INFO] - Step 2263, Loss: {'train/loss': 0.06618213653564453, 'train/grad_norm_pre_clip': 2.960984468460083, 'timing/data': 0.00031544407829642296, 'timing/model': 0.43412379594519734, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.42})
+[2026-07-28 11:19:34,555][starVLA.training.train_starvla][INFO] - Step 2264, Loss: {'train/loss': 0.05270634591579437, 'train/grad_norm_pre_clip': 1.569249153137207, 'timing/data': 0.00028135208413004875, 'timing/model': 0.4639046713709831, 'learning_rate/qwen_vl_interface': 9.336575557449641e-06, 'learning_rate/action_model': 9.336575557449641e-05, 'epoch': 0.42})
+[2026-07-28 11:19:40,836][starVLA.training.train_starvla][INFO] - Step 2265, Loss: {'train/loss': 0.03468991816043854, 'train/grad_norm_pre_clip': 1.5753722190856934, 'timing/data': 0.0003038090653717518, 'timing/model': 0.43415733287110925, 'learning_rate/qwen_vl_interface': 9.334623834896039e-06, 'learning_rate/action_model': 9.334623834896038e-05, 'epoch': 0.42})
+[2026-07-28 11:19:47,422][starVLA.training.train_starvla][INFO] - Step 2266, Loss: {'train/loss': 0.0448760911822319, 'train/grad_norm_pre_clip': 1.3423186540603638, 'timing/data': 0.0002982672303915024, 'timing/model': 0.4329205728136003, 'learning_rate/qwen_vl_interface': 9.332669461874376e-06, 'learning_rate/action_model': 9.332669461874374e-05, 'epoch': 0.42})
+[2026-07-28 11:19:53,749][starVLA.training.train_starvla][INFO] - Step 2267, Loss: {'train/loss': 0.09207107126712799, 'train/grad_norm_pre_clip': 2.797478675842285, 'timing/data': 0.0003118850290775299, 'timing/model': 0.45433699106797576, 'learning_rate/qwen_vl_interface': 9.330712439652824e-06, 'learning_rate/action_model': 9.330712439652823e-05, 'epoch': 0.42})
+[2026-07-28 11:20:00,346][starVLA.training.train_starvla][INFO] - Step 2268, Loss: {'train/loss': 0.08195631206035614, 'train/grad_norm_pre_clip': 2.060905694961548, 'timing/data': 0.0002793222665786743, 'timing/model': 0.43197504384443164, 'learning_rate/qwen_vl_interface': 9.328752769501272e-06, 'learning_rate/action_model': 9.328752769501272e-05, 'epoch': 0.42})
+[2026-07-28 11:20:07,219][starVLA.training.train_starvla][INFO] - Step 2269, Loss: {'train/loss': 0.0750521719455719, 'train/grad_norm_pre_clip': 6.451594352722168, 'timing/data': 0.00037309667095541954, 'timing/model': 0.7801828221417964, 'learning_rate/qwen_vl_interface': 9.326790452691332e-06, 'learning_rate/action_model': 9.326790452691332e-05, 'epoch': 0.42})
+[2026-07-28 11:20:13,721][starVLA.training.train_starvla][INFO] - Step 2270, Loss: {'train/loss': 0.07229771465063095, 'train/grad_norm_pre_clip': 2.5067765712738037, 'timing/data': 0.0003103637136518955, 'timing/model': 0.4327912759035826, 'learning_rate/qwen_vl_interface': 9.324825490496327e-06, 'learning_rate/action_model': 9.324825490496327e-05, 'epoch': 0.42})
+[2026-07-28 11:20:20,381][starVLA.training.train_starvla][INFO] - Step 2271, Loss: {'train/loss': 0.03974353149533272, 'train/grad_norm_pre_clip': 0.8059377670288086, 'timing/data': 0.0003056200221180916, 'timing/model': 0.6703255800530314, 'learning_rate/qwen_vl_interface': 9.322857884191303e-06, 'learning_rate/action_model': 9.322857884191302e-05, 'epoch': 0.42})
+[2026-07-28 11:20:26,716][starVLA.training.train_starvla][INFO] - Step 2272, Loss: {'train/loss': 0.06450580060482025, 'train/grad_norm_pre_clip': 4.215994358062744, 'timing/data': 0.0003646891564130783, 'timing/model': 0.4323178599588573, 'learning_rate/qwen_vl_interface': 9.320887635053013e-06, 'learning_rate/action_model': 9.320887635053014e-05, 'epoch': 0.42})
+[2026-07-28 11:20:33,449][starVLA.training.train_starvla][INFO] - Step 2273, Loss: {'train/loss': 0.08681202679872513, 'train/grad_norm_pre_clip': 1.6049884557724, 'timing/data': 0.00029826676473021507, 'timing/model': 0.4297347622923553, 'learning_rate/qwen_vl_interface': 9.318914744359936e-06, 'learning_rate/action_model': 9.318914744359937e-05, 'epoch': 0.42})
+[2026-07-28 11:20:39,634][starVLA.training.train_starvla][INFO] - Step 2274, Loss: {'train/loss': 0.058153171092271805, 'train/grad_norm_pre_clip': 1.5988051891326904, 'timing/data': 0.0002888408489525318, 'timing/model': 0.43030332401394844, 'learning_rate/qwen_vl_interface': 9.316939213392257e-06, 'learning_rate/action_model': 9.316939213392257e-05, 'epoch': 0.42})
+[2026-07-28 11:20:45,806][starVLA.training.train_starvla][INFO] - Step 2275, Loss: {'train/loss': 0.042981162667274475, 'train/grad_norm_pre_clip': 1.209439754486084, 'timing/data': 0.00027884822338819504, 'timing/model': 0.4303371673449874, 'learning_rate/qwen_vl_interface': 9.314961043431877e-06, 'learning_rate/action_model': 9.314961043431876e-05, 'epoch': 0.42})
+[2026-07-28 11:20:52,375][starVLA.training.train_starvla][INFO] - Step 2276, Loss: {'train/loss': 0.03239046782255173, 'train/grad_norm_pre_clip': 0.9474343657493591, 'timing/data': 0.0003054710105061531, 'timing/model': 0.6845170161686838, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.42})
+[2026-07-28 11:20:58,564][starVLA.training.train_starvla][INFO] - Step 2277, Loss: {'train/loss': 0.05249521881341934, 'train/grad_norm_pre_clip': 1.5967093706130981, 'timing/data': 0.00028724269941449165, 'timing/model': 0.42418390326201916, 'learning_rate/qwen_vl_interface': 9.310996791669175e-06, 'learning_rate/action_model': 9.310996791669175e-05, 'epoch': 0.42})
+[2026-07-28 11:21:04,703][starVLA.training.train_starvla][INFO] - Step 2278, Loss: {'train/loss': 0.05159838870167732, 'train/grad_norm_pre_clip': 1.3452816009521484, 'timing/data': 0.0003831186331808567, 'timing/model': 0.4298268919810653, 'learning_rate/qwen_vl_interface': 9.309010712439213e-06, 'learning_rate/action_model': 9.309010712439213e-05, 'epoch': 0.42})
+[2026-07-28 11:21:10,750][starVLA.training.train_starvla][INFO] - Step 2279, Loss: {'train/loss': 0.05740605667233467, 'train/grad_norm_pre_clip': 1.496951699256897, 'timing/data': 0.0003902888856828213, 'timing/model': 0.4310097610577941, 'learning_rate/qwen_vl_interface': 9.307021999361267e-06, 'learning_rate/action_model': 9.307021999361267e-05, 'epoch': 0.42})
+[2026-07-28 11:21:17,102][starVLA.training.train_starvla][INFO] - Step 2280, Loss: {'train/loss': 0.07447709143161774, 'train/grad_norm_pre_clip': 1.0964761972427368, 'timing/data': 0.00031928811222314835, 'timing/model': 0.7199153932742774, 'learning_rate/qwen_vl_interface': 9.305030653725794e-06, 'learning_rate/action_model': 9.305030653725794e-05, 'epoch': 0.42})
+[2026-07-28 11:21:23,219][starVLA.training.train_starvla][INFO] - Step 2281, Loss: {'train/loss': 0.03946594148874283, 'train/grad_norm_pre_clip': 1.011392593383789, 'timing/data': 0.0002875938080251217, 'timing/model': 0.4212938267737627, 'learning_rate/qwen_vl_interface': 9.303036676824954e-06, 'learning_rate/action_model': 9.303036676824953e-05, 'epoch': 0.42})
+[2026-07-28 11:21:29,542][starVLA.training.train_starvla][INFO] - Step 2282, Loss: {'train/loss': 0.03736858814954758, 'train/grad_norm_pre_clip': 1.4010201692581177, 'timing/data': 0.00029424019157886505, 'timing/model': 0.6720162061974406, 'learning_rate/qwen_vl_interface': 9.301040069952616e-06, 'learning_rate/action_model': 9.301040069952616e-05, 'epoch': 0.42})
+[2026-07-28 11:21:35,657][starVLA.training.train_starvla][INFO] - Step 2283, Loss: {'train/loss': 0.028905846178531647, 'train/grad_norm_pre_clip': 1.2925026416778564, 'timing/data': 0.0003405371680855751, 'timing/model': 0.4250062396749854, 'learning_rate/qwen_vl_interface': 9.299040834404362e-06, 'learning_rate/action_model': 9.299040834404361e-05, 'epoch': 0.42})
+[2026-07-28 11:21:41,777][starVLA.training.train_starvla][INFO] - Step 2284, Loss: {'train/loss': 0.0824606865644455, 'train/grad_norm_pre_clip': 2.916964292526245, 'timing/data': 0.00037092575803399086, 'timing/model': 0.44444638723507524, 'learning_rate/qwen_vl_interface': 9.297038971477468e-06, 'learning_rate/action_model': 9.297038971477466e-05, 'epoch': 0.42})
+[2026-07-28 11:21:47,949][starVLA.training.train_starvla][INFO] - Step 2285, Loss: {'train/loss': 0.04375109076499939, 'train/grad_norm_pre_clip': 1.8652235269546509, 'timing/data': 0.0003071990795433521, 'timing/model': 0.4233466568402946, 'learning_rate/qwen_vl_interface': 9.295034482470922e-06, 'learning_rate/action_model': 9.295034482470923e-05, 'epoch': 0.42})
+[2026-07-28 11:21:54,417][starVLA.training.train_starvla][INFO] - Step 2286, Loss: {'train/loss': 0.033776961266994476, 'train/grad_norm_pre_clip': 1.1642252206802368, 'timing/data': 0.00039136409759521484, 'timing/model': 0.4250669036991894, 'learning_rate/qwen_vl_interface': 9.293027368685418e-06, 'learning_rate/action_model': 9.293027368685419e-05, 'epoch': 0.42})
+[2026-07-28 11:22:00,768][starVLA.training.train_starvla][INFO] - Step 2287, Loss: {'train/loss': 0.049183353781700134, 'train/grad_norm_pre_clip': 1.1055642366409302, 'timing/data': 0.0004865718074142933, 'timing/model': 0.42537009017542005, 'learning_rate/qwen_vl_interface': 9.29101763142335e-06, 'learning_rate/action_model': 9.29101763142335e-05, 'epoch': 0.42})
+[2026-07-28 11:22:06,990][starVLA.training.train_starvla][INFO] - Step 2288, Loss: {'train/loss': 0.04467901214957237, 'train/grad_norm_pre_clip': 2.7430672645568848, 'timing/data': 0.00031673489138484, 'timing/model': 0.4318540068343282, 'learning_rate/qwen_vl_interface': 9.28900527198881e-06, 'learning_rate/action_model': 9.289005271988811e-05, 'epoch': 0.42})
+[2026-07-28 11:22:13,192][starVLA.training.train_starvla][INFO] - Step 2289, Loss: {'train/loss': 0.1114659383893013, 'train/grad_norm_pre_clip': 3.72995662689209, 'timing/data': 0.00029306206852197647, 'timing/model': 0.42310516303405166, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.42})
+[2026-07-28 11:22:19,758][starVLA.training.train_starvla][INFO] - Step 2290, Loss: {'train/loss': 0.07650893926620483, 'train/grad_norm_pre_clip': 1.8614321947097778, 'timing/data': 0.000299645122140646, 'timing/model': 0.4851133059710264, 'learning_rate/qwen_vl_interface': 9.284972691827226e-06, 'learning_rate/action_model': 9.284972691827224e-05, 'epoch': 0.42})
+[2026-07-28 11:22:26,237][starVLA.training.train_starvla][INFO] - Step 2291, Loss: {'train/loss': 0.07549693435430527, 'train/grad_norm_pre_clip': 2.3186306953430176, 'timing/data': 0.0003991308622062206, 'timing/model': 0.447856810875237, 'learning_rate/qwen_vl_interface': 9.282952473716873e-06, 'learning_rate/action_model': 9.282952473716874e-05, 'epoch': 0.42})
+[2026-07-28 11:22:32,386][starVLA.training.train_starvla][INFO] - Step 2292, Loss: {'train/loss': 0.05007682368159294, 'train/grad_norm_pre_clip': 1.203598141670227, 'timing/data': 0.000415189191699028, 'timing/model': 0.4265914149582386, 'learning_rate/qwen_vl_interface': 9.280929638667447e-06, 'learning_rate/action_model': 9.280929638667447e-05, 'epoch': 0.42})
+[2026-07-28 11:22:38,838][starVLA.training.train_starvla][INFO] - Step 2293, Loss: {'train/loss': 0.07200515270233154, 'train/grad_norm_pre_clip': 3.00663685798645, 'timing/data': 0.0003116368316113949, 'timing/model': 0.42638008715584874, 'learning_rate/qwen_vl_interface': 9.278904187991542e-06, 'learning_rate/action_model': 9.278904187991541e-05, 'epoch': 0.42})
+[2026-07-28 11:22:44,893][starVLA.training.train_starvla][INFO] - Step 2294, Loss: {'train/loss': 0.045913755893707275, 'train/grad_norm_pre_clip': 1.0897775888442993, 'timing/data': 0.00031647179275751114, 'timing/model': 0.42392778396606445, 'learning_rate/qwen_vl_interface': 9.27687612300345e-06, 'learning_rate/action_model': 9.27687612300345e-05, 'epoch': 0.42})
+[2026-07-28 11:22:50,993][starVLA.training.train_starvla][INFO] - Step 2295, Loss: {'train/loss': 0.08053680509328842, 'train/grad_norm_pre_clip': 2.296079635620117, 'timing/data': 0.00036187609657645226, 'timing/model': 0.4270533840171993, 'learning_rate/qwen_vl_interface': 9.27484544501916e-06, 'learning_rate/action_model': 9.274845445019159e-05, 'epoch': 0.42})
+[2026-07-28 11:22:57,500][starVLA.training.train_starvla][INFO] - Step 2296, Loss: {'train/loss': 0.06446252763271332, 'train/grad_norm_pre_clip': 2.65087890625, 'timing/data': 0.00026469025760889053, 'timing/model': 0.8418207350187004, 'learning_rate/qwen_vl_interface': 9.27281215535636e-06, 'learning_rate/action_model': 9.272812155356358e-05, 'epoch': 0.42})
+[2026-07-28 11:23:03,487][starVLA.training.train_starvla][INFO] - Step 2297, Loss: {'train/loss': 0.029393788427114487, 'train/grad_norm_pre_clip': 0.7323482036590576, 'timing/data': 0.00027003325521945953, 'timing/model': 0.42246745293959975, 'learning_rate/qwen_vl_interface': 9.270776255334425e-06, 'learning_rate/action_model': 9.270776255334425e-05, 'epoch': 0.42})
+[2026-07-28 11:23:09,583][starVLA.training.train_starvla][INFO] - Step 2298, Loss: {'train/loss': 0.04919593781232834, 'train/grad_norm_pre_clip': 2.047264575958252, 'timing/data': 0.00035894615575671196, 'timing/model': 0.42195392679423094, 'learning_rate/qwen_vl_interface': 9.268737746274432e-06, 'learning_rate/action_model': 9.268737746274432e-05, 'epoch': 0.42})
+[2026-07-28 11:23:15,588][starVLA.training.train_starvla][INFO] - Step 2299, Loss: {'train/loss': 0.03743162751197815, 'train/grad_norm_pre_clip': 1.457637906074524, 'timing/data': 0.0003099157474935055, 'timing/model': 0.4234814839437604, 'learning_rate/qwen_vl_interface': 9.266696629499146e-06, 'learning_rate/action_model': 9.266696629499145e-05, 'epoch': 0.42})
+[2026-07-28 11:23:21,927][starVLA.training.train_starvla][INFO] - Step 2300, Loss: {'train/loss': 0.06920835375785828, 'train/grad_norm_pre_clip': 1.3631327152252197, 'timing/data': 0.00028143683448433876, 'timing/model': 0.4245129288174212, 'learning_rate/qwen_vl_interface': 9.264652906333026e-06, 'learning_rate/action_model': 9.264652906333026e-05, 'epoch': 0.42})
+[2026-07-28 11:23:27,967][starVLA.training.train_starvla][INFO] - Step 2301, Loss: {'train/loss': 0.04894108325242996, 'train/grad_norm_pre_clip': 1.1070232391357422, 'timing/data': 0.0002785520628094673, 'timing/model': 0.42644407879561186, 'learning_rate/qwen_vl_interface': 9.26260657810222e-06, 'learning_rate/action_model': 9.262606578102221e-05, 'epoch': 0.42})
+[2026-07-28 11:23:33,992][starVLA.training.train_starvla][INFO] - Step 2302, Loss: {'train/loss': 0.0313388854265213, 'train/grad_norm_pre_clip': 2.1599862575531006, 'timing/data': 0.00028227875009179115, 'timing/model': 0.424984073266387, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.42})
+[2026-07-28 11:23:40,249][starVLA.training.train_starvla][INFO] - Step 2303, Loss: {'train/loss': 0.04777326434850693, 'train/grad_norm_pre_clip': 1.8641034364700317, 'timing/data': 0.00033819302916526794, 'timing/model': 0.4238753546960652, 'learning_rate/qwen_vl_interface': 9.258506111759608e-06, 'learning_rate/action_model': 9.258506111759609e-05, 'epoch': 0.42})
+[2026-07-28 11:23:46,561][starVLA.training.train_starvla][INFO] - Step 2304, Loss: {'train/loss': 0.032224107533693314, 'train/grad_norm_pre_clip': 1.1735185384750366, 'timing/data': 0.00029759109020233154, 'timing/model': 0.4205733397975564, 'learning_rate/qwen_vl_interface': 9.256451976308548e-06, 'learning_rate/action_model': 9.25645197630855e-05, 'epoch': 0.42})
+[2026-07-28 11:23:52,571][starVLA.training.train_starvla][INFO] - Step 2305, Loss: {'train/loss': 0.03843574598431587, 'train/grad_norm_pre_clip': 1.1530729532241821, 'timing/data': 0.00042758602648973465, 'timing/model': 0.429667426738888, 'learning_rate/qwen_vl_interface': 9.2543952411143e-06, 'learning_rate/action_model': 9.254395241114299e-05, 'epoch': 0.42})
+[2026-07-28 11:23:58,666][starVLA.training.train_starvla][INFO] - Step 2306, Loss: {'train/loss': 0.06217283010482788, 'train/grad_norm_pre_clip': 1.44013249874115, 'timing/data': 0.000320404302328825, 'timing/model': 0.4250305681489408, 'learning_rate/qwen_vl_interface': 9.252335907511452e-06, 'learning_rate/action_model': 9.252335907511453e-05, 'epoch': 0.42})
+[2026-07-28 11:24:05,237][starVLA.training.train_starvla][INFO] - Step 2307, Loss: {'train/loss': 0.03427668288350105, 'train/grad_norm_pre_clip': 0.6260960698127747, 'timing/data': 0.00032487697899341583, 'timing/model': 0.6673936946317554, 'learning_rate/qwen_vl_interface': 9.250273976836289e-06, 'learning_rate/action_model': 9.250273976836289e-05, 'epoch': 0.42})
+[2026-07-28 11:24:11,343][starVLA.training.train_starvla][INFO] - Step 2308, Loss: {'train/loss': 0.05184027552604675, 'train/grad_norm_pre_clip': 1.21171236038208, 'timing/data': 0.00029697176069021225, 'timing/model': 0.4241104503162205, 'learning_rate/qwen_vl_interface': 9.248209450426772e-06, 'learning_rate/action_model': 9.248209450426772e-05, 'epoch': 0.42})
+[2026-07-28 11:24:17,606][starVLA.training.train_starvla][INFO] - Step 2309, Loss: {'train/loss': 0.04246878996491432, 'train/grad_norm_pre_clip': 1.5919263362884521, 'timing/data': 0.000291536096483469, 'timing/model': 0.6745915180072188, 'learning_rate/qwen_vl_interface': 9.24614232962255e-06, 'learning_rate/action_model': 9.24614232962255e-05, 'epoch': 0.42})
+[2026-07-28 11:24:23,808][starVLA.training.train_starvla][INFO] - Step 2310, Loss: {'train/loss': 0.08293472230434418, 'train/grad_norm_pre_clip': 3.2544636726379395, 'timing/data': 0.00048055220395326614, 'timing/model': 0.43549513490870595, 'learning_rate/qwen_vl_interface': 9.244072615764957e-06, 'learning_rate/action_model': 9.244072615764956e-05, 'epoch': 0.42})
+[2026-07-28 11:24:30,192][starVLA.training.train_starvla][INFO] - Step 2311, Loss: {'train/loss': 0.03743524104356766, 'train/grad_norm_pre_clip': 1.3719006776809692, 'timing/data': 0.00027885381132364273, 'timing/model': 0.4291074159555137, 'learning_rate/qwen_vl_interface': 9.242000310197005e-06, 'learning_rate/action_model': 9.242000310197004e-05, 'epoch': 0.42})
+[2026-07-28 11:24:36,235][starVLA.training.train_starvla][INFO] - Step 2312, Loss: {'train/loss': 0.01829642988741398, 'train/grad_norm_pre_clip': 0.6880832314491272, 'timing/data': 0.0004654061049222946, 'timing/model': 0.4300519027747214, 'learning_rate/qwen_vl_interface': 9.239925414263392e-06, 'learning_rate/action_model': 9.239925414263392e-05, 'epoch': 0.42})
+[2026-07-28 11:24:42,344][starVLA.training.train_starvla][INFO] - Step 2313, Loss: {'train/loss': 0.07624611258506775, 'train/grad_norm_pre_clip': 1.152048110961914, 'timing/data': 0.0003546271473169327, 'timing/model': 0.4380699950270355, 'learning_rate/qwen_vl_interface': 9.237847929310497e-06, 'learning_rate/action_model': 9.237847929310496e-05, 'epoch': 0.42})
+[2026-07-28 11:24:48,734][starVLA.training.train_starvla][INFO] - Step 2314, Loss: {'train/loss': 0.030072689056396484, 'train/grad_norm_pre_clip': 1.0198767185211182, 'timing/data': 0.0003814329393208027, 'timing/model': 0.42960864305496216, 'learning_rate/qwen_vl_interface': 9.235767856686375e-06, 'learning_rate/action_model': 9.235767856686375e-05, 'epoch': 0.43})
+[2026-07-28 11:24:54,989][starVLA.training.train_starvla][INFO] - Step 2315, Loss: {'train/loss': 0.025924600660800934, 'train/grad_norm_pre_clip': 0.707006573677063, 'timing/data': 0.00035428302362561226, 'timing/model': 0.42925283312797546, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.43})
+[2026-07-28 11:25:01,057][starVLA.training.train_starvla][INFO] - Step 2316, Loss: {'train/loss': 0.04758483171463013, 'train/grad_norm_pre_clip': 1.904893398284912, 'timing/data': 0.0002876371145248413, 'timing/model': 0.4292828058823943, 'learning_rate/qwen_vl_interface': 9.23159995382508e-06, 'learning_rate/action_model': 9.231599953825079e-05, 'epoch': 0.43})
+[2026-07-28 11:25:07,394][starVLA.training.train_starvla][INFO] - Step 2317, Loss: {'train/loss': 0.034806761890649796, 'train/grad_norm_pre_clip': 1.1656020879745483, 'timing/data': 0.0003829970955848694, 'timing/model': 0.4318580226972699, 'learning_rate/qwen_vl_interface': 9.22951212629241e-06, 'learning_rate/action_model': 9.229512126292411e-05, 'epoch': 0.43})
+[2026-07-28 11:25:13,706][starVLA.training.train_starvla][INFO] - Step 2318, Loss: {'train/loss': 0.09957911819219589, 'train/grad_norm_pre_clip': 2.214141607284546, 'timing/data': 0.00035068532451987267, 'timing/model': 0.6678868108429015, 'learning_rate/qwen_vl_interface': 9.227421716497531e-06, 'learning_rate/action_model': 9.227421716497531e-05, 'epoch': 0.43})
+[2026-07-28 11:25:19,815][starVLA.training.train_starvla][INFO] - Step 2319, Loss: {'train/loss': 0.07614418864250183, 'train/grad_norm_pre_clip': 2.5673651695251465, 'timing/data': 0.00035118311643600464, 'timing/model': 0.4499440770596266, 'learning_rate/qwen_vl_interface': 9.225328725796879e-06, 'learning_rate/action_model': 9.22532872579688e-05, 'epoch': 0.43})
+[2026-07-28 11:25:25,825][starVLA.training.train_starvla][INFO] - Step 2320, Loss: {'train/loss': 0.05460912734270096, 'train/grad_norm_pre_clip': 3.3341164588928223, 'timing/data': 0.000300068873912096, 'timing/model': 0.42315654130652547, 'learning_rate/qwen_vl_interface': 9.223233155548579e-06, 'learning_rate/action_model': 9.223233155548578e-05, 'epoch': 0.43})
+[2026-07-28 11:25:32,159][starVLA.training.train_starvla][INFO] - Step 2321, Loss: {'train/loss': 0.06372398883104324, 'train/grad_norm_pre_clip': 1.158722996711731, 'timing/data': 0.000272498931735754, 'timing/model': 0.6668012570589781, 'learning_rate/qwen_vl_interface': 9.22113500711242e-06, 'learning_rate/action_model': 9.221135007112419e-05, 'epoch': 0.43})
+[2026-07-28 11:25:38,181][starVLA.training.train_starvla][INFO] - Step 2322, Loss: {'train/loss': 0.08155930042266846, 'train/grad_norm_pre_clip': 5.609338283538818, 'timing/data': 0.00038430094718933105, 'timing/model': 0.4203938031569123, 'learning_rate/qwen_vl_interface': 9.219034281849867e-06, 'learning_rate/action_model': 9.219034281849867e-05, 'epoch': 0.43})
+[2026-07-28 11:25:44,443][starVLA.training.train_starvla][INFO] - Step 2323, Loss: {'train/loss': 0.054544828832149506, 'train/grad_norm_pre_clip': 2.0113611221313477, 'timing/data': 0.00040867598727345467, 'timing/model': 0.6846576863899827, 'learning_rate/qwen_vl_interface': 9.216930981124062e-06, 'learning_rate/action_model': 9.216930981124062e-05, 'epoch': 0.43})
+[2026-07-28 11:25:50,602][starVLA.training.train_starvla][INFO] - Step 2324, Loss: {'train/loss': 0.07141198962926865, 'train/grad_norm_pre_clip': 1.643494963645935, 'timing/data': 0.0004179151728749275, 'timing/model': 0.42188544292002916, 'learning_rate/qwen_vl_interface': 9.214825106299808e-06, 'learning_rate/action_model': 9.214825106299808e-05, 'epoch': 0.43})
+[2026-07-28 11:25:56,737][starVLA.training.train_starvla][INFO] - Step 2325, Loss: {'train/loss': 0.11928682029247284, 'train/grad_norm_pre_clip': 2.448495626449585, 'timing/data': 0.0003053639084100723, 'timing/model': 0.4759324458427727, 'learning_rate/qwen_vl_interface': 9.212716658743592e-06, 'learning_rate/action_model': 9.212716658743592e-05, 'epoch': 0.43})
+[2026-07-28 11:26:02,878][starVLA.training.train_starvla][INFO] - Step 2326, Loss: {'train/loss': 0.03024996630847454, 'train/grad_norm_pre_clip': 1.3049308061599731, 'timing/data': 0.0003078239969909191, 'timing/model': 0.42734835017472506, 'learning_rate/qwen_vl_interface': 9.210605639823558e-06, 'learning_rate/action_model': 9.210605639823557e-05, 'epoch': 0.43})
+[2026-07-28 11:26:09,157][starVLA.training.train_starvla][INFO] - Step 2327, Loss: {'train/loss': 0.05450785160064697, 'train/grad_norm_pre_clip': 1.7873860597610474, 'timing/data': 0.000291415024548769, 'timing/model': 0.4342132182791829, 'learning_rate/qwen_vl_interface': 9.208492050909523e-06, 'learning_rate/action_model': 9.208492050909523e-05, 'epoch': 0.43})
+[2026-07-28 11:26:15,273][starVLA.training.train_starvla][INFO] - Step 2328, Loss: {'train/loss': 0.028045516461133957, 'train/grad_norm_pre_clip': 1.1538996696472168, 'timing/data': 0.0002753450535237789, 'timing/model': 0.4254563502036035, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.43})
+[2026-07-28 11:26:21,616][starVLA.training.train_starvla][INFO] - Step 2329, Loss: {'train/loss': 0.03161514550447464, 'train/grad_norm_pre_clip': 0.870990514755249, 'timing/data': 0.0002803127281367779, 'timing/model': 0.6656302260234952, 'learning_rate/qwen_vl_interface': 9.204257168587061e-06, 'learning_rate/action_model': 9.204257168587061e-05, 'epoch': 0.43})
+[2026-07-28 11:26:27,675][starVLA.training.train_starvla][INFO] - Step 2330, Loss: {'train/loss': 0.07034528255462646, 'train/grad_norm_pre_clip': 2.3834850788116455, 'timing/data': 0.0003371201455593109, 'timing/model': 0.4401452550664544, 'learning_rate/qwen_vl_interface': 9.2021358779266e-06, 'learning_rate/action_model': 9.202135877926601e-05, 'epoch': 0.43})
+[2026-07-28 11:26:34,017][starVLA.training.train_starvla][INFO] - Step 2331, Loss: {'train/loss': 0.10244767367839813, 'train/grad_norm_pre_clip': 2.1854336261749268, 'timing/data': 0.0004540281370282173, 'timing/model': 0.43124067410826683, 'learning_rate/qwen_vl_interface': 9.200012022768077e-06, 'learning_rate/action_model': 9.200012022768077e-05, 'epoch': 0.43})
+[2026-07-28 11:26:40,089][starVLA.training.train_starvla][INFO] - Step 2332, Loss: {'train/loss': 0.058578282594680786, 'train/grad_norm_pre_clip': 1.0147935152053833, 'timing/data': 0.0003749825991690159, 'timing/model': 0.42731485702097416, 'learning_rate/qwen_vl_interface': 9.19788560448964e-06, 'learning_rate/action_model': 9.197885604489639e-05, 'epoch': 0.43})
+[2026-07-28 11:26:46,182][starVLA.training.train_starvla][INFO] - Step 2333, Loss: {'train/loss': 0.0462816059589386, 'train/grad_norm_pre_clip': 1.1757999658584595, 'timing/data': 0.00037382496520876884, 'timing/model': 0.4253411418758333, 'learning_rate/qwen_vl_interface': 9.195756624471094e-06, 'learning_rate/action_model': 9.195756624471094e-05, 'epoch': 0.43})
+[2026-07-28 11:26:52,615][starVLA.training.train_starvla][INFO] - Step 2334, Loss: {'train/loss': 0.04761769622564316, 'train/grad_norm_pre_clip': 0.8373430967330933, 'timing/data': 0.0003731739707291126, 'timing/model': 0.4276869911700487, 'learning_rate/qwen_vl_interface': 9.193625084093912e-06, 'learning_rate/action_model': 9.193625084093911e-05, 'epoch': 0.43})
+[2026-07-28 11:26:58,688][starVLA.training.train_starvla][INFO] - Step 2335, Loss: {'train/loss': 0.04562368243932724, 'train/grad_norm_pre_clip': 0.9609456658363342, 'timing/data': 0.0003012591041624546, 'timing/model': 0.4380756220780313, 'learning_rate/qwen_vl_interface': 9.191490984741227e-06, 'learning_rate/action_model': 9.191490984741227e-05, 'epoch': 0.43})
+[2026-07-28 11:27:04,515][starVLA.training.train_starvla][INFO] - Step 2336, Loss: {'train/loss': 0.041623491793870926, 'train/grad_norm_pre_clip': 1.0644725561141968, 'timing/data': 0.00028763292357325554, 'timing/model': 0.4341176743619144, 'learning_rate/qwen_vl_interface': 9.189354327797835e-06, 'learning_rate/action_model': 9.189354327797836e-05, 'epoch': 0.43})
+[2026-07-28 11:27:10,753][starVLA.training.train_starvla][INFO] - Step 2337, Loss: {'train/loss': 0.0362287238240242, 'train/grad_norm_pre_clip': 0.9123069643974304, 'timing/data': 0.00034210318699479103, 'timing/model': 0.5973299471661448, 'learning_rate/qwen_vl_interface': 9.18721511465019e-06, 'learning_rate/action_model': 9.187215114650191e-05, 'epoch': 0.43})
+[2026-07-28 11:27:17,084][starVLA.training.train_starvla][INFO] - Step 2338, Loss: {'train/loss': 0.04785194620490074, 'train/grad_norm_pre_clip': 1.7645998001098633, 'timing/data': 0.00025221705436706543, 'timing/model': 0.42085618898272514, 'learning_rate/qwen_vl_interface': 9.185073346686402e-06, 'learning_rate/action_model': 9.185073346686401e-05, 'epoch': 0.43})
+[2026-07-28 11:27:23,105][starVLA.training.train_starvla][INFO] - Step 2339, Loss: {'train/loss': 0.057889074087142944, 'train/grad_norm_pre_clip': 2.4076895713806152, 'timing/data': 0.0003595859743654728, 'timing/model': 0.433750222902745, 'learning_rate/qwen_vl_interface': 9.182929025296242e-06, 'learning_rate/action_model': 9.182929025296242e-05, 'epoch': 0.43})
+[2026-07-28 11:27:29,175][starVLA.training.train_starvla][INFO] - Step 2340, Loss: {'train/loss': 0.10505884885787964, 'train/grad_norm_pre_clip': 2.1739373207092285, 'timing/data': 0.0003040512092411518, 'timing/model': 0.4297063290141523, 'learning_rate/qwen_vl_interface': 9.180782151871138e-06, 'learning_rate/action_model': 9.180782151871137e-05, 'epoch': 0.43})
+[2026-07-28 11:27:35,451][starVLA.training.train_starvla][INFO] - Step 2341, Loss: {'train/loss': 0.06329607218503952, 'train/grad_norm_pre_clip': 1.0492595434188843, 'timing/data': 0.0006353668868541718, 'timing/model': 0.4250465859659016, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.43})
+[2026-07-28 11:27:41,446][starVLA.training.train_starvla][INFO] - Step 2342, Loss: {'train/loss': 0.05714324116706848, 'train/grad_norm_pre_clip': 1.2165180444717407, 'timing/data': 0.0002734311856329441, 'timing/model': 0.42743443604558706, 'learning_rate/qwen_vl_interface': 9.176480754490078e-06, 'learning_rate/action_model': 9.176480754490078e-05, 'epoch': 0.43})
+[2026-07-28 11:27:47,721][starVLA.training.train_starvla][INFO] - Step 2343, Loss: {'train/loss': 0.022210467606782913, 'train/grad_norm_pre_clip': 1.6613426208496094, 'timing/data': 0.00036684004589915276, 'timing/model': 0.6870279088616371, 'learning_rate/qwen_vl_interface': 9.174326233325253e-06, 'learning_rate/action_model': 9.174326233325253e-05, 'epoch': 0.43})
+[2026-07-28 11:27:54,169][starVLA.training.train_starvla][INFO] - Step 2344, Loss: {'train/loss': 0.052843570709228516, 'train/grad_norm_pre_clip': 1.6853188276290894, 'timing/data': 0.00030254386365413666, 'timing/model': 0.6524182399734855, 'learning_rate/qwen_vl_interface': 9.17216916570774e-06, 'learning_rate/action_model': 9.172169165707741e-05, 'epoch': 0.43})
+[2026-07-28 11:28:00,151][starVLA.training.train_starvla][INFO] - Step 2345, Loss: {'train/loss': 0.06020880490541458, 'train/grad_norm_pre_clip': 1.6366225481033325, 'timing/data': 0.0002893209457397461, 'timing/model': 0.42475546803325415, 'learning_rate/qwen_vl_interface': 9.170009553037239e-06, 'learning_rate/action_model': 9.170009553037238e-05, 'epoch': 0.43})
+[2026-07-28 11:28:06,483][starVLA.training.train_starvla][INFO] - Step 2346, Loss: {'train/loss': 0.060340218245983124, 'train/grad_norm_pre_clip': 1.1889169216156006, 'timing/data': 0.000309675931930542, 'timing/model': 0.4266620809212327, 'learning_rate/qwen_vl_interface': 9.167847396715098e-06, 'learning_rate/action_model': 9.167847396715098e-05, 'epoch': 0.43})
+[2026-07-28 11:28:12,652][starVLA.training.train_starvla][INFO] - Step 2347, Loss: {'train/loss': 0.030614282935857773, 'train/grad_norm_pre_clip': 0.5819336175918579, 'timing/data': 0.00030856719240546227, 'timing/model': 0.43285259790718555, 'learning_rate/qwen_vl_interface': 9.165682698144317e-06, 'learning_rate/action_model': 9.165682698144317e-05, 'epoch': 0.43})
+[2026-07-28 11:28:18,673][starVLA.training.train_starvla][INFO] - Step 2348, Loss: {'train/loss': 0.03674941137433052, 'train/grad_norm_pre_clip': 1.4204847812652588, 'timing/data': 0.000317287165671587, 'timing/model': 0.4240959440357983, 'learning_rate/qwen_vl_interface': 9.163515458729545e-06, 'learning_rate/action_model': 9.163515458729545e-05, 'epoch': 0.43})
+[2026-07-28 11:28:24,925][starVLA.training.train_starvla][INFO] - Step 2349, Loss: {'train/loss': 0.047058142721652985, 'train/grad_norm_pre_clip': 1.4670530557632446, 'timing/data': 0.00039889803156256676, 'timing/model': 0.4219076163135469, 'learning_rate/qwen_vl_interface': 9.16134567987708e-06, 'learning_rate/action_model': 9.16134567987708e-05, 'epoch': 0.43})
+[2026-07-28 11:28:31,334][starVLA.training.train_starvla][INFO] - Step 2350, Loss: {'train/loss': 0.042488522827625275, 'train/grad_norm_pre_clip': 0.9096963405609131, 'timing/data': 0.00037800101563334465, 'timing/model': 0.4782593627460301, 'learning_rate/qwen_vl_interface': 9.159173362994868e-06, 'learning_rate/action_model': 9.159173362994869e-05, 'epoch': 0.43})
+[2026-07-28 11:28:37,505][starVLA.training.train_starvla][INFO] - Step 2351, Loss: {'train/loss': 0.04344932734966278, 'train/grad_norm_pre_clip': 1.186062216758728, 'timing/data': 0.0003061550669372082, 'timing/model': 0.4320174288004637, 'learning_rate/qwen_vl_interface': 9.156998509492503e-06, 'learning_rate/action_model': 9.156998509492503e-05, 'epoch': 0.43})
+[2026-07-28 11:28:43,575][starVLA.training.train_starvla][INFO] - Step 2352, Loss: {'train/loss': 0.03231119364500046, 'train/grad_norm_pre_clip': 0.9994655847549438, 'timing/data': 0.00030731596052646637, 'timing/model': 0.4306221790611744, 'learning_rate/qwen_vl_interface': 9.154821120781221e-06, 'learning_rate/action_model': 9.154821120781222e-05, 'epoch': 0.43})
+[2026-07-28 11:28:49,649][starVLA.training.train_starvla][INFO] - Step 2353, Loss: {'train/loss': 0.03914712741971016, 'train/grad_norm_pre_clip': 2.404980182647705, 'timing/data': 0.0003637690097093582, 'timing/model': 0.4238522411324084, 'learning_rate/qwen_vl_interface': 9.152641198273911e-06, 'learning_rate/action_model': 9.15264119827391e-05, 'epoch': 0.43})
+[2026-07-28 11:28:55,725][starVLA.training.train_starvla][INFO] - Step 2354, Loss: {'train/loss': 0.037483278661966324, 'train/grad_norm_pre_clip': 1.3044434785842896, 'timing/data': 0.00036892620846629143, 'timing/model': 0.6739906440488994, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.43})
+[2026-07-28 11:29:01,597][starVLA.training.train_starvla][INFO] - Step 2355, Loss: {'train/loss': 0.04383448138833046, 'train/grad_norm_pre_clip': 1.363057255744934, 'timing/data': 0.00045153917744755745, 'timing/model': 0.4673089240677655, 'learning_rate/qwen_vl_interface': 9.148273757530948e-06, 'learning_rate/action_model': 9.148273757530948e-05, 'epoch': 0.43})
+[2026-07-28 11:29:07,756][starVLA.training.train_starvla][INFO] - Step 2356, Loss: {'train/loss': 0.0677201896905899, 'train/grad_norm_pre_clip': 1.532666802406311, 'timing/data': 0.00035614194348454475, 'timing/model': 0.42332567228004336, 'learning_rate/qwen_vl_interface': 9.146086242129283e-06, 'learning_rate/action_model': 9.146086242129282e-05, 'epoch': 0.43})
+[2026-07-28 11:29:13,872][starVLA.training.train_starvla][INFO] - Step 2357, Loss: {'train/loss': 0.025299090892076492, 'train/grad_norm_pre_clip': 0.8835746645927429, 'timing/data': 0.0003047729842364788, 'timing/model': 0.4311468559317291, 'learning_rate/qwen_vl_interface': 9.143896198599553e-06, 'learning_rate/action_model': 9.143896198599553e-05, 'epoch': 0.43})
+[2026-07-28 11:29:20,231][starVLA.training.train_starvla][INFO] - Step 2358, Loss: {'train/loss': 0.02437898889183998, 'train/grad_norm_pre_clip': 0.9901434183120728, 'timing/data': 0.0003435802645981312, 'timing/model': 0.42176234582439065, 'learning_rate/qwen_vl_interface': 9.141703628362858e-06, 'learning_rate/action_model': 9.141703628362858e-05, 'epoch': 0.43})
+[2026-07-28 11:29:26,504][starVLA.training.train_starvla][INFO] - Step 2359, Loss: {'train/loss': 0.026214510202407837, 'train/grad_norm_pre_clip': 1.6179722547531128, 'timing/data': 0.00030002184212207794, 'timing/model': 0.42341476818546653, 'learning_rate/qwen_vl_interface': 9.139508532841928e-06, 'learning_rate/action_model': 9.139508532841927e-05, 'epoch': 0.43})
+[2026-07-28 11:29:32,560][starVLA.training.train_starvla][INFO] - Step 2360, Loss: {'train/loss': 0.07366922497749329, 'train/grad_norm_pre_clip': 3.239232063293457, 'timing/data': 0.00030917488038539886, 'timing/model': 0.4251194908283651, 'learning_rate/qwen_vl_interface': 9.137310913461139e-06, 'learning_rate/action_model': 9.137310913461139e-05, 'epoch': 0.43})
+[2026-07-28 11:29:38,703][starVLA.training.train_starvla][INFO] - Step 2361, Loss: {'train/loss': 0.017370257526636124, 'train/grad_norm_pre_clip': 1.1185977458953857, 'timing/data': 0.00035709328949451447, 'timing/model': 0.42783880373463035, 'learning_rate/qwen_vl_interface': 9.135110771646502e-06, 'learning_rate/action_model': 9.135110771646502e-05, 'epoch': 0.43})
+[2026-07-28 11:29:45,244][starVLA.training.train_starvla][INFO] - Step 2362, Loss: {'train/loss': 0.02660219557583332, 'train/grad_norm_pre_clip': 1.7978289127349854, 'timing/data': 0.000281329732388258, 'timing/model': 0.4199861320666969, 'learning_rate/qwen_vl_interface': 9.132908108825667e-06, 'learning_rate/action_model': 9.132908108825667e-05, 'epoch': 0.43})
+[2026-07-28 11:29:51,308][starVLA.training.train_starvla][INFO] - Step 2363, Loss: {'train/loss': 0.020171916112303734, 'train/grad_norm_pre_clip': 0.9513260126113892, 'timing/data': 0.00037620216608047485, 'timing/model': 0.42417625105008483, 'learning_rate/qwen_vl_interface': 9.130702926427915e-06, 'learning_rate/action_model': 9.130702926427914e-05, 'epoch': 0.43})
+[2026-07-28 11:29:57,352][starVLA.training.train_starvla][INFO] - Step 2364, Loss: {'train/loss': 0.06156199425458908, 'train/grad_norm_pre_clip': 1.4456623792648315, 'timing/data': 0.0003099110908806324, 'timing/model': 0.42304070573300123, 'learning_rate/qwen_vl_interface': 9.128495225884167e-06, 'learning_rate/action_model': 9.128495225884166e-05, 'epoch': 0.43})
+[2026-07-28 11:30:03,544][starVLA.training.train_starvla][INFO] - Step 2365, Loss: {'train/loss': 0.060290057212114334, 'train/grad_norm_pre_clip': 1.5654438734054565, 'timing/data': 0.00030088890343904495, 'timing/model': 0.602642169687897, 'learning_rate/qwen_vl_interface': 9.126285008626975e-06, 'learning_rate/action_model': 9.126285008626974e-05, 'epoch': 0.43})
+[2026-07-28 11:30:09,872][starVLA.training.train_starvla][INFO] - Step 2366, Loss: {'train/loss': 0.08378739655017853, 'train/grad_norm_pre_clip': 1.9945951700210571, 'timing/data': 0.0003626542165875435, 'timing/model': 0.42290865490213037, 'learning_rate/qwen_vl_interface': 9.124072276090524e-06, 'learning_rate/action_model': 9.124072276090524e-05, 'epoch': 0.43})
+[2026-07-28 11:30:15,947][starVLA.training.train_starvla][INFO] - Step 2367, Loss: {'train/loss': 0.06558840721845627, 'train/grad_norm_pre_clip': 2.9361133575439453, 'timing/data': 0.00030688103288412094, 'timing/model': 0.42325921123847365, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.43})
+[2026-07-28 11:30:21,997][starVLA.training.train_starvla][INFO] - Step 2368, Loss: {'train/loss': 0.04709357023239136, 'train/grad_norm_pre_clip': 1.746907353401184, 'timing/data': 0.0009278221987187862, 'timing/model': 0.4287102408707142, 'learning_rate/qwen_vl_interface': 9.119639270924754e-06, 'learning_rate/action_model': 9.119639270924754e-05, 'epoch': 0.43})
+[2026-07-28 11:30:28,301][starVLA.training.train_starvla][INFO] - Step 2369, Loss: {'train/loss': 0.09156152606010437, 'train/grad_norm_pre_clip': 3.6268672943115234, 'timing/data': 0.00041485298424959183, 'timing/model': 0.424864721018821, 'learning_rate/qwen_vl_interface': 9.11741900117196e-06, 'learning_rate/action_model': 9.11741900117196e-05, 'epoch': 0.44})
+[2026-07-28 11:30:34,377][starVLA.training.train_starvla][INFO] - Step 2370, Loss: {'train/loss': 0.0902310162782669, 'train/grad_norm_pre_clip': 2.763442277908325, 'timing/data': 0.00039138877764344215, 'timing/model': 0.42656355490908027, 'learning_rate/qwen_vl_interface': 9.115196221892969e-06, 'learning_rate/action_model': 9.115196221892969e-05, 'epoch': 0.44})
+[2026-07-28 11:30:40,772][starVLA.training.train_starvla][INFO] - Step 2371, Loss: {'train/loss': 0.02091187983751297, 'train/grad_norm_pre_clip': 1.4936420917510986, 'timing/data': 0.00030110403895378113, 'timing/model': 0.41986026987433434, 'learning_rate/qwen_vl_interface': 9.112970934530111e-06, 'learning_rate/action_model': 9.112970934530111e-05, 'epoch': 0.44})
+[2026-07-28 11:30:46,900][starVLA.training.train_starvla][INFO] - Step 2372, Loss: {'train/loss': 0.05000797659158707, 'train/grad_norm_pre_clip': 1.5288952589035034, 'timing/data': 0.0003069792874157429, 'timing/model': 0.4240048290230334, 'learning_rate/qwen_vl_interface': 9.110743140527353e-06, 'learning_rate/action_model': 9.110743140527352e-05, 'epoch': 0.44})
+[2026-07-28 11:30:53,295][starVLA.training.train_starvla][INFO] - Step 2373, Loss: {'train/loss': 0.04849879816174507, 'train/grad_norm_pre_clip': 1.1091482639312744, 'timing/data': 0.0003133933059871197, 'timing/model': 0.7378109698183835, 'learning_rate/qwen_vl_interface': 9.108512841330288e-06, 'learning_rate/action_model': 9.108512841330287e-05, 'epoch': 0.44})
+[2026-07-28 11:30:59,367][starVLA.training.train_starvla][INFO] - Step 2374, Loss: {'train/loss': 0.05187121406197548, 'train/grad_norm_pre_clip': 2.0124552249908447, 'timing/data': 0.00027199089527130127, 'timing/model': 0.421266783028841, 'learning_rate/qwen_vl_interface': 9.106280038386129e-06, 'learning_rate/action_model': 9.10628003838613e-05, 'epoch': 0.44})
+[2026-07-28 11:31:05,493][starVLA.training.train_starvla][INFO] - Step 2375, Loss: {'train/loss': 0.07559581100940704, 'train/grad_norm_pre_clip': 3.131307601928711, 'timing/data': 0.00029708119109272957, 'timing/model': 0.43962436309084296, 'learning_rate/qwen_vl_interface': 9.104044733143723e-06, 'learning_rate/action_model': 9.104044733143723e-05, 'epoch': 0.44})
+[2026-07-28 11:31:11,526][starVLA.training.train_starvla][INFO] - Step 2376, Loss: {'train/loss': 0.030945302918553352, 'train/grad_norm_pre_clip': 1.6817171573638916, 'timing/data': 0.0002884240821003914, 'timing/model': 0.43737451592460275, 'learning_rate/qwen_vl_interface': 9.10180692705353e-06, 'learning_rate/action_model': 9.10180692705353e-05, 'epoch': 0.44})
+[2026-07-28 11:31:17,804][starVLA.training.train_starvla][INFO] - Step 2377, Loss: {'train/loss': 0.0469999685883522, 'train/grad_norm_pre_clip': 2.956807851791382, 'timing/data': 0.0002769818529486656, 'timing/model': 0.42038998287171125, 'learning_rate/qwen_vl_interface': 9.099566621567643e-06, 'learning_rate/action_model': 9.099566621567642e-05, 'epoch': 0.44})
+[2026-07-28 11:31:23,877][starVLA.training.train_starvla][INFO] - Step 2378, Loss: {'train/loss': 0.09006836265325546, 'train/grad_norm_pre_clip': 2.4545798301696777, 'timing/data': 0.00036928290501236916, 'timing/model': 0.4399953498505056, 'learning_rate/qwen_vl_interface': 9.097323818139767e-06, 'learning_rate/action_model': 9.097323818139767e-05, 'epoch': 0.44})
+[2026-07-28 11:31:30,146][starVLA.training.train_starvla][INFO] - Step 2379, Loss: {'train/loss': 0.05445855110883713, 'train/grad_norm_pre_clip': 1.1094449758529663, 'timing/data': 0.00026122434064745903, 'timing/model': 0.42072686506435275, 'learning_rate/qwen_vl_interface': 9.095078518225237e-06, 'learning_rate/action_model': 9.095078518225236e-05, 'epoch': 0.44})
+[2026-07-28 11:31:36,274][starVLA.training.train_starvla][INFO] - Step 2380, Loss: {'train/loss': 0.05410129204392433, 'train/grad_norm_pre_clip': 1.760967493057251, 'timing/data': 0.0002996833063662052, 'timing/model': 0.42564959824085236, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.44})
+[2026-07-28 11:31:42,372][starVLA.training.train_starvla][INFO] - Step 2381, Loss: {'train/loss': 0.02455790527164936, 'train/grad_norm_pre_clip': 1.0022337436676025, 'timing/data': 0.0002797730267047882, 'timing/model': 0.4244912639260292, 'learning_rate/qwen_vl_interface': 9.09058043476563e-06, 'learning_rate/action_model': 9.09058043476563e-05, 'epoch': 0.44})
+[2026-07-28 11:31:48,673][starVLA.training.train_starvla][INFO] - Step 2382, Loss: {'train/loss': 0.05173696205019951, 'train/grad_norm_pre_clip': 2.0613574981689453, 'timing/data': 0.00027505122125148773, 'timing/model': 0.42240190505981445, 'learning_rate/qwen_vl_interface': 9.088327654139311e-06, 'learning_rate/action_model': 9.08832765413931e-05, 'epoch': 0.44})
+[2026-07-28 11:31:54,985][starVLA.training.train_starvla][INFO] - Step 2383, Loss: {'train/loss': 0.03415997698903084, 'train/grad_norm_pre_clip': 1.1267129182815552, 'timing/data': 0.0002684718929231167, 'timing/model': 0.42293302388861775, 'learning_rate/qwen_vl_interface': 9.08607238286385e-06, 'learning_rate/action_model': 9.08607238286385e-05, 'epoch': 0.44})
+[2026-07-28 11:32:00,984][starVLA.training.train_starvla][INFO] - Step 2384, Loss: {'train/loss': 0.06083561107516289, 'train/grad_norm_pre_clip': 1.6707276105880737, 'timing/data': 0.0003575901500880718, 'timing/model': 0.43110159086063504, 'learning_rate/qwen_vl_interface': 9.083814622402667e-06, 'learning_rate/action_model': 9.083814622402666e-05, 'epoch': 0.44})
+[2026-07-28 11:32:07,014][starVLA.training.train_starvla][INFO] - Step 2385, Loss: {'train/loss': 0.04317667335271835, 'train/grad_norm_pre_clip': 1.20181405544281, 'timing/data': 0.00037654489278793335, 'timing/model': 0.430200700648129, 'learning_rate/qwen_vl_interface': 9.0815543742208e-06, 'learning_rate/action_model': 9.081554374220799e-05, 'epoch': 0.44})
+[2026-07-28 11:32:13,494][starVLA.training.train_starvla][INFO] - Step 2386, Loss: {'train/loss': 0.08584655821323395, 'train/grad_norm_pre_clip': 1.8619805574417114, 'timing/data': 0.0002738218754529953, 'timing/model': 0.6577936466783285, 'learning_rate/qwen_vl_interface': 9.079291639784897e-06, 'learning_rate/action_model': 9.079291639784896e-05, 'epoch': 0.44})
+[2026-07-28 11:32:19,551][starVLA.training.train_starvla][INFO] - Step 2387, Loss: {'train/loss': 0.029911372810602188, 'train/grad_norm_pre_clip': 0.8574005961418152, 'timing/data': 0.0003724549897015095, 'timing/model': 0.4292490081861615, 'learning_rate/qwen_vl_interface': 9.077026420563223e-06, 'learning_rate/action_model': 9.077026420563223e-05, 'epoch': 0.44})
+[2026-07-28 11:32:25,626][starVLA.training.train_starvla][INFO] - Step 2388, Loss: {'train/loss': 0.07845836877822876, 'train/grad_norm_pre_clip': 2.8444015979766846, 'timing/data': 0.0002820049412548542, 'timing/model': 0.4293716410174966, 'learning_rate/qwen_vl_interface': 9.074758718025652e-06, 'learning_rate/action_model': 9.074758718025652e-05, 'epoch': 0.44})
+[2026-07-28 11:32:31,868][starVLA.training.train_starvla][INFO] - Step 2389, Loss: {'train/loss': 0.03323935344815254, 'train/grad_norm_pre_clip': 0.8129040002822876, 'timing/data': 0.00036581000313162804, 'timing/model': 0.4194590989500284, 'learning_rate/qwen_vl_interface': 9.072488533643678e-06, 'learning_rate/action_model': 9.072488533643677e-05, 'epoch': 0.44})
+[2026-07-28 11:32:38,100][starVLA.training.train_starvla][INFO] - Step 2390, Loss: {'train/loss': 0.02736988663673401, 'train/grad_norm_pre_clip': 0.64565110206604, 'timing/data': 0.00027651526033878326, 'timing/model': 0.4208586527965963, 'learning_rate/qwen_vl_interface': 9.070215868890391e-06, 'learning_rate/action_model': 9.070215868890391e-05, 'epoch': 0.44})
+[2026-07-28 11:32:44,149][starVLA.training.train_starvla][INFO] - Step 2391, Loss: {'train/loss': 0.06238620728254318, 'train/grad_norm_pre_clip': 1.7230985164642334, 'timing/data': 0.00035505741834640503, 'timing/model': 0.4213838651776314, 'learning_rate/qwen_vl_interface': 9.067940725240503e-06, 'learning_rate/action_model': 9.067940725240503e-05, 'epoch': 0.44})
+[2026-07-28 11:32:50,185][starVLA.training.train_starvla][INFO] - Step 2392, Loss: {'train/loss': 0.07265830039978027, 'train/grad_norm_pre_clip': 2.123512029647827, 'timing/data': 0.000279740896075964, 'timing/model': 0.4231642330996692, 'learning_rate/qwen_vl_interface': 9.06566310417033e-06, 'learning_rate/action_model': 9.06566310417033e-05, 'epoch': 0.44})
+[2026-07-28 11:32:56,302][starVLA.training.train_starvla][INFO] - Step 2393, Loss: {'train/loss': 0.04536431282758713, 'train/grad_norm_pre_clip': 1.8281646966934204, 'timing/data': 0.00031168991699814796, 'timing/model': 0.672228925395757, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.44})
+[2026-07-28 11:33:02,636][starVLA.training.train_starvla][INFO] - Step 2394, Loss: {'train/loss': 0.054439108818769455, 'train/grad_norm_pre_clip': 1.163639783859253, 'timing/data': 0.00027623679488897324, 'timing/model': 0.6594471191056073, 'learning_rate/qwen_vl_interface': 9.061100435682429e-06, 'learning_rate/action_model': 9.061100435682429e-05, 'epoch': 0.44})
+[2026-07-28 11:33:08,690][starVLA.training.train_starvla][INFO] - Step 2395, Loss: {'train/loss': 0.07851430028676987, 'train/grad_norm_pre_clip': 2.951138973236084, 'timing/data': 0.0008202390745282173, 'timing/model': 0.42196669103577733, 'learning_rate/qwen_vl_interface': 9.058815391225363e-06, 'learning_rate/action_model': 9.058815391225362e-05, 'epoch': 0.44})
+[2026-07-28 11:33:14,725][starVLA.training.train_starvla][INFO] - Step 2396, Loss: {'train/loss': 0.03515440225601196, 'train/grad_norm_pre_clip': 0.8521205186843872, 'timing/data': 0.00036850618198513985, 'timing/model': 0.423131522256881, 'learning_rate/qwen_vl_interface': 9.056527875269344e-06, 'learning_rate/action_model': 9.056527875269344e-05, 'epoch': 0.44})
+[2026-07-28 11:33:20,816][starVLA.training.train_starvla][INFO] - Step 2397, Loss: {'train/loss': 0.04395139589905739, 'train/grad_norm_pre_clip': 0.9285093545913696, 'timing/data': 0.00029940763488411903, 'timing/model': 0.4235113998875022, 'learning_rate/qwen_vl_interface': 9.054237889298714e-06, 'learning_rate/action_model': 9.054237889298714e-05, 'epoch': 0.44})
+[2026-07-28 11:33:26,852][starVLA.training.train_starvla][INFO] - Step 2398, Loss: {'train/loss': 0.04774351790547371, 'train/grad_norm_pre_clip': 1.3901796340942383, 'timing/data': 0.00035311607643961906, 'timing/model': 0.4231782560236752, 'learning_rate/qwen_vl_interface': 9.05194543479942e-06, 'learning_rate/action_model': 9.05194543479942e-05, 'epoch': 0.44})
+[2026-07-28 11:33:33,109][starVLA.training.train_starvla][INFO] - Step 2399, Loss: {'train/loss': 0.02365807816386223, 'train/grad_norm_pre_clip': 0.926590621471405, 'timing/data': 0.0003957711160182953, 'timing/model': 0.422010516282171, 'learning_rate/qwen_vl_interface': 9.049650513259007e-06, 'learning_rate/action_model': 9.049650513259008e-05, 'epoch': 0.44})
+[2026-07-28 11:33:39,356][starVLA.training.train_starvla][INFO] - Step 2400, Loss: {'train/loss': 0.05772650986909866, 'train/grad_norm_pre_clip': 1.4155398607254028, 'timing/data': 0.0002958257682621479, 'timing/model': 0.427093377802521, 'learning_rate/qwen_vl_interface': 9.047353126166632e-06, 'learning_rate/action_model': 9.047353126166632e-05, 'epoch': 0.44})
+[2026-07-28 11:33:45,377][starVLA.training.train_starvla][INFO] - Step 2401, Loss: {'train/loss': 0.04546641930937767, 'train/grad_norm_pre_clip': 2.742126941680908, 'timing/data': 0.0003572809509932995, 'timing/model': 0.42356179282069206, 'learning_rate/qwen_vl_interface': 9.045053275013038e-06, 'learning_rate/action_model': 9.045053275013038e-05, 'epoch': 0.44})
+[2026-07-28 11:33:51,165][starVLA.training.train_starvla][INFO] - Step 2402, Loss: {'train/loss': 0.05702507495880127, 'train/grad_norm_pre_clip': 0.8835076093673706, 'timing/data': 0.0003002420999109745, 'timing/model': 0.4231680282391608, 'learning_rate/qwen_vl_interface': 9.042750961290578e-06, 'learning_rate/action_model': 9.042750961290578e-05, 'epoch': 0.44})
+[2026-07-28 11:33:57,221][starVLA.training.train_starvla][INFO] - Step 2403, Loss: {'train/loss': 0.03896665573120117, 'train/grad_norm_pre_clip': 2.0916025638580322, 'timing/data': 0.00028788531199097633, 'timing/model': 0.42419293290004134, 'learning_rate/qwen_vl_interface': 9.040446186493195e-06, 'learning_rate/action_model': 9.040446186493194e-05, 'epoch': 0.44})
+[2026-07-28 11:34:03,535][starVLA.training.train_starvla][INFO] - Step 2404, Loss: {'train/loss': 0.06727451831102371, 'train/grad_norm_pre_clip': 1.2636470794677734, 'timing/data': 0.00029558269307017326, 'timing/model': 0.41991380183026195, 'learning_rate/qwen_vl_interface': 9.038138952116435e-06, 'learning_rate/action_model': 9.038138952116435e-05, 'epoch': 0.44})
+[2026-07-28 11:34:09,843][starVLA.training.train_starvla][INFO] - Step 2405, Loss: {'train/loss': 0.04066592454910278, 'train/grad_norm_pre_clip': 0.8926246166229248, 'timing/data': 0.00033922214061021805, 'timing/model': 0.4201479749754071, 'learning_rate/qwen_vl_interface': 9.035829259657437e-06, 'learning_rate/action_model': 9.035829259657436e-05, 'epoch': 0.44})
+[2026-07-28 11:34:15,867][starVLA.training.train_starvla][INFO] - Step 2406, Loss: {'train/loss': 0.043458886444568634, 'train/grad_norm_pre_clip': 1.5436936616897583, 'timing/data': 0.0003788466565310955, 'timing/model': 0.4236142998561263, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.44})
+[2026-07-28 11:34:22,497][starVLA.training.train_starvla][INFO] - Step 2407, Loss: {'train/loss': 0.059150390326976776, 'train/grad_norm_pre_clip': 1.7831863164901733, 'timing/data': 0.00035033514723181725, 'timing/model': 0.7750501977279782, 'learning_rate/qwen_vl_interface': 9.031202506489254e-06, 'learning_rate/action_model': 9.031202506489253e-05, 'epoch': 0.44})
+[2026-07-28 11:34:28,894][starVLA.training.train_starvla][INFO] - Step 2408, Loss: {'train/loss': 0.045359473675489426, 'train/grad_norm_pre_clip': 1.46902334690094, 'timing/data': 0.0002888222225010395, 'timing/model': 0.43359525315463543, 'learning_rate/qwen_vl_interface': 9.02888544878232e-06, 'learning_rate/action_model': 9.028885448782319e-05, 'epoch': 0.44})
+[2026-07-28 11:34:35,288][starVLA.training.train_starvla][INFO] - Step 2409, Loss: {'train/loss': 0.06678112596273422, 'train/grad_norm_pre_clip': 2.711555004119873, 'timing/data': 0.00031794095411896706, 'timing/model': 0.4338272390887141, 'learning_rate/qwen_vl_interface': 9.026565938997644e-06, 'learning_rate/action_model': 9.026565938997645e-05, 'epoch': 0.44})
+[2026-07-28 11:34:41,768][starVLA.training.train_starvla][INFO] - Step 2410, Loss: {'train/loss': 0.040180277079343796, 'train/grad_norm_pre_clip': 0.9927855730056763, 'timing/data': 0.00031133508309721947, 'timing/model': 0.43206810392439365, 'learning_rate/qwen_vl_interface': 9.024243978640334e-06, 'learning_rate/action_model': 9.024243978640333e-05, 'epoch': 0.44})
+[2026-07-28 11:34:47,917][starVLA.training.train_starvla][INFO] - Step 2411, Loss: {'train/loss': 0.04543609172105789, 'train/grad_norm_pre_clip': 1.454954743385315, 'timing/data': 0.00039588427171111107, 'timing/model': 0.42903850693255663, 'learning_rate/qwen_vl_interface': 9.02191956921708e-06, 'learning_rate/action_model': 9.02191956921708e-05, 'epoch': 0.44})
+[2026-07-28 11:34:54,301][starVLA.training.train_starvla][INFO] - Step 2412, Loss: {'train/loss': 0.04004399850964546, 'train/grad_norm_pre_clip': 0.985685408115387, 'timing/data': 0.0003456082195043564, 'timing/model': 0.42211753921583295, 'learning_rate/qwen_vl_interface': 9.019592712236172e-06, 'learning_rate/action_model': 9.019592712236171e-05, 'epoch': 0.44})
+[2026-07-28 11:35:00,385][starVLA.training.train_starvla][INFO] - Step 2413, Loss: {'train/loss': 0.07522013783454895, 'train/grad_norm_pre_clip': 2.8069605827331543, 'timing/data': 0.0003850311040878296, 'timing/model': 0.42501628724858165, 'learning_rate/qwen_vl_interface': 9.017263409207478e-06, 'learning_rate/action_model': 9.017263409207478e-05, 'epoch': 0.44})
+[2026-07-28 11:35:06,790][starVLA.training.train_starvla][INFO] - Step 2414, Loss: {'train/loss': 0.037593092769384384, 'train/grad_norm_pre_clip': 0.8990358710289001, 'timing/data': 0.0002600299194455147, 'timing/model': 0.42037621326744556, 'learning_rate/qwen_vl_interface': 9.014931661642455e-06, 'learning_rate/action_model': 9.014931661642455e-05, 'epoch': 0.44})
+[2026-07-28 11:35:12,883][starVLA.training.train_starvla][INFO] - Step 2415, Loss: {'train/loss': 0.028277665376663208, 'train/grad_norm_pre_clip': 0.5425125956535339, 'timing/data': 0.0002928599715232849, 'timing/model': 0.4267976079136133, 'learning_rate/qwen_vl_interface': 9.012597471054154e-06, 'learning_rate/action_model': 9.012597471054154e-05, 'epoch': 0.44})
+[2026-07-28 11:35:19,071][starVLA.training.train_starvla][INFO] - Step 2416, Loss: {'train/loss': 0.041364721953868866, 'train/grad_norm_pre_clip': 0.9529377222061157, 'timing/data': 0.0002879118546843529, 'timing/model': 0.4273419179953635, 'learning_rate/qwen_vl_interface': 9.010260838957206e-06, 'learning_rate/action_model': 9.010260838957205e-05, 'epoch': 0.44})
+[2026-07-28 11:35:25,595][starVLA.training.train_starvla][INFO] - Step 2417, Loss: {'train/loss': 0.05447400361299515, 'train/grad_norm_pre_clip': 1.2188453674316406, 'timing/data': 0.00029480503872036934, 'timing/model': 0.6627581380307674, 'learning_rate/qwen_vl_interface': 9.007921766867823e-06, 'learning_rate/action_model': 9.007921766867822e-05, 'epoch': 0.44})
+[2026-07-28 11:35:31,637][starVLA.training.train_starvla][INFO] - Step 2418, Loss: {'train/loss': 0.05745358765125275, 'train/grad_norm_pre_clip': 1.8838387727737427, 'timing/data': 0.0003847377374768257, 'timing/model': 0.42362123914062977, 'learning_rate/qwen_vl_interface': 9.005580256303804e-06, 'learning_rate/action_model': 9.005580256303803e-05, 'epoch': 0.44})
+[2026-07-28 11:35:37,640][starVLA.training.train_starvla][INFO] - Step 2419, Loss: {'train/loss': 0.03319296985864639, 'train/grad_norm_pre_clip': 1.0312119722366333, 'timing/data': 0.00027757976204156876, 'timing/model': 0.4252726030535996, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.44})
+[2026-07-28 11:35:43,664][starVLA.training.train_starvla][INFO] - Step 2420, Loss: {'train/loss': 0.043254222720861435, 'train/grad_norm_pre_clip': 1.375590443611145, 'timing/data': 0.00030018994584679604, 'timing/model': 0.4234679979272187, 'learning_rate/qwen_vl_interface': 9.000889925830964e-06, 'learning_rate/action_model': 9.000889925830965e-05, 'epoch': 0.44})
+[2026-07-28 11:35:49,784][starVLA.training.train_starvla][INFO] - Step 2421, Loss: {'train/loss': 0.05893498659133911, 'train/grad_norm_pre_clip': 1.5878182649612427, 'timing/data': 0.00029131677001714706, 'timing/model': 0.6759900688193738, 'learning_rate/qwen_vl_interface': 8.998541108965648e-06, 'learning_rate/action_model': 8.998541108965648e-05, 'epoch': 0.44})
+[2026-07-28 11:35:56,083][starVLA.training.train_starvla][INFO] - Step 2422, Loss: {'train/loss': 0.0455707423388958, 'train/grad_norm_pre_clip': 1.7122315168380737, 'timing/data': 0.0003440449945628643, 'timing/model': 0.4295435743406415, 'learning_rate/qwen_vl_interface': 8.996189859712704e-06, 'learning_rate/action_model': 8.996189859712703e-05, 'epoch': 0.44})
+[2026-07-28 11:36:02,122][starVLA.training.train_starvla][INFO] - Step 2423, Loss: {'train/loss': 0.06609416007995605, 'train/grad_norm_pre_clip': 2.21856951713562, 'timing/data': 0.0003014220856130123, 'timing/model': 0.4249123870395124, 'learning_rate/qwen_vl_interface': 8.993836179597833e-06, 'learning_rate/action_model': 8.993836179597833e-05, 'epoch': 0.45})
+[2026-07-28 11:36:08,187][starVLA.training.train_starvla][INFO] - Step 2424, Loss: {'train/loss': 0.053002506494522095, 'train/grad_norm_pre_clip': 1.8523354530334473, 'timing/data': 0.0003019031137228012, 'timing/model': 0.4252221407368779, 'learning_rate/qwen_vl_interface': 8.99148007014831e-06, 'learning_rate/action_model': 8.99148007014831e-05, 'epoch': 0.45})
+[2026-07-28 11:36:14,465][starVLA.training.train_starvla][INFO] - Step 2425, Loss: {'train/loss': 0.01920657977461815, 'train/grad_norm_pre_clip': 0.7562743425369263, 'timing/data': 0.0003062840551137924, 'timing/model': 0.4246116429567337, 'learning_rate/qwen_vl_interface': 8.989121532892989e-06, 'learning_rate/action_model': 8.989121532892988e-05, 'epoch': 0.45})
+[2026-07-28 11:36:20,299][starVLA.training.train_starvla][INFO] - Step 2426, Loss: {'train/loss': 0.018096495419740677, 'train/grad_norm_pre_clip': 1.418067455291748, 'timing/data': 0.00031297188252210617, 'timing/model': 0.4245977089740336, 'learning_rate/qwen_vl_interface': 8.986760569362302e-06, 'learning_rate/action_model': 8.986760569362301e-05, 'epoch': 0.45})
+[2026-07-28 11:36:26,373][starVLA.training.train_starvla][INFO] - Step 2427, Loss: {'train/loss': 0.041891440749168396, 'train/grad_norm_pre_clip': 1.0497469902038574, 'timing/data': 0.00029534008353948593, 'timing/model': 0.425392447039485, 'learning_rate/qwen_vl_interface': 8.984397181088251e-06, 'learning_rate/action_model': 8.984397181088251e-05, 'epoch': 0.45})
+[2026-07-28 11:36:32,373][starVLA.training.train_starvla][INFO] - Step 2428, Loss: {'train/loss': 0.0413152351975441, 'train/grad_norm_pre_clip': 1.6401934623718262, 'timing/data': 0.0002822079695761204, 'timing/model': 0.4232630906626582, 'learning_rate/qwen_vl_interface': 8.982031369604412e-06, 'learning_rate/action_model': 8.982031369604411e-05, 'epoch': 0.45})
+[2026-07-28 11:36:38,836][starVLA.training.train_starvla][INFO] - Step 2429, Loss: {'train/loss': 0.02165241353213787, 'train/grad_norm_pre_clip': 0.630938708782196, 'timing/data': 0.0002758852206170559, 'timing/model': 0.6594562097452581, 'learning_rate/qwen_vl_interface': 8.979663136445936e-06, 'learning_rate/action_model': 8.979663136445935e-05, 'epoch': 0.45})
+[2026-07-28 11:36:44,887][starVLA.training.train_starvla][INFO] - Step 2430, Loss: {'train/loss': 0.05041271448135376, 'train/grad_norm_pre_clip': 1.3199464082717896, 'timing/data': 0.00037411320954561234, 'timing/model': 0.4426787830889225, 'learning_rate/qwen_vl_interface': 8.977292483149542e-06, 'learning_rate/action_model': 8.977292483149542e-05, 'epoch': 0.45})
+[2026-07-28 11:36:51,053][starVLA.training.train_starvla][INFO] - Step 2431, Loss: {'train/loss': 0.05246146768331528, 'train/grad_norm_pre_clip': 1.927483320236206, 'timing/data': 0.00036608800292015076, 'timing/model': 0.42321239598095417, 'learning_rate/qwen_vl_interface': 8.974919411253523e-06, 'learning_rate/action_model': 8.974919411253523e-05, 'epoch': 0.45})
+[2026-07-28 11:36:57,343][starVLA.training.train_starvla][INFO] - Step 2432, Loss: {'train/loss': 0.041130244731903076, 'train/grad_norm_pre_clip': 1.5012742280960083, 'timing/data': 0.00039973994717001915, 'timing/model': 0.4195981412194669, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.45})
+[2026-07-28 11:37:03,340][starVLA.training.train_starvla][INFO] - Step 2433, Loss: {'train/loss': 0.04549267143011093, 'train/grad_norm_pre_clip': 1.7279719114303589, 'timing/data': 0.0002997857518494129, 'timing/model': 0.4231075020506978, 'learning_rate/qwen_vl_interface': 8.970166017823621e-06, 'learning_rate/action_model': 8.97016601782362e-05, 'epoch': 0.45})
+[2026-07-28 11:37:09,363][starVLA.training.train_starvla][INFO] - Step 2434, Loss: {'train/loss': 0.02764991484582424, 'train/grad_norm_pre_clip': 1.240066409111023, 'timing/data': 0.00035377731546759605, 'timing/model': 0.42383407335728407, 'learning_rate/qwen_vl_interface': 8.96778569937416e-06, 'learning_rate/action_model': 8.96778569937416e-05, 'epoch': 0.45})
+[2026-07-28 11:37:15,589][starVLA.training.train_starvla][INFO] - Step 2435, Loss: {'train/loss': 0.052570514380931854, 'train/grad_norm_pre_clip': 1.3123135566711426, 'timing/data': 0.00031229807063937187, 'timing/model': 0.4207526966929436, 'learning_rate/qwen_vl_interface': 8.965402968493922e-06, 'learning_rate/action_model': 8.965402968493922e-05, 'epoch': 0.45})
+[2026-07-28 11:37:21,854][starVLA.training.train_starvla][INFO] - Step 2436, Loss: {'train/loss': 0.05681613087654114, 'train/grad_norm_pre_clip': 2.4422502517700195, 'timing/data': 0.0003113136626780033, 'timing/model': 0.6653143889270723, 'learning_rate/qwen_vl_interface': 8.96301782672904e-06, 'learning_rate/action_model': 8.963017826729038e-05, 'epoch': 0.45})
+[2026-07-28 11:37:27,869][starVLA.training.train_starvla][INFO] - Step 2437, Loss: {'train/loss': 0.03587682545185089, 'train/grad_norm_pre_clip': 1.2844346761703491, 'timing/data': 0.00029456615447998047, 'timing/model': 0.42440292425453663, 'learning_rate/qwen_vl_interface': 8.960630275627197e-06, 'learning_rate/action_model': 8.960630275627197e-05, 'epoch': 0.45})
+[2026-07-28 11:37:34,025][starVLA.training.train_starvla][INFO] - Step 2438, Loss: {'train/loss': 0.03580884635448456, 'train/grad_norm_pre_clip': 1.2068932056427002, 'timing/data': 0.00030402280390262604, 'timing/model': 0.4241074249148369, 'learning_rate/qwen_vl_interface': 8.958240316737655e-06, 'learning_rate/action_model': 8.958240316737656e-05, 'epoch': 0.45})
+[2026-07-28 11:37:40,333][starVLA.training.train_starvla][INFO] - Step 2439, Loss: {'train/loss': 0.07294623553752899, 'train/grad_norm_pre_clip': 2.477994680404663, 'timing/data': 0.00038560666143894196, 'timing/model': 0.422250262927264, 'learning_rate/qwen_vl_interface': 8.955847951611233e-06, 'learning_rate/action_model': 8.955847951611232e-05, 'epoch': 0.45})
+[2026-07-28 11:37:46,342][starVLA.training.train_starvla][INFO] - Step 2440, Loss: {'train/loss': 0.0845041275024414, 'train/grad_norm_pre_clip': 2.0260961055755615, 'timing/data': 0.0002815062180161476, 'timing/model': 0.42523892875760794, 'learning_rate/qwen_vl_interface': 8.953453181800307e-06, 'learning_rate/action_model': 8.953453181800306e-05, 'epoch': 0.45})
+[2026-07-28 11:37:52,612][starVLA.training.train_starvla][INFO] - Step 2441, Loss: {'train/loss': 0.03915979340672493, 'train/grad_norm_pre_clip': 0.7905049324035645, 'timing/data': 0.0002884496934711933, 'timing/model': 0.6666664960794151, 'learning_rate/qwen_vl_interface': 8.951056008858819e-06, 'learning_rate/action_model': 8.95105600885882e-05, 'epoch': 0.45})
+[2026-07-28 11:37:58,834][starVLA.training.train_starvla][INFO] - Step 2442, Loss: {'train/loss': 0.03496210649609566, 'train/grad_norm_pre_clip': 1.4129751920700073, 'timing/data': 0.0002745380625128746, 'timing/model': 0.6492806891910732, 'learning_rate/qwen_vl_interface': 8.94865643434227e-06, 'learning_rate/action_model': 8.948656434342269e-05, 'epoch': 0.45})
+[2026-07-28 11:38:04,619][starVLA.training.train_starvla][INFO] - Step 2443, Loss: {'train/loss': 0.06291848421096802, 'train/grad_norm_pre_clip': 1.727639079093933, 'timing/data': 0.000305855181068182, 'timing/model': 0.4278569929301739, 'learning_rate/qwen_vl_interface': 8.946254459807717e-06, 'learning_rate/action_model': 8.946254459807717e-05, 'epoch': 0.45})
+[2026-07-28 11:38:10,632][starVLA.training.train_starvla][INFO] - Step 2444, Loss: {'train/loss': 0.07507935166358948, 'train/grad_norm_pre_clip': 1.5574299097061157, 'timing/data': 0.00029643718153238297, 'timing/model': 0.4254620219580829, 'learning_rate/qwen_vl_interface': 8.943850086813773e-06, 'learning_rate/action_model': 8.943850086813773e-05, 'epoch': 0.45})
+[2026-07-28 11:38:16,656][starVLA.training.train_starvla][INFO] - Step 2445, Loss: {'train/loss': 0.07357971370220184, 'train/grad_norm_pre_clip': 1.2011092901229858, 'timing/data': 0.0003027091734111309, 'timing/model': 0.42893955018371344, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.45})
+[2026-07-28 11:38:22,727][starVLA.training.train_starvla][INFO] - Step 2446, Loss: {'train/loss': 0.0767749696969986, 'train/grad_norm_pre_clip': 3.635939359664917, 'timing/data': 0.0003675348125398159, 'timing/model': 0.4293891447596252, 'learning_rate/qwen_vl_interface': 8.93903415168996e-06, 'learning_rate/action_model': 8.93903415168996e-05, 'epoch': 0.45})
+[2026-07-28 11:38:29,254][starVLA.training.train_starvla][INFO] - Step 2447, Loss: {'train/loss': 0.03332337737083435, 'train/grad_norm_pre_clip': 1.323874831199646, 'timing/data': 0.0002990942448377609, 'timing/model': 0.6610567993484437, 'learning_rate/qwen_vl_interface': 8.9366225926851e-06, 'learning_rate/action_model': 8.936622592685098e-05, 'epoch': 0.45})
+[2026-07-28 11:38:35,437][starVLA.training.train_starvla][INFO] - Step 2448, Loss: {'train/loss': 0.10147881507873535, 'train/grad_norm_pre_clip': 2.6338675022125244, 'timing/data': 0.0003194729797542095, 'timing/model': 0.43606538185849786, 'learning_rate/qwen_vl_interface': 8.934208641470863e-06, 'learning_rate/action_model': 8.934208641470863e-05, 'epoch': 0.45})
+[2026-07-28 11:38:41,497][starVLA.training.train_starvla][INFO] - Step 2449, Loss: {'train/loss': 0.0628456398844719, 'train/grad_norm_pre_clip': 2.0171165466308594, 'timing/data': 0.0006552892737090588, 'timing/model': 0.4261782681569457, 'learning_rate/qwen_vl_interface': 8.931792299613637e-06, 'learning_rate/action_model': 8.931792299613636e-05, 'epoch': 0.45})
+[2026-07-28 11:38:47,398][starVLA.training.train_starvla][INFO] - Step 2450, Loss: {'train/loss': 0.03846186399459839, 'train/grad_norm_pre_clip': 0.7141152024269104, 'timing/data': 0.0003046928904950619, 'timing/model': 0.4289203891530633, 'learning_rate/qwen_vl_interface': 8.92937356868136e-06, 'learning_rate/action_model': 8.92937356868136e-05, 'epoch': 0.45})
+[2026-07-28 11:38:53,322][starVLA.training.train_starvla][INFO] - Step 2451, Loss: {'train/loss': 0.04077932983636856, 'train/grad_norm_pre_clip': 0.7996580600738525, 'timing/data': 0.0003494131378829479, 'timing/model': 0.43565852101892233, 'learning_rate/qwen_vl_interface': 8.926952450243522e-06, 'learning_rate/action_model': 8.926952450243522e-05, 'epoch': 0.45})
+[2026-07-28 11:38:59,977][starVLA.training.train_starvla][INFO] - Step 2452, Loss: {'train/loss': 0.03163542598485947, 'train/grad_norm_pre_clip': 0.6514078974723816, 'timing/data': 0.0003713849000632763, 'timing/model': 0.6616919627413154, 'learning_rate/qwen_vl_interface': 8.924528945871157e-06, 'learning_rate/action_model': 8.924528945871156e-05, 'epoch': 0.45})
+[2026-07-28 11:39:06,102][starVLA.training.train_starvla][INFO] - Step 2453, Loss: {'train/loss': 0.04903324693441391, 'train/grad_norm_pre_clip': 0.6141575574874878, 'timing/data': 0.00039909128099679947, 'timing/model': 0.4274933161213994, 'learning_rate/qwen_vl_interface': 8.922103057136857e-06, 'learning_rate/action_model': 8.922103057136856e-05, 'epoch': 0.45})
+[2026-07-28 11:39:12,252][starVLA.training.train_starvla][INFO] - Step 2454, Loss: {'train/loss': 0.03237432613968849, 'train/grad_norm_pre_clip': 0.7703694701194763, 'timing/data': 0.0002791536971926689, 'timing/model': 0.43057823134586215, 'learning_rate/qwen_vl_interface': 8.919674785614747e-06, 'learning_rate/action_model': 8.919674785614747e-05, 'epoch': 0.45})
+[2026-07-28 11:39:18,282][starVLA.training.train_starvla][INFO] - Step 2455, Loss: {'train/loss': 0.0653318390250206, 'train/grad_norm_pre_clip': 1.0563080310821533, 'timing/data': 0.0002836771309375763, 'timing/model': 0.4251871230080724, 'learning_rate/qwen_vl_interface': 8.91724413288051e-06, 'learning_rate/action_model': 8.91724413288051e-05, 'epoch': 0.45})
+[2026-07-28 11:39:24,581][starVLA.training.train_starvla][INFO] - Step 2456, Loss: {'train/loss': 0.04709343612194061, 'train/grad_norm_pre_clip': 0.9792847633361816, 'timing/data': 0.0005164826288819313, 'timing/model': 0.4228684906847775, 'learning_rate/qwen_vl_interface': 8.91481110051137e-06, 'learning_rate/action_model': 8.91481110051137e-05, 'epoch': 0.45})
+[2026-07-28 11:39:30,931][starVLA.training.train_starvla][INFO] - Step 2457, Loss: {'train/loss': 0.01981549896299839, 'train/grad_norm_pre_clip': 0.7322392463684082, 'timing/data': 0.00031236233189702034, 'timing/model': 0.4241023799404502, 'learning_rate/qwen_vl_interface': 8.912375690086097e-06, 'learning_rate/action_model': 8.912375690086096e-05, 'epoch': 0.45})
+[2026-07-28 11:39:37,029][starVLA.training.train_starvla][INFO] - Step 2458, Loss: {'train/loss': 0.04000119864940643, 'train/grad_norm_pre_clip': 0.9809094667434692, 'timing/data': 0.0002808370627462864, 'timing/model': 0.42121588066220284, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.45})
+[2026-07-28 11:39:42,905][starVLA.training.train_starvla][INFO] - Step 2459, Loss: {'train/loss': 0.02564859576523304, 'train/grad_norm_pre_clip': 0.9262771606445312, 'timing/data': 0.000463136937469244, 'timing/model': 0.4545206492766738, 'learning_rate/qwen_vl_interface': 8.907497741389932e-06, 'learning_rate/action_model': 8.907497741389932e-05, 'epoch': 0.45})
+[2026-07-28 11:39:49,069][starVLA.training.train_starvla][INFO] - Step 2460, Loss: {'train/loss': 0.028338849544525146, 'train/grad_norm_pre_clip': 0.9700018763542175, 'timing/data': 0.0004576016217470169, 'timing/model': 0.42174790520220995, 'learning_rate/qwen_vl_interface': 8.905055206284289e-06, 'learning_rate/action_model': 8.905055206284288e-05, 'epoch': 0.45})
+[2026-07-28 11:39:55,380][starVLA.training.train_starvla][INFO] - Step 2461, Loss: {'train/loss': 0.037611223757267, 'train/grad_norm_pre_clip': 0.8687759637832642, 'timing/data': 0.0003776359371840954, 'timing/model': 0.6841066689230502, 'learning_rate/qwen_vl_interface': 8.902610299453005e-06, 'learning_rate/action_model': 8.902610299453004e-05, 'epoch': 0.45})
+[2026-07-28 11:40:01,438][starVLA.training.train_starvla][INFO] - Step 2462, Loss: {'train/loss': 0.05420566722750664, 'train/grad_norm_pre_clip': 1.0680898427963257, 'timing/data': 0.0003131008706986904, 'timing/model': 0.4217271599918604, 'learning_rate/qwen_vl_interface': 8.900163022482554e-06, 'learning_rate/action_model': 8.900163022482553e-05, 'epoch': 0.45})
+[2026-07-28 11:40:07,625][starVLA.training.train_starvla][INFO] - Step 2463, Loss: {'train/loss': 0.034110840409994125, 'train/grad_norm_pre_clip': 1.145790457725525, 'timing/data': 0.000273772981017828, 'timing/model': 0.6003363882191479, 'learning_rate/qwen_vl_interface': 8.897713376960946e-06, 'learning_rate/action_model': 8.897713376960946e-05, 'epoch': 0.45})
+[2026-07-28 11:40:13,649][starVLA.training.train_starvla][INFO] - Step 2464, Loss: {'train/loss': 0.033748745918273926, 'train/grad_norm_pre_clip': 0.8048017621040344, 'timing/data': 0.0003074025735259056, 'timing/model': 0.43461857503280044, 'learning_rate/qwen_vl_interface': 8.895261364477731e-06, 'learning_rate/action_model': 8.895261364477731e-05, 'epoch': 0.45})
+[2026-07-28 11:40:19,673][starVLA.training.train_starvla][INFO] - Step 2465, Loss: {'train/loss': 0.026462361216545105, 'train/grad_norm_pre_clip': 1.0423128604888916, 'timing/data': 0.0004301457665860653, 'timing/model': 0.4234078018926084, 'learning_rate/qwen_vl_interface': 8.892806986623993e-06, 'learning_rate/action_model': 8.892806986623992e-05, 'epoch': 0.45})
+[2026-07-28 11:40:25,763][starVLA.training.train_starvla][INFO] - Step 2466, Loss: {'train/loss': 0.06548953801393509, 'train/grad_norm_pre_clip': 2.0798251628875732, 'timing/data': 0.0003628488630056381, 'timing/model': 0.42319145519286394, 'learning_rate/qwen_vl_interface': 8.890350244992349e-06, 'learning_rate/action_model': 8.890350244992348e-05, 'epoch': 0.45})
+[2026-07-28 11:40:32,017][starVLA.training.train_starvla][INFO] - Step 2467, Loss: {'train/loss': 0.03389152139425278, 'train/grad_norm_pre_clip': 1.2424359321594238, 'timing/data': 0.0003390917554497719, 'timing/model': 0.42060460709035397, 'learning_rate/qwen_vl_interface': 8.887891141176954e-06, 'learning_rate/action_model': 8.887891141176954e-05, 'epoch': 0.45})
+[2026-07-28 11:40:38,309][starVLA.training.train_starvla][INFO] - Step 2468, Loss: {'train/loss': 0.15622907876968384, 'train/grad_norm_pre_clip': 2.3424103260040283, 'timing/data': 0.00027698418125510216, 'timing/model': 0.41960202204063535, 'learning_rate/qwen_vl_interface': 8.885429676773495e-06, 'learning_rate/action_model': 8.885429676773495e-05, 'epoch': 0.45})
+[2026-07-28 11:40:44,344][starVLA.training.train_starvla][INFO] - Step 2469, Loss: {'train/loss': 0.0455947145819664, 'train/grad_norm_pre_clip': 2.286196231842041, 'timing/data': 0.0011196630075573921, 'timing/model': 0.42717460496351123, 'learning_rate/qwen_vl_interface': 8.882965853379188e-06, 'learning_rate/action_model': 8.882965853379187e-05, 'epoch': 0.45})
+[2026-07-28 11:40:50,735][starVLA.training.train_starvla][INFO] - Step 2470, Loss: {'train/loss': 0.04723551869392395, 'train/grad_norm_pre_clip': 2.053400993347168, 'timing/data': 0.0002976441755890846, 'timing/model': 0.6887392308562994, 'learning_rate/qwen_vl_interface': 8.88049967259278e-06, 'learning_rate/action_model': 8.880499672592778e-05, 'epoch': 0.45})
+[2026-07-28 11:40:56,813][starVLA.training.train_starvla][INFO] - Step 2471, Loss: {'train/loss': 0.033558838069438934, 'train/grad_norm_pre_clip': 2.2039597034454346, 'timing/data': 0.00036182813346385956, 'timing/model': 0.4313734988681972, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.45})
+[2026-07-28 11:41:03,158][starVLA.training.train_starvla][INFO] - Step 2472, Loss: {'train/loss': 0.039585988968610764, 'train/grad_norm_pre_clip': 1.5078412294387817, 'timing/data': 0.00038767140358686447, 'timing/model': 0.6663534957915545, 'learning_rate/qwen_vl_interface': 8.875560245246307e-06, 'learning_rate/action_model': 8.875560245246307e-05, 'epoch': 0.45})
+[2026-07-28 11:41:09,213][starVLA.training.train_starvla][INFO] - Step 2473, Loss: {'train/loss': 0.0700790286064148, 'train/grad_norm_pre_clip': 2.1247432231903076, 'timing/data': 0.00039363978430628777, 'timing/model': 0.42905200319364667, 'learning_rate/qwen_vl_interface': 8.873087001891381e-06, 'learning_rate/action_model': 8.87308700189138e-05, 'epoch': 0.45})
+[2026-07-28 11:41:15,316][starVLA.training.train_starvla][INFO] - Step 2474, Loss: {'train/loss': 0.03670814260840416, 'train/grad_norm_pre_clip': 1.4898029565811157, 'timing/data': 0.00029009999707341194, 'timing/model': 0.4270071960054338, 'learning_rate/qwen_vl_interface': 8.870611407554635e-06, 'learning_rate/action_model': 8.870611407554636e-05, 'epoch': 0.45})
+[2026-07-28 11:41:21,698][starVLA.training.train_starvla][INFO] - Step 2475, Loss: {'train/loss': 0.05582105368375778, 'train/grad_norm_pre_clip': 0.9943280220031738, 'timing/data': 0.0002690940164029598, 'timing/model': 0.4209640100598335, 'learning_rate/qwen_vl_interface': 8.868133463842456e-06, 'learning_rate/action_model': 8.868133463842457e-05, 'epoch': 0.45})
+[2026-07-28 11:41:27,940][starVLA.training.train_starvla][INFO] - Step 2476, Loss: {'train/loss': 0.056167252361774445, 'train/grad_norm_pre_clip': 1.6947320699691772, 'timing/data': 0.00045008398592472076, 'timing/model': 0.42263722606003284, 'learning_rate/qwen_vl_interface': 8.865653172362755e-06, 'learning_rate/action_model': 8.865653172362754e-05, 'epoch': 0.45})
+[2026-07-28 11:41:34,032][starVLA.training.train_starvla][INFO] - Step 2477, Loss: {'train/loss': 0.037201881408691406, 'train/grad_norm_pre_clip': 0.7482700347900391, 'timing/data': 0.0002971598878502846, 'timing/model': 0.42310407431796193, 'learning_rate/qwen_vl_interface': 8.863170534724962e-06, 'learning_rate/action_model': 8.863170534724961e-05, 'epoch': 0.45})
+[2026-07-28 11:41:40,175][starVLA.training.train_starvla][INFO] - Step 2478, Loss: {'train/loss': 0.058194831013679504, 'train/grad_norm_pre_clip': 2.6749284267425537, 'timing/data': 0.00030768197029829025, 'timing/model': 0.4227620651945472, 'learning_rate/qwen_vl_interface': 8.860685552540038e-06, 'learning_rate/action_model': 8.860685552540037e-05, 'epoch': 0.46})
+[2026-07-28 11:41:45,960][starVLA.training.train_starvla][INFO] - Step 2479, Loss: {'train/loss': 0.02053053118288517, 'train/grad_norm_pre_clip': 0.6414200067520142, 'timing/data': 0.00035097356885671616, 'timing/model': 0.42251727590337396, 'learning_rate/qwen_vl_interface': 8.858198227420461e-06, 'learning_rate/action_model': 8.85819822742046e-05, 'epoch': 0.46})
+[2026-07-28 11:41:52,578][starVLA.training.train_starvla][INFO] - Step 2480, Loss: {'train/loss': 0.04260384663939476, 'train/grad_norm_pre_clip': 1.1425710916519165, 'timing/data': 0.00028251809999346733, 'timing/model': 0.7139737191610038, 'learning_rate/qwen_vl_interface': 8.855708560980225e-06, 'learning_rate/action_model': 8.855708560980225e-05, 'epoch': 0.46})
+[2026-07-28 11:41:58,667][starVLA.training.train_starvla][INFO] - Step 2481, Loss: {'train/loss': 0.023192942142486572, 'train/grad_norm_pre_clip': 0.6984266638755798, 'timing/data': 0.00031692907214164734, 'timing/model': 0.42228781897574663, 'learning_rate/qwen_vl_interface': 8.853216554834852e-06, 'learning_rate/action_model': 8.85321655483485e-05, 'epoch': 0.46})
+[2026-07-28 11:42:04,802][starVLA.training.train_starvla][INFO] - Step 2482, Loss: {'train/loss': 0.06259384751319885, 'train/grad_norm_pre_clip': 2.8618645668029785, 'timing/data': 0.0003716382198035717, 'timing/model': 0.4289626111276448, 'learning_rate/qwen_vl_interface': 8.850722210601376e-06, 'learning_rate/action_model': 8.850722210601377e-05, 'epoch': 0.46})
+[2026-07-28 11:42:11,249][starVLA.training.train_starvla][INFO] - Step 2483, Loss: {'train/loss': 0.06860317289829254, 'train/grad_norm_pre_clip': 2.021876335144043, 'timing/data': 0.0002864580601453781, 'timing/model': 0.6836766940541565, 'learning_rate/qwen_vl_interface': 8.848225529898351e-06, 'learning_rate/action_model': 8.84822552989835e-05, 'epoch': 0.46})
+[2026-07-28 11:42:17,392][starVLA.training.train_starvla][INFO] - Step 2484, Loss: {'train/loss': 0.05127275735139847, 'train/grad_norm_pre_clip': 1.4544767141342163, 'timing/data': 0.0004054899327456951, 'timing/model': 0.42115206783637404, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.46})
+[2026-07-28 11:42:23,679][starVLA.training.train_starvla][INFO] - Step 2485, Loss: {'train/loss': 0.029173044487833977, 'train/grad_norm_pre_clip': 1.1337578296661377, 'timing/data': 0.00029049068689346313, 'timing/model': 0.42551187705248594, 'learning_rate/qwen_vl_interface': 8.84322516556544e-06, 'learning_rate/action_model': 8.84322516556544e-05, 'epoch': 0.46})
+[2026-07-28 11:42:29,745][starVLA.training.train_starvla][INFO] - Step 2486, Loss: {'train/loss': 0.06659676134586334, 'train/grad_norm_pre_clip': 2.310529947280884, 'timing/data': 0.00038076983764767647, 'timing/model': 0.42245856998488307, 'learning_rate/qwen_vl_interface': 8.840721485180237e-06, 'learning_rate/action_model': 8.840721485180236e-05, 'epoch': 0.46})
+[2026-07-28 11:42:35,738][starVLA.training.train_starvla][INFO] - Step 2487, Loss: {'train/loss': 0.05373413860797882, 'train/grad_norm_pre_clip': 1.6762405633926392, 'timing/data': 0.00027683423832058907, 'timing/model': 0.4233655841089785, 'learning_rate/qwen_vl_interface': 8.838215474814848e-06, 'learning_rate/action_model': 8.838215474814847e-05, 'epoch': 0.46})
+[2026-07-28 11:42:42,386][starVLA.training.train_starvla][INFO] - Step 2488, Loss: {'train/loss': 0.051916077733039856, 'train/grad_norm_pre_clip': 2.0909297466278076, 'timing/data': 0.0004032091237604618, 'timing/model': 0.7214707750827074, 'learning_rate/qwen_vl_interface': 8.835707136095392e-06, 'learning_rate/action_model': 8.835707136095391e-05, 'epoch': 0.46})
+[2026-07-28 11:42:48,548][starVLA.training.train_starvla][INFO] - Step 2489, Loss: {'train/loss': 0.04397940635681152, 'train/grad_norm_pre_clip': 1.3899495601654053, 'timing/data': 0.00040555186569690704, 'timing/model': 0.42708835611119866, 'learning_rate/qwen_vl_interface': 8.833196470649505e-06, 'learning_rate/action_model': 8.833196470649505e-05, 'epoch': 0.46})
+[2026-07-28 11:42:54,703][starVLA.training.train_starvla][INFO] - Step 2490, Loss: {'train/loss': 0.051690131425857544, 'train/grad_norm_pre_clip': 1.512688398361206, 'timing/data': 0.0003008069470524788, 'timing/model': 0.424137304071337, 'learning_rate/qwen_vl_interface': 8.83068348010633e-06, 'learning_rate/action_model': 8.83068348010633e-05, 'epoch': 0.46})
+[2026-07-28 11:43:00,898][starVLA.training.train_starvla][INFO] - Step 2491, Loss: {'train/loss': 0.06586995720863342, 'train/grad_norm_pre_clip': 1.7913151979446411, 'timing/data': 0.0003085578791797161, 'timing/model': 0.5971694779582322, 'learning_rate/qwen_vl_interface': 8.82816816609652e-06, 'learning_rate/action_model': 8.828168166096521e-05, 'epoch': 0.46})
+[2026-07-28 11:43:07,062][starVLA.training.train_starvla][INFO] - Step 2492, Loss: {'train/loss': 0.039168015122413635, 'train/grad_norm_pre_clip': 1.0904440879821777, 'timing/data': 0.00037092482671141624, 'timing/model': 0.42325562378391623, 'learning_rate/qwen_vl_interface': 8.825650530252238e-06, 'learning_rate/action_model': 8.825650530252238e-05, 'epoch': 0.46})
+[2026-07-28 11:43:13,089][starVLA.training.train_starvla][INFO] - Step 2493, Loss: {'train/loss': 0.02950749360024929, 'train/grad_norm_pre_clip': 0.814272940158844, 'timing/data': 0.00031022215262055397, 'timing/model': 0.4292847909964621, 'learning_rate/qwen_vl_interface': 8.823130574207144e-06, 'learning_rate/action_model': 8.823130574207143e-05, 'epoch': 0.46})
+[2026-07-28 11:43:19,444][starVLA.training.train_starvla][INFO] - Step 2494, Loss: {'train/loss': 0.05501879006624222, 'train/grad_norm_pre_clip': 1.0778679847717285, 'timing/data': 0.0003694361075758934, 'timing/model': 0.7199732949957252, 'learning_rate/qwen_vl_interface': 8.82060829959641e-06, 'learning_rate/action_model': 8.820608299596411e-05, 'epoch': 0.46})
+[2026-07-28 11:43:25,503][starVLA.training.train_starvla][INFO] - Step 2495, Loss: {'train/loss': 0.044658929109573364, 'train/grad_norm_pre_clip': 0.8825191855430603, 'timing/data': 0.00035566603764891624, 'timing/model': 0.42353323893621564, 'learning_rate/qwen_vl_interface': 8.81808370805672e-06, 'learning_rate/action_model': 8.818083708056721e-05, 'epoch': 0.46})
+[2026-07-28 11:43:31,600][starVLA.training.train_starvla][INFO] - Step 2496, Loss: {'train/loss': 0.06618533283472061, 'train/grad_norm_pre_clip': 2.6608898639678955, 'timing/data': 0.000289407093077898, 'timing/model': 0.42265927605330944, 'learning_rate/qwen_vl_interface': 8.815556801226251e-06, 'learning_rate/action_model': 8.815556801226251e-05, 'epoch': 0.46})
+[2026-07-28 11:43:37,362][starVLA.training.train_starvla][INFO] - Step 2497, Loss: {'train/loss': 0.04257307946681976, 'train/grad_norm_pre_clip': 1.3376188278198242, 'timing/data': 0.0003717769868671894, 'timing/model': 0.42165627190843225, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.46})
+[2026-07-28 11:43:43,625][starVLA.training.train_starvla][INFO] - Step 2498, Loss: {'train/loss': 0.03920412063598633, 'train/grad_norm_pre_clip': 1.8110144138336182, 'timing/data': 0.0003151828423142433, 'timing/model': 0.6803049389272928, 'learning_rate/qwen_vl_interface': 8.810496048253203e-06, 'learning_rate/action_model': 8.810496048253202e-05, 'epoch': 0.46})
+[2026-07-28 11:43:49,963][starVLA.training.train_starvla][INFO] - Step 2499, Loss: {'train/loss': 0.04776393994688988, 'train/grad_norm_pre_clip': 1.1845476627349854, 'timing/data': 0.0004586828872561455, 'timing/model': 0.4221297479234636, 'learning_rate/qwen_vl_interface': 8.80796220539449e-06, 'learning_rate/action_model': 8.80796220539449e-05, 'epoch': 0.46})
+[2026-07-28 11:43:56,065][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2500
+[2026-07-28 11:44:01,035][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2500
+[2026-07-28 11:51:09,579][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2500
+[2026-07-28 11:51:09,581][starVLA.training.train_starvla][INFO] - Step 2500, Loss: {'train/loss': 0.03452875837683678, 'train/grad_norm_pre_clip': 1.1020092964172363, 'eval/action_loss/seconds': 4.967179197818041, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.04455208778381348, 'eval/latency_3/loss': 0.04455208778381348, 'eval/flappy/loss': 0.04455208778381348, 'eval/flappy/latency_3/loss': 0.04455208778381348, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 275.22000324577095, 'rl_games_eval/mid_train/mean_length': 2236.8, 'rl_games_eval/mid_train/std_reward': 170.87475320509728, 'rl_games_eval/mid_train/std_length': 1376.9918155167081, 'rl_games_eval/mid_train/macro_mean_reward': 275.22000324577095, 'rl_games_eval/mid_train/macro_mean_length': 2236.8, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 275.22000324577095, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2236.8, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 170.87475320509728, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1376.9918155167081, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 275.22000324577095, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0012028589844703674, 'timing/model': 0.42414007103070617, 'learning_rate/qwen_vl_interface': 8.805426053812731e-06, 'learning_rate/action_model': 8.805426053812731e-05, 'epoch': 0.46})
+[2026-07-28 11:51:09,583][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_2500_state
+[2026-07-28 11:51:09,583][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 11:51:46,858][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_2500_state/pytorch_model
+[2026-07-28 11:51:46,859][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_2500_state/sampler.bin
+[2026-07-28 11:51:46,859][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_2500_state/sampler_1.bin
+[2026-07-28 11:51:46,861][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_2500_state/random_states_0.pkl
+[2026-07-28 11:51:56,107][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 2500
+[2026-07-28 11:51:56,108][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 11:51:56,130][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 11:52:02,693][starVLA.training.train_starvla][INFO] - Step 2501, Loss: {'train/loss': 0.06012168899178505, 'train/grad_norm_pre_clip': 1.6819136142730713, 'timing/data': 0.0003890213556587696, 'timing/model': 0.6368050668388605, 'learning_rate/qwen_vl_interface': 8.802887595153607e-06, 'learning_rate/action_model': 8.802887595153606e-05, 'epoch': 0.46})
+[2026-07-28 11:52:08,763][starVLA.training.train_starvla][INFO] - Step 2502, Loss: {'train/loss': 0.043111950159072876, 'train/grad_norm_pre_clip': 1.8830537796020508, 'timing/data': 0.0003936667926609516, 'timing/model': 0.4222569838166237, 'learning_rate/qwen_vl_interface': 8.800346831064294e-06, 'learning_rate/action_model': 8.800346831064294e-05, 'epoch': 0.46})
+[2026-07-28 11:52:14,810][starVLA.training.train_starvla][INFO] - Step 2503, Loss: {'train/loss': 0.08556903153657913, 'train/grad_norm_pre_clip': 1.5640044212341309, 'timing/data': 0.00048318691551685333, 'timing/model': 0.42127696378156543, 'learning_rate/qwen_vl_interface': 8.797803763193471e-06, 'learning_rate/action_model': 8.797803763193471e-05, 'epoch': 0.46})
+[2026-07-28 11:52:20,681][starVLA.training.train_starvla][INFO] - Step 2504, Loss: {'train/loss': 0.04687490314245224, 'train/grad_norm_pre_clip': 1.3497390747070312, 'timing/data': 0.0003839717246592045, 'timing/model': 0.42222866881638765, 'learning_rate/qwen_vl_interface': 8.795258393191302e-06, 'learning_rate/action_model': 8.795258393191301e-05, 'epoch': 0.46})
+[2026-07-28 11:52:26,932][starVLA.training.train_starvla][INFO] - Step 2505, Loss: {'train/loss': 0.041064877063035965, 'train/grad_norm_pre_clip': 1.680142879486084, 'timing/data': 0.00030603911727666855, 'timing/model': 0.4243078879080713, 'learning_rate/qwen_vl_interface': 8.792710722709451e-06, 'learning_rate/action_model': 8.792710722709451e-05, 'epoch': 0.46})
+[2026-07-28 11:52:33,217][starVLA.training.train_starvla][INFO] - Step 2506, Loss: {'train/loss': 0.0659179612994194, 'train/grad_norm_pre_clip': 1.9660166501998901, 'timing/data': 0.00027097389101982117, 'timing/model': 0.4211587831377983, 'learning_rate/qwen_vl_interface': 8.790160753401076e-06, 'learning_rate/action_model': 8.790160753401077e-05, 'epoch': 0.46})
+[2026-07-28 11:52:39,256][starVLA.training.train_starvla][INFO] - Step 2507, Loss: {'train/loss': 0.046467069536447525, 'train/grad_norm_pre_clip': 1.1235824823379517, 'timing/data': 0.00036910269409418106, 'timing/model': 0.43175233667716384, 'learning_rate/qwen_vl_interface': 8.787608486920824e-06, 'learning_rate/action_model': 8.787608486920824e-05, 'epoch': 0.46})
+[2026-07-28 11:52:45,349][starVLA.training.train_starvla][INFO] - Step 2508, Loss: {'train/loss': 0.05755523592233658, 'train/grad_norm_pre_clip': 1.3056541681289673, 'timing/data': 0.0003087129443883896, 'timing/model': 0.42216921504586935, 'learning_rate/qwen_vl_interface': 8.78505392492483e-06, 'learning_rate/action_model': 8.78505392492483e-05, 'epoch': 0.46})
+[2026-07-28 11:52:51,623][starVLA.training.train_starvla][INFO] - Step 2509, Loss: {'train/loss': 0.020398898050189018, 'train/grad_norm_pre_clip': 0.617340087890625, 'timing/data': 0.00031375186517834663, 'timing/model': 0.42172985011711717, 'learning_rate/qwen_vl_interface': 8.782497069070725e-06, 'learning_rate/action_model': 8.782497069070725e-05, 'epoch': 0.46})
+[2026-07-28 11:52:57,947][starVLA.training.train_starvla][INFO] - Step 2510, Loss: {'train/loss': 0.04451367259025574, 'train/grad_norm_pre_clip': 1.6552797555923462, 'timing/data': 0.00030686287209391594, 'timing/model': 0.6837677820585668, 'learning_rate/qwen_vl_interface': 8.779937921017624e-06, 'learning_rate/action_model': 8.779937921017623e-05, 'epoch': 0.46})
+[2026-07-28 11:53:03,971][starVLA.training.train_starvla][INFO] - Step 2511, Loss: {'train/loss': 0.06251931190490723, 'train/grad_norm_pre_clip': 1.9932358264923096, 'timing/data': 0.0002933517098426819, 'timing/model': 0.43282674066722393, 'learning_rate/qwen_vl_interface': 8.777376482426129e-06, 'learning_rate/action_model': 8.777376482426128e-05, 'epoch': 0.46})
+[2026-07-28 11:53:09,972][starVLA.training.train_starvla][INFO] - Step 2512, Loss: {'train/loss': 0.04300898313522339, 'train/grad_norm_pre_clip': 0.7793982028961182, 'timing/data': 0.00029082782566547394, 'timing/model': 0.4240567102096975, 'learning_rate/qwen_vl_interface': 8.774812754958332e-06, 'learning_rate/action_model': 8.774812754958331e-05, 'epoch': 0.46})
+[2026-07-28 11:53:15,973][starVLA.training.train_starvla][INFO] - Step 2513, Loss: {'train/loss': 0.034043844789266586, 'train/grad_norm_pre_clip': 0.7748385071754456, 'timing/data': 0.00038209883496165276, 'timing/model': 0.42058621207252145, 'learning_rate/qwen_vl_interface': 8.772246740277805e-06, 'learning_rate/action_model': 8.772246740277805e-05, 'epoch': 0.46})
+[2026-07-28 11:53:22,458][starVLA.training.train_starvla][INFO] - Step 2514, Loss: {'train/loss': 0.05976714566349983, 'train/grad_norm_pre_clip': 1.194327712059021, 'timing/data': 0.0002546869218349457, 'timing/model': 0.6593713508918881, 'learning_rate/qwen_vl_interface': 8.769678440049608e-06, 'learning_rate/action_model': 8.769678440049608e-05, 'epoch': 0.46})
+[2026-07-28 11:53:28,446][starVLA.training.train_starvla][INFO] - Step 2515, Loss: {'train/loss': 0.07640472054481506, 'train/grad_norm_pre_clip': 1.3994992971420288, 'timing/data': 0.0002697110176086426, 'timing/model': 0.4202901762910187, 'learning_rate/qwen_vl_interface': 8.767107855940287e-06, 'learning_rate/action_model': 8.767107855940286e-05, 'epoch': 0.46})
+[2026-07-28 11:53:34,429][starVLA.training.train_starvla][INFO] - Step 2516, Loss: {'train/loss': 0.035326529294252396, 'train/grad_norm_pre_clip': 0.9310262203216553, 'timing/data': 0.00028899870812892914, 'timing/model': 0.4225836400873959, 'learning_rate/qwen_vl_interface': 8.764534989617861e-06, 'learning_rate/action_model': 8.76453498961786e-05, 'epoch': 0.46})
+[2026-07-28 11:53:40,415][starVLA.training.train_starvla][INFO] - Step 2517, Loss: {'train/loss': 0.0266472939401865, 'train/grad_norm_pre_clip': 1.1317830085754395, 'timing/data': 0.00028963061049580574, 'timing/model': 0.4233369040302932, 'learning_rate/qwen_vl_interface': 8.761959842751837e-06, 'learning_rate/action_model': 8.761959842751837e-05, 'epoch': 0.46})
+[2026-07-28 11:53:46,654][starVLA.training.train_starvla][INFO] - Step 2518, Loss: {'train/loss': 0.023675408214330673, 'train/grad_norm_pre_clip': 0.9584681987762451, 'timing/data': 0.0002777581103146076, 'timing/model': 0.420466439332813, 'learning_rate/qwen_vl_interface': 8.7593824170132e-06, 'learning_rate/action_model': 8.7593824170132e-05, 'epoch': 0.46})
+[2026-07-28 11:53:52,921][starVLA.training.train_starvla][INFO] - Step 2519, Loss: {'train/loss': 0.030766140669584274, 'train/grad_norm_pre_clip': 1.0205669403076172, 'timing/data': 0.0003587030805647373, 'timing/model': 0.4280168451368809, 'learning_rate/qwen_vl_interface': 8.756802714074412e-06, 'learning_rate/action_model': 8.756802714074412e-05, 'epoch': 0.46})
+[2026-07-28 11:53:58,944][starVLA.training.train_starvla][INFO] - Step 2520, Loss: {'train/loss': 0.06166534125804901, 'train/grad_norm_pre_clip': 1.4161185026168823, 'timing/data': 0.00028995517641305923, 'timing/model': 0.42153114080429077, 'learning_rate/qwen_vl_interface': 8.75422073560942e-06, 'learning_rate/action_model': 8.754220735609419e-05, 'epoch': 0.46})
+[2026-07-28 11:54:05,199][starVLA.training.train_starvla][INFO] - Step 2521, Loss: {'train/loss': 0.018954943865537643, 'train/grad_norm_pre_clip': 1.286790132522583, 'timing/data': 0.0003336300142109394, 'timing/model': 0.6645927703939378, 'learning_rate/qwen_vl_interface': 8.751636483293635e-06, 'learning_rate/action_model': 8.751636483293634e-05, 'epoch': 0.46})
+[2026-07-28 11:54:11,437][starVLA.training.train_starvla][INFO] - Step 2522, Loss: {'train/loss': 0.07731649279594421, 'train/grad_norm_pre_clip': 2.3885338306427, 'timing/data': 0.00029265088960528374, 'timing/model': 0.4209502600133419, 'learning_rate/qwen_vl_interface': 8.749049958803952e-06, 'learning_rate/action_model': 8.749049958803951e-05, 'epoch': 0.46})
+[2026-07-28 11:54:17,425][starVLA.training.train_starvla][INFO] - Step 2523, Loss: {'train/loss': 0.048773083835840225, 'train/grad_norm_pre_clip': 2.1470091342926025, 'timing/data': 0.00028653768822550774, 'timing/model': 0.42141695972532034, 'learning_rate/qwen_vl_interface': 8.74646116381874e-06, 'learning_rate/action_model': 8.74646116381874e-05, 'epoch': 0.46})
+[2026-07-28 11:54:23,421][starVLA.training.train_starvla][INFO] - Step 2524, Loss: {'train/loss': 0.023162567988038063, 'train/grad_norm_pre_clip': 1.1946451663970947, 'timing/data': 0.0002902289852499962, 'timing/model': 0.42183906864374876, 'learning_rate/qwen_vl_interface': 8.74387010001784e-06, 'learning_rate/action_model': 8.743870100017838e-05, 'epoch': 0.46})
+[2026-07-28 11:54:29,393][starVLA.training.train_starvla][INFO] - Step 2525, Loss: {'train/loss': 0.07011725753545761, 'train/grad_norm_pre_clip': 2.3992667198181152, 'timing/data': 0.00029081711545586586, 'timing/model': 0.42184783704578876, 'learning_rate/qwen_vl_interface': 8.741276769082562e-06, 'learning_rate/action_model': 8.741276769082562e-05, 'epoch': 0.46})
+[2026-07-28 11:54:35,386][starVLA.training.train_starvla][INFO] - Step 2526, Loss: {'train/loss': 0.05164216458797455, 'train/grad_norm_pre_clip': 1.3115782737731934, 'timing/data': 0.0002957489341497421, 'timing/model': 0.6631654789671302, 'learning_rate/qwen_vl_interface': 8.738681172695694e-06, 'learning_rate/action_model': 8.738681172695694e-05, 'epoch': 0.46})
+[2026-07-28 11:54:41,134][starVLA.training.train_starvla][INFO] - Step 2527, Loss: {'train/loss': 0.050632644444704056, 'train/grad_norm_pre_clip': 1.9352185726165771, 'timing/data': 0.0002804831601679325, 'timing/model': 0.42026784317567945, 'learning_rate/qwen_vl_interface': 8.73608331254149e-06, 'learning_rate/action_model': 8.736083312541489e-05, 'epoch': 0.46})
+[2026-07-28 11:54:46,934][starVLA.training.train_starvla][INFO] - Step 2528, Loss: {'train/loss': 0.05476352199912071, 'train/grad_norm_pre_clip': 1.3473674058914185, 'timing/data': 0.00036592409014701843, 'timing/model': 0.4227954000234604, 'learning_rate/qwen_vl_interface': 8.733483190305671e-06, 'learning_rate/action_model': 8.733483190305671e-05, 'epoch': 0.46})
+[2026-07-28 11:54:52,936][starVLA.training.train_starvla][INFO] - Step 2529, Loss: {'train/loss': 0.04408358782529831, 'train/grad_norm_pre_clip': 2.1591286659240723, 'timing/data': 0.0004274076782166958, 'timing/model': 0.42300936486572027, 'learning_rate/qwen_vl_interface': 8.730880807675431e-06, 'learning_rate/action_model': 8.73088080767543e-05, 'epoch': 0.46})
+[2026-07-28 11:54:58,999][starVLA.training.train_starvla][INFO] - Step 2530, Loss: {'train/loss': 0.04124894738197327, 'train/grad_norm_pre_clip': 1.509520411491394, 'timing/data': 0.0008468129672110081, 'timing/model': 0.42649099603295326, 'learning_rate/qwen_vl_interface': 8.728276166339426e-06, 'learning_rate/action_model': 8.728276166339426e-05, 'epoch': 0.46})
+[2026-07-28 11:55:05,004][starVLA.training.train_starvla][INFO] - Step 2531, Loss: {'train/loss': 0.05204705893993378, 'train/grad_norm_pre_clip': 2.436706066131592, 'timing/data': 0.00029824115335941315, 'timing/model': 0.42601135605946183, 'learning_rate/qwen_vl_interface': 8.725669267987779e-06, 'learning_rate/action_model': 8.725669267987778e-05, 'epoch': 0.46})
+[2026-07-28 11:55:11,486][starVLA.training.train_starvla][INFO] - Step 2532, Loss: {'train/loss': 0.0560523197054863, 'train/grad_norm_pre_clip': 1.5285323858261108, 'timing/data': 0.0002803499810397625, 'timing/model': 0.6802263441495597, 'learning_rate/qwen_vl_interface': 8.723060114312079e-06, 'learning_rate/action_model': 8.723060114312079e-05, 'epoch': 0.47})
+[2026-07-28 11:55:17,494][starVLA.training.train_starvla][INFO] - Step 2533, Loss: {'train/loss': 0.027042414993047714, 'train/grad_norm_pre_clip': 0.845980703830719, 'timing/data': 0.00030001066625118256, 'timing/model': 0.42277257377281785, 'learning_rate/qwen_vl_interface': 8.720448707005377e-06, 'learning_rate/action_model': 8.720448707005376e-05, 'epoch': 0.47})
+[2026-07-28 11:55:23,617][starVLA.training.train_starvla][INFO] - Step 2534, Loss: {'train/loss': 0.03633935749530792, 'train/grad_norm_pre_clip': 1.6232329607009888, 'timing/data': 0.0003086179494857788, 'timing/model': 0.4239451503381133, 'learning_rate/qwen_vl_interface': 8.717835047762188e-06, 'learning_rate/action_model': 8.717835047762187e-05, 'epoch': 0.47})
+[2026-07-28 11:55:29,638][starVLA.training.train_starvla][INFO] - Step 2535, Loss: {'train/loss': 0.044551778584718704, 'train/grad_norm_pre_clip': 1.6293920278549194, 'timing/data': 0.00036921724677085876, 'timing/model': 0.4230730850249529, 'learning_rate/qwen_vl_interface': 8.715219138278486e-06, 'learning_rate/action_model': 8.715219138278486e-05, 'epoch': 0.47})
+[2026-07-28 11:55:35,900][starVLA.training.train_starvla][INFO] - Step 2536, Loss: {'train/loss': 0.05105689913034439, 'train/grad_norm_pre_clip': 2.082146406173706, 'timing/data': 0.00028943363577127457, 'timing/model': 0.42164745833724737, 'learning_rate/qwen_vl_interface': 8.712600980251706e-06, 'learning_rate/action_model': 8.712600980251705e-05, 'epoch': 0.47})
+[2026-07-28 11:55:42,136][starVLA.training.train_starvla][INFO] - Step 2537, Loss: {'train/loss': 0.044663019478321075, 'train/grad_norm_pre_clip': 1.7056057453155518, 'timing/data': 0.00030915113165974617, 'timing/model': 0.4212893471121788, 'learning_rate/qwen_vl_interface': 8.709980575380741e-06, 'learning_rate/action_model': 8.709980575380741e-05, 'epoch': 0.47})
+[2026-07-28 11:55:48,174][starVLA.training.train_starvla][INFO] - Step 2538, Loss: {'train/loss': 0.022572673857212067, 'train/grad_norm_pre_clip': 0.9496048092842102, 'timing/data': 0.00036949897184967995, 'timing/model': 0.42312924610450864, 'learning_rate/qwen_vl_interface': 8.707357925365949e-06, 'learning_rate/action_model': 8.707357925365948e-05, 'epoch': 0.47})
+[2026-07-28 11:55:54,406][starVLA.training.train_starvla][INFO] - Step 2539, Loss: {'train/loss': 0.038771532475948334, 'train/grad_norm_pre_clip': 1.2440526485443115, 'timing/data': 0.00039253104478120804, 'timing/model': 0.42537503503262997, 'learning_rate/qwen_vl_interface': 8.704733031909134e-06, 'learning_rate/action_model': 8.704733031909133e-05, 'epoch': 0.47})
+[2026-07-28 11:56:00,435][starVLA.training.train_starvla][INFO] - Step 2540, Loss: {'train/loss': 0.060692355036735535, 'train/grad_norm_pre_clip': 1.7165143489837646, 'timing/data': 0.00032025808468461037, 'timing/model': 0.4275171421468258, 'learning_rate/qwen_vl_interface': 8.70210589671356e-06, 'learning_rate/action_model': 8.702105896713561e-05, 'epoch': 0.47})
+[2026-07-28 11:56:06,770][starVLA.training.train_starvla][INFO] - Step 2541, Loss: {'train/loss': 0.045804522931575775, 'train/grad_norm_pre_clip': 1.049448013305664, 'timing/data': 0.0003648260608315468, 'timing/model': 0.4234182653017342, 'learning_rate/qwen_vl_interface': 8.699476521483952e-06, 'learning_rate/action_model': 8.699476521483952e-05, 'epoch': 0.47})
+[2026-07-28 11:56:13,032][starVLA.training.train_starvla][INFO] - Step 2542, Loss: {'train/loss': 0.04902813211083412, 'train/grad_norm_pre_clip': 1.8993360996246338, 'timing/data': 0.00031247688457369804, 'timing/model': 0.5986190447583795, 'learning_rate/qwen_vl_interface': 8.696844907926478e-06, 'learning_rate/action_model': 8.696844907926477e-05, 'epoch': 0.47})
+[2026-07-28 11:56:19,089][starVLA.training.train_starvla][INFO] - Step 2543, Loss: {'train/loss': 0.05846552550792694, 'train/grad_norm_pre_clip': 1.4109125137329102, 'timing/data': 0.0003099278546869755, 'timing/model': 0.42816911404952407, 'learning_rate/qwen_vl_interface': 8.694211057748766e-06, 'learning_rate/action_model': 8.694211057748765e-05, 'epoch': 0.47})
+[2026-07-28 11:56:25,146][starVLA.training.train_starvla][INFO] - Step 2544, Loss: {'train/loss': 0.0411352664232254, 'train/grad_norm_pre_clip': 1.3111847639083862, 'timing/data': 0.0004114150069653988, 'timing/model': 0.42663749819621444, 'learning_rate/qwen_vl_interface': 8.69157497265989e-06, 'learning_rate/action_model': 8.69157497265989e-05, 'epoch': 0.47})
+[2026-07-28 11:56:31,178][starVLA.training.train_starvla][INFO] - Step 2545, Loss: {'train/loss': 0.08335912227630615, 'train/grad_norm_pre_clip': 1.3453624248504639, 'timing/data': 0.0003821989521384239, 'timing/model': 0.4227611683309078, 'learning_rate/qwen_vl_interface': 8.68893665437038e-06, 'learning_rate/action_model': 8.688936654370378e-05, 'epoch': 0.47})
+[2026-07-28 11:56:37,469][starVLA.training.train_starvla][INFO] - Step 2546, Loss: {'train/loss': 0.03862939774990082, 'train/grad_norm_pre_clip': 1.1402658224105835, 'timing/data': 0.00039028609171509743, 'timing/model': 0.6681531011126935, 'learning_rate/qwen_vl_interface': 8.68629610459221e-06, 'learning_rate/action_model': 8.686296104592209e-05, 'epoch': 0.47})
+[2026-07-28 11:56:43,460][starVLA.training.train_starvla][INFO] - Step 2547, Loss: {'train/loss': 0.030583560466766357, 'train/grad_norm_pre_clip': 0.6864684224128723, 'timing/data': 0.00028163567185401917, 'timing/model': 0.42180248023942113, 'learning_rate/qwen_vl_interface': 8.683653325038805e-06, 'learning_rate/action_model': 8.683653325038804e-05, 'epoch': 0.47})
+[2026-07-28 11:56:49,473][starVLA.training.train_starvla][INFO] - Step 2548, Loss: {'train/loss': 0.06716008484363556, 'train/grad_norm_pre_clip': 1.5272830724716187, 'timing/data': 0.0003213980235159397, 'timing/model': 0.42367215221747756, 'learning_rate/qwen_vl_interface': 8.681008317425037e-06, 'learning_rate/action_model': 8.681008317425037e-05, 'epoch': 0.47})
+[2026-07-28 11:56:55,474][starVLA.training.train_starvla][INFO] - Step 2549, Loss: {'train/loss': 0.053534362465143204, 'train/grad_norm_pre_clip': 1.5500565767288208, 'timing/data': 0.00037235720083117485, 'timing/model': 0.422451444901526, 'learning_rate/qwen_vl_interface': 8.678361083467221e-06, 'learning_rate/action_model': 8.67836108346722e-05, 'epoch': 0.47})
+[2026-07-28 11:57:01,715][starVLA.training.train_starvla][INFO] - Step 2550, Loss: {'train/loss': 0.03864490985870361, 'train/grad_norm_pre_clip': 0.6387372016906738, 'timing/data': 0.00029548490419983864, 'timing/model': 0.42476840829476714, 'learning_rate/qwen_vl_interface': 8.67571162488312e-06, 'learning_rate/action_model': 8.67571162488312e-05, 'epoch': 0.47})
+[2026-07-28 11:57:07,720][starVLA.training.train_starvla][INFO] - Step 2551, Loss: {'train/loss': 0.042595379054546356, 'train/grad_norm_pre_clip': 0.801651656627655, 'timing/data': 0.00030037015676498413, 'timing/model': 0.4250037488527596, 'learning_rate/qwen_vl_interface': 8.673059943391937e-06, 'learning_rate/action_model': 8.673059943391936e-05, 'epoch': 0.47})
+[2026-07-28 11:57:13,908][starVLA.training.train_starvla][INFO] - Step 2552, Loss: {'train/loss': 0.059354268014431, 'train/grad_norm_pre_clip': 1.4780001640319824, 'timing/data': 0.00030632782727479935, 'timing/model': 0.42861267970874906, 'learning_rate/qwen_vl_interface': 8.670406040714321e-06, 'learning_rate/action_model': 8.670406040714321e-05, 'epoch': 0.47})
+[2026-07-28 11:57:20,137][starVLA.training.train_starvla][INFO] - Step 2553, Loss: {'train/loss': 0.03808189183473587, 'train/grad_norm_pre_clip': 1.1262965202331543, 'timing/data': 0.00036302488297224045, 'timing/model': 0.4228294463828206, 'learning_rate/qwen_vl_interface': 8.667749918572361e-06, 'learning_rate/action_model': 8.667749918572361e-05, 'epoch': 0.47})
+[2026-07-28 11:57:26,228][starVLA.training.train_starvla][INFO] - Step 2554, Loss: {'train/loss': 0.03869389370083809, 'train/grad_norm_pre_clip': 0.7947772741317749, 'timing/data': 0.0003794729709625244, 'timing/model': 0.4223515121266246, 'learning_rate/qwen_vl_interface': 8.665091578689584e-06, 'learning_rate/action_model': 8.665091578689584e-05, 'epoch': 0.47})
+[2026-07-28 11:57:32,498][starVLA.training.train_starvla][INFO] - Step 2555, Loss: {'train/loss': 0.03679969906806946, 'train/grad_norm_pre_clip': 0.9022274017333984, 'timing/data': 0.0004965621046721935, 'timing/model': 0.6703154779970646, 'learning_rate/qwen_vl_interface': 8.662431022790958e-06, 'learning_rate/action_model': 8.662431022790958e-05, 'epoch': 0.47})
+[2026-07-28 11:57:38,503][starVLA.training.train_starvla][INFO] - Step 2556, Loss: {'train/loss': 0.043615106493234634, 'train/grad_norm_pre_clip': 1.33829927444458, 'timing/data': 0.000357125885784626, 'timing/model': 0.4217882198281586, 'learning_rate/qwen_vl_interface': 8.65976825260289e-06, 'learning_rate/action_model': 8.659768252602888e-05, 'epoch': 0.47})
+[2026-07-28 11:57:44,751][starVLA.training.train_starvla][INFO] - Step 2557, Loss: {'train/loss': 0.06009766086935997, 'train/grad_norm_pre_clip': 2.0146372318267822, 'timing/data': 0.0007253140211105347, 'timing/model': 0.6770268608815968, 'learning_rate/qwen_vl_interface': 8.657103269853222e-06, 'learning_rate/action_model': 8.657103269853222e-05, 'epoch': 0.47})
+[2026-07-28 11:57:50,772][starVLA.training.train_starvla][INFO] - Step 2558, Loss: {'train/loss': 0.04062166437506676, 'train/grad_norm_pre_clip': 1.8746168613433838, 'timing/data': 0.0003026360645890236, 'timing/model': 0.42376770498231053, 'learning_rate/qwen_vl_interface': 8.654436076271232e-06, 'learning_rate/action_model': 8.654436076271232e-05, 'epoch': 0.47})
+[2026-07-28 11:57:56,784][starVLA.training.train_starvla][INFO] - Step 2559, Loss: {'train/loss': 0.034410759806632996, 'train/grad_norm_pre_clip': 1.6976826190948486, 'timing/data': 0.00031029898673295975, 'timing/model': 0.4250131929293275, 'learning_rate/qwen_vl_interface': 8.651766673587635e-06, 'learning_rate/action_model': 8.651766673587635e-05, 'epoch': 0.47})
+[2026-07-28 11:58:03,119][starVLA.training.train_starvla][INFO] - Step 2560, Loss: {'train/loss': 0.05155690014362335, 'train/grad_norm_pre_clip': 1.4522384405136108, 'timing/data': 0.0003427388146519661, 'timing/model': 0.46229458041489124, 'learning_rate/qwen_vl_interface': 8.649095063534576e-06, 'learning_rate/action_model': 8.649095063534576e-05, 'epoch': 0.47})
+[2026-07-28 11:58:09,374][starVLA.training.train_starvla][INFO] - Step 2561, Loss: {'train/loss': 0.02889365702867508, 'train/grad_norm_pre_clip': 0.7489849328994751, 'timing/data': 0.0003265300765633583, 'timing/model': 0.4203958949074149, 'learning_rate/qwen_vl_interface': 8.646421247845633e-06, 'learning_rate/action_model': 8.646421247845633e-05, 'epoch': 0.47})
+[2026-07-28 11:58:15,414][starVLA.training.train_starvla][INFO] - Step 2562, Loss: {'train/loss': 0.03132129833102226, 'train/grad_norm_pre_clip': 1.1132726669311523, 'timing/data': 0.00035634590312838554, 'timing/model': 0.42443526396527886, 'learning_rate/qwen_vl_interface': 8.643745228255816e-06, 'learning_rate/action_model': 8.643745228255817e-05, 'epoch': 0.47})
+[2026-07-28 11:58:21,673][starVLA.training.train_starvla][INFO] - Step 2563, Loss: {'train/loss': 0.026310179382562637, 'train/grad_norm_pre_clip': 0.9873867034912109, 'timing/data': 0.00029238639399409294, 'timing/model': 0.42241359828040004, 'learning_rate/qwen_vl_interface': 8.641067006501568e-06, 'learning_rate/action_model': 8.641067006501567e-05, 'epoch': 0.47})
+[2026-07-28 11:58:27,651][starVLA.training.train_starvla][INFO] - Step 2564, Loss: {'train/loss': 0.04256110638380051, 'train/grad_norm_pre_clip': 0.8416929841041565, 'timing/data': 0.0002667349763214588, 'timing/model': 0.4219516678713262, 'learning_rate/qwen_vl_interface': 8.638386584320753e-06, 'learning_rate/action_model': 8.638386584320752e-05, 'epoch': 0.47})
+[2026-07-28 11:58:33,897][starVLA.training.train_starvla][INFO] - Step 2565, Loss: {'train/loss': 0.025079548358917236, 'train/grad_norm_pre_clip': 0.6723427176475525, 'timing/data': 0.0002757278271019459, 'timing/model': 0.42045527417212725, 'learning_rate/qwen_vl_interface': 8.635703963452673e-06, 'learning_rate/action_model': 8.635703963452673e-05, 'epoch': 0.47})
+[2026-07-28 11:58:39,883][starVLA.training.train_starvla][INFO] - Step 2566, Loss: {'train/loss': 0.03834104537963867, 'train/grad_norm_pre_clip': 1.8411232233047485, 'timing/data': 0.0003917221911251545, 'timing/model': 0.42142276326194406, 'learning_rate/qwen_vl_interface': 8.633019145638048e-06, 'learning_rate/action_model': 8.633019145638048e-05, 'epoch': 0.47})
+[2026-07-28 11:58:46,111][starVLA.training.train_starvla][INFO] - Step 2567, Loss: {'train/loss': 0.018890921026468277, 'train/grad_norm_pre_clip': 0.5137806534767151, 'timing/data': 0.0002993559464812279, 'timing/model': 0.6565489009954035, 'learning_rate/qwen_vl_interface': 8.630332132619029e-06, 'learning_rate/action_model': 8.630332132619028e-05, 'epoch': 0.47})
+[2026-07-28 11:58:52,346][starVLA.training.train_starvla][INFO] - Step 2568, Loss: {'train/loss': 0.03862103074789047, 'train/grad_norm_pre_clip': 2.394578456878662, 'timing/data': 0.00027032429352402687, 'timing/model': 0.6654210118576884, 'learning_rate/qwen_vl_interface': 8.627642926139187e-06, 'learning_rate/action_model': 8.627642926139187e-05, 'epoch': 0.47})
+[2026-07-28 11:58:58,314][starVLA.training.train_starvla][INFO] - Step 2569, Loss: {'train/loss': 0.025368250906467438, 'train/grad_norm_pre_clip': 1.128685712814331, 'timing/data': 0.0003398568369448185, 'timing/model': 0.4202180439606309, 'learning_rate/qwen_vl_interface': 8.624951527943522e-06, 'learning_rate/action_model': 8.624951527943522e-05, 'epoch': 0.47})
+[2026-07-28 11:59:04,088][starVLA.training.train_starvla][INFO] - Step 2570, Loss: {'train/loss': 0.035800494253635406, 'train/grad_norm_pre_clip': 1.5209312438964844, 'timing/data': 0.00030999723821878433, 'timing/model': 0.4229607293382287, 'learning_rate/qwen_vl_interface': 8.622257939778453e-06, 'learning_rate/action_model': 8.622257939778453e-05, 'epoch': 0.47})
+[2026-07-28 11:59:10,096][starVLA.training.train_starvla][INFO] - Step 2571, Loss: {'train/loss': 0.02308093011379242, 'train/grad_norm_pre_clip': 0.770776629447937, 'timing/data': 0.0003909207880496979, 'timing/model': 0.4221999808214605, 'learning_rate/qwen_vl_interface': 8.619562163391818e-06, 'learning_rate/action_model': 8.619562163391817e-05, 'epoch': 0.47})
+[2026-07-28 11:59:16,119][starVLA.training.train_starvla][INFO] - Step 2572, Loss: {'train/loss': 0.03823943808674812, 'train/grad_norm_pre_clip': 2.175196886062622, 'timing/data': 0.0003140401095151901, 'timing/model': 0.42242301907390356, 'learning_rate/qwen_vl_interface': 8.616864200532877e-06, 'learning_rate/action_model': 8.616864200532877e-05, 'epoch': 0.47})
+[2026-07-28 11:59:22,627][starVLA.training.train_starvla][INFO] - Step 2573, Loss: {'train/loss': 0.048961006104946136, 'train/grad_norm_pre_clip': 2.0425937175750732, 'timing/data': 0.00034467875957489014, 'timing/model': 0.6744384597986937, 'learning_rate/qwen_vl_interface': 8.61416405295231e-06, 'learning_rate/action_model': 8.61416405295231e-05, 'epoch': 0.47})
+[2026-07-28 11:59:28,620][starVLA.training.train_starvla][INFO] - Step 2574, Loss: {'train/loss': 0.07468681037425995, 'train/grad_norm_pre_clip': 3.5403265953063965, 'timing/data': 0.0002602674067020416, 'timing/model': 0.4216648358851671, 'learning_rate/qwen_vl_interface': 8.611461722402214e-06, 'learning_rate/action_model': 8.611461722402215e-05, 'epoch': 0.47})
+[2026-07-28 11:59:34,631][starVLA.training.train_starvla][INFO] - Step 2575, Loss: {'train/loss': 0.07670512795448303, 'train/grad_norm_pre_clip': 2.100295066833496, 'timing/data': 0.0002923193387687206, 'timing/model': 0.4218146149069071, 'learning_rate/qwen_vl_interface': 8.608757210636101e-06, 'learning_rate/action_model': 8.608757210636101e-05, 'epoch': 0.47})
+[2026-07-28 11:59:40,695][starVLA.training.train_starvla][INFO] - Step 2576, Loss: {'train/loss': 0.10345237702131271, 'train/grad_norm_pre_clip': 3.140089511871338, 'timing/data': 0.0003475709818303585, 'timing/model': 0.42124689323827624, 'learning_rate/qwen_vl_interface': 8.606050519408901e-06, 'learning_rate/action_model': 8.606050519408901e-05, 'epoch': 0.47})
+[2026-07-28 11:59:46,914][starVLA.training.train_starvla][INFO] - Step 2577, Loss: {'train/loss': 0.04562783241271973, 'train/grad_norm_pre_clip': 1.8657690286636353, 'timing/data': 0.00042321300134062767, 'timing/model': 0.4231641632504761, 'learning_rate/qwen_vl_interface': 8.603341650476953e-06, 'learning_rate/action_model': 8.603341650476952e-05, 'epoch': 0.47})
+[2026-07-28 11:59:52,896][starVLA.training.train_starvla][INFO] - Step 2578, Loss: {'train/loss': 0.029238495975732803, 'train/grad_norm_pre_clip': 0.8534988760948181, 'timing/data': 0.0003014160320162773, 'timing/model': 0.4227397278882563, 'learning_rate/qwen_vl_interface': 8.600630605598016e-06, 'learning_rate/action_model': 8.600630605598015e-05, 'epoch': 0.47})
+[2026-07-28 11:59:59,164][starVLA.training.train_starvla][INFO] - Step 2579, Loss: {'train/loss': 0.07433368265628815, 'train/grad_norm_pre_clip': 3.0091841220855713, 'timing/data': 0.00030559906736016273, 'timing/model': 0.6730019832029939, 'learning_rate/qwen_vl_interface': 8.597917386531255e-06, 'learning_rate/action_model': 8.597917386531254e-05, 'epoch': 0.47})
+[2026-07-28 12:00:05,461][starVLA.training.train_starvla][INFO] - Step 2580, Loss: {'train/loss': 0.04537498205900192, 'train/grad_norm_pre_clip': 1.3280292749404907, 'timing/data': 0.0003582988865673542, 'timing/model': 0.420736881904304, 'learning_rate/qwen_vl_interface': 8.595201995037252e-06, 'learning_rate/action_model': 8.595201995037251e-05, 'epoch': 0.47})
+[2026-07-28 12:00:11,207][starVLA.training.train_starvla][INFO] - Step 2581, Loss: {'train/loss': 0.07042188197374344, 'train/grad_norm_pre_clip': 2.4936890602111816, 'timing/data': 0.00036482978612184525, 'timing/model': 0.42313593067228794, 'learning_rate/qwen_vl_interface': 8.59248443287799e-06, 'learning_rate/action_model': 8.592484432877989e-05, 'epoch': 0.47})
+[2026-07-28 12:00:17,212][starVLA.training.train_starvla][INFO] - Step 2582, Loss: {'train/loss': 0.03172293305397034, 'train/grad_norm_pre_clip': 1.1544748544692993, 'timing/data': 0.00038454215973615646, 'timing/model': 0.4225853490643203, 'learning_rate/qwen_vl_interface': 8.589764701816868e-06, 'learning_rate/action_model': 8.589764701816868e-05, 'epoch': 0.47})
+[2026-07-28 12:00:23,396][starVLA.training.train_starvla][INFO] - Step 2583, Loss: {'train/loss': 0.0549241378903389, 'train/grad_norm_pre_clip': 2.261695146560669, 'timing/data': 0.000353383831679821, 'timing/model': 0.5985807729884982, 'learning_rate/qwen_vl_interface': 8.587042803618693e-06, 'learning_rate/action_model': 8.587042803618693e-05, 'epoch': 0.47})
+[2026-07-28 12:00:29,626][starVLA.training.train_starvla][INFO] - Step 2584, Loss: {'train/loss': 0.04882211610674858, 'train/grad_norm_pre_clip': 2.5359063148498535, 'timing/data': 0.000770071055740118, 'timing/model': 0.420825089327991, 'learning_rate/qwen_vl_interface': 8.584318740049672e-06, 'learning_rate/action_model': 8.584318740049672e-05, 'epoch': 0.47})
+[2026-07-28 12:00:35,613][starVLA.training.train_starvla][INFO] - Step 2585, Loss: {'train/loss': 0.04369489848613739, 'train/grad_norm_pre_clip': 0.9999691247940063, 'timing/data': 0.0003633042797446251, 'timing/model': 0.4219868020154536, 'learning_rate/qwen_vl_interface': 8.581592512877419e-06, 'learning_rate/action_model': 8.581592512877418e-05, 'epoch': 0.47})
+[2026-07-28 12:00:41,636][starVLA.training.train_starvla][INFO] - Step 2586, Loss: {'train/loss': 0.03970923274755478, 'train/grad_norm_pre_clip': 0.9722529053688049, 'timing/data': 0.00034116674214601517, 'timing/model': 0.42412049090489745, 'learning_rate/qwen_vl_interface': 8.578864123870955e-06, 'learning_rate/action_model': 8.578864123870955e-05, 'epoch': 0.47})
+[2026-07-28 12:00:47,858][starVLA.training.train_starvla][INFO] - Step 2587, Loss: {'train/loss': 0.0225913617759943, 'train/grad_norm_pre_clip': 0.592366635799408, 'timing/data': 0.00033644912764430046, 'timing/model': 0.4233257710002363, 'learning_rate/qwen_vl_interface': 8.576133574800703e-06, 'learning_rate/action_model': 8.576133574800703e-05, 'epoch': 0.48})
+[2026-07-28 12:00:54,103][starVLA.training.train_starvla][INFO] - Step 2588, Loss: {'train/loss': 0.10140335559844971, 'train/grad_norm_pre_clip': 1.7008384466171265, 'timing/data': 0.0003025028854608536, 'timing/model': 0.42036852799355984, 'learning_rate/qwen_vl_interface': 8.573400867438486e-06, 'learning_rate/action_model': 8.573400867438486e-05, 'epoch': 0.48})
+[2026-07-28 12:01:00,147][starVLA.training.train_starvla][INFO] - Step 2589, Loss: {'train/loss': 0.08462908864021301, 'train/grad_norm_pre_clip': 2.2518539428710938, 'timing/data': 0.0003676731139421463, 'timing/model': 0.42353207897394896, 'learning_rate/qwen_vl_interface': 8.570666003557528e-06, 'learning_rate/action_model': 8.570666003557528e-05, 'epoch': 0.48})
+[2026-07-28 12:01:06,164][starVLA.training.train_starvla][INFO] - Step 2590, Loss: {'train/loss': 0.04336262494325638, 'train/grad_norm_pre_clip': 1.2251836061477661, 'timing/data': 0.0003159088082611561, 'timing/model': 0.4222791800275445, 'learning_rate/qwen_vl_interface': 8.567928984932452e-06, 'learning_rate/action_model': 8.567928984932452e-05, 'epoch': 0.48})
+[2026-07-28 12:01:12,433][starVLA.training.train_starvla][INFO] - Step 2591, Loss: {'train/loss': 0.03804117441177368, 'train/grad_norm_pre_clip': 1.1530495882034302, 'timing/data': 0.0004267762415111065, 'timing/model': 0.4272735225968063, 'learning_rate/qwen_vl_interface': 8.565189813339278e-06, 'learning_rate/action_model': 8.565189813339278e-05, 'epoch': 0.48})
+[2026-07-28 12:01:18,441][starVLA.training.train_starvla][INFO] - Step 2592, Loss: {'train/loss': 0.04754507169127464, 'train/grad_norm_pre_clip': 1.145311951637268, 'timing/data': 0.0003576776944100857, 'timing/model': 0.4215965559706092, 'learning_rate/qwen_vl_interface': 8.562448490555429e-06, 'learning_rate/action_model': 8.562448490555429e-05, 'epoch': 0.48})
+[2026-07-28 12:01:24,708][starVLA.training.train_starvla][INFO] - Step 2593, Loss: {'train/loss': 0.04276499152183533, 'train/grad_norm_pre_clip': 2.18188214302063, 'timing/data': 0.0003708438016474247, 'timing/model': 0.6675614342093468, 'learning_rate/qwen_vl_interface': 8.559705018359716e-06, 'learning_rate/action_model': 8.559705018359716e-05, 'epoch': 0.48})
+[2026-07-28 12:01:30,687][starVLA.training.train_starvla][INFO] - Step 2594, Loss: {'train/loss': 0.05164598673582077, 'train/grad_norm_pre_clip': 0.9201080799102783, 'timing/data': 0.0003366628661751747, 'timing/model': 0.42161515401676297, 'learning_rate/qwen_vl_interface': 8.55695939853235e-06, 'learning_rate/action_model': 8.55695939853235e-05, 'epoch': 0.48})
+[2026-07-28 12:01:36,668][starVLA.training.train_starvla][INFO] - Step 2595, Loss: {'train/loss': 0.04328636825084686, 'train/grad_norm_pre_clip': 2.2630066871643066, 'timing/data': 0.0003657117486000061, 'timing/model': 0.42081320099532604, 'learning_rate/qwen_vl_interface': 8.554211632854932e-06, 'learning_rate/action_model': 8.554211632854932e-05, 'epoch': 0.48})
+[2026-07-28 12:01:42,889][starVLA.training.train_starvla][INFO] - Step 2596, Loss: {'train/loss': 0.03782286494970322, 'train/grad_norm_pre_clip': 1.5200282335281372, 'timing/data': 0.0003071008250117302, 'timing/model': 0.6596255209296942, 'learning_rate/qwen_vl_interface': 8.551461723110461e-06, 'learning_rate/action_model': 8.551461723110461e-05, 'epoch': 0.48})
+[2026-07-28 12:01:48,889][starVLA.training.train_starvla][INFO] - Step 2597, Loss: {'train/loss': 0.02287846803665161, 'train/grad_norm_pre_clip': 0.7710000872612, 'timing/data': 0.0003001578152179718, 'timing/model': 0.42273073410615325, 'learning_rate/qwen_vl_interface': 8.548709671083321e-06, 'learning_rate/action_model': 8.54870967108332e-05, 'epoch': 0.48})
+[2026-07-28 12:01:54,897][starVLA.training.train_starvla][INFO] - Step 2598, Loss: {'train/loss': 0.08468751609325409, 'train/grad_norm_pre_clip': 2.622553825378418, 'timing/data': 0.0002797497436404228, 'timing/model': 0.42316537303850055, 'learning_rate/qwen_vl_interface': 8.545955478559288e-06, 'learning_rate/action_model': 8.545955478559288e-05, 'epoch': 0.48})
+[2026-07-28 12:02:00,900][starVLA.training.train_starvla][INFO] - Step 2599, Loss: {'train/loss': 0.059050098061561584, 'train/grad_norm_pre_clip': 2.203535556793213, 'timing/data': 0.0003136522136628628, 'timing/model': 0.4234593091532588, 'learning_rate/qwen_vl_interface': 8.54319914732553e-06, 'learning_rate/action_model': 8.543199147325529e-05, 'epoch': 0.48})
+[2026-07-28 12:02:07,169][starVLA.training.train_starvla][INFO] - Step 2600, Loss: {'train/loss': 0.04021134227514267, 'train/grad_norm_pre_clip': 1.3190234899520874, 'timing/data': 0.00029387883841991425, 'timing/model': 0.42194825131446123, 'learning_rate/qwen_vl_interface': 8.540440679170599e-06, 'learning_rate/action_model': 8.540440679170598e-05, 'epoch': 0.48})
+[2026-07-28 12:02:13,397][starVLA.training.train_starvla][INFO] - Step 2601, Loss: {'train/loss': 0.05358948931097984, 'train/grad_norm_pre_clip': 1.8243991136550903, 'timing/data': 0.00034137070178985596, 'timing/model': 0.4220664673484862, 'learning_rate/qwen_vl_interface': 8.537680075884433e-06, 'learning_rate/action_model': 8.537680075884433e-05, 'epoch': 0.48})
+[2026-07-28 12:02:19,400][starVLA.training.train_starvla][INFO] - Step 2602, Loss: {'train/loss': 0.0824059545993805, 'train/grad_norm_pre_clip': 1.6971464157104492, 'timing/data': 0.00031212298199534416, 'timing/model': 0.4239585460163653, 'learning_rate/qwen_vl_interface': 8.53491733925836e-06, 'learning_rate/action_model': 8.53491733925836e-05, 'epoch': 0.48})
+[2026-07-28 12:02:25,401][starVLA.training.train_starvla][INFO] - Step 2603, Loss: {'train/loss': 0.038623515516519547, 'train/grad_norm_pre_clip': 1.2503875494003296, 'timing/data': 0.0003599459305405617, 'timing/model': 0.421479654032737, 'learning_rate/qwen_vl_interface': 8.532152471085092e-06, 'learning_rate/action_model': 8.53215247108509e-05, 'epoch': 0.48})
+[2026-07-28 12:02:31,636][starVLA.training.train_starvla][INFO] - Step 2604, Loss: {'train/loss': 0.033385325223207474, 'train/grad_norm_pre_clip': 0.8131814002990723, 'timing/data': 0.00037274695932865143, 'timing/model': 0.6786567200906575, 'learning_rate/qwen_vl_interface': 8.529385473158717e-06, 'learning_rate/action_model': 8.529385473158717e-05, 'epoch': 0.48})
+[2026-07-28 12:02:37,599][starVLA.training.train_starvla][INFO] - Step 2605, Loss: {'train/loss': 0.031874872744083405, 'train/grad_norm_pre_clip': 0.8051066994667053, 'timing/data': 0.0003599608317017555, 'timing/model': 0.4221968650817871, 'learning_rate/qwen_vl_interface': 8.526616347274712e-06, 'learning_rate/action_model': 8.526616347274712e-05, 'epoch': 0.48})
+[2026-07-28 12:02:43,594][starVLA.training.train_starvla][INFO] - Step 2606, Loss: {'train/loss': 0.04487880319356918, 'train/grad_norm_pre_clip': 1.025938868522644, 'timing/data': 0.00029692891985177994, 'timing/model': 0.42175514670088887, 'learning_rate/qwen_vl_interface': 8.523845095229935e-06, 'learning_rate/action_model': 8.523845095229935e-05, 'epoch': 0.48})
+[2026-07-28 12:02:49,344][starVLA.training.train_starvla][INFO] - Step 2607, Loss: {'train/loss': 0.07265844941139221, 'train/grad_norm_pre_clip': 1.2575438022613525, 'timing/data': 0.0003066207282245159, 'timing/model': 0.4224530467763543, 'learning_rate/qwen_vl_interface': 8.521071718822616e-06, 'learning_rate/action_model': 8.521071718822616e-05, 'epoch': 0.48})
+[2026-07-28 12:02:55,559][starVLA.training.train_starvla][INFO] - Step 2608, Loss: {'train/loss': 0.017247632145881653, 'train/grad_norm_pre_clip': 0.756932258605957, 'timing/data': 0.0003227498382329941, 'timing/model': 0.6525170658715069, 'learning_rate/qwen_vl_interface': 8.518296219852375e-06, 'learning_rate/action_model': 8.518296219852375e-05, 'epoch': 0.48})
+[2026-07-28 12:03:01,534][starVLA.training.train_starvla][INFO] - Step 2609, Loss: {'train/loss': 0.029028750956058502, 'train/grad_norm_pre_clip': 0.7612327337265015, 'timing/data': 0.00031805504113435745, 'timing/model': 0.41881314758211374, 'learning_rate/qwen_vl_interface': 8.515518600120201e-06, 'learning_rate/action_model': 8.5155186001202e-05, 'epoch': 0.48})
+[2026-07-28 12:03:07,505][starVLA.training.train_starvla][INFO] - Step 2610, Loss: {'train/loss': 0.024164997041225433, 'train/grad_norm_pre_clip': 1.3510172367095947, 'timing/data': 0.00036793993785977364, 'timing/model': 0.42065615206956863, 'learning_rate/qwen_vl_interface': 8.512738861428458e-06, 'learning_rate/action_model': 8.512738861428458e-05, 'epoch': 0.48})
+[2026-07-28 12:03:13,476][starVLA.training.train_starvla][INFO] - Step 2611, Loss: {'train/loss': 0.025877662003040314, 'train/grad_norm_pre_clip': 0.6156503558158875, 'timing/data': 0.000517191831022501, 'timing/model': 0.4201902928762138, 'learning_rate/qwen_vl_interface': 8.509957005580892e-06, 'learning_rate/action_model': 8.509957005580892e-05, 'epoch': 0.48})
+[2026-07-28 12:03:19,744][starVLA.training.train_starvla][INFO] - Step 2612, Loss: {'train/loss': 0.0683230385184288, 'train/grad_norm_pre_clip': 2.196768283843994, 'timing/data': 0.0003323289565742016, 'timing/model': 0.42545991390943527, 'learning_rate/qwen_vl_interface': 8.507173034382617e-06, 'learning_rate/action_model': 8.507173034382617e-05, 'epoch': 0.48})
+[2026-07-28 12:03:25,773][starVLA.training.train_starvla][INFO] - Step 2613, Loss: {'train/loss': 0.023764291778206825, 'train/grad_norm_pre_clip': 0.9814245700836182, 'timing/data': 0.0003413748927414417, 'timing/model': 0.41982142720371485, 'learning_rate/qwen_vl_interface': 8.504386949640124e-06, 'learning_rate/action_model': 8.504386949640124e-05, 'epoch': 0.48})
+[2026-07-28 12:03:31,985][starVLA.training.train_starvla][INFO] - Step 2614, Loss: {'train/loss': 0.034285906702280045, 'train/grad_norm_pre_clip': 1.052642583847046, 'timing/data': 0.00035335682332515717, 'timing/model': 0.42179904878139496, 'learning_rate/qwen_vl_interface': 8.501598753161273e-06, 'learning_rate/action_model': 8.501598753161272e-05, 'epoch': 0.48})
+[2026-07-28 12:03:38,255][starVLA.training.train_starvla][INFO] - Step 2615, Loss: {'train/loss': 0.04767372086644173, 'train/grad_norm_pre_clip': 1.2122201919555664, 'timing/data': 0.0002897861413657665, 'timing/model': 0.6753106042742729, 'learning_rate/qwen_vl_interface': 8.498808446755291e-06, 'learning_rate/action_model': 8.498808446755291e-05, 'epoch': 0.48})
+[2026-07-28 12:03:44,287][starVLA.training.train_starvla][INFO] - Step 2616, Loss: {'train/loss': 0.0465068444609642, 'train/grad_norm_pre_clip': 1.595455288887024, 'timing/data': 0.00036404794082045555, 'timing/model': 0.4308638130314648, 'learning_rate/qwen_vl_interface': 8.49601603223278e-06, 'learning_rate/action_model': 8.49601603223278e-05, 'epoch': 0.48})
+[2026-07-28 12:03:50,490][starVLA.training.train_starvla][INFO] - Step 2617, Loss: {'train/loss': 0.048070721328258514, 'train/grad_norm_pre_clip': 2.1836013793945312, 'timing/data': 0.0004723896272480488, 'timing/model': 0.4359860587865114, 'learning_rate/qwen_vl_interface': 8.493221511405708e-06, 'learning_rate/action_model': 8.493221511405707e-05, 'epoch': 0.48})
+[2026-07-28 12:03:56,847][starVLA.training.train_starvla][INFO] - Step 2618, Loss: {'train/loss': 0.021893247961997986, 'train/grad_norm_pre_clip': 0.8130244016647339, 'timing/data': 0.00032088207080960274, 'timing/model': 0.42860190477222204, 'learning_rate/qwen_vl_interface': 8.490424886087407e-06, 'learning_rate/action_model': 8.490424886087407e-05, 'epoch': 0.48})
+[2026-07-28 12:04:03,030][starVLA.training.train_starvla][INFO] - Step 2619, Loss: {'train/loss': 0.02446538582444191, 'train/grad_norm_pre_clip': 0.7672992944717407, 'timing/data': 0.0002783159725368023, 'timing/model': 0.4226834373548627, 'learning_rate/qwen_vl_interface': 8.487626158092578e-06, 'learning_rate/action_model': 8.487626158092579e-05, 'epoch': 0.48})
+[2026-07-28 12:04:09,435][starVLA.training.train_starvla][INFO] - Step 2620, Loss: {'train/loss': 0.0859067514538765, 'train/grad_norm_pre_clip': 1.9185653924942017, 'timing/data': 0.00037191063165664673, 'timing/model': 0.7485736911185086, 'learning_rate/qwen_vl_interface': 8.484825329237287e-06, 'learning_rate/action_model': 8.484825329237287e-05, 'epoch': 0.48})
+[2026-07-28 12:04:15,840][starVLA.training.train_starvla][INFO] - Step 2621, Loss: {'train/loss': 0.022488612681627274, 'train/grad_norm_pre_clip': 0.8948540091514587, 'timing/data': 0.00037475908175110817, 'timing/model': 0.4245143886655569, 'learning_rate/qwen_vl_interface': 8.48202240133896e-06, 'learning_rate/action_model': 8.482022401338959e-05, 'epoch': 0.48})
+[2026-07-28 12:04:21,903][starVLA.training.train_starvla][INFO] - Step 2622, Loss: {'train/loss': 0.07327990978956223, 'train/grad_norm_pre_clip': 2.6579082012176514, 'timing/data': 0.0003031785599887371, 'timing/model': 0.42301279399544, 'learning_rate/qwen_vl_interface': 8.479217376216385e-06, 'learning_rate/action_model': 8.479217376216385e-05, 'epoch': 0.48})
+[2026-07-28 12:04:28,125][starVLA.training.train_starvla][INFO] - Step 2623, Loss: {'train/loss': 0.0448274202644825, 'train/grad_norm_pre_clip': 1.4045168161392212, 'timing/data': 0.0003243759274482727, 'timing/model': 0.44574724696576595, 'learning_rate/qwen_vl_interface': 8.476410255689714e-06, 'learning_rate/action_model': 8.476410255689713e-05, 'epoch': 0.48})
+[2026-07-28 12:04:34,419][starVLA.training.train_starvla][INFO] - Step 2624, Loss: {'train/loss': 0.04783753678202629, 'train/grad_norm_pre_clip': 1.9528390169143677, 'timing/data': 0.00035019591450691223, 'timing/model': 0.6020784806460142, 'learning_rate/qwen_vl_interface': 8.473601041580457e-06, 'learning_rate/action_model': 8.473601041580457e-05, 'epoch': 0.48})
+[2026-07-28 12:04:40,526][starVLA.training.train_starvla][INFO] - Step 2625, Loss: {'train/loss': 0.04577669873833656, 'train/grad_norm_pre_clip': 1.5241707563400269, 'timing/data': 0.0003667958080768585, 'timing/model': 0.42038527596741915, 'learning_rate/qwen_vl_interface': 8.470789735711483e-06, 'learning_rate/action_model': 8.470789735711483e-05, 'epoch': 0.48})
+[2026-07-28 12:04:46,752][starVLA.training.train_starvla][INFO] - Step 2626, Loss: {'train/loss': 0.04874624311923981, 'train/grad_norm_pre_clip': 0.795835554599762, 'timing/data': 0.0002968483604490757, 'timing/model': 0.6648388518951833, 'learning_rate/qwen_vl_interface': 8.467976339907015e-06, 'learning_rate/action_model': 8.467976339907016e-05, 'epoch': 0.48})
+[2026-07-28 12:04:52,794][starVLA.training.train_starvla][INFO] - Step 2627, Loss: {'train/loss': 0.04756347835063934, 'train/grad_norm_pre_clip': 1.0391793251037598, 'timing/data': 0.0004973919130861759, 'timing/model': 0.42342429608106613, 'learning_rate/qwen_vl_interface': 8.465160855992639e-06, 'learning_rate/action_model': 8.465160855992638e-05, 'epoch': 0.48})
+[2026-07-28 12:04:59,170][starVLA.training.train_starvla][INFO] - Step 2628, Loss: {'train/loss': 0.03359924629330635, 'train/grad_norm_pre_clip': 0.6891571283340454, 'timing/data': 0.00029951706528663635, 'timing/model': 0.426406464073807, 'learning_rate/qwen_vl_interface': 8.462343285795289e-06, 'learning_rate/action_model': 8.462343285795288e-05, 'epoch': 0.48})
+[2026-07-28 12:05:05,260][starVLA.training.train_starvla][INFO] - Step 2629, Loss: {'train/loss': 0.03654675558209419, 'train/grad_norm_pre_clip': 1.516856074333191, 'timing/data': 0.0003895997069776058, 'timing/model': 0.43059025425463915, 'learning_rate/qwen_vl_interface': 8.459523631143259e-06, 'learning_rate/action_model': 8.459523631143258e-05, 'epoch': 0.48})
+[2026-07-28 12:05:11,410][starVLA.training.train_starvla][INFO] - Step 2630, Loss: {'train/loss': 0.036036618053913116, 'train/grad_norm_pre_clip': 0.9196744561195374, 'timing/data': 0.0002794242464005947, 'timing/model': 0.4277622629888356, 'learning_rate/qwen_vl_interface': 8.456701893866188e-06, 'learning_rate/action_model': 8.456701893866187e-05, 'epoch': 0.48})
+[2026-07-28 12:05:17,901][starVLA.training.train_starvla][INFO] - Step 2631, Loss: {'train/loss': 0.013223640620708466, 'train/grad_norm_pre_clip': 0.6035428047180176, 'timing/data': 0.0004139319062232971, 'timing/model': 0.6494743349030614, 'learning_rate/qwen_vl_interface': 8.453878075795072e-06, 'learning_rate/action_model': 8.453878075795072e-05, 'epoch': 0.48})
+[2026-07-28 12:05:23,703][starVLA.training.train_starvla][INFO] - Step 2632, Loss: {'train/loss': 0.021030930802226067, 'train/grad_norm_pre_clip': 0.7410115003585815, 'timing/data': 0.00031495001167058945, 'timing/model': 0.4247240349650383, 'learning_rate/qwen_vl_interface': 8.451052178762258e-06, 'learning_rate/action_model': 8.451052178762258e-05, 'epoch': 0.48})
+[2026-07-28 12:05:29,810][starVLA.training.train_starvla][INFO] - Step 2633, Loss: {'train/loss': 0.03514470160007477, 'train/grad_norm_pre_clip': 0.9799618124961853, 'timing/data': 0.00030474504455924034, 'timing/model': 0.4276715777814388, 'learning_rate/qwen_vl_interface': 8.448224204601439e-06, 'learning_rate/action_model': 8.448224204601438e-05, 'epoch': 0.48})
+[2026-07-28 12:05:35,809][starVLA.training.train_starvla][INFO] - Step 2634, Loss: {'train/loss': 0.05490661785006523, 'train/grad_norm_pre_clip': 1.293957233428955, 'timing/data': 0.0003113369457423687, 'timing/model': 0.4218571623787284, 'learning_rate/qwen_vl_interface': 8.445394155147655e-06, 'learning_rate/action_model': 8.445394155147653e-05, 'epoch': 0.48})
+[2026-07-28 12:05:42,258][starVLA.training.train_starvla][INFO] - Step 2635, Loss: {'train/loss': 0.03798236325383186, 'train/grad_norm_pre_clip': 1.4808969497680664, 'timing/data': 0.0002884902060031891, 'timing/model': 0.41970303608104587, 'learning_rate/qwen_vl_interface': 8.442562032237295e-06, 'learning_rate/action_model': 8.442562032237296e-05, 'epoch': 0.48})
+[2026-07-28 12:05:48,242][starVLA.training.train_starvla][INFO] - Step 2636, Loss: {'train/loss': 0.04930800944566727, 'train/grad_norm_pre_clip': 1.3814359903335571, 'timing/data': 0.000295781996101141, 'timing/model': 0.4225460132583976, 'learning_rate/qwen_vl_interface': 8.439727837708094e-06, 'learning_rate/action_model': 8.439727837708095e-05, 'epoch': 0.48})
+[2026-07-28 12:05:54,254][starVLA.training.train_starvla][INFO] - Step 2637, Loss: {'train/loss': 0.062464311718940735, 'train/grad_norm_pre_clip': 2.1461548805236816, 'timing/data': 0.0002894601784646511, 'timing/model': 0.4221029062755406, 'learning_rate/qwen_vl_interface': 8.436891573399132e-06, 'learning_rate/action_model': 8.436891573399131e-05, 'epoch': 0.48})
+[2026-07-28 12:06:00,520][starVLA.training.train_starvla][INFO] - Step 2638, Loss: {'train/loss': 0.029747307300567627, 'train/grad_norm_pre_clip': 0.9522417187690735, 'timing/data': 0.0004192190244793892, 'timing/model': 0.4256053688004613, 'learning_rate/qwen_vl_interface': 8.434053241150826e-06, 'learning_rate/action_model': 8.434053241150825e-05, 'epoch': 0.48})
+[2026-07-28 12:06:06,514][starVLA.training.train_starvla][INFO] - Step 2639, Loss: {'train/loss': 0.03112161159515381, 'train/grad_norm_pre_clip': 0.8302960991859436, 'timing/data': 0.0003113020211458206, 'timing/model': 0.4236551942303777, 'learning_rate/qwen_vl_interface': 8.43121284280494e-06, 'learning_rate/action_model': 8.431212842804939e-05, 'epoch': 0.48})
+[2026-07-28 12:06:12,763][starVLA.training.train_starvla][INFO] - Step 2640, Loss: {'train/loss': 0.04374879598617554, 'train/grad_norm_pre_clip': 1.329392433166504, 'timing/data': 0.00036774342879652977, 'timing/model': 0.6656570951454341, 'learning_rate/qwen_vl_interface': 8.428370380204577e-06, 'learning_rate/action_model': 8.428370380204577e-05, 'epoch': 0.48})
+[2026-07-28 12:06:19,015][starVLA.training.train_starvla][INFO] - Step 2641, Loss: {'train/loss': 0.041247472167015076, 'train/grad_norm_pre_clip': 1.1009550094604492, 'timing/data': 0.00026668794453144073, 'timing/model': 0.6827596942894161, 'learning_rate/qwen_vl_interface': 8.425525855194184e-06, 'learning_rate/action_model': 8.425525855194183e-05, 'epoch': 0.49})
+[2026-07-28 12:06:24,803][starVLA.training.train_starvla][INFO] - Step 2642, Loss: {'train/loss': 0.04274995997548103, 'train/grad_norm_pre_clip': 0.9410973191261292, 'timing/data': 0.0003047296777367592, 'timing/model': 0.4222865947522223, 'learning_rate/qwen_vl_interface': 8.422679269619538e-06, 'learning_rate/action_model': 8.422679269619538e-05, 'epoch': 0.49})
+[2026-07-28 12:06:30,874][starVLA.training.train_starvla][INFO] - Step 2643, Loss: {'train/loss': 0.06037468835711479, 'train/grad_norm_pre_clip': 1.4404295682907104, 'timing/data': 0.00032341014593839645, 'timing/model': 0.4265453410334885, 'learning_rate/qwen_vl_interface': 8.419830625327758e-06, 'learning_rate/action_model': 8.419830625327758e-05, 'epoch': 0.49})
+[2026-07-28 12:06:36,946][starVLA.training.train_starvla][INFO] - Step 2644, Loss: {'train/loss': 0.032146573066711426, 'train/grad_norm_pre_clip': 2.1227805614471436, 'timing/data': 0.00041366880759596825, 'timing/model': 0.42357560293748975, 'learning_rate/qwen_vl_interface': 8.416979924167303e-06, 'learning_rate/action_model': 8.416979924167303e-05, 'epoch': 0.49})
+[2026-07-28 12:06:43,164][starVLA.training.train_starvla][INFO] - Step 2645, Loss: {'train/loss': 0.042162250727415085, 'train/grad_norm_pre_clip': 1.089741826057434, 'timing/data': 0.00033272430300712585, 'timing/model': 0.41937406407669187, 'learning_rate/qwen_vl_interface': 8.414127167987957e-06, 'learning_rate/action_model': 8.414127167987957e-05, 'epoch': 0.49})
+[2026-07-28 12:06:49,149][starVLA.training.train_starvla][INFO] - Step 2646, Loss: {'train/loss': 0.06780534982681274, 'train/grad_norm_pre_clip': 2.542436361312866, 'timing/data': 0.00027709314599633217, 'timing/model': 0.4290464078076184, 'learning_rate/qwen_vl_interface': 8.411272358640843e-06, 'learning_rate/action_model': 8.411272358640844e-05, 'epoch': 0.49})
+[2026-07-28 12:06:55,385][starVLA.training.train_starvla][INFO] - Step 2647, Loss: {'train/loss': 0.023008622229099274, 'train/grad_norm_pre_clip': 0.536013662815094, 'timing/data': 0.00026103388518095016, 'timing/model': 0.4210874200798571, 'learning_rate/qwen_vl_interface': 8.408415497978418e-06, 'learning_rate/action_model': 8.408415497978418e-05, 'epoch': 0.49})
+[2026-07-28 12:07:01,619][starVLA.training.train_starvla][INFO] - Step 2648, Loss: {'train/loss': 0.043539613485336304, 'train/grad_norm_pre_clip': 1.116827368736267, 'timing/data': 0.00034713419154286385, 'timing/model': 0.4200615338049829, 'learning_rate/qwen_vl_interface': 8.405556587854468e-06, 'learning_rate/action_model': 8.405556587854467e-05, 'epoch': 0.49})
+[2026-07-28 12:07:07,627][starVLA.training.train_starvla][INFO] - Step 2649, Loss: {'train/loss': 0.01462157815694809, 'train/grad_norm_pre_clip': 0.3177559971809387, 'timing/data': 0.00030736904591321945, 'timing/model': 0.43591053830459714, 'learning_rate/qwen_vl_interface': 8.402695630124106e-06, 'learning_rate/action_model': 8.402695630124106e-05, 'epoch': 0.49})
+[2026-07-28 12:07:13,705][starVLA.training.train_starvla][INFO] - Step 2650, Loss: {'train/loss': 0.10760122537612915, 'train/grad_norm_pre_clip': 5.403721332550049, 'timing/data': 0.0003738026134669781, 'timing/model': 0.42366570932790637, 'learning_rate/qwen_vl_interface': 8.399832626643778e-06, 'learning_rate/action_model': 8.399832626643778e-05, 'epoch': 0.49})
+[2026-07-28 12:07:20,085][starVLA.training.train_starvla][INFO] - Step 2651, Loss: {'train/loss': 0.07022802531719208, 'train/grad_norm_pre_clip': 2.810784101486206, 'timing/data': 0.000390367116779089, 'timing/model': 0.6702338708564639, 'learning_rate/qwen_vl_interface': 8.396967579271256e-06, 'learning_rate/action_model': 8.396967579271257e-05, 'epoch': 0.49})
+[2026-07-28 12:07:26,065][starVLA.training.train_starvla][INFO] - Step 2652, Loss: {'train/loss': 0.05057377740740776, 'train/grad_norm_pre_clip': 1.3962541818618774, 'timing/data': 0.00030279718339443207, 'timing/model': 0.42312489077448845, 'learning_rate/qwen_vl_interface': 8.394100489865638e-06, 'learning_rate/action_model': 8.394100489865638e-05, 'epoch': 0.49})
+[2026-07-28 12:07:31,872][starVLA.training.train_starvla][INFO] - Step 2653, Loss: {'train/loss': 0.055195026099681854, 'train/grad_norm_pre_clip': 1.0864380598068237, 'timing/data': 0.00028868578374385834, 'timing/model': 0.4241574048064649, 'learning_rate/qwen_vl_interface': 8.391231360287348e-06, 'learning_rate/action_model': 8.391231360287347e-05, 'epoch': 0.49})
+[2026-07-28 12:07:37,881][starVLA.training.train_starvla][INFO] - Step 2654, Loss: {'train/loss': 0.07104793190956116, 'train/grad_norm_pre_clip': 2.857253313064575, 'timing/data': 0.0002896408550441265, 'timing/model': 0.4224487650208175, 'learning_rate/qwen_vl_interface': 8.388360192398131e-06, 'learning_rate/action_model': 8.38836019239813e-05, 'epoch': 0.49})
+[2026-07-28 12:07:44,169][starVLA.training.train_starvla][INFO] - Step 2655, Loss: {'train/loss': 0.06766818463802338, 'train/grad_norm_pre_clip': 2.757017135620117, 'timing/data': 0.0002805311232805252, 'timing/model': 0.42003600066527724, 'learning_rate/qwen_vl_interface': 8.385486988061058e-06, 'learning_rate/action_model': 8.385486988061058e-05, 'epoch': 0.49})
+[2026-07-28 12:07:50,530][starVLA.training.train_starvla][INFO] - Step 2656, Loss: {'train/loss': 0.02313837967813015, 'train/grad_norm_pre_clip': 0.5330297350883484, 'timing/data': 0.0002750619314610958, 'timing/model': 0.4202357763424516, 'learning_rate/qwen_vl_interface': 8.38261174914052e-06, 'learning_rate/action_model': 8.38261174914052e-05, 'epoch': 0.49})
+[2026-07-28 12:07:56,580][starVLA.training.train_starvla][INFO] - Step 2657, Loss: {'train/loss': 0.06441520154476166, 'train/grad_norm_pre_clip': 2.1463372707366943, 'timing/data': 0.0004081549122929573, 'timing/model': 0.4269197629764676, 'learning_rate/qwen_vl_interface': 8.379734477502229e-06, 'learning_rate/action_model': 8.379734477502228e-05, 'epoch': 0.49})
+[2026-07-28 12:08:02,642][starVLA.training.train_starvla][INFO] - Step 2658, Loss: {'train/loss': 0.030295897275209427, 'train/grad_norm_pre_clip': 0.6465699672698975, 'timing/data': 0.00029488885775208473, 'timing/model': 0.439103907905519, 'learning_rate/qwen_vl_interface': 8.376855175013214e-06, 'learning_rate/action_model': 8.376855175013213e-05, 'epoch': 0.49})
+[2026-07-28 12:08:08,977][starVLA.training.train_starvla][INFO] - Step 2659, Loss: {'train/loss': 0.04620283842086792, 'train/grad_norm_pre_clip': 1.135895013809204, 'timing/data': 0.00027458788827061653, 'timing/model': 0.42167288390919566, 'learning_rate/qwen_vl_interface': 8.373973843541822e-06, 'learning_rate/action_model': 8.373973843541822e-05, 'epoch': 0.49})
+[2026-07-28 12:08:14,975][starVLA.training.train_starvla][INFO] - Step 2660, Loss: {'train/loss': 0.03734728321433067, 'train/grad_norm_pre_clip': 0.5617222189903259, 'timing/data': 0.0002997959963977337, 'timing/model': 0.4255356057547033, 'learning_rate/qwen_vl_interface': 8.371090484957722e-06, 'learning_rate/action_model': 8.371090484957722e-05, 'epoch': 0.49})
+[2026-07-28 12:08:21,228][starVLA.training.train_starvla][INFO] - Step 2661, Loss: {'train/loss': 0.03601953014731407, 'train/grad_norm_pre_clip': 1.4758418798446655, 'timing/data': 0.00039263488724827766, 'timing/model': 0.6566287903115153, 'learning_rate/qwen_vl_interface': 8.368205101131891e-06, 'learning_rate/action_model': 8.368205101131891e-05, 'epoch': 0.49})
+[2026-07-28 12:08:27,505][starVLA.training.train_starvla][INFO] - Step 2662, Loss: {'train/loss': 0.04953285679221153, 'train/grad_norm_pre_clip': 1.02009916305542, 'timing/data': 0.00034120911732316017, 'timing/model': 0.6951366243883967, 'learning_rate/qwen_vl_interface': 8.365317693936624e-06, 'learning_rate/action_model': 8.365317693936623e-05, 'epoch': 0.49})
+[2026-07-28 12:08:33,472][starVLA.training.train_starvla][INFO] - Step 2663, Loss: {'train/loss': 0.02273186668753624, 'train/grad_norm_pre_clip': 1.0721036195755005, 'timing/data': 0.00029289210215210915, 'timing/model': 0.42210486717522144, 'learning_rate/qwen_vl_interface': 8.362428265245526e-06, 'learning_rate/action_model': 8.362428265245526e-05, 'epoch': 0.49})
+[2026-07-28 12:08:39,446][starVLA.training.train_starvla][INFO] - Step 2664, Loss: {'train/loss': 0.07613629102706909, 'train/grad_norm_pre_clip': 1.7035316228866577, 'timing/data': 0.0002812221646308899, 'timing/model': 0.4207348069176078, 'learning_rate/qwen_vl_interface': 8.359536816933518e-06, 'learning_rate/action_model': 8.359536816933518e-05, 'epoch': 0.49})
+[2026-07-28 12:08:45,416][starVLA.training.train_starvla][INFO] - Step 2665, Loss: {'train/loss': 0.0391865111887455, 'train/grad_norm_pre_clip': 1.864089012145996, 'timing/data': 0.0005162176676094532, 'timing/model': 0.4231144390068948, 'learning_rate/qwen_vl_interface': 8.356643350876832e-06, 'learning_rate/action_model': 8.356643350876832e-05, 'epoch': 0.49})
+[2026-07-28 12:08:51,639][starVLA.training.train_starvla][INFO] - Step 2666, Loss: {'train/loss': 0.07493966072797775, 'train/grad_norm_pre_clip': 1.7193598747253418, 'timing/data': 0.0002792561426758766, 'timing/model': 0.4209503619931638, 'learning_rate/qwen_vl_interface': 8.353747868953002e-06, 'learning_rate/action_model': 8.353747868953003e-05, 'epoch': 0.49})
+[2026-07-28 12:08:57,850][starVLA.training.train_starvla][INFO] - Step 2667, Loss: {'train/loss': 0.029566435143351555, 'train/grad_norm_pre_clip': 1.3853988647460938, 'timing/data': 0.00026258500292897224, 'timing/model': 0.6631120252422988, 'learning_rate/qwen_vl_interface': 8.350850373040876e-06, 'learning_rate/action_model': 8.350850373040876e-05, 'epoch': 0.49})
+[2026-07-28 12:09:03,857][starVLA.training.train_starvla][INFO] - Step 2668, Loss: {'train/loss': 0.04510117322206497, 'train/grad_norm_pre_clip': 1.074365258216858, 'timing/data': 0.0002590622752904892, 'timing/model': 0.41951544489711523, 'learning_rate/qwen_vl_interface': 8.34795086502061e-06, 'learning_rate/action_model': 8.347950865020609e-05, 'epoch': 0.49})
+[2026-07-28 12:09:10,106][starVLA.training.train_starvla][INFO] - Step 2669, Loss: {'train/loss': 0.05945334583520889, 'train/grad_norm_pre_clip': 2.468390703201294, 'timing/data': 0.00032120896503329277, 'timing/model': 0.4206749889999628, 'learning_rate/qwen_vl_interface': 8.34504934677366e-06, 'learning_rate/action_model': 8.34504934677366e-05, 'epoch': 0.49})
+[2026-07-28 12:09:16,181][starVLA.training.train_starvla][INFO] - Step 2670, Loss: {'train/loss': 0.03393806144595146, 'train/grad_norm_pre_clip': 0.7683002948760986, 'timing/data': 0.0003565722145140171, 'timing/model': 0.43094681203365326, 'learning_rate/qwen_vl_interface': 8.342145820182795e-06, 'learning_rate/action_model': 8.342145820182793e-05, 'epoch': 0.49})
+[2026-07-28 12:09:21,996][starVLA.training.train_starvla][INFO] - Step 2671, Loss: {'train/loss': 0.03439854085445404, 'train/grad_norm_pre_clip': 0.9811422228813171, 'timing/data': 0.00037952372804284096, 'timing/model': 0.4233912779018283, 'learning_rate/qwen_vl_interface': 8.339240287132073e-06, 'learning_rate/action_model': 8.339240287132073e-05, 'epoch': 0.49})
+[2026-07-28 12:09:28,015][starVLA.training.train_starvla][INFO] - Step 2672, Loss: {'train/loss': 0.023034963756799698, 'train/grad_norm_pre_clip': 1.1097406148910522, 'timing/data': 0.0002765129320323467, 'timing/model': 0.42399179423227906, 'learning_rate/qwen_vl_interface': 8.336332749506872e-06, 'learning_rate/action_model': 8.336332749506872e-05, 'epoch': 0.49})
+[2026-07-28 12:09:34,347][starVLA.training.train_starvla][INFO] - Step 2673, Loss: {'train/loss': 0.052587367594242096, 'train/grad_norm_pre_clip': 1.746290683746338, 'timing/data': 0.0003192787989974022, 'timing/model': 0.6667094239965081, 'learning_rate/qwen_vl_interface': 8.333423209193857e-06, 'learning_rate/action_model': 8.333423209193856e-05, 'epoch': 0.49})
+[2026-07-28 12:09:40,159][starVLA.training.train_starvla][INFO] - Step 2674, Loss: {'train/loss': 0.04457610100507736, 'train/grad_norm_pre_clip': 1.5099437236785889, 'timing/data': 0.00029875198379158974, 'timing/model': 0.4226285978220403, 'learning_rate/qwen_vl_interface': 8.330511668080997e-06, 'learning_rate/action_model': 8.330511668080997e-05, 'epoch': 0.49})
+[2026-07-28 12:09:46,170][starVLA.training.train_starvla][INFO] - Step 2675, Loss: {'train/loss': 0.04011300578713417, 'train/grad_norm_pre_clip': 0.8115183115005493, 'timing/data': 0.0003050612285733223, 'timing/model': 0.4237250378355384, 'learning_rate/qwen_vl_interface': 8.32759812805756e-06, 'learning_rate/action_model': 8.32759812805756e-05, 'epoch': 0.49})
+[2026-07-28 12:09:52,503][starVLA.training.train_starvla][INFO] - Step 2676, Loss: {'train/loss': 0.0431898832321167, 'train/grad_norm_pre_clip': 2.2892954349517822, 'timing/data': 0.0006342683918774128, 'timing/model': 0.4643724630586803, 'learning_rate/qwen_vl_interface': 8.324682591014111e-06, 'learning_rate/action_model': 8.32468259101411e-05, 'epoch': 0.49})
+[2026-07-28 12:09:58,913][starVLA.training.train_starvla][INFO] - Step 2677, Loss: {'train/loss': 0.0382925346493721, 'train/grad_norm_pre_clip': 1.3716634511947632, 'timing/data': 0.00031523313373327255, 'timing/model': 0.43113552033901215, 'learning_rate/qwen_vl_interface': 8.32176505884251e-06, 'learning_rate/action_model': 8.32176505884251e-05, 'epoch': 0.49})
+[2026-07-28 12:10:05,568][starVLA.training.train_starvla][INFO] - Step 2678, Loss: {'train/loss': 0.06218574196100235, 'train/grad_norm_pre_clip': 1.0993379354476929, 'timing/data': 0.00033143721520900726, 'timing/model': 0.4298287509009242, 'learning_rate/qwen_vl_interface': 8.318845533435913e-06, 'learning_rate/action_model': 8.318845533435913e-05, 'epoch': 0.49})
+[2026-07-28 12:10:12,091][starVLA.training.train_starvla][INFO] - Step 2679, Loss: {'train/loss': 0.0167537834495306, 'train/grad_norm_pre_clip': 0.5159543752670288, 'timing/data': 0.00030730897560715675, 'timing/model': 0.42861463967710733, 'learning_rate/qwen_vl_interface': 8.315924016688767e-06, 'learning_rate/action_model': 8.315924016688766e-05, 'epoch': 0.49})
+[2026-07-28 12:10:18,367][starVLA.training.train_starvla][INFO] - Step 2680, Loss: {'train/loss': 0.02253975346684456, 'train/grad_norm_pre_clip': 0.6250663995742798, 'timing/data': 0.00034695304930210114, 'timing/model': 0.4318128228187561, 'learning_rate/qwen_vl_interface': 8.313000510496811e-06, 'learning_rate/action_model': 8.313000510496811e-05, 'epoch': 0.49})
+[2026-07-28 12:10:24,590][starVLA.training.train_starvla][INFO] - Step 2681, Loss: {'train/loss': 0.06061926111578941, 'train/grad_norm_pre_clip': 1.6396136283874512, 'timing/data': 0.0003065369091928005, 'timing/model': 0.43250121595337987, 'learning_rate/qwen_vl_interface': 8.31007501675708e-06, 'learning_rate/action_model': 8.310075016757079e-05, 'epoch': 0.49})
+[2026-07-28 12:10:31,240][starVLA.training.train_starvla][INFO] - Step 2682, Loss: {'train/loss': 0.031030939891934395, 'train/grad_norm_pre_clip': 0.7424171566963196, 'timing/data': 0.0003082309849560261, 'timing/model': 0.43286766903474927, 'learning_rate/qwen_vl_interface': 8.307147537367894e-06, 'learning_rate/action_model': 8.307147537367893e-05, 'epoch': 0.49})
+[2026-07-28 12:10:37,737][starVLA.training.train_starvla][INFO] - Step 2683, Loss: {'train/loss': 0.013023968786001205, 'train/grad_norm_pre_clip': 0.5729008316993713, 'timing/data': 0.0003086850047111511, 'timing/model': 0.4335671290755272, 'learning_rate/qwen_vl_interface': 8.30421807422886e-06, 'learning_rate/action_model': 8.30421807422886e-05, 'epoch': 0.49})
+[2026-07-28 12:10:43,958][starVLA.training.train_starvla][INFO] - Step 2684, Loss: {'train/loss': 0.03167286887764931, 'train/grad_norm_pre_clip': 1.7992910146713257, 'timing/data': 0.00032225390896201134, 'timing/model': 0.43081143125891685, 'learning_rate/qwen_vl_interface': 8.301286629240876e-06, 'learning_rate/action_model': 8.301286629240876e-05, 'epoch': 0.49})
+[2026-07-28 12:10:50,028][starVLA.training.train_starvla][INFO] - Step 2685, Loss: {'train/loss': 0.04664850980043411, 'train/grad_norm_pre_clip': 2.0139639377593994, 'timing/data': 0.0002974807284772396, 'timing/model': 0.432057524099946, 'learning_rate/qwen_vl_interface': 8.298353204306127e-06, 'learning_rate/action_model': 8.298353204306127e-05, 'epoch': 0.49})
+[2026-07-28 12:10:56,193][starVLA.training.train_starvla][INFO] - Step 2686, Loss: {'train/loss': 0.06692568957805634, 'train/grad_norm_pre_clip': 2.2440688610076904, 'timing/data': 0.0003383960574865341, 'timing/model': 0.428497729357332, 'learning_rate/qwen_vl_interface': 8.295417801328081e-06, 'learning_rate/action_model': 8.29541780132808e-05, 'epoch': 0.49})
+[2026-07-28 12:11:02,781][starVLA.training.train_starvla][INFO] - Step 2687, Loss: {'train/loss': 0.05729302391409874, 'train/grad_norm_pre_clip': 1.8160005807876587, 'timing/data': 0.0004275050014257431, 'timing/model': 0.7524056420661509, 'learning_rate/qwen_vl_interface': 8.292480422211486e-06, 'learning_rate/action_model': 8.292480422211485e-05, 'epoch': 0.49})
+[2026-07-28 12:11:09,066][starVLA.training.train_starvla][INFO] - Step 2688, Loss: {'train/loss': 0.06148315966129303, 'train/grad_norm_pre_clip': 1.424208641052246, 'timing/data': 0.0004299166612327099, 'timing/model': 0.43403236381709576, 'learning_rate/qwen_vl_interface': 8.289541068862377e-06, 'learning_rate/action_model': 8.289541068862377e-05, 'epoch': 0.49})
+[2026-07-28 12:11:15,420][starVLA.training.train_starvla][INFO] - Step 2689, Loss: {'train/loss': 0.06418959051370621, 'train/grad_norm_pre_clip': 1.4002937078475952, 'timing/data': 0.00030848802998661995, 'timing/model': 0.604860161896795, 'learning_rate/qwen_vl_interface': 8.286599743188068e-06, 'learning_rate/action_model': 8.286599743188069e-05, 'epoch': 0.49})
+[2026-07-28 12:11:21,583][starVLA.training.train_starvla][INFO] - Step 2690, Loss: {'train/loss': 0.04816950857639313, 'train/grad_norm_pre_clip': 1.5833263397216797, 'timing/data': 0.0004113009199500084, 'timing/model': 0.43069408973678946, 'learning_rate/qwen_vl_interface': 8.283656447097154e-06, 'learning_rate/action_model': 8.283656447097153e-05, 'epoch': 0.49})
+[2026-07-28 12:11:27,754][starVLA.training.train_starvla][INFO] - Step 2691, Loss: {'train/loss': 0.030285144224762917, 'train/grad_norm_pre_clip': 1.753616452217102, 'timing/data': 0.00038451096042990685, 'timing/model': 0.43595915380865335, 'learning_rate/qwen_vl_interface': 8.280711182499507e-06, 'learning_rate/action_model': 8.280711182499506e-05, 'epoch': 0.49})
+[2026-07-28 12:11:33,940][starVLA.training.train_starvla][INFO] - Step 2692, Loss: {'train/loss': 0.046779509633779526, 'train/grad_norm_pre_clip': 1.400595784187317, 'timing/data': 0.0008159196004271507, 'timing/model': 0.42698149802163243, 'learning_rate/qwen_vl_interface': 8.277763951306277e-06, 'learning_rate/action_model': 8.277763951306277e-05, 'epoch': 0.49})
+[2026-07-28 12:11:40,358][starVLA.training.train_starvla][INFO] - Step 2693, Loss: {'train/loss': 0.05869288742542267, 'train/grad_norm_pre_clip': 1.2178475856781006, 'timing/data': 0.00033807195723056793, 'timing/model': 0.4335081116296351, 'learning_rate/qwen_vl_interface': 8.274814755429892e-06, 'learning_rate/action_model': 8.274814755429891e-05, 'epoch': 0.49})
+[2026-07-28 12:11:47,050][starVLA.training.train_starvla][INFO] - Step 2694, Loss: {'train/loss': 0.04529327154159546, 'train/grad_norm_pre_clip': 1.2230726480484009, 'timing/data': 0.00032919924706220627, 'timing/model': 0.42983871791511774, 'learning_rate/qwen_vl_interface': 8.271863596784053e-06, 'learning_rate/action_model': 8.271863596784051e-05, 'epoch': 0.49})
+[2026-07-28 12:11:53,190][starVLA.training.train_starvla][INFO] - Step 2695, Loss: {'train/loss': 0.027707593515515327, 'train/grad_norm_pre_clip': 0.6180218458175659, 'timing/data': 0.0004560723900794983, 'timing/model': 0.43207281781360507, 'learning_rate/qwen_vl_interface': 8.268910477283732e-06, 'learning_rate/action_model': 8.26891047728373e-05, 'epoch': 0.5})
+[2026-07-28 12:11:59,413][starVLA.training.train_starvla][INFO] - Step 2696, Loss: {'train/loss': 0.02429591491818428, 'train/grad_norm_pre_clip': 0.40989959239959717, 'timing/data': 0.00030546030029654503, 'timing/model': 0.4501092708669603, 'learning_rate/qwen_vl_interface': 8.265955398845177e-06, 'learning_rate/action_model': 8.265955398845177e-05, 'epoch': 0.5})
+[2026-07-28 12:12:05,864][starVLA.training.train_starvla][INFO] - Step 2697, Loss: {'train/loss': 0.04979131370782852, 'train/grad_norm_pre_clip': 0.9974826574325562, 'timing/data': 0.0003101471811532974, 'timing/model': 0.6632173070684075, 'learning_rate/qwen_vl_interface': 8.26299836338591e-06, 'learning_rate/action_model': 8.26299836338591e-05, 'epoch': 0.5})
+[2026-07-28 12:12:12,340][starVLA.training.train_starvla][INFO] - Step 2698, Loss: {'train/loss': 0.03215363994240761, 'train/grad_norm_pre_clip': 0.7770006060600281, 'timing/data': 0.0003089890815317631, 'timing/model': 0.7224600911140442, 'learning_rate/qwen_vl_interface': 8.260039372824715e-06, 'learning_rate/action_model': 8.260039372824715e-05, 'epoch': 0.5})
+[2026-07-28 12:12:18,552][starVLA.training.train_starvla][INFO] - Step 2699, Loss: {'train/loss': 0.08160969614982605, 'train/grad_norm_pre_clip': 1.8993757963180542, 'timing/data': 0.0005069500766694546, 'timing/model': 0.4380391361191869, 'learning_rate/qwen_vl_interface': 8.257078429081652e-06, 'learning_rate/action_model': 8.257078429081651e-05, 'epoch': 0.5})
+[2026-07-28 12:12:24,576][starVLA.training.train_starvla][INFO] - Step 2700, Loss: {'train/loss': 0.052264824509620667, 'train/grad_norm_pre_clip': 1.3949906826019287, 'timing/data': 0.00037987902760505676, 'timing/model': 0.44245778769254684, 'learning_rate/qwen_vl_interface': 8.254115534078041e-06, 'learning_rate/action_model': 8.254115534078042e-05, 'epoch': 0.5})
+[2026-07-28 12:12:30,836][starVLA.training.train_starvla][INFO] - Step 2701, Loss: {'train/loss': 0.08574950695037842, 'train/grad_norm_pre_clip': 1.7537469863891602, 'timing/data': 0.00034602684900164604, 'timing/model': 0.43563003465533257, 'learning_rate/qwen_vl_interface': 8.25115068973648e-06, 'learning_rate/action_model': 8.25115068973648e-05, 'epoch': 0.5})
+[2026-07-28 12:12:37,366][starVLA.training.train_starvla][INFO] - Step 2702, Loss: {'train/loss': 0.04838084056973457, 'train/grad_norm_pre_clip': 1.6069624423980713, 'timing/data': 0.00030516600236296654, 'timing/model': 0.4317201632075012, 'learning_rate/qwen_vl_interface': 8.248183897980815e-06, 'learning_rate/action_model': 8.248183897980815e-05, 'epoch': 0.5})
+[2026-07-28 12:12:43,849][starVLA.training.train_starvla][INFO] - Step 2703, Loss: {'train/loss': 0.08765946328639984, 'train/grad_norm_pre_clip': 2.715263843536377, 'timing/data': 0.0003376849927008152, 'timing/model': 0.43423130875453353, 'learning_rate/qwen_vl_interface': 8.24521516073617e-06, 'learning_rate/action_model': 8.24521516073617e-05, 'epoch': 0.5})
+[2026-07-28 12:12:50,057][starVLA.training.train_starvla][INFO] - Step 2704, Loss: {'train/loss': 0.05561119318008423, 'train/grad_norm_pre_clip': 1.071560263633728, 'timing/data': 0.0003079483285546303, 'timing/model': 0.4284740719012916, 'learning_rate/qwen_vl_interface': 8.24224447992893e-06, 'learning_rate/action_model': 8.242244479928928e-05, 'epoch': 0.5})
+[2026-07-28 12:12:56,207][starVLA.training.train_starvla][INFO] - Step 2705, Loss: {'train/loss': 0.03189128637313843, 'train/grad_norm_pre_clip': 0.7103023529052734, 'timing/data': 0.000332775991410017, 'timing/model': 0.4319441788829863, 'learning_rate/qwen_vl_interface': 8.23927185748673e-06, 'learning_rate/action_model': 8.239271857486729e-05, 'epoch': 0.5})
+[2026-07-28 12:13:02,709][starVLA.training.train_starvla][INFO] - Step 2706, Loss: {'train/loss': 0.023601669818162918, 'train/grad_norm_pre_clip': 0.6723728775978088, 'timing/data': 0.00031366292387247086, 'timing/model': 0.42810151912271976, 'learning_rate/qwen_vl_interface': 8.236297295338474e-06, 'learning_rate/action_model': 8.236297295338474e-05, 'epoch': 0.5})
+[2026-07-28 12:13:09,178][starVLA.training.train_starvla][INFO] - Step 2707, Loss: {'train/loss': 0.05305492505431175, 'train/grad_norm_pre_clip': 1.666999340057373, 'timing/data': 0.0003373827785253525, 'timing/model': 0.42852549720555544, 'learning_rate/qwen_vl_interface': 8.233320795414327e-06, 'learning_rate/action_model': 8.233320795414327e-05, 'epoch': 0.5})
+[2026-07-28 12:13:15,381][starVLA.training.train_starvla][INFO] - Step 2708, Loss: {'train/loss': 0.0480465441942215, 'train/grad_norm_pre_clip': 1.2493194341659546, 'timing/data': 0.00032199686393141747, 'timing/model': 0.4334359932690859, 'learning_rate/qwen_vl_interface': 8.230342359645705e-06, 'learning_rate/action_model': 8.230342359645705e-05, 'epoch': 0.5})
+[2026-07-28 12:13:21,792][starVLA.training.train_starvla][INFO] - Step 2709, Loss: {'train/loss': 0.055083997547626495, 'train/grad_norm_pre_clip': 1.4621226787567139, 'timing/data': 0.00033197086304426193, 'timing/model': 0.7161809979006648, 'learning_rate/qwen_vl_interface': 8.227361989965285e-06, 'learning_rate/action_model': 8.227361989965284e-05, 'epoch': 0.5})
+[2026-07-28 12:13:28,157][starVLA.training.train_starvla][INFO] - Step 2710, Loss: {'train/loss': 0.09301160275936127, 'train/grad_norm_pre_clip': 2.449591875076294, 'timing/data': 0.0002959594130516052, 'timing/model': 0.6686000409536064, 'learning_rate/qwen_vl_interface': 8.22437968830699e-06, 'learning_rate/action_model': 8.22437968830699e-05, 'epoch': 0.5})
+[2026-07-28 12:13:34,274][starVLA.training.train_starvla][INFO] - Step 2711, Loss: {'train/loss': 0.042112112045288086, 'train/grad_norm_pre_clip': 1.489496111869812, 'timing/data': 0.00028457120060920715, 'timing/model': 0.4288626699708402, 'learning_rate/qwen_vl_interface': 8.221395456606012e-06, 'learning_rate/action_model': 8.221395456606011e-05, 'epoch': 0.5})
+[2026-07-28 12:13:40,390][starVLA.training.train_starvla][INFO] - Step 2712, Loss: {'train/loss': 0.03663886338472366, 'train/grad_norm_pre_clip': 0.6123939156532288, 'timing/data': 0.000491805374622345, 'timing/model': 0.4296615435741842, 'learning_rate/qwen_vl_interface': 8.21840929679878e-06, 'learning_rate/action_model': 8.218409296798778e-05, 'epoch': 0.5})
+[2026-07-28 12:13:46,545][starVLA.training.train_starvla][INFO] - Step 2713, Loss: {'train/loss': 0.06383933126926422, 'train/grad_norm_pre_clip': 1.0985850095748901, 'timing/data': 0.0003143511712551117, 'timing/model': 0.4249511999078095, 'learning_rate/qwen_vl_interface': 8.215421210822982e-06, 'learning_rate/action_model': 8.215421210822982e-05, 'epoch': 0.5})
+[2026-07-28 12:13:52,719][starVLA.training.train_starvla][INFO] - Step 2714, Loss: {'train/loss': 0.06710493564605713, 'train/grad_norm_pre_clip': 3.095858335494995, 'timing/data': 0.00028704898431897163, 'timing/model': 0.6766744200140238, 'learning_rate/qwen_vl_interface': 8.212431200617557e-06, 'learning_rate/action_model': 8.212431200617556e-05, 'epoch': 0.5})
+[2026-07-28 12:13:58,546][starVLA.training.train_starvla][INFO] - Step 2715, Loss: {'train/loss': 0.03300098702311516, 'train/grad_norm_pre_clip': 1.0381876230239868, 'timing/data': 0.000330441165715456, 'timing/model': 0.4280229830183089, 'learning_rate/qwen_vl_interface': 8.209439268122689e-06, 'learning_rate/action_model': 8.209439268122688e-05, 'epoch': 0.5})
+[2026-07-28 12:14:04,378][starVLA.training.train_starvla][INFO] - Step 2716, Loss: {'train/loss': 0.04591796547174454, 'train/grad_norm_pre_clip': 1.2974588871002197, 'timing/data': 0.0003623180091381073, 'timing/model': 0.428678372874856, 'learning_rate/qwen_vl_interface': 8.206445415279808e-06, 'learning_rate/action_model': 8.206445415279807e-05, 'epoch': 0.5})
+[2026-07-28 12:14:10,467][starVLA.training.train_starvla][INFO] - Step 2717, Loss: {'train/loss': 0.034709587693214417, 'train/grad_norm_pre_clip': 0.7999362349510193, 'timing/data': 0.0003878418356180191, 'timing/model': 0.43801618507131934, 'learning_rate/qwen_vl_interface': 8.203449644031595e-06, 'learning_rate/action_model': 8.203449644031596e-05, 'epoch': 0.5})
+[2026-07-28 12:14:16,523][starVLA.training.train_starvla][INFO] - Step 2718, Loss: {'train/loss': 0.058265767991542816, 'train/grad_norm_pre_clip': 1.330884337425232, 'timing/data': 0.0003704531118273735, 'timing/model': 0.4332635961472988, 'learning_rate/qwen_vl_interface': 8.200451956321974e-06, 'learning_rate/action_model': 8.200451956321974e-05, 'epoch': 0.5})
+[2026-07-28 12:14:22,574][starVLA.training.train_starvla][INFO] - Step 2719, Loss: {'train/loss': 0.03338714316487312, 'train/grad_norm_pre_clip': 1.2493380308151245, 'timing/data': 0.0007088868878781796, 'timing/model': 0.4254251252859831, 'learning_rate/qwen_vl_interface': 8.197452354096112e-06, 'learning_rate/action_model': 8.197452354096112e-05, 'epoch': 0.5})
+[2026-07-28 12:14:29,196][starVLA.training.train_starvla][INFO] - Step 2720, Loss: {'train/loss': 0.08726153522729874, 'train/grad_norm_pre_clip': 1.494480848312378, 'timing/data': 0.0002843323163688183, 'timing/model': 0.7197581618092954, 'learning_rate/qwen_vl_interface': 8.194450839300418e-06, 'learning_rate/action_model': 8.194450839300417e-05, 'epoch': 0.5})
+[2026-07-28 12:14:35,309][starVLA.training.train_starvla][INFO] - Step 2721, Loss: {'train/loss': 0.03571256250143051, 'train/grad_norm_pre_clip': 2.38238263130188, 'timing/data': 0.00026861298829317093, 'timing/model': 0.42388255102559924, 'learning_rate/qwen_vl_interface': 8.191447413882539e-06, 'learning_rate/action_model': 8.191447413882539e-05, 'epoch': 0.5})
+[2026-07-28 12:14:41,420][starVLA.training.train_starvla][INFO] - Step 2722, Loss: {'train/loss': 0.03873880207538605, 'train/grad_norm_pre_clip': 1.3845759630203247, 'timing/data': 0.0002912641502916813, 'timing/model': 0.42712846398353577, 'learning_rate/qwen_vl_interface': 8.188442079791369e-06, 'learning_rate/action_model': 8.188442079791369e-05, 'epoch': 0.5})
+[2026-07-28 12:14:47,513][starVLA.training.train_starvla][INFO] - Step 2723, Loss: {'train/loss': 0.04179486632347107, 'train/grad_norm_pre_clip': 2.47322940826416, 'timing/data': 0.00029591284692287445, 'timing/model': 0.42639116011559963, 'learning_rate/qwen_vl_interface': 8.185434838977036e-06, 'learning_rate/action_model': 8.185434838977037e-05, 'epoch': 0.5})
+[2026-07-28 12:14:53,825][starVLA.training.train_starvla][INFO] - Step 2724, Loss: {'train/loss': 0.037914078682661057, 'train/grad_norm_pre_clip': 0.9884044528007507, 'timing/data': 0.00029050372540950775, 'timing/model': 0.4254529387690127, 'learning_rate/qwen_vl_interface': 8.182425693390905e-06, 'learning_rate/action_model': 8.182425693390905e-05, 'epoch': 0.5})
+[2026-07-28 12:15:00,214][starVLA.training.train_starvla][INFO] - Step 2725, Loss: {'train/loss': 0.03840094804763794, 'train/grad_norm_pre_clip': 0.9525720477104187, 'timing/data': 0.00029107602313160896, 'timing/model': 0.4229564219713211, 'learning_rate/qwen_vl_interface': 8.179414644985582e-06, 'learning_rate/action_model': 8.179414644985582e-05, 'epoch': 0.5})
+[2026-07-28 12:15:06,305][starVLA.training.train_starvla][INFO] - Step 2726, Loss: {'train/loss': 0.03863023966550827, 'train/grad_norm_pre_clip': 1.3298916816711426, 'timing/data': 0.0002885316498577595, 'timing/model': 0.43000267585739493, 'learning_rate/qwen_vl_interface': 8.176401695714897e-06, 'learning_rate/action_model': 8.176401695714895e-05, 'epoch': 0.5})
+[2026-07-28 12:15:12,590][starVLA.training.train_starvla][INFO] - Step 2727, Loss: {'train/loss': 0.028825068846344948, 'train/grad_norm_pre_clip': 0.9132347702980042, 'timing/data': 0.00027866801247000694, 'timing/model': 0.4225133457221091, 'learning_rate/qwen_vl_interface': 8.173386847533922e-06, 'learning_rate/action_model': 8.173386847533921e-05, 'epoch': 0.5})
+[2026-07-28 12:15:18,639][starVLA.training.train_starvla][INFO] - Step 2728, Loss: {'train/loss': 0.03521503135561943, 'train/grad_norm_pre_clip': 1.4148460626602173, 'timing/data': 0.00029122503474354744, 'timing/model': 0.4260344170033932, 'learning_rate/qwen_vl_interface': 8.170370102398962e-06, 'learning_rate/action_model': 8.17037010239896e-05, 'epoch': 0.5})
+[2026-07-28 12:15:24,985][starVLA.training.train_starvla][INFO] - Step 2729, Loss: {'train/loss': 0.046495355665683746, 'train/grad_norm_pre_clip': 1.1178512573242188, 'timing/data': 0.00034179212525486946, 'timing/model': 0.4223244721069932, 'learning_rate/qwen_vl_interface': 8.167351462267548e-06, 'learning_rate/action_model': 8.167351462267548e-05, 'epoch': 0.5})
+[2026-07-28 12:15:31,048][starVLA.training.train_starvla][INFO] - Step 2730, Loss: {'train/loss': 0.0444505475461483, 'train/grad_norm_pre_clip': 1.2533509731292725, 'timing/data': 0.00029133306816220284, 'timing/model': 0.42451584432274103, 'learning_rate/qwen_vl_interface': 8.164330929098444e-06, 'learning_rate/action_model': 8.164330929098443e-05, 'epoch': 0.5})
+[2026-07-28 12:15:37,354][starVLA.training.train_starvla][INFO] - Step 2731, Loss: {'train/loss': 0.018455615267157555, 'train/grad_norm_pre_clip': 0.5007951259613037, 'timing/data': 0.0003618909977376461, 'timing/model': 0.43335088388994336, 'learning_rate/qwen_vl_interface': 8.16130850485164e-06, 'learning_rate/action_model': 8.161308504851639e-05, 'epoch': 0.5})
+[2026-07-28 12:15:43,450][starVLA.training.train_starvla][INFO] - Step 2732, Loss: {'train/loss': 0.06206031143665314, 'train/grad_norm_pre_clip': 3.189847707748413, 'timing/data': 0.0002857698127627373, 'timing/model': 0.4260240951552987, 'learning_rate/qwen_vl_interface': 8.158284191488355e-06, 'learning_rate/action_model': 8.158284191488355e-05, 'epoch': 0.5})
+[2026-07-28 12:15:49,590][starVLA.training.train_starvla][INFO] - Step 2733, Loss: {'train/loss': 0.04143974184989929, 'train/grad_norm_pre_clip': 1.4626537561416626, 'timing/data': 0.0002867328003048897, 'timing/model': 0.4254533792845905, 'learning_rate/qwen_vl_interface': 8.155257990971032e-06, 'learning_rate/action_model': 8.155257990971031e-05, 'epoch': 0.5})
+[2026-07-28 12:15:55,893][starVLA.training.train_starvla][INFO] - Step 2734, Loss: {'train/loss': 0.040788255631923676, 'train/grad_norm_pre_clip': 2.2829763889312744, 'timing/data': 0.00034898286685347557, 'timing/model': 0.6862699659541249, 'learning_rate/qwen_vl_interface': 8.152229905263341e-06, 'learning_rate/action_model': 8.152229905263342e-05, 'epoch': 0.5})
+[2026-07-28 12:16:02,218][starVLA.training.train_starvla][INFO] - Step 2735, Loss: {'train/loss': 0.04611952602863312, 'train/grad_norm_pre_clip': 0.9700338840484619, 'timing/data': 0.0003269929438829422, 'timing/model': 0.42519752820953727, 'learning_rate/qwen_vl_interface': 8.149199936330175e-06, 'learning_rate/action_model': 8.149199936330174e-05, 'epoch': 0.5})
+[2026-07-28 12:16:08,310][starVLA.training.train_starvla][INFO] - Step 2736, Loss: {'train/loss': 0.02558639459311962, 'train/grad_norm_pre_clip': 1.0181033611297607, 'timing/data': 0.0002917270176112652, 'timing/model': 0.4456292549148202, 'learning_rate/qwen_vl_interface': 8.146168086137646e-06, 'learning_rate/action_model': 8.146168086137646e-05, 'epoch': 0.5})
+[2026-07-28 12:16:14,337][starVLA.training.train_starvla][INFO] - Step 2737, Loss: {'train/loss': 0.08686769008636475, 'train/grad_norm_pre_clip': 1.5518951416015625, 'timing/data': 0.000337985809892416, 'timing/model': 0.42581188678741455, 'learning_rate/qwen_vl_interface': 8.143134356653087e-06, 'learning_rate/action_model': 8.143134356653087e-05, 'epoch': 0.5})
+[2026-07-28 12:16:20,162][starVLA.training.train_starvla][INFO] - Step 2738, Loss: {'train/loss': 0.04625251889228821, 'train/grad_norm_pre_clip': 3.541167736053467, 'timing/data': 0.00028735212981700897, 'timing/model': 0.430241531226784, 'learning_rate/qwen_vl_interface': 8.140098749845056e-06, 'learning_rate/action_model': 8.140098749845055e-05, 'epoch': 0.5})
+[2026-07-28 12:16:26,026][starVLA.training.train_starvla][INFO] - Step 2739, Loss: {'train/loss': 0.04432510584592819, 'train/grad_norm_pre_clip': 1.3511401414871216, 'timing/data': 0.0002865847200155258, 'timing/model': 0.43074263678863645, 'learning_rate/qwen_vl_interface': 8.137061267683323e-06, 'learning_rate/action_model': 8.137061267683323e-05, 'epoch': 0.5})
+[2026-07-28 12:16:32,244][starVLA.training.train_starvla][INFO] - Step 2740, Loss: {'train/loss': 0.06057398393750191, 'train/grad_norm_pre_clip': 1.4832185506820679, 'timing/data': 0.00035247066989541054, 'timing/model': 0.6639980240724981, 'learning_rate/qwen_vl_interface': 8.134021912138879e-06, 'learning_rate/action_model': 8.134021912138879e-05, 'epoch': 0.5})
+[2026-07-28 12:16:38,327][starVLA.training.train_starvla][INFO] - Step 2741, Loss: {'train/loss': 0.04704709351062775, 'train/grad_norm_pre_clip': 1.7739003896713257, 'timing/data': 0.0002699699252843857, 'timing/model': 0.42603075318038464, 'learning_rate/qwen_vl_interface': 8.130980685183927e-06, 'learning_rate/action_model': 8.130980685183926e-05, 'epoch': 0.5})
+[2026-07-28 12:16:44,192][starVLA.training.train_starvla][INFO] - Step 2742, Loss: {'train/loss': 0.03764265775680542, 'train/grad_norm_pre_clip': 1.6341753005981445, 'timing/data': 0.00035258103162050247, 'timing/model': 0.424133971799165, 'learning_rate/qwen_vl_interface': 8.127937588791887e-06, 'learning_rate/action_model': 8.127937588791886e-05, 'epoch': 0.5})
+[2026-07-28 12:16:50,258][starVLA.training.train_starvla][INFO] - Step 2743, Loss: {'train/loss': 0.06466954946517944, 'train/grad_norm_pre_clip': 1.1107393503189087, 'timing/data': 0.00029413122683763504, 'timing/model': 0.4299688250757754, 'learning_rate/qwen_vl_interface': 8.12489262493739e-06, 'learning_rate/action_model': 8.12489262493739e-05, 'epoch': 0.5})
+[2026-07-28 12:16:56,331][starVLA.training.train_starvla][INFO] - Step 2744, Loss: {'train/loss': 0.05219652131199837, 'train/grad_norm_pre_clip': 0.6669777631759644, 'timing/data': 0.0003397120162844658, 'timing/model': 0.4318392379209399, 'learning_rate/qwen_vl_interface': 8.121845795596282e-06, 'learning_rate/action_model': 8.121845795596282e-05, 'epoch': 0.5})
+[2026-07-28 12:17:02,922][starVLA.training.train_starvla][INFO] - Step 2745, Loss: {'train/loss': 0.03668712079524994, 'train/grad_norm_pre_clip': 0.9312259554862976, 'timing/data': 0.0003551510162651539, 'timing/model': 0.6729947822168469, 'learning_rate/qwen_vl_interface': 8.118797102745614e-06, 'learning_rate/action_model': 8.118797102745615e-05, 'epoch': 0.5})
+[2026-07-28 12:17:09,006][starVLA.training.train_starvla][INFO] - Step 2746, Loss: {'train/loss': 0.027425549924373627, 'train/grad_norm_pre_clip': 0.7238184213638306, 'timing/data': 0.0005598627030849457, 'timing/model': 0.43027306720614433, 'learning_rate/qwen_vl_interface': 8.115746548363655e-06, 'learning_rate/action_model': 8.115746548363655e-05, 'epoch': 0.5})
+[2026-07-28 12:17:15,124][starVLA.training.train_starvla][INFO] - Step 2747, Loss: {'train/loss': 0.03756962716579437, 'train/grad_norm_pre_clip': 1.0955753326416016, 'timing/data': 0.0003929128870368004, 'timing/model': 0.4292998597957194, 'learning_rate/qwen_vl_interface': 8.112694134429873e-06, 'learning_rate/action_model': 8.112694134429873e-05, 'epoch': 0.5})
+[2026-07-28 12:17:21,523][starVLA.training.train_starvla][INFO] - Step 2748, Loss: {'train/loss': 0.035985592752695084, 'train/grad_norm_pre_clip': 1.1864291429519653, 'timing/data': 0.00036192918196320534, 'timing/model': 0.42713208869099617, 'learning_rate/qwen_vl_interface': 8.109639862924947e-06, 'learning_rate/action_model': 8.109639862924945e-05, 'epoch': 0.5})
+[2026-07-28 12:17:27,857][starVLA.training.train_starvla][INFO] - Step 2749, Loss: {'train/loss': 0.040810585021972656, 'train/grad_norm_pre_clip': 2.31498384475708, 'timing/data': 0.00027587777003645897, 'timing/model': 0.42264543985947967, 'learning_rate/qwen_vl_interface': 8.10658373583076e-06, 'learning_rate/action_model': 8.106583735830761e-05, 'epoch': 0.5})
+[2026-07-28 12:17:33,937][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 2750
+[2026-07-28 12:17:39,095][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 2750
+[2026-07-28 12:26:05,712][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 2750
+[2026-07-28 12:26:05,714][starVLA.training.train_starvla][INFO] - Step 2750, Loss: {'train/loss': 0.04168377071619034, 'train/grad_norm_pre_clip': 0.8490076661109924, 'eval/action_loss/seconds': 5.1552121406421065, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.03266891837120056, 'eval/latency_3/loss': 0.03266891837120056, 'eval/flappy/loss': 0.03266891837120056, 'eval/flappy/latency_3/loss': 0.03266891837120056, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 391.8000046081841, 'rl_games_eval/mid_train/mean_length': 3175.55, 'rl_games_eval/mid_train/std_reward': 112.64322126288126, 'rl_games_eval/mid_train/std_length': 906.7831314597776, 'rl_games_eval/mid_train/macro_mean_reward': 391.8000046081841, 'rl_games_eval/mid_train/macro_mean_length': 3175.55, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 391.8000046081841, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 3175.55, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 112.64322126288126, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 906.7831314597776, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 391.8000046081841, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00028537213802337646, 'timing/model': 0.42482467787340283, 'learning_rate/qwen_vl_interface': 8.1035257551304e-06, 'learning_rate/action_model': 8.1035257551304e-05, 'epoch': 0.51})
+[2026-07-28 12:26:12,184][starVLA.training.train_starvla][INFO] - Step 2751, Loss: {'train/loss': 0.03624250739812851, 'train/grad_norm_pre_clip': 0.7226808667182922, 'timing/data': 0.0003679618239402771, 'timing/model': 0.6619743620976806, 'learning_rate/qwen_vl_interface': 8.100465922808158e-06, 'learning_rate/action_model': 8.100465922808159e-05, 'epoch': 0.51})
+[2026-07-28 12:26:18,344][starVLA.training.train_starvla][INFO] - Step 2752, Loss: {'train/loss': 0.055828068405389786, 'train/grad_norm_pre_clip': 0.7782729268074036, 'timing/data': 0.000291263684630394, 'timing/model': 0.44167609605938196, 'learning_rate/qwen_vl_interface': 8.097404240849527e-06, 'learning_rate/action_model': 8.097404240849527e-05, 'epoch': 0.51})
+[2026-07-28 12:26:24,414][starVLA.training.train_starvla][INFO] - Step 2753, Loss: {'train/loss': 0.05733244866132736, 'train/grad_norm_pre_clip': 1.5287293195724487, 'timing/data': 0.0003506150096654892, 'timing/model': 0.432369542773813, 'learning_rate/qwen_vl_interface': 8.094340711241198e-06, 'learning_rate/action_model': 8.094340711241198e-05, 'epoch': 0.51})
+[2026-07-28 12:26:30,495][starVLA.training.train_starvla][INFO] - Step 2754, Loss: {'train/loss': 0.045124538242816925, 'train/grad_norm_pre_clip': 1.1197618246078491, 'timing/data': 0.00027533574029803276, 'timing/model': 0.4258997980505228, 'learning_rate/qwen_vl_interface': 8.09127533597106e-06, 'learning_rate/action_model': 8.091275335971061e-05, 'epoch': 0.51})
+[2026-07-28 12:26:37,061][starVLA.training.train_starvla][INFO] - Step 2755, Loss: {'train/loss': 0.03375483676791191, 'train/grad_norm_pre_clip': 0.840231716632843, 'timing/data': 0.0002754777669906616, 'timing/model': 0.6645417278632522, 'learning_rate/qwen_vl_interface': 8.088208117028205e-06, 'learning_rate/action_model': 8.088208117028206e-05, 'epoch': 0.51})
+[2026-07-28 12:26:43,104][starVLA.training.train_starvla][INFO] - Step 2756, Loss: {'train/loss': 0.0590890534222126, 'train/grad_norm_pre_clip': 1.5121703147888184, 'timing/data': 0.0002767699770629406, 'timing/model': 0.4268912277184427, 'learning_rate/qwen_vl_interface': 8.085139056402918e-06, 'learning_rate/action_model': 8.085139056402917e-05, 'epoch': 0.51})
+[2026-07-28 12:26:49,181][starVLA.training.train_starvla][INFO] - Step 2757, Loss: {'train/loss': 0.023390688002109528, 'train/grad_norm_pre_clip': 0.6807359457015991, 'timing/data': 0.0003051762469112873, 'timing/model': 0.4262210773304105, 'learning_rate/qwen_vl_interface': 8.082068156086675e-06, 'learning_rate/action_model': 8.082068156086675e-05, 'epoch': 0.51})
+[2026-07-28 12:26:55,500][starVLA.training.train_starvla][INFO] - Step 2758, Loss: {'train/loss': 0.05783012881875038, 'train/grad_norm_pre_clip': 2.2595298290252686, 'timing/data': 0.0002913479693233967, 'timing/model': 0.6681342208757997, 'learning_rate/qwen_vl_interface': 8.078995418072156e-06, 'learning_rate/action_model': 8.078995418072157e-05, 'epoch': 0.51})
+[2026-07-28 12:27:01,542][starVLA.training.train_starvla][INFO] - Step 2759, Loss: {'train/loss': 0.06036495044827461, 'train/grad_norm_pre_clip': 2.227689266204834, 'timing/data': 0.00027113500982522964, 'timing/model': 0.42498929612338543, 'learning_rate/qwen_vl_interface': 8.075920844353223e-06, 'learning_rate/action_model': 8.075920844353223e-05, 'epoch': 0.51})
+[2026-07-28 12:27:07,609][starVLA.training.train_starvla][INFO] - Step 2760, Loss: {'train/loss': 0.02639211341738701, 'train/grad_norm_pre_clip': 0.8628944158554077, 'timing/data': 0.00038910098373889923, 'timing/model': 0.43612638395279646, 'learning_rate/qwen_vl_interface': 8.072844436924934e-06, 'learning_rate/action_model': 8.072844436924934e-05, 'epoch': 0.51})
+[2026-07-28 12:27:13,657][starVLA.training.train_starvla][INFO] - Step 2761, Loss: {'train/loss': 0.023629657924175262, 'train/grad_norm_pre_clip': 0.8941152691841125, 'timing/data': 0.00027268892154097557, 'timing/model': 0.42526858719065785, 'learning_rate/qwen_vl_interface': 8.069766197783537e-06, 'learning_rate/action_model': 8.069766197783537e-05, 'epoch': 0.51})
+[2026-07-28 12:27:19,956][starVLA.training.train_starvla][INFO] - Step 2762, Loss: {'train/loss': 0.04304599016904831, 'train/grad_norm_pre_clip': 2.8054606914520264, 'timing/data': 0.00027874577790498734, 'timing/model': 0.657707488629967, 'learning_rate/qwen_vl_interface': 8.066686128926468e-06, 'learning_rate/action_model': 8.066686128926468e-05, 'epoch': 0.51})
+[2026-07-28 12:27:26,245][starVLA.training.train_starvla][INFO] - Step 2763, Loss: {'train/loss': 0.045147985219955444, 'train/grad_norm_pre_clip': 0.9208291172981262, 'timing/data': 0.0002822084352374077, 'timing/model': 0.6848837868310511, 'learning_rate/qwen_vl_interface': 8.063604232352352e-06, 'learning_rate/action_model': 8.06360423235235e-05, 'epoch': 0.51})
+[2026-07-28 12:27:32,260][starVLA.training.train_starvla][INFO] - Step 2764, Loss: {'train/loss': 0.04063429683446884, 'train/grad_norm_pre_clip': 1.3526197671890259, 'timing/data': 0.00026355916634202003, 'timing/model': 0.4243310699239373, 'learning_rate/qwen_vl_interface': 8.060520510060996e-06, 'learning_rate/action_model': 8.060520510060995e-05, 'epoch': 0.51})
+[2026-07-28 12:27:38,240][starVLA.training.train_starvla][INFO] - Step 2765, Loss: {'train/loss': 0.05787091329693794, 'train/grad_norm_pre_clip': 2.166949510574341, 'timing/data': 0.00027617672458291054, 'timing/model': 0.42014643363654613, 'learning_rate/qwen_vl_interface': 8.057434964053391e-06, 'learning_rate/action_model': 8.057434964053392e-05, 'epoch': 0.51})
+[2026-07-28 12:27:44,623][starVLA.training.train_starvla][INFO] - Step 2766, Loss: {'train/loss': 0.04392518475651741, 'train/grad_norm_pre_clip': 1.177902340888977, 'timing/data': 0.00030555296689271927, 'timing/model': 0.44139676820486784, 'learning_rate/qwen_vl_interface': 8.05434759633172e-06, 'learning_rate/action_model': 8.054347596331721e-05, 'epoch': 0.51})
+[2026-07-28 12:27:50,607][starVLA.training.train_starvla][INFO] - Step 2767, Loss: {'train/loss': 0.021609794348478317, 'train/grad_norm_pre_clip': 0.6322976350784302, 'timing/data': 0.00026927608996629715, 'timing/model': 0.42151780938729644, 'learning_rate/qwen_vl_interface': 8.05125840889934e-06, 'learning_rate/action_model': 8.05125840889934e-05, 'epoch': 0.51})
+[2026-07-28 12:27:56,607][starVLA.training.train_starvla][INFO] - Step 2768, Loss: {'train/loss': 0.045413579791784286, 'train/grad_norm_pre_clip': 1.5634320974349976, 'timing/data': 0.0002795690670609474, 'timing/model': 0.4227326330728829, 'learning_rate/qwen_vl_interface': 8.048167403760792e-06, 'learning_rate/action_model': 8.048167403760792e-05, 'epoch': 0.51})
+[2026-07-28 12:28:03,072][starVLA.training.train_starvla][INFO] - Step 2769, Loss: {'train/loss': 0.022469595074653625, 'train/grad_norm_pre_clip': 0.6396575570106506, 'timing/data': 0.00027860328555107117, 'timing/model': 0.6603785930201411, 'learning_rate/qwen_vl_interface': 8.045074582921797e-06, 'learning_rate/action_model': 8.045074582921797e-05, 'epoch': 0.51})
+[2026-07-28 12:28:09,084][starVLA.training.train_starvla][INFO] - Step 2770, Loss: {'train/loss': 0.02837805449962616, 'train/grad_norm_pre_clip': 1.0097005367279053, 'timing/data': 0.000381366815418005, 'timing/model': 0.4250442902557552, 'learning_rate/qwen_vl_interface': 8.041979948389248e-06, 'learning_rate/action_model': 8.041979948389248e-05, 'epoch': 0.51})
+[2026-07-28 12:28:15,147][starVLA.training.train_starvla][INFO] - Step 2771, Loss: {'train/loss': 0.053813133388757706, 'train/grad_norm_pre_clip': 1.9650095701217651, 'timing/data': 0.0002861609682440758, 'timing/model': 0.4234956088475883, 'learning_rate/qwen_vl_interface': 8.038883502171223e-06, 'learning_rate/action_model': 8.038883502171223e-05, 'epoch': 0.51})
+[2026-07-28 12:28:21,472][starVLA.training.train_starvla][INFO] - Step 2772, Loss: {'train/loss': 0.08336634933948517, 'train/grad_norm_pre_clip': 1.3979928493499756, 'timing/data': 0.0002722619101405144, 'timing/model': 0.6621544379740953, 'learning_rate/qwen_vl_interface': 8.035785246276973e-06, 'learning_rate/action_model': 8.035785246276973e-05, 'epoch': 0.51})
+[2026-07-28 12:28:27,524][starVLA.training.train_starvla][INFO] - Step 2773, Loss: {'train/loss': 0.02754964306950569, 'train/grad_norm_pre_clip': 0.4979175329208374, 'timing/data': 0.0007516872137784958, 'timing/model': 0.4228996499441564, 'learning_rate/qwen_vl_interface': 8.03268518271692e-06, 'learning_rate/action_model': 8.032685182716919e-05, 'epoch': 0.51})
+[2026-07-28 12:28:33,797][starVLA.training.train_starvla][INFO] - Step 2774, Loss: {'train/loss': 0.06633538752794266, 'train/grad_norm_pre_clip': 1.4780055284500122, 'timing/data': 0.0002821902744472027, 'timing/model': 0.4246505359187722, 'learning_rate/qwen_vl_interface': 8.029583313502664e-06, 'learning_rate/action_model': 8.029583313502664e-05, 'epoch': 0.51})
+[2026-07-28 12:28:39,899][starVLA.training.train_starvla][INFO] - Step 2775, Loss: {'train/loss': 0.05416873097419739, 'train/grad_norm_pre_clip': 2.451353073120117, 'timing/data': 0.00035774894058704376, 'timing/model': 0.42634887620806694, 'learning_rate/qwen_vl_interface': 8.026479640646974e-06, 'learning_rate/action_model': 8.026479640646973e-05, 'epoch': 0.51})
+[2026-07-28 12:28:46,196][starVLA.training.train_starvla][INFO] - Step 2776, Loss: {'train/loss': 0.053545594215393066, 'train/grad_norm_pre_clip': 1.0327824354171753, 'timing/data': 0.00026994384825229645, 'timing/model': 0.4213397200219333, 'learning_rate/qwen_vl_interface': 8.023374166163787e-06, 'learning_rate/action_model': 8.023374166163787e-05, 'epoch': 0.51})
+[2026-07-28 12:28:52,197][starVLA.training.train_starvla][INFO] - Step 2777, Loss: {'train/loss': 0.045241065323352814, 'train/grad_norm_pre_clip': 1.3240724802017212, 'timing/data': 0.00028633372858166695, 'timing/model': 0.4232898959890008, 'learning_rate/qwen_vl_interface': 8.020266892068214e-06, 'learning_rate/action_model': 8.020266892068214e-05, 'epoch': 0.51})
+[2026-07-28 12:28:58,502][starVLA.training.train_starvla][INFO] - Step 2778, Loss: {'train/loss': 0.047482702881097794, 'train/grad_norm_pre_clip': 1.359161376953125, 'timing/data': 0.0002810000441968441, 'timing/model': 0.42323497496545315, 'learning_rate/qwen_vl_interface': 8.017157820376532e-06, 'learning_rate/action_model': 8.017157820376532e-05, 'epoch': 0.51})
+[2026-07-28 12:29:04,825][starVLA.training.train_starvla][INFO] - Step 2779, Loss: {'train/loss': 0.05810730904340744, 'train/grad_norm_pre_clip': 1.5740965604782104, 'timing/data': 0.0003488278016448021, 'timing/model': 0.42276517022401094, 'learning_rate/qwen_vl_interface': 8.014046953106183e-06, 'learning_rate/action_model': 8.014046953106182e-05, 'epoch': 0.51})
+[2026-07-28 12:29:10,921][starVLA.training.train_starvla][INFO] - Step 2780, Loss: {'train/loss': 0.051155947148799896, 'train/grad_norm_pre_clip': 1.4958628416061401, 'timing/data': 0.0002755173482000828, 'timing/model': 0.43988386960700154, 'learning_rate/qwen_vl_interface': 8.010934292275775e-06, 'learning_rate/action_model': 8.010934292275775e-05, 'epoch': 0.51})
+[2026-07-28 12:29:16,906][starVLA.training.train_starvla][INFO] - Step 2781, Loss: {'train/loss': 0.050350915640592575, 'train/grad_norm_pre_clip': 1.5606071949005127, 'timing/data': 0.00026416731998324394, 'timing/model': 0.4210634776391089, 'learning_rate/qwen_vl_interface': 8.007819839905079e-06, 'learning_rate/action_model': 8.007819839905078e-05, 'epoch': 0.51})
+[2026-07-28 12:29:23,075][starVLA.training.train_starvla][INFO] - Step 2782, Loss: {'train/loss': 0.05428021401166916, 'train/grad_norm_pre_clip': 1.9924079179763794, 'timing/data': 0.0002582930028438568, 'timing/model': 0.5945708528161049, 'learning_rate/qwen_vl_interface': 8.004703598015032e-06, 'learning_rate/action_model': 8.004703598015031e-05, 'epoch': 0.51})
+[2026-07-28 12:29:29,331][starVLA.training.train_starvla][INFO] - Step 2783, Loss: {'train/loss': 0.05735643953084946, 'train/grad_norm_pre_clip': 3.3553993701934814, 'timing/data': 0.00028886133804917336, 'timing/model': 0.6665418641641736, 'learning_rate/qwen_vl_interface': 8.001585568627724e-06, 'learning_rate/action_model': 8.001585568627725e-05, 'epoch': 0.51})
+[2026-07-28 12:29:35,369][starVLA.training.train_starvla][INFO] - Step 2784, Loss: {'train/loss': 0.045139484107494354, 'train/grad_norm_pre_clip': 1.9126622676849365, 'timing/data': 0.0002628699876368046, 'timing/model': 0.4242987297475338, 'learning_rate/qwen_vl_interface': 7.998465753766418e-06, 'learning_rate/action_model': 7.998465753766418e-05, 'epoch': 0.51})
+[2026-07-28 12:29:41,445][starVLA.training.train_starvla][INFO] - Step 2785, Loss: {'train/loss': 0.044751524925231934, 'train/grad_norm_pre_clip': 1.9969760179519653, 'timing/data': 0.00031889695674180984, 'timing/model': 0.42600862169638276, 'learning_rate/qwen_vl_interface': 7.995344155455523e-06, 'learning_rate/action_model': 7.995344155455522e-05, 'epoch': 0.51})
+[2026-07-28 12:29:47,798][starVLA.training.train_starvla][INFO] - Step 2786, Loss: {'train/loss': 0.03694581612944603, 'train/grad_norm_pre_clip': 0.921589195728302, 'timing/data': 0.00029586395248770714, 'timing/model': 0.42435584822669625, 'learning_rate/qwen_vl_interface': 7.992220775720609e-06, 'learning_rate/action_model': 7.992220775720609e-05, 'epoch': 0.51})
+[2026-07-28 12:29:53,851][starVLA.training.train_starvla][INFO] - Step 2787, Loss: {'train/loss': 0.08057355880737305, 'train/grad_norm_pre_clip': 1.668774127960205, 'timing/data': 0.00033553503453731537, 'timing/model': 0.42972803907468915, 'learning_rate/qwen_vl_interface': 7.989095616588409e-06, 'learning_rate/action_model': 7.989095616588408e-05, 'epoch': 0.51})
+[2026-07-28 12:29:59,913][starVLA.training.train_starvla][INFO] - Step 2788, Loss: {'train/loss': 0.04300473630428314, 'train/grad_norm_pre_clip': 1.1412677764892578, 'timing/data': 0.0002728649415075779, 'timing/model': 0.4247335223481059, 'learning_rate/qwen_vl_interface': 7.985968680086799e-06, 'learning_rate/action_model': 7.985968680086798e-05, 'epoch': 0.51})
+[2026-07-28 12:30:05,968][starVLA.training.train_starvla][INFO] - Step 2789, Loss: {'train/loss': 0.04416613280773163, 'train/grad_norm_pre_clip': 1.3142551183700562, 'timing/data': 0.0002802531234920025, 'timing/model': 0.4233529828488827, 'learning_rate/qwen_vl_interface': 7.982839968244816e-06, 'learning_rate/action_model': 7.982839968244816e-05, 'epoch': 0.51})
+[2026-07-28 12:30:12,477][starVLA.training.train_starvla][INFO] - Step 2790, Loss: {'train/loss': 0.033456429839134216, 'train/grad_norm_pre_clip': 0.7099816203117371, 'timing/data': 0.00028127990663051605, 'timing/model': 0.4291759612970054, 'learning_rate/qwen_vl_interface': 7.97970948309265e-06, 'learning_rate/action_model': 7.979709483092649e-05, 'epoch': 0.51})
+[2026-07-28 12:30:18,555][starVLA.training.train_starvla][INFO] - Step 2791, Loss: {'train/loss': 0.05441843718290329, 'train/grad_norm_pre_clip': 2.3212294578552246, 'timing/data': 0.00028769997879862785, 'timing/model': 0.4274791991338134, 'learning_rate/qwen_vl_interface': 7.976577226661635e-06, 'learning_rate/action_model': 7.976577226661634e-05, 'epoch': 0.51})
+[2026-07-28 12:30:24,622][starVLA.training.train_starvla][INFO] - Step 2792, Loss: {'train/loss': 0.04153134673833847, 'train/grad_norm_pre_clip': 1.0277005434036255, 'timing/data': 0.0002828598953783512, 'timing/model': 0.4222649480216205, 'learning_rate/qwen_vl_interface': 7.973443200984259e-06, 'learning_rate/action_model': 7.973443200984259e-05, 'epoch': 0.51})
+[2026-07-28 12:30:30,648][starVLA.training.train_starvla][INFO] - Step 2793, Loss: {'train/loss': 0.055843085050582886, 'train/grad_norm_pre_clip': 0.816453218460083, 'timing/data': 0.0003338712267577648, 'timing/model': 0.4269438460469246, 'learning_rate/qwen_vl_interface': 7.970307408094158e-06, 'learning_rate/action_model': 7.970307408094157e-05, 'epoch': 0.51})
+[2026-07-28 12:30:36,929][starVLA.training.train_starvla][INFO] - Step 2794, Loss: {'train/loss': 0.05359687656164169, 'train/grad_norm_pre_clip': 1.6342039108276367, 'timing/data': 0.0002859043888747692, 'timing/model': 0.6643008259125054, 'learning_rate/qwen_vl_interface': 7.967169850026113e-06, 'learning_rate/action_model': 7.967169850026113e-05, 'epoch': 0.51})
+[2026-07-28 12:30:43,158][starVLA.training.train_starvla][INFO] - Step 2795, Loss: {'train/loss': 0.028222747147083282, 'train/grad_norm_pre_clip': 1.1334073543548584, 'timing/data': 0.00035265693441033363, 'timing/model': 0.6646488178521395, 'learning_rate/qwen_vl_interface': 7.964030528816055e-06, 'learning_rate/action_model': 7.964030528816055e-05, 'epoch': 0.51})
+[2026-07-28 12:30:49,261][starVLA.training.train_starvla][INFO] - Step 2796, Loss: {'train/loss': 0.08290725201368332, 'train/grad_norm_pre_clip': 1.981850266456604, 'timing/data': 0.00037442799657583237, 'timing/model': 0.4271365748718381, 'learning_rate/qwen_vl_interface': 7.960889446501052e-06, 'learning_rate/action_model': 7.960889446501052e-05, 'epoch': 0.51})
+[2026-07-28 12:30:55,285][starVLA.training.train_starvla][INFO] - Step 2797, Loss: {'train/loss': 0.048513010144233704, 'train/grad_norm_pre_clip': 1.9127202033996582, 'timing/data': 0.000280498992651701, 'timing/model': 0.42465727822855115, 'learning_rate/qwen_vl_interface': 7.957746605119317e-06, 'learning_rate/action_model': 7.957746605119318e-05, 'epoch': 0.51})
+[2026-07-28 12:31:01,542][starVLA.training.train_starvla][INFO] - Step 2798, Loss: {'train/loss': 0.04779907688498497, 'train/grad_norm_pre_clip': 2.1181976795196533, 'timing/data': 0.0002706018276512623, 'timing/model': 0.4230702151544392, 'learning_rate/qwen_vl_interface': 7.954602006710209e-06, 'learning_rate/action_model': 7.954602006710209e-05, 'epoch': 0.51})
+[2026-07-28 12:31:07,575][starVLA.training.train_starvla][INFO] - Step 2799, Loss: {'train/loss': 0.020967232063412666, 'train/grad_norm_pre_clip': 0.8640602827072144, 'timing/data': 0.0003641219809651375, 'timing/model': 0.42612067330628633, 'learning_rate/qwen_vl_interface': 7.951455653314222e-06, 'learning_rate/action_model': 7.951455653314221e-05, 'epoch': 0.51})
+[2026-07-28 12:31:13,833][starVLA.training.train_starvla][INFO] - Step 2800, Loss: {'train/loss': 0.06559107452630997, 'train/grad_norm_pre_clip': 1.1935722827911377, 'timing/data': 0.00048720603808760643, 'timing/model': 0.4235420967452228, 'learning_rate/qwen_vl_interface': 7.94830754697299e-06, 'learning_rate/action_model': 7.94830754697299e-05, 'epoch': 0.51})
+[2026-07-28 12:31:19,859][starVLA.training.train_starvla][INFO] - Step 2801, Loss: {'train/loss': 0.052438005805015564, 'train/grad_norm_pre_clip': 0.9033046364784241, 'timing/data': 0.000261601060628891, 'timing/model': 0.42575893830507994, 'learning_rate/qwen_vl_interface': 7.945157689729287e-06, 'learning_rate/action_model': 7.945157689729287e-05, 'epoch': 0.51})
+[2026-07-28 12:31:26,143][starVLA.training.train_starvla][INFO] - Step 2802, Loss: {'train/loss': 0.024504568427801132, 'train/grad_norm_pre_clip': 0.5557762384414673, 'timing/data': 0.0002753036096692085, 'timing/model': 0.4222609642893076, 'learning_rate/qwen_vl_interface': 7.94200608362702e-06, 'learning_rate/action_model': 7.94200608362702e-05, 'epoch': 0.51})
+[2026-07-28 12:31:32,505][starVLA.training.train_starvla][INFO] - Step 2803, Loss: {'train/loss': 0.06483176350593567, 'train/grad_norm_pre_clip': 1.2202516794204712, 'timing/data': 0.00033925101161003113, 'timing/model': 0.42383740609511733, 'learning_rate/qwen_vl_interface': 7.938852730711232e-06, 'learning_rate/action_model': 7.938852730711231e-05, 'epoch': 0.51})
+[2026-07-28 12:31:38,577][starVLA.training.train_starvla][INFO] - Step 2804, Loss: {'train/loss': 0.028673779219388962, 'train/grad_norm_pre_clip': 0.9541547298431396, 'timing/data': 0.0003391760401427746, 'timing/model': 0.4278482822701335, 'learning_rate/qwen_vl_interface': 7.935697633028099e-06, 'learning_rate/action_model': 7.935697633028099e-05, 'epoch': 0.52})
+[2026-07-28 12:31:44,941][starVLA.training.train_starvla][INFO] - Step 2805, Loss: {'train/loss': 0.04892749711871147, 'train/grad_norm_pre_clip': 1.0992987155914307, 'timing/data': 0.0002607479691505432, 'timing/model': 0.72451944462955, 'learning_rate/qwen_vl_interface': 7.932540792624931e-06, 'learning_rate/action_model': 7.93254079262493e-05, 'epoch': 0.52})
+[2026-07-28 12:31:51,049][starVLA.training.train_starvla][INFO] - Step 2806, Loss: {'train/loss': 0.035001952201128006, 'train/grad_norm_pre_clip': 1.4221469163894653, 'timing/data': 0.00027953460812568665, 'timing/model': 0.476769361179322, 'learning_rate/qwen_vl_interface': 7.929382211550166e-06, 'learning_rate/action_model': 7.929382211550165e-05, 'epoch': 0.52})
+[2026-07-28 12:31:57,305][starVLA.training.train_starvla][INFO] - Step 2807, Loss: {'train/loss': 0.05443158745765686, 'train/grad_norm_pre_clip': 2.013657808303833, 'timing/data': 0.00028565991669893265, 'timing/model': 0.666456107981503, 'learning_rate/qwen_vl_interface': 7.926221891853373e-06, 'learning_rate/action_model': 7.926221891853372e-05, 'epoch': 0.52})
+[2026-07-28 12:32:03,397][starVLA.training.train_starvla][INFO] - Step 2808, Loss: {'train/loss': 0.048887237906455994, 'train/grad_norm_pre_clip': 1.2989786863327026, 'timing/data': 0.00026842672377824783, 'timing/model': 0.4486463740468025, 'learning_rate/qwen_vl_interface': 7.923059835585251e-06, 'learning_rate/action_model': 7.923059835585251e-05, 'epoch': 0.52})
+[2026-07-28 12:32:09,491][starVLA.training.train_starvla][INFO] - Step 2809, Loss: {'train/loss': 0.08262433856725693, 'train/grad_norm_pre_clip': 2.4685981273651123, 'timing/data': 0.00032135890796780586, 'timing/model': 0.42505351081490517, 'learning_rate/qwen_vl_interface': 7.919896044797622e-06, 'learning_rate/action_model': 7.91989604479762e-05, 'epoch': 0.52})
+[2026-07-28 12:32:15,581][starVLA.training.train_starvla][INFO] - Step 2810, Loss: {'train/loss': 0.023122239857912064, 'train/grad_norm_pre_clip': 0.7395282983779907, 'timing/data': 0.0002889418974518776, 'timing/model': 0.4257953022606671, 'learning_rate/qwen_vl_interface': 7.916730521543436e-06, 'learning_rate/action_model': 7.916730521543436e-05, 'epoch': 0.52})
+[2026-07-28 12:32:21,897][starVLA.training.train_starvla][INFO] - Step 2811, Loss: {'train/loss': 0.04395846277475357, 'train/grad_norm_pre_clip': 1.2690303325653076, 'timing/data': 0.0003593349829316139, 'timing/model': 0.42141592502593994, 'learning_rate/qwen_vl_interface': 7.913563267876765e-06, 'learning_rate/action_model': 7.913563267876766e-05, 'epoch': 0.52})
+[2026-07-28 12:32:28,011][starVLA.training.train_starvla][INFO] - Step 2812, Loss: {'train/loss': 0.04728584736585617, 'train/grad_norm_pre_clip': 2.017547607421875, 'timing/data': 0.00029355334118008614, 'timing/model': 0.42714070389047265, 'learning_rate/qwen_vl_interface': 7.910394285852811e-06, 'learning_rate/action_model': 7.910394285852811e-05, 'epoch': 0.52})
+[2026-07-28 12:32:34,295][starVLA.training.train_starvla][INFO] - Step 2813, Loss: {'train/loss': 0.054185885936021805, 'train/grad_norm_pre_clip': 2.099870204925537, 'timing/data': 0.00029577407985925674, 'timing/model': 0.4227596214041114, 'learning_rate/qwen_vl_interface': 7.907223577527886e-06, 'learning_rate/action_model': 7.907223577527886e-05, 'epoch': 0.52})
+[2026-07-28 12:32:40,348][starVLA.training.train_starvla][INFO] - Step 2814, Loss: {'train/loss': 0.11220022290945053, 'train/grad_norm_pre_clip': 2.4376487731933594, 'timing/data': 0.00035815127193927765, 'timing/model': 0.42735505383461714, 'learning_rate/qwen_vl_interface': 7.904051144959432e-06, 'learning_rate/action_model': 7.904051144959431e-05, 'epoch': 0.52})
+[2026-07-28 12:32:46,399][starVLA.training.train_starvla][INFO] - Step 2815, Loss: {'train/loss': 0.04903997480869293, 'train/grad_norm_pre_clip': 1.027962327003479, 'timing/data': 0.00033696508035063744, 'timing/model': 0.4249651492573321, 'learning_rate/qwen_vl_interface': 7.900876990206002e-06, 'learning_rate/action_model': 7.900876990206002e-05, 'epoch': 0.52})
+[2026-07-28 12:32:52,219][starVLA.training.train_starvla][INFO] - Step 2816, Loss: {'train/loss': 0.03285505250096321, 'train/grad_norm_pre_clip': 0.7797791957855225, 'timing/data': 0.0003510918468236923, 'timing/model': 0.42725956765934825, 'learning_rate/qwen_vl_interface': 7.897701115327276e-06, 'learning_rate/action_model': 7.897701115327276e-05, 'epoch': 0.52})
+[2026-07-28 12:32:58,771][starVLA.training.train_starvla][INFO] - Step 2817, Loss: {'train/loss': 0.040683746337890625, 'train/grad_norm_pre_clip': 0.9095293879508972, 'timing/data': 0.00027492130175232887, 'timing/model': 0.4255949081853032, 'learning_rate/qwen_vl_interface': 7.894523522384043e-06, 'learning_rate/action_model': 7.894523522384042e-05, 'epoch': 0.52})
+[2026-07-28 12:33:04,872][starVLA.training.train_starvla][INFO] - Step 2818, Loss: {'train/loss': 0.05280648544430733, 'train/grad_norm_pre_clip': 1.3215311765670776, 'timing/data': 0.00028484221547842026, 'timing/model': 0.4281298629939556, 'learning_rate/qwen_vl_interface': 7.891344213438205e-06, 'learning_rate/action_model': 7.891344213438204e-05, 'epoch': 0.52})
+[2026-07-28 12:33:10,952][starVLA.training.train_starvla][INFO] - Step 2819, Loss: {'train/loss': 0.07258918881416321, 'train/grad_norm_pre_clip': 1.887539029121399, 'timing/data': 0.00027155177667737007, 'timing/model': 0.4275025920942426, 'learning_rate/qwen_vl_interface': 7.888163190552783e-06, 'learning_rate/action_model': 7.888163190552783e-05, 'epoch': 0.52})
+[2026-07-28 12:33:17,238][starVLA.training.train_starvla][INFO] - Step 2820, Loss: {'train/loss': 0.035411663353443146, 'train/grad_norm_pre_clip': 1.5754791498184204, 'timing/data': 0.00028577307239174843, 'timing/model': 0.42332029109820724, 'learning_rate/qwen_vl_interface': 7.88498045579191e-06, 'learning_rate/action_model': 7.88498045579191e-05, 'epoch': 0.52})
+[2026-07-28 12:33:23,271][starVLA.training.train_starvla][INFO] - Step 2821, Loss: {'train/loss': 0.033549461513757706, 'train/grad_norm_pre_clip': 0.8426619172096252, 'timing/data': 0.0002805618569254875, 'timing/model': 0.4224150748923421, 'learning_rate/qwen_vl_interface': 7.881796011220827e-06, 'learning_rate/action_model': 7.881796011220827e-05, 'epoch': 0.52})
+[2026-07-28 12:33:29,514][starVLA.training.train_starvla][INFO] - Step 2822, Loss: {'train/loss': 0.03940058872103691, 'train/grad_norm_pre_clip': 1.2916942834854126, 'timing/data': 0.0002601561136543751, 'timing/model': 0.4242978277616203, 'learning_rate/qwen_vl_interface': 7.878609858905885e-06, 'learning_rate/action_model': 7.878609858905886e-05, 'epoch': 0.52})
+[2026-07-28 12:33:35,725][starVLA.training.train_starvla][INFO] - Step 2823, Loss: {'train/loss': 0.032563552260398865, 'train/grad_norm_pre_clip': 1.2584309577941895, 'timing/data': 0.00027759093791246414, 'timing/model': 0.6062428848817945, 'learning_rate/qwen_vl_interface': 7.875422000914544e-06, 'learning_rate/action_model': 7.875422000914544e-05, 'epoch': 0.52})
+[2026-07-28 12:33:41,760][starVLA.training.train_starvla][INFO] - Step 2824, Loss: {'train/loss': 0.041243985295295715, 'train/grad_norm_pre_clip': 2.0823075771331787, 'timing/data': 0.00027913693338632584, 'timing/model': 0.42624458810314536, 'learning_rate/qwen_vl_interface': 7.872232439315371e-06, 'learning_rate/action_model': 7.87223243931537e-05, 'epoch': 0.52})
+[2026-07-28 12:33:47,821][starVLA.training.train_starvla][INFO] - Step 2825, Loss: {'train/loss': 0.042626477777957916, 'train/grad_norm_pre_clip': 1.2643626928329468, 'timing/data': 0.00034064706414937973, 'timing/model': 0.423854349181056, 'learning_rate/qwen_vl_interface': 7.869041176178038e-06, 'learning_rate/action_model': 7.869041176178037e-05, 'epoch': 0.52})
+[2026-07-28 12:33:54,065][starVLA.training.train_starvla][INFO] - Step 2826, Loss: {'train/loss': 0.060841333121061325, 'train/grad_norm_pre_clip': 1.7920138835906982, 'timing/data': 0.0002625398337841034, 'timing/model': 0.4216474099084735, 'learning_rate/qwen_vl_interface': 7.865848213573316e-06, 'learning_rate/action_model': 7.865848213573317e-05, 'epoch': 0.52})
+[2026-07-28 12:34:00,307][starVLA.training.train_starvla][INFO] - Step 2827, Loss: {'train/loss': 0.05548489838838577, 'train/grad_norm_pre_clip': 2.2352969646453857, 'timing/data': 0.0008607557974755764, 'timing/model': 0.4231193750165403, 'learning_rate/qwen_vl_interface': 7.862653553573092e-06, 'learning_rate/action_model': 7.862653553573092e-05, 'epoch': 0.52})
+[2026-07-28 12:34:06,341][starVLA.training.train_starvla][INFO] - Step 2828, Loss: {'train/loss': 0.029611850157380104, 'train/grad_norm_pre_clip': 1.193170189857483, 'timing/data': 0.0003737430088222027, 'timing/model': 0.42404119577258825, 'learning_rate/qwen_vl_interface': 7.859457198250342e-06, 'learning_rate/action_model': 7.859457198250342e-05, 'epoch': 0.52})
+[2026-07-28 12:34:12,617][starVLA.training.train_starvla][INFO] - Step 2829, Loss: {'train/loss': 0.048072248697280884, 'train/grad_norm_pre_clip': 1.7806174755096436, 'timing/data': 0.00037136394530534744, 'timing/model': 0.4225654322654009, 'learning_rate/qwen_vl_interface': 7.856259149679146e-06, 'learning_rate/action_model': 7.856259149679146e-05, 'epoch': 0.52})
+[2026-07-28 12:34:18,869][starVLA.training.train_starvla][INFO] - Step 2830, Loss: {'train/loss': 0.07599794119596481, 'train/grad_norm_pre_clip': 2.2646000385284424, 'timing/data': 0.0002682018093764782, 'timing/model': 0.6741224657744169, 'learning_rate/qwen_vl_interface': 7.853059409934685e-06, 'learning_rate/action_model': 7.853059409934685e-05, 'epoch': 0.52})
+[2026-07-28 12:34:24,872][starVLA.training.train_starvla][INFO] - Step 2831, Loss: {'train/loss': 0.05641232058405876, 'train/grad_norm_pre_clip': 1.0980771780014038, 'timing/data': 0.0002571609802544117, 'timing/model': 0.4243146008811891, 'learning_rate/qwen_vl_interface': 7.849857981093231e-06, 'learning_rate/action_model': 7.84985798109323e-05, 'epoch': 0.52})
+[2026-07-28 12:34:30,930][starVLA.training.train_starvla][INFO] - Step 2832, Loss: {'train/loss': 0.05212979018688202, 'train/grad_norm_pre_clip': 1.509032130241394, 'timing/data': 0.0003943503834307194, 'timing/model': 0.42466623801738024, 'learning_rate/qwen_vl_interface': 7.84665486523216e-06, 'learning_rate/action_model': 7.846654865232159e-05, 'epoch': 0.52})
+[2026-07-28 12:34:37,196][starVLA.training.train_starvla][INFO] - Step 2833, Loss: {'train/loss': 0.02347012609243393, 'train/grad_norm_pre_clip': 0.7592364549636841, 'timing/data': 0.0002818489447236061, 'timing/model': 0.6609913958236575, 'learning_rate/qwen_vl_interface': 7.84345006442994e-06, 'learning_rate/action_model': 7.843450064429939e-05, 'epoch': 0.52})
+[2026-07-28 12:34:43,427][starVLA.training.train_starvla][INFO] - Step 2834, Loss: {'train/loss': 0.03479449078440666, 'train/grad_norm_pre_clip': 2.162032127380371, 'timing/data': 0.0003454601392149925, 'timing/model': 0.4269369780085981, 'learning_rate/qwen_vl_interface': 7.840243580766129e-06, 'learning_rate/action_model': 7.840243580766128e-05, 'epoch': 0.52})
+[2026-07-28 12:34:49,497][starVLA.training.train_starvla][INFO] - Step 2835, Loss: {'train/loss': 0.0955270528793335, 'train/grad_norm_pre_clip': 1.7331098318099976, 'timing/data': 0.00027071498334407806, 'timing/model': 0.4244000008329749, 'learning_rate/qwen_vl_interface': 7.83703541632138e-06, 'learning_rate/action_model': 7.83703541632138e-05, 'epoch': 0.52})
+[2026-07-28 12:34:55,555][starVLA.training.train_starvla][INFO] - Step 2836, Loss: {'train/loss': 0.032824695110321045, 'train/grad_norm_pre_clip': 1.8419790267944336, 'timing/data': 0.00026757363229990005, 'timing/model': 0.42773533798754215, 'learning_rate/qwen_vl_interface': 7.833825573177435e-06, 'learning_rate/action_model': 7.833825573177436e-05, 'epoch': 0.52})
+[2026-07-28 12:35:01,843][starVLA.training.train_starvla][INFO] - Step 2837, Loss: {'train/loss': 0.05322445183992386, 'train/grad_norm_pre_clip': 1.626007080078125, 'timing/data': 0.00027227820828557014, 'timing/model': 0.4235586547292769, 'learning_rate/qwen_vl_interface': 7.83061405341713e-06, 'learning_rate/action_model': 7.83061405341713e-05, 'epoch': 0.52})
+[2026-07-28 12:35:08,123][starVLA.training.train_starvla][INFO] - Step 2838, Loss: {'train/loss': 0.059350017458200455, 'train/grad_norm_pre_clip': 1.5207834243774414, 'timing/data': 0.0003396961838006973, 'timing/model': 0.4250146052800119, 'learning_rate/qwen_vl_interface': 7.827400859124384e-06, 'learning_rate/action_model': 7.827400859124384e-05, 'epoch': 0.52})
+[2026-07-28 12:35:14,154][starVLA.training.train_starvla][INFO] - Step 2839, Loss: {'train/loss': 0.06387832760810852, 'train/grad_norm_pre_clip': 1.6182820796966553, 'timing/data': 0.0003135404549539089, 'timing/model': 0.42258848063647747, 'learning_rate/qwen_vl_interface': 7.8241859923842e-06, 'learning_rate/action_model': 7.8241859923842e-05, 'epoch': 0.52})
+[2026-07-28 12:35:20,170][starVLA.training.train_starvla][INFO] - Step 2840, Loss: {'train/loss': 0.04112086445093155, 'train/grad_norm_pre_clip': 0.6944811940193176, 'timing/data': 0.00033984892070293427, 'timing/model': 0.4238037788309157, 'learning_rate/qwen_vl_interface': 7.820969455282676e-06, 'learning_rate/action_model': 7.820969455282675e-05, 'epoch': 0.52})
+[2026-07-28 12:35:26,440][starVLA.training.train_starvla][INFO] - Step 2841, Loss: {'train/loss': 0.03185412660241127, 'train/grad_norm_pre_clip': 0.9328694939613342, 'timing/data': 0.00027529289945960045, 'timing/model': 0.42358474200591445, 'learning_rate/qwen_vl_interface': 7.817751249906982e-06, 'learning_rate/action_model': 7.817751249906983e-05, 'epoch': 0.52})
+[2026-07-28 12:35:32,475][starVLA.training.train_starvla][INFO] - Step 2842, Loss: {'train/loss': 0.026733800768852234, 'train/grad_norm_pre_clip': 1.7135955095291138, 'timing/data': 0.0002708509564399719, 'timing/model': 0.42487930227071047, 'learning_rate/qwen_vl_interface': 7.814531378345381e-06, 'learning_rate/action_model': 7.814531378345381e-05, 'epoch': 0.52})
+[2026-07-28 12:35:38,756][starVLA.training.train_starvla][INFO] - Step 2843, Loss: {'train/loss': 0.034725528210401535, 'train/grad_norm_pre_clip': 1.3676577806472778, 'timing/data': 0.0003465181216597557, 'timing/model': 0.6575839519500732, 'learning_rate/qwen_vl_interface': 7.811309842687213e-06, 'learning_rate/action_model': 7.811309842687212e-05, 'epoch': 0.52})
+[2026-07-28 12:35:45,027][starVLA.training.train_starvla][INFO] - Step 2844, Loss: {'train/loss': 0.018368832767009735, 'train/grad_norm_pre_clip': 0.6023229360580444, 'timing/data': 0.0002734307199716568, 'timing/model': 0.42284091701731086, 'learning_rate/qwen_vl_interface': 7.808086645022895e-06, 'learning_rate/action_model': 7.808086645022895e-05, 'epoch': 0.52})
+[2026-07-28 12:35:51,072][starVLA.training.train_starvla][INFO] - Step 2845, Loss: {'train/loss': 0.05904857814311981, 'train/grad_norm_pre_clip': 1.2734167575836182, 'timing/data': 0.00028699589893221855, 'timing/model': 0.4255209448747337, 'learning_rate/qwen_vl_interface': 7.804861787443924e-06, 'learning_rate/action_model': 7.804861787443923e-05, 'epoch': 0.52})
+[2026-07-28 12:35:57,129][starVLA.training.train_starvla][INFO] - Step 2846, Loss: {'train/loss': 0.03341042995452881, 'train/grad_norm_pre_clip': 1.0448284149169922, 'timing/data': 0.00026759086176753044, 'timing/model': 0.4263982488773763, 'learning_rate/qwen_vl_interface': 7.801635272042879e-06, 'learning_rate/action_model': 7.801635272042877e-05, 'epoch': 0.52})
+[2026-07-28 12:36:03,197][starVLA.training.train_starvla][INFO] - Step 2847, Loss: {'train/loss': 0.051171306520700455, 'train/grad_norm_pre_clip': 1.8642975091934204, 'timing/data': 0.0002552662044763565, 'timing/model': 0.4264474231749773, 'learning_rate/qwen_vl_interface': 7.798407100913407e-06, 'learning_rate/action_model': 7.798407100913407e-05, 'epoch': 0.52})
+[2026-07-28 12:36:09,714][starVLA.training.train_starvla][INFO] - Step 2848, Loss: {'train/loss': 0.0672706663608551, 'train/grad_norm_pre_clip': 1.5971860885620117, 'timing/data': 0.00027492595836520195, 'timing/model': 0.6789730102755129, 'learning_rate/qwen_vl_interface': 7.795177276150234e-06, 'learning_rate/action_model': 7.795177276150234e-05, 'epoch': 0.52})
+[2026-07-28 12:36:15,774][starVLA.training.train_starvla][INFO] - Step 2849, Loss: {'train/loss': 0.03607412427663803, 'train/grad_norm_pre_clip': 1.0537617206573486, 'timing/data': 0.00027373991906642914, 'timing/model': 0.42621930688619614, 'learning_rate/qwen_vl_interface': 7.79194579984916e-06, 'learning_rate/action_model': 7.791945799849161e-05, 'epoch': 0.52})
+[2026-07-28 12:36:21,814][starVLA.training.train_starvla][INFO] - Step 2850, Loss: {'train/loss': 0.037549905478954315, 'train/grad_norm_pre_clip': 1.3426443338394165, 'timing/data': 0.0002612350508570671, 'timing/model': 0.42591874999925494, 'learning_rate/qwen_vl_interface': 7.788712674107054e-06, 'learning_rate/action_model': 7.788712674107055e-05, 'epoch': 0.52})
+[2026-07-28 12:36:28,085][starVLA.training.train_starvla][INFO] - Step 2851, Loss: {'train/loss': 0.04863428324460983, 'train/grad_norm_pre_clip': 1.946446418762207, 'timing/data': 0.00033892691135406494, 'timing/model': 0.4238620181567967, 'learning_rate/qwen_vl_interface': 7.785477901021859e-06, 'learning_rate/action_model': 7.785477901021859e-05, 'epoch': 0.52})
+[2026-07-28 12:36:34,380][starVLA.training.train_starvla][INFO] - Step 2852, Loss: {'train/loss': 0.033875614404678345, 'train/grad_norm_pre_clip': 1.4750792980194092, 'timing/data': 0.0002841097302734852, 'timing/model': 0.4217364569194615, 'learning_rate/qwen_vl_interface': 7.78224148269258e-06, 'learning_rate/action_model': 7.78224148269258e-05, 'epoch': 0.52})
+[2026-07-28 12:36:40,407][starVLA.training.train_starvla][INFO] - Step 2853, Loss: {'train/loss': 0.02502601407468319, 'train/grad_norm_pre_clip': 1.1658939123153687, 'timing/data': 0.0002670362591743469, 'timing/model': 0.4281251379288733, 'learning_rate/qwen_vl_interface': 7.779003421219293e-06, 'learning_rate/action_model': 7.779003421219293e-05, 'epoch': 0.52})
+[2026-07-28 12:36:46,625][starVLA.training.train_starvla][INFO] - Step 2854, Loss: {'train/loss': 0.039200372993946075, 'train/grad_norm_pre_clip': 1.0971359014511108, 'timing/data': 0.00027325330302119255, 'timing/model': 0.6010617800056934, 'learning_rate/qwen_vl_interface': 7.775763718703144e-06, 'learning_rate/action_model': 7.775763718703144e-05, 'epoch': 0.52})
+[2026-07-28 12:36:52,681][starVLA.training.train_starvla][INFO] - Step 2855, Loss: {'train/loss': 0.060821328312158585, 'train/grad_norm_pre_clip': 1.7029855251312256, 'timing/data': 0.0002611312083899975, 'timing/model': 0.42585864290595055, 'learning_rate/qwen_vl_interface': 7.772522377246339e-06, 'learning_rate/action_model': 7.77252237724634e-05, 'epoch': 0.52})
+[2026-07-28 12:36:58,944][starVLA.training.train_starvla][INFO] - Step 2856, Loss: {'train/loss': 0.02550727128982544, 'train/grad_norm_pre_clip': 1.0643097162246704, 'timing/data': 0.00026627304032444954, 'timing/model': 0.4196190070360899, 'learning_rate/qwen_vl_interface': 7.769279398952152e-06, 'learning_rate/action_model': 7.769279398952151e-05, 'epoch': 0.52})
+[2026-07-28 12:37:04,998][starVLA.training.train_starvla][INFO] - Step 2857, Loss: {'train/loss': 0.034476298838853836, 'train/grad_norm_pre_clip': 0.871538519859314, 'timing/data': 0.00028566597029566765, 'timing/model': 0.4433512012474239, 'learning_rate/qwen_vl_interface': 7.766034785924914e-06, 'learning_rate/action_model': 7.766034785924912e-05, 'epoch': 0.52})
+[2026-07-28 12:37:11,271][starVLA.training.train_starvla][INFO] - Step 2858, Loss: {'train/loss': 0.02357359230518341, 'train/grad_norm_pre_clip': 0.7987403273582458, 'timing/data': 0.00028169574216008186, 'timing/model': 0.42566120345145464, 'learning_rate/qwen_vl_interface': 7.762788540270018e-06, 'learning_rate/action_model': 7.762788540270018e-05, 'epoch': 0.52})
+[2026-07-28 12:37:17,561][starVLA.training.train_starvla][INFO] - Step 2859, Loss: {'train/loss': 0.03793582692742348, 'train/grad_norm_pre_clip': 1.406298041343689, 'timing/data': 0.0003496161662042141, 'timing/model': 0.43204592866823077, 'learning_rate/qwen_vl_interface': 7.759540664093919e-06, 'learning_rate/action_model': 7.759540664093919e-05, 'epoch': 0.53})
+[2026-07-28 12:37:23,670][starVLA.training.train_starvla][INFO] - Step 2860, Loss: {'train/loss': 0.07313154637813568, 'train/grad_norm_pre_clip': 1.7224459648132324, 'timing/data': 0.0002674073912203312, 'timing/model': 0.4239842537790537, 'learning_rate/qwen_vl_interface': 7.756291159504132e-06, 'learning_rate/action_model': 7.756291159504132e-05, 'epoch': 0.53})
+[2026-07-28 12:37:29,704][starVLA.training.train_starvla][INFO] - Step 2861, Loss: {'train/loss': 0.10183212906122208, 'train/grad_norm_pre_clip': 2.065458297729492, 'timing/data': 0.00026702880859375, 'timing/model': 0.4280730071477592, 'learning_rate/qwen_vl_interface': 7.75304002860922e-06, 'learning_rate/action_model': 7.75304002860922e-05, 'epoch': 0.53})
+[2026-07-28 12:37:35,964][starVLA.training.train_starvla][INFO] - Step 2862, Loss: {'train/loss': 0.07385292649269104, 'train/grad_norm_pre_clip': 3.19692325592041, 'timing/data': 0.0002720053307712078, 'timing/model': 0.42197962990030646, 'learning_rate/qwen_vl_interface': 7.749787273518807e-06, 'learning_rate/action_model': 7.749787273518807e-05, 'epoch': 0.53})
+[2026-07-28 12:37:42,234][starVLA.training.train_starvla][INFO] - Step 2863, Loss: {'train/loss': 0.04274909198284149, 'train/grad_norm_pre_clip': 1.6844937801361084, 'timing/data': 0.0002576727420091629, 'timing/model': 0.6624463177286088, 'learning_rate/qwen_vl_interface': 7.746532896343575e-06, 'learning_rate/action_model': 7.746532896343575e-05, 'epoch': 0.53})
+[2026-07-28 12:37:48,250][starVLA.training.train_starvla][INFO] - Step 2864, Loss: {'train/loss': 0.03620287775993347, 'train/grad_norm_pre_clip': 1.3004765510559082, 'timing/data': 0.00032200058922171593, 'timing/model': 0.4243424688465893, 'learning_rate/qwen_vl_interface': 7.743276899195249e-06, 'learning_rate/action_model': 7.743276899195249e-05, 'epoch': 0.53})
+[2026-07-28 12:37:54,050][starVLA.training.train_starvla][INFO] - Step 2865, Loss: {'train/loss': 0.051913995295763016, 'train/grad_norm_pre_clip': 2.1270411014556885, 'timing/data': 0.0002657361328601837, 'timing/model': 0.42511474480852485, 'learning_rate/qwen_vl_interface': 7.740019284186612e-06, 'learning_rate/action_model': 7.740019284186612e-05, 'epoch': 0.53})
+[2026-07-28 12:38:00,101][starVLA.training.train_starvla][INFO] - Step 2866, Loss: {'train/loss': 0.08845211565494537, 'train/grad_norm_pre_clip': 3.1024646759033203, 'timing/data': 0.0002612578682601452, 'timing/model': 0.4216307899914682, 'learning_rate/qwen_vl_interface': 7.736760053431496e-06, 'learning_rate/action_model': 7.736760053431495e-05, 'epoch': 0.53})
+[2026-07-28 12:38:06,567][starVLA.training.train_starvla][INFO] - Step 2867, Loss: {'train/loss': 0.015293500386178493, 'train/grad_norm_pre_clip': 0.9844099879264832, 'timing/data': 0.0002484000287950039, 'timing/model': 0.6592054651118815, 'learning_rate/qwen_vl_interface': 7.73349920904478e-06, 'learning_rate/action_model': 7.73349920904478e-05, 'epoch': 0.53})
+[2026-07-28 12:38:12,570][starVLA.training.train_starvla][INFO] - Step 2868, Loss: {'train/loss': 0.027198977768421173, 'train/grad_norm_pre_clip': 0.6445983648300171, 'timing/data': 0.00025398610159754753, 'timing/model': 0.4233076283708215, 'learning_rate/qwen_vl_interface': 7.73023675314239e-06, 'learning_rate/action_model': 7.73023675314239e-05, 'epoch': 0.53})
+[2026-07-28 12:38:18,610][starVLA.training.train_starvla][INFO] - Step 2869, Loss: {'train/loss': 0.02507740817964077, 'train/grad_norm_pre_clip': 1.1027369499206543, 'timing/data': 0.00032848678529262543, 'timing/model': 0.42593066627159715, 'learning_rate/qwen_vl_interface': 7.726972687841298e-06, 'learning_rate/action_model': 7.726972687841299e-05, 'epoch': 0.53})
+[2026-07-28 12:38:24,626][starVLA.training.train_starvla][INFO] - Step 2870, Loss: {'train/loss': 0.023971758782863617, 'train/grad_norm_pre_clip': 0.9329346418380737, 'timing/data': 0.0002596229314804077, 'timing/model': 0.4242041427642107, 'learning_rate/qwen_vl_interface': 7.723707015259522e-06, 'learning_rate/action_model': 7.723707015259522e-05, 'epoch': 0.53})
+[2026-07-28 12:38:30,882][starVLA.training.train_starvla][INFO] - Step 2871, Loss: {'train/loss': 0.05727072060108185, 'train/grad_norm_pre_clip': 1.5192912817001343, 'timing/data': 0.00026045413687825203, 'timing/model': 0.4232229213230312, 'learning_rate/qwen_vl_interface': 7.72043973751612e-06, 'learning_rate/action_model': 7.720439737516119e-05, 'epoch': 0.53})
+[2026-07-28 12:38:37,215][starVLA.training.train_starvla][INFO] - Step 2872, Loss: {'train/loss': 0.04379698634147644, 'train/grad_norm_pre_clip': 2.350499391555786, 'timing/data': 0.0003260783851146698, 'timing/model': 0.43361614318564534, 'learning_rate/qwen_vl_interface': 7.717170856731192e-06, 'learning_rate/action_model': 7.717170856731193e-05, 'epoch': 0.53})
+[2026-07-28 12:38:43,299][starVLA.training.train_starvla][INFO] - Step 2873, Loss: {'train/loss': 0.05048012733459473, 'train/grad_norm_pre_clip': 0.7761784791946411, 'timing/data': 0.0003347308374941349, 'timing/model': 0.42578380601480603, 'learning_rate/qwen_vl_interface': 7.713900375025883e-06, 'learning_rate/action_model': 7.713900375025882e-05, 'epoch': 0.53})
+[2026-07-28 12:38:49,679][starVLA.training.train_starvla][INFO] - Step 2874, Loss: {'train/loss': 0.037555042654275894, 'train/grad_norm_pre_clip': 1.3471980094909668, 'timing/data': 0.0002809753641486168, 'timing/model': 0.422115963883698, 'learning_rate/qwen_vl_interface': 7.71062829452237e-06, 'learning_rate/action_model': 7.710628294522369e-05, 'epoch': 0.53})
+[2026-07-28 12:38:55,979][starVLA.training.train_starvla][INFO] - Step 2875, Loss: {'train/loss': 0.030150484293699265, 'train/grad_norm_pre_clip': 1.1314051151275635, 'timing/data': 0.0002679917961359024, 'timing/model': 0.42088514380156994, 'learning_rate/qwen_vl_interface': 7.707354617343871e-06, 'learning_rate/action_model': 7.707354617343871e-05, 'epoch': 0.53})
+[2026-07-28 12:39:01,994][starVLA.training.train_starvla][INFO] - Step 2876, Loss: {'train/loss': 0.06441257148981094, 'train/grad_norm_pre_clip': 1.6169853210449219, 'timing/data': 0.0003496822901070118, 'timing/model': 0.42516720993444324, 'learning_rate/qwen_vl_interface': 7.704079345614639e-06, 'learning_rate/action_model': 7.704079345614639e-05, 'epoch': 0.53})
+[2026-07-28 12:39:08,023][starVLA.training.train_starvla][INFO] - Step 2877, Loss: {'train/loss': 0.0506778284907341, 'train/grad_norm_pre_clip': 0.959889829158783, 'timing/data': 0.00027177901938557625, 'timing/model': 0.4313325993716717, 'learning_rate/qwen_vl_interface': 7.700802481459964e-06, 'learning_rate/action_model': 7.700802481459964e-05, 'epoch': 0.53})
+[2026-07-28 12:39:14,313][starVLA.training.train_starvla][INFO] - Step 2878, Loss: {'train/loss': 0.035128187388181686, 'train/grad_norm_pre_clip': 0.8201842904090881, 'timing/data': 0.0003065471537411213, 'timing/model': 0.6694717658683658, 'learning_rate/qwen_vl_interface': 7.697524027006168e-06, 'learning_rate/action_model': 7.697524027006167e-05, 'epoch': 0.53})
+[2026-07-28 12:39:20,627][starVLA.training.train_starvla][INFO] - Step 2879, Loss: {'train/loss': 0.016288377344608307, 'train/grad_norm_pre_clip': 0.5605166554450989, 'timing/data': 0.00026608770713210106, 'timing/model': 0.6635385728441179, 'learning_rate/qwen_vl_interface': 7.694243984380603e-06, 'learning_rate/action_model': 7.694243984380603e-05, 'epoch': 0.53})
+[2026-07-28 12:39:26,673][starVLA.training.train_starvla][INFO] - Step 2880, Loss: {'train/loss': 0.04342079162597656, 'train/grad_norm_pre_clip': 1.316413164138794, 'timing/data': 0.00032428791746497154, 'timing/model': 0.4230695003643632, 'learning_rate/qwen_vl_interface': 7.690962355711651e-06, 'learning_rate/action_model': 7.690962355711652e-05, 'epoch': 0.53})
+[2026-07-28 12:39:32,729][starVLA.training.train_starvla][INFO] - Step 2881, Loss: {'train/loss': 0.04414482042193413, 'train/grad_norm_pre_clip': 1.1595981121063232, 'timing/data': 0.00048609310761094093, 'timing/model': 0.4261905453167856, 'learning_rate/qwen_vl_interface': 7.68767914312873e-06, 'learning_rate/action_model': 7.68767914312873e-05, 'epoch': 0.53})
+[2026-07-28 12:39:38,822][starVLA.training.train_starvla][INFO] - Step 2882, Loss: {'train/loss': 0.05568212270736694, 'train/grad_norm_pre_clip': 1.3182728290557861, 'timing/data': 0.0003482028841972351, 'timing/model': 0.4297054377384484, 'learning_rate/qwen_vl_interface': 7.684394348762278e-06, 'learning_rate/action_model': 7.684394348762277e-05, 'epoch': 0.53})
+[2026-07-28 12:39:45,106][starVLA.training.train_starvla][INFO] - Step 2883, Loss: {'train/loss': 0.054783523082733154, 'train/grad_norm_pre_clip': 2.067676305770874, 'timing/data': 0.0005263900384306908, 'timing/model': 0.6675338079221547, 'learning_rate/qwen_vl_interface': 7.681107974743763e-06, 'learning_rate/action_model': 7.681107974743764e-05, 'epoch': 0.53})
+[2026-07-28 12:39:51,210][starVLA.training.train_starvla][INFO] - Step 2884, Loss: {'train/loss': 0.04139642044901848, 'train/grad_norm_pre_clip': 2.193782329559326, 'timing/data': 0.0002599591389298439, 'timing/model': 0.42471745423972607, 'learning_rate/qwen_vl_interface': 7.677820023205676e-06, 'learning_rate/action_model': 7.677820023205677e-05, 'epoch': 0.53})
+[2026-07-28 12:39:57,488][starVLA.training.train_starvla][INFO] - Step 2885, Loss: {'train/loss': 0.024156108498573303, 'train/grad_norm_pre_clip': 0.7300538420677185, 'timing/data': 0.0003115851432085037, 'timing/model': 0.42091594403609633, 'learning_rate/qwen_vl_interface': 7.674530496281536e-06, 'learning_rate/action_model': 7.674530496281536e-05, 'epoch': 0.53})
+[2026-07-28 12:40:03,527][starVLA.training.train_starvla][INFO] - Step 2886, Loss: {'train/loss': 0.024082738906145096, 'train/grad_norm_pre_clip': 0.5197408199310303, 'timing/data': 0.0002726549282670021, 'timing/model': 0.42551354272291064, 'learning_rate/qwen_vl_interface': 7.67123939610588e-06, 'learning_rate/action_model': 7.67123939610588e-05, 'epoch': 0.53})
+[2026-07-28 12:40:09,513][starVLA.training.train_starvla][INFO] - Step 2887, Loss: {'train/loss': 0.03505239263176918, 'train/grad_norm_pre_clip': 0.7991612553596497, 'timing/data': 0.00035528698936104774, 'timing/model': 0.4272698760032654, 'learning_rate/qwen_vl_interface': 7.667946724814264e-06, 'learning_rate/action_model': 7.667946724814265e-05, 'epoch': 0.53})
+[2026-07-28 12:40:15,644][starVLA.training.train_starvla][INFO] - Step 2888, Loss: {'train/loss': 0.04503896087408066, 'train/grad_norm_pre_clip': 1.6081764698028564, 'timing/data': 0.00028119515627622604, 'timing/model': 0.43272119481116533, 'learning_rate/qwen_vl_interface': 7.664652484543271e-06, 'learning_rate/action_model': 7.664652484543271e-05, 'epoch': 0.53})
+[2026-07-28 12:40:22,200][starVLA.training.train_starvla][INFO] - Step 2889, Loss: {'train/loss': 0.02936437726020813, 'train/grad_norm_pre_clip': 1.1799967288970947, 'timing/data': 0.00035596592351794243, 'timing/model': 0.6777297677472234, 'learning_rate/qwen_vl_interface': 7.661356677430499e-06, 'learning_rate/action_model': 7.661356677430499e-05, 'epoch': 0.53})
+[2026-07-28 12:40:28,303][starVLA.training.train_starvla][INFO] - Step 2890, Loss: {'train/loss': 0.029181282967329025, 'train/grad_norm_pre_clip': 0.9253875017166138, 'timing/data': 0.0002630278468132019, 'timing/model': 0.4231488611549139, 'learning_rate/qwen_vl_interface': 7.658059305614556e-06, 'learning_rate/action_model': 7.658059305614555e-05, 'epoch': 0.53})
+[2026-07-28 12:40:34,338][starVLA.training.train_starvla][INFO] - Step 2891, Loss: {'train/loss': 0.03170955181121826, 'train/grad_norm_pre_clip': 0.6445611119270325, 'timing/data': 0.0002774070017039776, 'timing/model': 0.4313934468664229, 'learning_rate/qwen_vl_interface': 7.654760371235073e-06, 'learning_rate/action_model': 7.654760371235074e-05, 'epoch': 0.53})
+[2026-07-28 12:40:40,643][starVLA.training.train_starvla][INFO] - Step 2892, Loss: {'train/loss': 0.05044228956103325, 'train/grad_norm_pre_clip': 1.7075088024139404, 'timing/data': 0.00026803184300661087, 'timing/model': 0.4237912050448358, 'learning_rate/qwen_vl_interface': 7.651459876432694e-06, 'learning_rate/action_model': 7.651459876432694e-05, 'epoch': 0.53})
+[2026-07-28 12:40:46,718][starVLA.training.train_starvla][INFO] - Step 2893, Loss: {'train/loss': 0.031604159623384476, 'train/grad_norm_pre_clip': 1.1134227514266968, 'timing/data': 0.0002713249996304512, 'timing/model': 0.4278130168095231, 'learning_rate/qwen_vl_interface': 7.648157823349074e-06, 'learning_rate/action_model': 7.648157823349074e-05, 'epoch': 0.53})
+[2026-07-28 12:40:53,037][starVLA.training.train_starvla][INFO] - Step 2894, Loss: {'train/loss': 0.04891273379325867, 'train/grad_norm_pre_clip': 1.7279715538024902, 'timing/data': 0.00028162170201539993, 'timing/model': 0.43043879978358746, 'learning_rate/qwen_vl_interface': 7.64485421412688e-06, 'learning_rate/action_model': 7.644854214126879e-05, 'epoch': 0.53})
+[2026-07-28 12:40:59,306][starVLA.training.train_starvla][INFO] - Step 2895, Loss: {'train/loss': 0.05528177320957184, 'train/grad_norm_pre_clip': 2.44036865234375, 'timing/data': 0.0002714828588068485, 'timing/model': 0.6075510531663895, 'learning_rate/qwen_vl_interface': 7.641549050909784e-06, 'learning_rate/action_model': 7.641549050909784e-05, 'epoch': 0.53})
+[2026-07-28 12:41:05,393][starVLA.training.train_starvla][INFO] - Step 2896, Loss: {'train/loss': 0.03730452060699463, 'train/grad_norm_pre_clip': 1.5335291624069214, 'timing/data': 0.0003747171722352505, 'timing/model': 0.434809235855937, 'learning_rate/qwen_vl_interface': 7.638242335842476e-06, 'learning_rate/action_model': 7.638242335842476e-05, 'epoch': 0.53})
+[2026-07-28 12:41:11,458][starVLA.training.train_starvla][INFO] - Step 2897, Loss: {'train/loss': 0.04146847128868103, 'train/grad_norm_pre_clip': 1.8806767463684082, 'timing/data': 0.00027630897238850594, 'timing/model': 0.4280153219588101, 'learning_rate/qwen_vl_interface': 7.634934071070643e-06, 'learning_rate/action_model': 7.634934071070643e-05, 'epoch': 0.53})
+[2026-07-28 12:41:17,757][starVLA.training.train_starvla][INFO] - Step 2898, Loss: {'train/loss': 0.02818257547914982, 'train/grad_norm_pre_clip': 0.7413269877433777, 'timing/data': 0.0002623051404953003, 'timing/model': 0.4321473389863968, 'learning_rate/qwen_vl_interface': 7.631624258740985e-06, 'learning_rate/action_model': 7.631624258740984e-05, 'epoch': 0.53})
+[2026-07-28 12:41:24,072][starVLA.training.train_starvla][INFO] - Step 2899, Loss: {'train/loss': 0.044713981449604034, 'train/grad_norm_pre_clip': 1.4922715425491333, 'timing/data': 0.00024877535179257393, 'timing/model': 0.43314956594258547, 'learning_rate/qwen_vl_interface': 7.628312901001203e-06, 'learning_rate/action_model': 7.628312901001204e-05, 'epoch': 0.53})
+[2026-07-28 12:41:30,132][starVLA.training.train_starvla][INFO] - Step 2900, Loss: {'train/loss': 0.05853835865855217, 'train/grad_norm_pre_clip': 1.0907413959503174, 'timing/data': 0.0003354102373123169, 'timing/model': 0.42539263097569346, 'learning_rate/qwen_vl_interface': 7.625e-06, 'learning_rate/action_model': 7.625e-05, 'epoch': 0.53})
+[2026-07-28 12:41:36,200][starVLA.training.train_starvla][INFO] - Step 2901, Loss: {'train/loss': 0.020283803343772888, 'train/grad_norm_pre_clip': 0.5990360379219055, 'timing/data': 0.00027247006073594093, 'timing/model': 0.4297456401400268, 'learning_rate/qwen_vl_interface': 7.621685557887083e-06, 'learning_rate/action_model': 7.621685557887083e-05, 'epoch': 0.53})
+[2026-07-28 12:41:42,289][starVLA.training.train_starvla][INFO] - Step 2902, Loss: {'train/loss': 0.04035794362425804, 'train/grad_norm_pre_clip': 1.0989961624145508, 'timing/data': 0.0003681001253426075, 'timing/model': 0.43076768377795815, 'learning_rate/qwen_vl_interface': 7.618369576813157e-06, 'learning_rate/action_model': 7.618369576813157e-05, 'epoch': 0.53})
+[2026-07-28 12:41:48,709][starVLA.training.train_starvla][INFO] - Step 2903, Loss: {'train/loss': 0.03814972937107086, 'train/grad_norm_pre_clip': 2.449808120727539, 'timing/data': 0.0002770889550447464, 'timing/model': 0.6790473987348378, 'learning_rate/qwen_vl_interface': 7.615052058929924e-06, 'learning_rate/action_model': 7.615052058929923e-05, 'epoch': 0.53})
+[2026-07-28 12:41:54,583][starVLA.training.train_starvla][INFO] - Step 2904, Loss: {'train/loss': 0.03561076894402504, 'train/grad_norm_pre_clip': 1.0260738134384155, 'timing/data': 0.0002472861669957638, 'timing/model': 0.42172308219596744, 'learning_rate/qwen_vl_interface': 7.611733006390087e-06, 'learning_rate/action_model': 7.611733006390086e-05, 'epoch': 0.53})
+[2026-07-28 12:42:00,634][starVLA.training.train_starvla][INFO] - Step 2905, Loss: {'train/loss': 0.05146321654319763, 'train/grad_norm_pre_clip': 2.8123252391815186, 'timing/data': 0.000353299081325531, 'timing/model': 0.4378360491245985, 'learning_rate/qwen_vl_interface': 7.608412421347344e-06, 'learning_rate/action_model': 7.608412421347344e-05, 'epoch': 0.53})
+[2026-07-28 12:42:06,735][starVLA.training.train_starvla][INFO] - Step 2906, Loss: {'train/loss': 0.05991197004914284, 'train/grad_norm_pre_clip': 1.1685020923614502, 'timing/data': 0.00033060135319828987, 'timing/model': 0.42715279990807176, 'learning_rate/qwen_vl_interface': 7.605090305956384e-06, 'learning_rate/action_model': 7.605090305956383e-05, 'epoch': 0.53})
+[2026-07-28 12:42:12,887][starVLA.training.train_starvla][INFO] - Step 2907, Loss: {'train/loss': 0.03078506514430046, 'train/grad_norm_pre_clip': 1.871119499206543, 'timing/data': 0.00026464229449629784, 'timing/model': 0.4285728787072003, 'learning_rate/qwen_vl_interface': 7.6017666623728954e-06, 'learning_rate/action_model': 7.601766662372896e-05, 'epoch': 0.53})
+[2026-07-28 12:42:18,950][starVLA.training.train_starvla][INFO] - Step 2908, Loss: {'train/loss': 0.039871424436569214, 'train/grad_norm_pre_clip': 1.4490481615066528, 'timing/data': 0.0005520437844097614, 'timing/model': 0.42787360306829214, 'learning_rate/qwen_vl_interface': 7.598441492753551e-06, 'learning_rate/action_model': 7.598441492753551e-05, 'epoch': 0.53})
+[2026-07-28 12:42:25,219][starVLA.training.train_starvla][INFO] - Step 2909, Loss: {'train/loss': 0.03399454057216644, 'train/grad_norm_pre_clip': 0.755642831325531, 'timing/data': 0.0002556680701673031, 'timing/model': 0.42174692125990987, 'learning_rate/qwen_vl_interface': 7.595114799256018e-06, 'learning_rate/action_model': 7.595114799256017e-05, 'epoch': 0.53})
+[2026-07-28 12:42:31,468][starVLA.training.train_starvla][INFO] - Step 2910, Loss: {'train/loss': 0.03468218818306923, 'train/grad_norm_pre_clip': 0.5500509142875671, 'timing/data': 0.00026440806686878204, 'timing/model': 0.4242196506820619, 'learning_rate/qwen_vl_interface': 7.591786584038954e-06, 'learning_rate/action_model': 7.591786584038954e-05, 'epoch': 0.53})
+[2026-07-28 12:42:37,517][starVLA.training.train_starvla][INFO] - Step 2911, Loss: {'train/loss': 0.05704226344823837, 'train/grad_norm_pre_clip': 1.0802687406539917, 'timing/data': 0.00026456406340003014, 'timing/model': 0.4292619680054486, 'learning_rate/qwen_vl_interface': 7.588456849261999e-06, 'learning_rate/action_model': 7.588456849261998e-05, 'epoch': 0.53})
+[2026-07-28 12:42:43,574][starVLA.training.train_starvla][INFO] - Step 2912, Loss: {'train/loss': 0.06200157105922699, 'train/grad_norm_pre_clip': 1.194776177406311, 'timing/data': 0.00037764897570014, 'timing/model': 0.4315350390970707, 'learning_rate/qwen_vl_interface': 7.585125597085784e-06, 'learning_rate/action_model': 7.585125597085783e-05, 'epoch': 0.53})
+[2026-07-28 12:42:49,891][starVLA.training.train_starvla][INFO] - Step 2913, Loss: {'train/loss': 0.0485198087990284, 'train/grad_norm_pre_clip': 2.265974760055542, 'timing/data': 0.000267155934125185, 'timing/model': 0.4263323862105608, 'learning_rate/qwen_vl_interface': 7.58179282967192e-06, 'learning_rate/action_model': 7.58179282967192e-05, 'epoch': 0.54})
+[2026-07-28 12:42:56,203][starVLA.training.train_starvla][INFO] - Step 2914, Loss: {'train/loss': 0.03769548237323761, 'train/grad_norm_pre_clip': 1.135589361190796, 'timing/data': 0.0002827756106853485, 'timing/model': 0.6709357341751456, 'learning_rate/qwen_vl_interface': 7.578458549183005e-06, 'learning_rate/action_model': 7.578458549183005e-05, 'epoch': 0.54})
+[2026-07-28 12:43:02,263][starVLA.training.train_starvla][INFO] - Step 2915, Loss: {'train/loss': 0.09126278758049011, 'train/grad_norm_pre_clip': 2.4897356033325195, 'timing/data': 0.00034384801983833313, 'timing/model': 0.4299778379499912, 'learning_rate/qwen_vl_interface': 7.575122757782617e-06, 'learning_rate/action_model': 7.575122757782617e-05, 'epoch': 0.54})
+[2026-07-28 12:43:08,581][starVLA.training.train_starvla][INFO] - Step 2916, Loss: {'train/loss': 0.06459371745586395, 'train/grad_norm_pre_clip': 2.03205943107605, 'timing/data': 0.0003623072989284992, 'timing/model': 0.6696694432757795, 'learning_rate/qwen_vl_interface': 7.571785457635316e-06, 'learning_rate/action_model': 7.571785457635317e-05, 'epoch': 0.54})
+[2026-07-28 12:43:14,704][starVLA.training.train_starvla][INFO] - Step 2917, Loss: {'train/loss': 0.06575240194797516, 'train/grad_norm_pre_clip': 1.402085304260254, 'timing/data': 0.0003771246410906315, 'timing/model': 0.4337595426477492, 'learning_rate/qwen_vl_interface': 7.568446650906637e-06, 'learning_rate/action_model': 7.568446650906636e-05, 'epoch': 0.54})
+[2026-07-28 12:43:21,014][starVLA.training.train_starvla][INFO] - Step 2918, Loss: {'train/loss': 0.02912958711385727, 'train/grad_norm_pre_clip': 0.9236387610435486, 'timing/data': 0.00026626279577612877, 'timing/model': 0.42676588566973805, 'learning_rate/qwen_vl_interface': 7.565106339763099e-06, 'learning_rate/action_model': 7.565106339763099e-05, 'epoch': 0.54})
+[2026-07-28 12:43:27,074][starVLA.training.train_starvla][INFO] - Step 2919, Loss: {'train/loss': 0.04873036965727806, 'train/grad_norm_pre_clip': 0.9025545716285706, 'timing/data': 0.0002791937440633774, 'timing/model': 0.4279582751914859, 'learning_rate/qwen_vl_interface': 7.561764526372191e-06, 'learning_rate/action_model': 7.561764526372191e-05, 'epoch': 0.54})
+[2026-07-28 12:43:33,522][starVLA.training.train_starvla][INFO] - Step 2920, Loss: {'train/loss': 0.01848367229104042, 'train/grad_norm_pre_clip': 0.668838620185852, 'timing/data': 0.00026620691642165184, 'timing/model': 0.4279574858956039, 'learning_rate/qwen_vl_interface': 7.5584212129023796e-06, 'learning_rate/action_model': 7.558421212902379e-05, 'epoch': 0.54})
+[2026-07-28 12:43:39,575][starVLA.training.train_starvla][INFO] - Step 2921, Loss: {'train/loss': 0.045425161719322205, 'train/grad_norm_pre_clip': 1.9410158395767212, 'timing/data': 0.00027376599609851837, 'timing/model': 0.4269940787926316, 'learning_rate/qwen_vl_interface': 7.555076401523106e-06, 'learning_rate/action_model': 7.555076401523105e-05, 'epoch': 0.54})
+[2026-07-28 12:43:45,847][starVLA.training.train_starvla][INFO] - Step 2922, Loss: {'train/loss': 0.017800286412239075, 'train/grad_norm_pre_clip': 0.7936335206031799, 'timing/data': 0.0002755420282483101, 'timing/model': 0.6601107828319073, 'learning_rate/qwen_vl_interface': 7.551730094404782e-06, 'learning_rate/action_model': 7.551730094404781e-05, 'epoch': 0.54})
+[2026-07-28 12:43:52,139][starVLA.training.train_starvla][INFO] - Step 2923, Loss: {'train/loss': 0.036533206701278687, 'train/grad_norm_pre_clip': 2.043691873550415, 'timing/data': 0.00037165405228734016, 'timing/model': 0.42362347384914756, 'learning_rate/qwen_vl_interface': 7.5483822937187876e-06, 'learning_rate/action_model': 7.548382293718787e-05, 'epoch': 0.54})
+[2026-07-28 12:43:58,237][starVLA.training.train_starvla][INFO] - Step 2924, Loss: {'train/loss': 0.05888301879167557, 'train/grad_norm_pre_clip': 1.8809887170791626, 'timing/data': 0.0003262897953391075, 'timing/model': 0.4561447179876268, 'learning_rate/qwen_vl_interface': 7.545033001637478e-06, 'learning_rate/action_model': 7.545033001637479e-05, 'epoch': 0.54})
+[2026-07-28 12:44:04,326][starVLA.training.train_starvla][INFO] - Step 2925, Loss: {'train/loss': 0.04551016166806221, 'train/grad_norm_pre_clip': 2.369450807571411, 'timing/data': 0.0002832487225532532, 'timing/model': 0.45113550312817097, 'learning_rate/qwen_vl_interface': 7.54168222033417e-06, 'learning_rate/action_model': 7.54168222033417e-05, 'epoch': 0.54})
+[2026-07-28 12:44:10,767][starVLA.training.train_starvla][INFO] - Step 2926, Loss: {'train/loss': 0.056912779808044434, 'train/grad_norm_pre_clip': 1.3326627016067505, 'timing/data': 0.0003666616976261139, 'timing/model': 0.5969890598207712, 'learning_rate/qwen_vl_interface': 7.538329951983148e-06, 'learning_rate/action_model': 7.538329951983147e-05, 'epoch': 0.54})
+[2026-07-28 12:44:16,814][starVLA.training.train_starvla][INFO] - Step 2927, Loss: {'train/loss': 0.021603435277938843, 'train/grad_norm_pre_clip': 0.8411546349525452, 'timing/data': 0.000275379978120327, 'timing/model': 0.4241945859976113, 'learning_rate/qwen_vl_interface': 7.534976198759667e-06, 'learning_rate/action_model': 7.534976198759666e-05, 'epoch': 0.54})
+[2026-07-28 12:44:22,851][starVLA.training.train_starvla][INFO] - Step 2928, Loss: {'train/loss': 0.03309600055217743, 'train/grad_norm_pre_clip': 1.8089838027954102, 'timing/data': 0.00027155037969350815, 'timing/model': 0.430506425909698, 'learning_rate/qwen_vl_interface': 7.531620962839935e-06, 'learning_rate/action_model': 7.531620962839935e-05, 'epoch': 0.54})
+[2026-07-28 12:44:28,904][starVLA.training.train_starvla][INFO] - Step 2929, Loss: {'train/loss': 0.04740125685930252, 'train/grad_norm_pre_clip': 1.5107543468475342, 'timing/data': 0.00031076883897185326, 'timing/model': 0.42711076186969876, 'learning_rate/qwen_vl_interface': 7.5282642464011325e-06, 'learning_rate/action_model': 7.528264246401133e-05, 'epoch': 0.54})
+[2026-07-28 12:44:35,205][starVLA.training.train_starvla][INFO] - Step 2930, Loss: {'train/loss': 0.02407778799533844, 'train/grad_norm_pre_clip': 0.8895838260650635, 'timing/data': 0.0002586040645837784, 'timing/model': 0.4235817976295948, 'learning_rate/qwen_vl_interface': 7.524906051621397e-06, 'learning_rate/action_model': 7.524906051621396e-05, 'epoch': 0.54})
+[2026-07-28 12:44:41,067][starVLA.training.train_starvla][INFO] - Step 2931, Loss: {'train/loss': 0.012264305725693703, 'train/grad_norm_pre_clip': 0.9120172262191772, 'timing/data': 0.000271758995950222, 'timing/model': 0.42872093990445137, 'learning_rate/qwen_vl_interface': 7.521546380679822e-06, 'learning_rate/action_model': 7.521546380679822e-05, 'epoch': 0.54})
+[2026-07-28 12:44:47,154][starVLA.training.train_starvla][INFO] - Step 2932, Loss: {'train/loss': 0.04348244145512581, 'train/grad_norm_pre_clip': 1.6126502752304077, 'timing/data': 0.00028632907196879387, 'timing/model': 0.4323313543573022, 'learning_rate/qwen_vl_interface': 7.518185235756461e-06, 'learning_rate/action_model': 7.518185235756461e-05, 'epoch': 0.54})
+[2026-07-28 12:44:53,501][starVLA.training.train_starvla][INFO] - Step 2933, Loss: {'train/loss': 0.035549648106098175, 'train/grad_norm_pre_clip': 1.011282205581665, 'timing/data': 0.00026727886870503426, 'timing/model': 0.4224067502655089, 'learning_rate/qwen_vl_interface': 7.514822619032328e-06, 'learning_rate/action_model': 7.514822619032327e-05, 'epoch': 0.54})
+[2026-07-28 12:44:59,777][starVLA.training.train_starvla][INFO] - Step 2934, Loss: {'train/loss': 0.06439762562513351, 'train/grad_norm_pre_clip': 1.55135977268219, 'timing/data': 0.00027264002710580826, 'timing/model': 0.6924443701282144, 'learning_rate/qwen_vl_interface': 7.511458532689385e-06, 'learning_rate/action_model': 7.511458532689385e-05, 'epoch': 0.54})
+[2026-07-28 12:45:05,849][starVLA.training.train_starvla][INFO] - Step 2935, Loss: {'train/loss': 0.020099706947803497, 'train/grad_norm_pre_clip': 1.0279806852340698, 'timing/data': 0.0003214329481124878, 'timing/model': 0.42461230605840683, 'learning_rate/qwen_vl_interface': 7.508092978910553e-06, 'learning_rate/action_model': 7.508092978910552e-05, 'epoch': 0.54})
+[2026-07-28 12:45:11,937][starVLA.training.train_starvla][INFO] - Step 2936, Loss: {'train/loss': 0.040021032094955444, 'train/grad_norm_pre_clip': 2.4638383388519287, 'timing/data': 0.00028092507272958755, 'timing/model': 0.4266427787952125, 'learning_rate/qwen_vl_interface': 7.5047259598797015e-06, 'learning_rate/action_model': 7.504725959879701e-05, 'epoch': 0.54})
+[2026-07-28 12:45:18,299][starVLA.training.train_starvla][INFO] - Step 2937, Loss: {'train/loss': 0.017221547663211823, 'train/grad_norm_pre_clip': 1.113244891166687, 'timing/data': 0.00033741630613803864, 'timing/model': 0.6688189366832376, 'learning_rate/qwen_vl_interface': 7.501357477781654e-06, 'learning_rate/action_model': 7.501357477781654e-05, 'epoch': 0.54})
+[2026-07-28 12:45:24,377][starVLA.training.train_starvla][INFO] - Step 2938, Loss: {'train/loss': 0.053167231380939484, 'train/grad_norm_pre_clip': 1.9476044178009033, 'timing/data': 0.0002726907841861248, 'timing/model': 0.42369086714461446, 'learning_rate/qwen_vl_interface': 7.497987534802185e-06, 'learning_rate/action_model': 7.497987534802184e-05, 'epoch': 0.54})
+[2026-07-28 12:45:30,457][starVLA.training.train_starvla][INFO] - Step 2939, Loss: {'train/loss': 0.03660393878817558, 'train/grad_norm_pre_clip': 0.7237223982810974, 'timing/data': 0.0003640619106590748, 'timing/model': 0.4248805860988796, 'learning_rate/qwen_vl_interface': 7.494616133128008e-06, 'learning_rate/action_model': 7.494616133128008e-05, 'epoch': 0.54})
+[2026-07-28 12:45:36,724][starVLA.training.train_starvla][INFO] - Step 2940, Loss: {'train/loss': 0.03419048339128494, 'train/grad_norm_pre_clip': 1.8739674091339111, 'timing/data': 0.0003527989611029625, 'timing/model': 0.4199376478791237, 'learning_rate/qwen_vl_interface': 7.49124327494679e-06, 'learning_rate/action_model': 7.49124327494679e-05, 'epoch': 0.54})
+[2026-07-28 12:45:42,808][starVLA.training.train_starvla][INFO] - Step 2941, Loss: {'train/loss': 0.04915209859609604, 'train/grad_norm_pre_clip': 2.3955187797546387, 'timing/data': 0.00033767614513635635, 'timing/model': 0.4489973266609013, 'learning_rate/qwen_vl_interface': 7.487868962447147e-06, 'learning_rate/action_model': 7.487868962447147e-05, 'epoch': 0.54})
+[2026-07-28 12:45:49,133][starVLA.training.train_starvla][INFO] - Step 2942, Loss: {'train/loss': 0.038187719881534576, 'train/grad_norm_pre_clip': 1.4848663806915283, 'timing/data': 0.00034388666972517967, 'timing/model': 0.6725207720883191, 'learning_rate/qwen_vl_interface': 7.48449319781863e-06, 'learning_rate/action_model': 7.484493197818629e-05, 'epoch': 0.54})
+[2026-07-28 12:45:55,258][starVLA.training.train_starvla][INFO] - Step 2943, Loss: {'train/loss': 0.020878802984952927, 'train/grad_norm_pre_clip': 1.661020040512085, 'timing/data': 0.00026361318305134773, 'timing/model': 0.4243356231600046, 'learning_rate/qwen_vl_interface': 7.481115983251735e-06, 'learning_rate/action_model': 7.481115983251735e-05, 'epoch': 0.54})
+[2026-07-28 12:46:01,574][starVLA.training.train_starvla][INFO] - Step 2944, Loss: {'train/loss': 0.06960293650627136, 'train/grad_norm_pre_clip': 3.261375904083252, 'timing/data': 0.00035405997186899185, 'timing/model': 0.42145861499011517, 'learning_rate/qwen_vl_interface': 7.477737320937902e-06, 'learning_rate/action_model': 7.477737320937902e-05, 'epoch': 0.54})
+[2026-07-28 12:46:07,680][starVLA.training.train_starvla][INFO] - Step 2945, Loss: {'train/loss': 0.04451340436935425, 'train/grad_norm_pre_clip': 1.6491771936416626, 'timing/data': 0.0002682269550859928, 'timing/model': 0.4265073360875249, 'learning_rate/qwen_vl_interface': 7.474357213069507e-06, 'learning_rate/action_model': 7.474357213069507e-05, 'epoch': 0.54})
+[2026-07-28 12:46:13,792][starVLA.training.train_starvla][INFO] - Step 2946, Loss: {'train/loss': 0.04961654543876648, 'train/grad_norm_pre_clip': 2.153156042098999, 'timing/data': 0.0002850070595741272, 'timing/model': 0.43196650920435786, 'learning_rate/qwen_vl_interface': 7.470975661839868e-06, 'learning_rate/action_model': 7.470975661839868e-05, 'epoch': 0.54})
+[2026-07-28 12:46:20,112][starVLA.training.train_starvla][INFO] - Step 2947, Loss: {'train/loss': 0.06284633278846741, 'train/grad_norm_pre_clip': 2.0981643199920654, 'timing/data': 0.0002718959003686905, 'timing/model': 0.661906884983182, 'learning_rate/qwen_vl_interface': 7.4675926694432336e-06, 'learning_rate/action_model': 7.467592669443233e-05, 'epoch': 0.54})
+[2026-07-28 12:46:26,416][starVLA.training.train_starvla][INFO] - Step 2948, Loss: {'train/loss': 0.02455880679190159, 'train/grad_norm_pre_clip': 0.9397866725921631, 'timing/data': 0.00028131576254963875, 'timing/model': 0.6696833441965282, 'learning_rate/qwen_vl_interface': 7.4642082380747905e-06, 'learning_rate/action_model': 7.46420823807479e-05, 'epoch': 0.54})
+[2026-07-28 12:46:32,508][starVLA.training.train_starvla][INFO] - Step 2949, Loss: {'train/loss': 0.04000522941350937, 'train/grad_norm_pre_clip': 1.3697928190231323, 'timing/data': 0.0002637980505824089, 'timing/model': 0.4238644209690392, 'learning_rate/qwen_vl_interface': 7.460822369930664e-06, 'learning_rate/action_model': 7.460822369930664e-05, 'epoch': 0.54})
+[2026-07-28 12:46:38,554][starVLA.training.train_starvla][INFO] - Step 2950, Loss: {'train/loss': 0.04093684256076813, 'train/grad_norm_pre_clip': 1.5022177696228027, 'timing/data': 0.0002758209593594074, 'timing/model': 0.4276797422207892, 'learning_rate/qwen_vl_interface': 7.457435067207901e-06, 'learning_rate/action_model': 7.4574350672079e-05, 'epoch': 0.54})
+[2026-07-28 12:46:44,844][starVLA.training.train_starvla][INFO] - Step 2951, Loss: {'train/loss': 0.036459844559431076, 'train/grad_norm_pre_clip': 1.7276710271835327, 'timing/data': 0.0002606641501188278, 'timing/model': 0.42330086324363947, 'learning_rate/qwen_vl_interface': 7.454046332104491e-06, 'learning_rate/action_model': 7.45404633210449e-05, 'epoch': 0.54})
+[2026-07-28 12:46:50,923][starVLA.training.train_starvla][INFO] - Step 2952, Loss: {'train/loss': 0.04654601216316223, 'train/grad_norm_pre_clip': 0.9025406837463379, 'timing/data': 0.0002833842299878597, 'timing/model': 0.42639415711164474, 'learning_rate/qwen_vl_interface': 7.450656166819344e-06, 'learning_rate/action_model': 7.450656166819344e-05, 'epoch': 0.54})
+[2026-07-28 12:46:57,196][starVLA.training.train_starvla][INFO] - Step 2953, Loss: {'train/loss': 0.029284272342920303, 'train/grad_norm_pre_clip': 0.9811108112335205, 'timing/data': 0.0002515669912099838, 'timing/model': 0.42136860312893987, 'learning_rate/qwen_vl_interface': 7.447264573552303e-06, 'learning_rate/action_model': 7.447264573552302e-05, 'epoch': 0.54})
+[2026-07-28 12:47:03,500][starVLA.training.train_starvla][INFO] - Step 2954, Loss: {'train/loss': 0.0804172158241272, 'train/grad_norm_pre_clip': 1.7394016981124878, 'timing/data': 0.0003596632741391659, 'timing/model': 0.42337811877951026, 'learning_rate/qwen_vl_interface': 7.443871554504135e-06, 'learning_rate/action_model': 7.443871554504134e-05, 'epoch': 0.54})
+[2026-07-28 12:47:09,581][starVLA.training.train_starvla][INFO] - Step 2955, Loss: {'train/loss': 0.07138242572546005, 'train/grad_norm_pre_clip': 2.065462827682495, 'timing/data': 0.000271426048129797, 'timing/model': 0.42654835991561413, 'learning_rate/qwen_vl_interface': 7.440477111876535e-06, 'learning_rate/action_model': 7.440477111876535e-05, 'epoch': 0.54})
+[2026-07-28 12:47:15,613][starVLA.training.train_starvla][INFO] - Step 2956, Loss: {'train/loss': 0.01845048926770687, 'train/grad_norm_pre_clip': 0.5693857669830322, 'timing/data': 0.0002774321474134922, 'timing/model': 0.42480797693133354, 'learning_rate/qwen_vl_interface': 7.437081247872118e-06, 'learning_rate/action_model': 7.437081247872118e-05, 'epoch': 0.54})
+[2026-07-28 12:47:22,088][starVLA.training.train_starvla][INFO] - Step 2957, Loss: {'train/loss': 0.029964100569486618, 'train/grad_norm_pre_clip': 0.7119914293289185, 'timing/data': 0.00034453533589839935, 'timing/model': 0.5894843828864396, 'learning_rate/qwen_vl_interface': 7.433683964694425e-06, 'learning_rate/action_model': 7.433683964694425e-05, 'epoch': 0.54})
+[2026-07-28 12:47:28,168][starVLA.training.train_starvla][INFO] - Step 2958, Loss: {'train/loss': 0.03785150498151779, 'train/grad_norm_pre_clip': 0.9896271824836731, 'timing/data': 0.0003400901332497597, 'timing/model': 0.4226039117202163, 'learning_rate/qwen_vl_interface': 7.430285264547913e-06, 'learning_rate/action_model': 7.430285264547913e-05, 'epoch': 0.54})
+[2026-07-28 12:47:33,943][starVLA.training.train_starvla][INFO] - Step 2959, Loss: {'train/loss': 0.05895812809467316, 'train/grad_norm_pre_clip': 1.0776784420013428, 'timing/data': 0.0003360612317919731, 'timing/model': 0.42502940399572253, 'learning_rate/qwen_vl_interface': 7.426885149637964e-06, 'learning_rate/action_model': 7.426885149637965e-05, 'epoch': 0.54})
+[2026-07-28 12:47:39,764][starVLA.training.train_starvla][INFO] - Step 2960, Loss: {'train/loss': 0.03537718206644058, 'train/grad_norm_pre_clip': 0.731997013092041, 'timing/data': 0.000268684234470129, 'timing/model': 0.42747396416962147, 'learning_rate/qwen_vl_interface': 7.42348362217088e-06, 'learning_rate/action_model': 7.423483622170879e-05, 'epoch': 0.54})
+[2026-07-28 12:47:46,073][starVLA.training.train_starvla][INFO] - Step 2961, Loss: {'train/loss': 0.0628584623336792, 'train/grad_norm_pre_clip': 2.4373395442962646, 'timing/data': 0.00029411306604743004, 'timing/model': 0.44087261613458395, 'learning_rate/qwen_vl_interface': 7.420080684353865e-06, 'learning_rate/action_model': 7.420080684353865e-05, 'epoch': 0.54})
+[2026-07-28 12:47:52,365][starVLA.training.train_starvla][INFO] - Step 2962, Loss: {'train/loss': 0.02351538836956024, 'train/grad_norm_pre_clip': 1.3336365222930908, 'timing/data': 0.0007727458141744137, 'timing/model': 0.6642700773663819, 'learning_rate/qwen_vl_interface': 7.416676338395057e-06, 'learning_rate/action_model': 7.416676338395057e-05, 'epoch': 0.54})
+[2026-07-28 12:47:58,415][starVLA.training.train_starvla][INFO] - Step 2963, Loss: {'train/loss': 0.051756590604782104, 'train/grad_norm_pre_clip': 1.4411531686782837, 'timing/data': 0.0002523609437048435, 'timing/model': 0.42238264391198754, 'learning_rate/qwen_vl_interface': 7.413270586503496e-06, 'learning_rate/action_model': 7.413270586503496e-05, 'epoch': 0.54})
+[2026-07-28 12:48:04,436][starVLA.training.train_starvla][INFO] - Step 2964, Loss: {'train/loss': 0.04758215323090553, 'train/grad_norm_pre_clip': 1.5012098550796509, 'timing/data': 0.00027219904586672783, 'timing/model': 0.4251219960860908, 'learning_rate/qwen_vl_interface': 7.409863430889137e-06, 'learning_rate/action_model': 7.409863430889136e-05, 'epoch': 0.54})
+[2026-07-28 12:48:10,776][starVLA.training.train_starvla][INFO] - Step 2965, Loss: {'train/loss': 0.038315270096063614, 'train/grad_norm_pre_clip': 2.3121860027313232, 'timing/data': 0.0002652709372341633, 'timing/model': 0.6650576433166862, 'learning_rate/qwen_vl_interface': 7.406454873762848e-06, 'learning_rate/action_model': 7.406454873762848e-05, 'epoch': 0.54})
+[2026-07-28 12:48:16,786][starVLA.training.train_starvla][INFO] - Step 2966, Loss: {'train/loss': 0.018905363976955414, 'train/grad_norm_pre_clip': 0.8316324949264526, 'timing/data': 0.0003381171263754368, 'timing/model': 0.4206764670088887, 'learning_rate/qwen_vl_interface': 7.403044917336404e-06, 'learning_rate/action_model': 7.403044917336404e-05, 'epoch': 0.54})
+[2026-07-28 12:48:22,798][starVLA.training.train_starvla][INFO] - Step 2967, Loss: {'train/loss': 0.03091668151319027, 'train/grad_norm_pre_clip': 1.0591717958450317, 'timing/data': 0.00028016790747642517, 'timing/model': 0.42689606500789523, 'learning_rate/qwen_vl_interface': 7.399633563822487e-06, 'learning_rate/action_model': 7.399633563822487e-05, 'epoch': 0.54})
+[2026-07-28 12:48:28,870][starVLA.training.train_starvla][INFO] - Step 2968, Loss: {'train/loss': 0.03648696094751358, 'train/grad_norm_pre_clip': 1.4260220527648926, 'timing/data': 0.0003790631890296936, 'timing/model': 0.4331299839541316, 'learning_rate/qwen_vl_interface': 7.396220815434691e-06, 'learning_rate/action_model': 7.39622081543469e-05, 'epoch': 0.55})
+[2026-07-28 12:48:35,168][starVLA.training.train_starvla][INFO] - Step 2969, Loss: {'train/loss': 0.03060711920261383, 'train/grad_norm_pre_clip': 1.311395287513733, 'timing/data': 0.0002555488608777523, 'timing/model': 0.4242057492956519, 'learning_rate/qwen_vl_interface': 7.392806674387506e-06, 'learning_rate/action_model': 7.392806674387506e-05, 'epoch': 0.55})
+[2026-07-28 12:48:41,314][starVLA.training.train_starvla][INFO] - Step 2970, Loss: {'train/loss': 0.030509017407894135, 'train/grad_norm_pre_clip': 1.1740809679031372, 'timing/data': 0.0003577559255063534, 'timing/model': 0.4424158278852701, 'learning_rate/qwen_vl_interface': 7.389391142896336e-06, 'learning_rate/action_model': 7.389391142896335e-05, 'epoch': 0.55})
+[2026-07-28 12:48:47,589][starVLA.training.train_starvla][INFO] - Step 2971, Loss: {'train/loss': 0.023323699831962585, 'train/grad_norm_pre_clip': 1.256772518157959, 'timing/data': 0.0004018428735435009, 'timing/model': 0.4237931207753718, 'learning_rate/qwen_vl_interface': 7.38597422317748e-06, 'learning_rate/action_model': 7.385974223177479e-05, 'epoch': 0.55})
+[2026-07-28 12:48:53,856][starVLA.training.train_starvla][INFO] - Step 2972, Loss: {'train/loss': 0.061711013317108154, 'train/grad_norm_pre_clip': 2.5527596473693848, 'timing/data': 0.00035283109173178673, 'timing/model': 0.43149447720497847, 'learning_rate/qwen_vl_interface': 7.382555917448141e-06, 'learning_rate/action_model': 7.382555917448141e-05, 'epoch': 0.55})
+[2026-07-28 12:49:00,277][starVLA.training.train_starvla][INFO] - Step 2973, Loss: {'train/loss': 0.03929543122649193, 'train/grad_norm_pre_clip': 2.7839014530181885, 'timing/data': 0.00027030613273382187, 'timing/model': 0.6600532340817153, 'learning_rate/qwen_vl_interface': 7.379136227926418e-06, 'learning_rate/action_model': 7.379136227926418e-05, 'epoch': 0.55})
+[2026-07-28 12:49:06,276][starVLA.training.train_starvla][INFO] - Step 2974, Loss: {'train/loss': 0.04472516104578972, 'train/grad_norm_pre_clip': 2.3874223232269287, 'timing/data': 0.0002665002830326557, 'timing/model': 0.42620425019413233, 'learning_rate/qwen_vl_interface': 7.3757151568313144e-06, 'learning_rate/action_model': 7.375715156831315e-05, 'epoch': 0.55})
+[2026-07-28 12:49:12,609][starVLA.training.train_starvla][INFO] - Step 2975, Loss: {'train/loss': 0.04757971316576004, 'train/grad_norm_pre_clip': 1.6242694854736328, 'timing/data': 0.0002606282941997051, 'timing/model': 0.4210575632750988, 'learning_rate/qwen_vl_interface': 7.3722927063827235e-06, 'learning_rate/action_model': 7.372292706382723e-05, 'epoch': 0.55})
+[2026-07-28 12:49:18,636][starVLA.training.train_starvla][INFO] - Step 2976, Loss: {'train/loss': 0.014768943190574646, 'train/grad_norm_pre_clip': 1.0594931840896606, 'timing/data': 0.000272905919700861, 'timing/model': 0.42594690900295973, 'learning_rate/qwen_vl_interface': 7.368868878801437e-06, 'learning_rate/action_model': 7.368868878801437e-05, 'epoch': 0.55})
+[2026-07-28 12:49:24,883][starVLA.training.train_starvla][INFO] - Step 2977, Loss: {'train/loss': 0.060565073043107986, 'train/grad_norm_pre_clip': 2.698657512664795, 'timing/data': 0.00024757999926805496, 'timing/model': 0.42024572286754847, 'learning_rate/qwen_vl_interface': 7.365443676309139e-06, 'learning_rate/action_model': 7.365443676309139e-05, 'epoch': 0.55})
+[2026-07-28 12:49:31,102][starVLA.training.train_starvla][INFO] - Step 2978, Loss: {'train/loss': 0.044218793511390686, 'train/grad_norm_pre_clip': 1.8827908039093018, 'timing/data': 0.00027771666646003723, 'timing/model': 0.6593338092789054, 'learning_rate/qwen_vl_interface': 7.362017101128406e-06, 'learning_rate/action_model': 7.362017101128406e-05, 'epoch': 0.55})
+[2026-07-28 12:49:37,255][starVLA.training.train_starvla][INFO] - Step 2979, Loss: {'train/loss': 0.022070113569498062, 'train/grad_norm_pre_clip': 1.9092915058135986, 'timing/data': 0.0002714451402425766, 'timing/model': 0.4314601323567331, 'learning_rate/qwen_vl_interface': 7.358589155482707e-06, 'learning_rate/action_model': 7.358589155482707e-05, 'epoch': 0.55})
+[2026-07-28 12:49:43,260][starVLA.training.train_starvla][INFO] - Step 2980, Loss: {'train/loss': 0.09630945324897766, 'train/grad_norm_pre_clip': 3.906163215637207, 'timing/data': 0.0002630767412483692, 'timing/model': 0.44031552597880363, 'learning_rate/qwen_vl_interface': 7.355159841596397e-06, 'learning_rate/action_model': 7.355159841596397e-05, 'epoch': 0.55})
+[2026-07-28 12:49:49,478][starVLA.training.train_starvla][INFO] - Step 2981, Loss: {'train/loss': 0.0315624475479126, 'train/grad_norm_pre_clip': 0.8663227558135986, 'timing/data': 0.0003288812004029751, 'timing/model': 0.42122937226668, 'learning_rate/qwen_vl_interface': 7.351729161694722e-06, 'learning_rate/action_model': 7.351729161694722e-05, 'epoch': 0.55})
+[2026-07-28 12:49:55,767][starVLA.training.train_starvla][INFO] - Step 2982, Loss: {'train/loss': 0.09519942104816437, 'train/grad_norm_pre_clip': 1.8152726888656616, 'timing/data': 0.00035665882751345634, 'timing/model': 0.4237745641730726, 'learning_rate/qwen_vl_interface': 7.348297118003811e-06, 'learning_rate/action_model': 7.348297118003811e-05, 'epoch': 0.55})
+[2026-07-28 12:50:01,818][starVLA.training.train_starvla][INFO] - Step 2983, Loss: {'train/loss': 0.057109154760837555, 'train/grad_norm_pre_clip': 1.4674153327941895, 'timing/data': 0.0002783113159239292, 'timing/model': 0.43606136180460453, 'learning_rate/qwen_vl_interface': 7.3448637127506815e-06, 'learning_rate/action_model': 7.344863712750682e-05, 'epoch': 0.55})
+[2026-07-28 12:50:08,203][starVLA.training.train_starvla][INFO] - Step 2984, Loss: {'train/loss': 0.03393974155187607, 'train/grad_norm_pre_clip': 1.2095800638198853, 'timing/data': 0.0003523556515574455, 'timing/model': 0.6856379262171686, 'learning_rate/qwen_vl_interface': 7.341428948163232e-06, 'learning_rate/action_model': 7.341428948163231e-05, 'epoch': 0.55})
+[2026-07-28 12:50:14,502][starVLA.training.train_starvla][INFO] - Step 2985, Loss: {'train/loss': 0.06800000369548798, 'train/grad_norm_pre_clip': 1.8967469930648804, 'timing/data': 0.00034957611933350563, 'timing/model': 0.42255648085847497, 'learning_rate/qwen_vl_interface': 7.337992826470245e-06, 'learning_rate/action_model': 7.337992826470245e-05, 'epoch': 0.55})
+[2026-07-28 12:50:20,593][starVLA.training.train_starvla][INFO] - Step 2986, Loss: {'train/loss': 0.04752951115369797, 'train/grad_norm_pre_clip': 1.5876847505569458, 'timing/data': 0.00035928282886743546, 'timing/model': 0.43145846016705036, 'learning_rate/qwen_vl_interface': 7.334555349901381e-06, 'learning_rate/action_model': 7.33455534990138e-05, 'epoch': 0.55})
+[2026-07-28 12:50:26,687][starVLA.training.train_starvla][INFO] - Step 2987, Loss: {'train/loss': 0.06737373769283295, 'train/grad_norm_pre_clip': 1.9101253747940063, 'timing/data': 0.00036014895886182785, 'timing/model': 0.4261480700224638, 'learning_rate/qwen_vl_interface': 7.331116520687184e-06, 'learning_rate/action_model': 7.331116520687184e-05, 'epoch': 0.55})
+[2026-07-28 12:50:32,741][starVLA.training.train_starvla][INFO] - Step 2988, Loss: {'train/loss': 0.034316472709178925, 'train/grad_norm_pre_clip': 0.8713910579681396, 'timing/data': 0.0002685859799385071, 'timing/model': 0.4256123770028353, 'learning_rate/qwen_vl_interface': 7.32767634105907e-06, 'learning_rate/action_model': 7.32767634105907e-05, 'epoch': 0.55})
+[2026-07-28 12:50:39,079][starVLA.training.train_starvla][INFO] - Step 2989, Loss: {'train/loss': 0.04176716133952141, 'train/grad_norm_pre_clip': 1.1191606521606445, 'timing/data': 0.0005005858838558197, 'timing/model': 0.6779965003952384, 'learning_rate/qwen_vl_interface': 7.324234813249334e-06, 'learning_rate/action_model': 7.324234813249335e-05, 'epoch': 0.55})
+[2026-07-28 12:50:44,890][starVLA.training.train_starvla][INFO] - Step 2990, Loss: {'train/loss': 0.018844200298190117, 'train/grad_norm_pre_clip': 0.5412821769714355, 'timing/data': 0.0002558259293437004, 'timing/model': 0.4282878949306905, 'learning_rate/qwen_vl_interface': 7.320791939491149e-06, 'learning_rate/action_model': 7.320791939491149e-05, 'epoch': 0.55})
+[2026-07-28 12:50:50,723][starVLA.training.train_starvla][INFO] - Step 2991, Loss: {'train/loss': 0.04856259003281593, 'train/grad_norm_pre_clip': 1.1463265419006348, 'timing/data': 0.0003572329878807068, 'timing/model': 0.43381907511502504, 'learning_rate/qwen_vl_interface': 7.317347722018557e-06, 'learning_rate/action_model': 7.317347722018557e-05, 'epoch': 0.55})
+[2026-07-28 12:50:56,763][starVLA.training.train_starvla][INFO] - Step 2992, Loss: {'train/loss': 0.03544269874691963, 'train/grad_norm_pre_clip': 0.5798368453979492, 'timing/data': 0.0002838396467268467, 'timing/model': 0.42549649393185973, 'learning_rate/qwen_vl_interface': 7.313902163066472e-06, 'learning_rate/action_model': 7.313902163066471e-05, 'epoch': 0.55})
+[2026-07-28 12:51:02,819][starVLA.training.train_starvla][INFO] - Step 2993, Loss: {'train/loss': 0.030931923538446426, 'train/grad_norm_pre_clip': 0.944329023361206, 'timing/data': 0.0002670660614967346, 'timing/model': 0.432378479745239, 'learning_rate/qwen_vl_interface': 7.3104552648706805e-06, 'learning_rate/action_model': 7.310455264870681e-05, 'epoch': 0.55})
+[2026-07-28 12:51:09,000][starVLA.training.train_starvla][INFO] - Step 2994, Loss: {'train/loss': 0.0474417582154274, 'train/grad_norm_pre_clip': 1.3342173099517822, 'timing/data': 0.0002719629555940628, 'timing/model': 0.4462283100001514, 'learning_rate/qwen_vl_interface': 7.307007029667839e-06, 'learning_rate/action_model': 7.30700702966784e-05, 'epoch': 0.55})
+[2026-07-28 12:51:15,550][starVLA.training.train_starvla][INFO] - Step 2995, Loss: {'train/loss': 0.06812003254890442, 'train/grad_norm_pre_clip': 2.3517346382141113, 'timing/data': 0.0002740100026130676, 'timing/model': 0.6527161761187017, 'learning_rate/qwen_vl_interface': 7.303557459695468e-06, 'learning_rate/action_model': 7.303557459695469e-05, 'epoch': 0.55})
+[2026-07-28 12:51:21,623][starVLA.training.train_starvla][INFO] - Step 2996, Loss: {'train/loss': 0.04481155797839165, 'train/grad_norm_pre_clip': 1.793549656867981, 'timing/data': 0.00027375388890504837, 'timing/model': 0.43118742294609547, 'learning_rate/qwen_vl_interface': 7.300106557191956e-06, 'learning_rate/action_model': 7.300106557191956e-05, 'epoch': 0.55})
+[2026-07-28 12:51:27,773][starVLA.training.train_starvla][INFO] - Step 2997, Loss: {'train/loss': 0.03802266716957092, 'train/grad_norm_pre_clip': 0.9071022868156433, 'timing/data': 0.0002682730555534363, 'timing/model': 0.4326882599852979, 'learning_rate/qwen_vl_interface': 7.296654324396557e-06, 'learning_rate/action_model': 7.296654324396556e-05, 'epoch': 0.55})
+[2026-07-28 12:51:33,910][starVLA.training.train_starvla][INFO] - Step 2998, Loss: {'train/loss': 0.055662401020526886, 'train/grad_norm_pre_clip': 1.171133041381836, 'timing/data': 0.00035139312967658043, 'timing/model': 0.43073859019204974, 'learning_rate/qwen_vl_interface': 7.2932007635493875e-06, 'learning_rate/action_model': 7.293200763549387e-05, 'epoch': 0.55})
+[2026-07-28 12:51:40,300][starVLA.training.train_starvla][INFO] - Step 2999, Loss: {'train/loss': 0.05245096981525421, 'train/grad_norm_pre_clip': 1.1315077543258667, 'timing/data': 0.00032736407592892647, 'timing/model': 0.4231615848839283, 'learning_rate/qwen_vl_interface': 7.289745876891427e-06, 'learning_rate/action_model': 7.289745876891426e-05, 'epoch': 0.55})
+[2026-07-28 12:51:46,613][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3000
+[2026-07-28 12:51:51,708][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3000
+[2026-07-28 12:59:31,494][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3000
+[2026-07-28 12:59:31,495][starVLA.training.train_starvla][INFO] - Step 3000, Loss: {'train/loss': 0.01758541911840439, 'train/grad_norm_pre_clip': 0.6382648944854736, 'eval/action_loss/seconds': 5.089873878285289, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.04101668670773506, 'eval/latency_3/loss': 0.04101668670773506, 'eval/flappy/loss': 0.04101668670773506, 'eval/flappy/latency_3/loss': 0.04101668670773506, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 358.67500422038137, 'rl_games_eval/mid_train/mean_length': 2908.4, 'rl_games_eval/mid_train/std_reward': 133.2495864541459, 'rl_games_eval/mid_train/std_length': 1073.4027855376564, 'rl_games_eval/mid_train/macro_mean_reward': 358.67500422038137, 'rl_games_eval/mid_train/macro_mean_length': 2908.4, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 358.67500422038137, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2908.4, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 133.2495864541459, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1073.4027855376564, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 358.67500422038137, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00025711115449666977, 'timing/model': 0.4239437421783805, 'learning_rate/qwen_vl_interface': 7.286289666664508e-06, 'learning_rate/action_model': 7.286289666664508e-05, 'epoch': 0.55})
+[2026-07-28 12:59:31,498][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state
+[2026-07-28 12:59:31,498][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 13:00:08,699][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state/pytorch_model
+[2026-07-28 13:00:08,700][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state/sampler.bin
+[2026-07-28 13:00:08,700][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state/sampler_1.bin
+[2026-07-28 13:00:08,707][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state/random_states_0.pkl
+[2026-07-28 13:00:16,762][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3000
+[2026-07-28 13:00:16,762][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 13:00:16,784][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 13:00:23,102][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.040024712681770325, 'train/grad_norm_pre_clip': 1.098411202430725, 'timing/data': 0.00028838589787483215, 'timing/model': 0.4303189888596535, 'learning_rate/qwen_vl_interface': 7.282832135111334e-06, 'learning_rate/action_model': 7.282832135111334e-05, 'epoch': 0.55})
+[2026-07-28 13:00:29,414][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.031510818749666214, 'train/grad_norm_pre_clip': 1.154933214187622, 'timing/data': 0.00028490694239735603, 'timing/model': 0.4241685848683119, 'learning_rate/qwen_vl_interface': 7.2793732844754584e-06, 'learning_rate/action_model': 7.279373284475459e-05, 'epoch': 0.55})
+[2026-07-28 13:00:35,853][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.016214191913604736, 'train/grad_norm_pre_clip': 0.6057056784629822, 'timing/data': 0.0002595977857708931, 'timing/model': 0.42361055314540863, 'learning_rate/qwen_vl_interface': 7.275913117001291e-06, 'learning_rate/action_model': 7.27591311700129e-05, 'epoch': 0.55})
+[2026-07-28 13:00:41,972][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.019159574061632156, 'train/grad_norm_pre_clip': 0.8773205280303955, 'timing/data': 0.00044270092621445656, 'timing/model': 0.42221916280686855, 'learning_rate/qwen_vl_interface': 7.272451634934096e-06, 'learning_rate/action_model': 7.272451634934096e-05, 'epoch': 0.55})
+[2026-07-28 13:00:48,312][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.03277231380343437, 'train/grad_norm_pre_clip': 0.7782281637191772, 'timing/data': 0.000288425013422966, 'timing/model': 0.42163505032658577, 'learning_rate/qwen_vl_interface': 7.268988840519994e-06, 'learning_rate/action_model': 7.268988840519993e-05, 'epoch': 0.55})
+[2026-07-28 13:00:54,631][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.10092486441135406, 'train/grad_norm_pre_clip': 5.4435038566589355, 'timing/data': 0.0002800910733640194, 'timing/model': 0.4262900981120765, 'learning_rate/qwen_vl_interface': 7.265524736005951e-06, 'learning_rate/action_model': 7.26552473600595e-05, 'epoch': 0.55})
+[2026-07-28 13:01:00,759][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.04519142955541611, 'train/grad_norm_pre_clip': 1.5306326150894165, 'timing/data': 0.0002713501453399658, 'timing/model': 0.4536785837262869, 'learning_rate/qwen_vl_interface': 7.262059323639792e-06, 'learning_rate/action_model': 7.262059323639792e-05, 'epoch': 0.55})
+[2026-07-28 13:01:06,958][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.0699818953871727, 'train/grad_norm_pre_clip': 3.2102444171905518, 'timing/data': 0.000299213919788599, 'timing/model': 0.42379677714779973, 'learning_rate/qwen_vl_interface': 7.25859260567018e-06, 'learning_rate/action_model': 7.258592605670179e-05, 'epoch': 0.55})
+[2026-07-28 13:01:13,546][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.03781649097800255, 'train/grad_norm_pre_clip': 1.5494494438171387, 'timing/data': 0.0002697720192372799, 'timing/model': 0.8384364889934659, 'learning_rate/qwen_vl_interface': 7.2551245843466334e-06, 'learning_rate/action_model': 7.255124584346633e-05, 'epoch': 0.55})
+[2026-07-28 13:01:19,807][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.08635637164115906, 'train/grad_norm_pre_clip': 4.011074066162109, 'timing/data': 0.0002720719203352928, 'timing/model': 0.6527333231642842, 'learning_rate/qwen_vl_interface': 7.251655261919515e-06, 'learning_rate/action_model': 7.251655261919515e-05, 'epoch': 0.55})
+[2026-07-28 13:01:25,647][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.026263803243637085, 'train/grad_norm_pre_clip': 0.8384131193161011, 'timing/data': 0.0002802531234920025, 'timing/model': 0.4336664159782231, 'learning_rate/qwen_vl_interface': 7.248184640640027e-06, 'learning_rate/action_model': 7.248184640640026e-05, 'epoch': 0.55})
+[2026-07-28 13:01:31,549][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.04215363785624504, 'train/grad_norm_pre_clip': 3.0853354930877686, 'timing/data': 0.00034615909680724144, 'timing/model': 0.41794836707413197, 'learning_rate/qwen_vl_interface': 7.24471272276022e-06, 'learning_rate/action_model': 7.24471272276022e-05, 'epoch': 0.55})
+[2026-07-28 13:01:37,506][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.057443130761384964, 'train/grad_norm_pre_clip': 2.2932050228118896, 'timing/data': 0.00036794785410165787, 'timing/model': 0.41801176592707634, 'learning_rate/qwen_vl_interface': 7.241239510532985e-06, 'learning_rate/action_model': 7.241239510532984e-05, 'epoch': 0.55})
+[2026-07-28 13:01:43,643][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.04572812467813492, 'train/grad_norm_pre_clip': 1.8174740076065063, 'timing/data': 0.00028717517852783203, 'timing/model': 0.4161430737003684, 'learning_rate/qwen_vl_interface': 7.237765006212049e-06, 'learning_rate/action_model': 7.237765006212048e-05, 'epoch': 0.55})
+[2026-07-28 13:01:49,935][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.07848434150218964, 'train/grad_norm_pre_clip': 2.5772314071655273, 'timing/data': 0.000338131096214056, 'timing/model': 0.41796224704012275, 'learning_rate/qwen_vl_interface': 7.2342892120519805e-06, 'learning_rate/action_model': 7.23428921205198e-05, 'epoch': 0.55})
+[2026-07-28 13:01:55,926][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.03906841576099396, 'train/grad_norm_pre_clip': 0.9673107862472534, 'timing/data': 0.00047735264524817467, 'timing/model': 0.41651881113648415, 'learning_rate/qwen_vl_interface': 7.230812130308187e-06, 'learning_rate/action_model': 7.230812130308186e-05, 'epoch': 0.55})
+[2026-07-28 13:02:01,849][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.054592493921518326, 'train/grad_norm_pre_clip': 2.3331801891326904, 'timing/data': 0.00029562506824731827, 'timing/model': 0.4188384870067239, 'learning_rate/qwen_vl_interface': 7.227333763236907e-06, 'learning_rate/action_model': 7.227333763236906e-05, 'epoch': 0.55})
+[2026-07-28 13:02:07,770][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.04693462327122688, 'train/grad_norm_pre_clip': 1.055706262588501, 'timing/data': 0.00030561909079551697, 'timing/model': 0.41710493015125394, 'learning_rate/qwen_vl_interface': 7.223854113095214e-06, 'learning_rate/action_model': 7.223854113095214e-05, 'epoch': 0.55})
+[2026-07-28 13:02:14,021][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.041330087929964066, 'train/grad_norm_pre_clip': 0.6995704174041748, 'timing/data': 0.00027256784960627556, 'timing/model': 0.41814032290130854, 'learning_rate/qwen_vl_interface': 7.220373182141021e-06, 'learning_rate/action_model': 7.22037318214102e-05, 'epoch': 0.55})
+[2026-07-28 13:02:19,957][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.041756317019462585, 'train/grad_norm_pre_clip': 1.1386030912399292, 'timing/data': 0.000292779877781868, 'timing/model': 0.4190133139491081, 'learning_rate/qwen_vl_interface': 7.216890972633057e-06, 'learning_rate/action_model': 7.216890972633057e-05, 'epoch': 0.55})
+[2026-07-28 13:02:25,929][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.031248988583683968, 'train/grad_norm_pre_clip': 1.4428578615188599, 'timing/data': 0.00027476996183395386, 'timing/model': 0.41695470781996846, 'learning_rate/qwen_vl_interface': 7.213407486830899e-06, 'learning_rate/action_model': 7.213407486830899e-05, 'epoch': 0.55})
+[2026-07-28 13:02:32,105][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.05336301401257515, 'train/grad_norm_pre_clip': 1.3331588506698608, 'timing/data': 0.0002778051421046257, 'timing/model': 0.4153245519846678, 'learning_rate/qwen_vl_interface': 7.209922726994937e-06, 'learning_rate/action_model': 7.209922726994936e-05, 'epoch': 0.56})
+[2026-07-28 13:02:38,052][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.022434396669268608, 'train/grad_norm_pre_clip': 0.917888343334198, 'timing/data': 0.00028265733271837234, 'timing/model': 0.41684533935040236, 'learning_rate/qwen_vl_interface': 7.206436695386396e-06, 'learning_rate/action_model': 7.206436695386396e-05, 'epoch': 0.56})
+[2026-07-28 13:02:43,992][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.031301502138376236, 'train/grad_norm_pre_clip': 1.1238452196121216, 'timing/data': 0.00035004690289497375, 'timing/model': 0.4165753913111985, 'learning_rate/qwen_vl_interface': 7.202949394267326e-06, 'learning_rate/action_model': 7.202949394267326e-05, 'epoch': 0.56})
+[2026-07-28 13:02:50,141][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.03608445078134537, 'train/grad_norm_pre_clip': 1.1228232383728027, 'timing/data': 0.00034093810245394707, 'timing/model': 0.41811355762183666, 'learning_rate/qwen_vl_interface': 7.199460825900596e-06, 'learning_rate/action_model': 7.199460825900596e-05, 'epoch': 0.56})
+[2026-07-28 13:02:56,081][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.04599073529243469, 'train/grad_norm_pre_clip': 1.548029899597168, 'timing/data': 0.0002836277708411217, 'timing/model': 0.4162238510325551, 'learning_rate/qwen_vl_interface': 7.1959709925499e-06, 'learning_rate/action_model': 7.1959709925499e-05, 'epoch': 0.56})
+[2026-07-28 13:03:01,989][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.04144877940416336, 'train/grad_norm_pre_clip': 0.8069280385971069, 'timing/data': 0.00037188176065683365, 'timing/model': 0.4179938738234341, 'learning_rate/qwen_vl_interface': 7.1924798964797545e-06, 'learning_rate/action_model': 7.192479896479754e-05, 'epoch': 0.56})
+[2026-07-28 13:03:07,924][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.03487461805343628, 'train/grad_norm_pre_clip': 0.9424321055412292, 'timing/data': 0.0003407970070838928, 'timing/model': 0.4160717809572816, 'learning_rate/qwen_vl_interface': 7.188987539955495e-06, 'learning_rate/action_model': 7.188987539955494e-05, 'epoch': 0.56})
+[2026-07-28 13:03:13,621][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.0573539212346077, 'train/grad_norm_pre_clip': 3.0392444133758545, 'timing/data': 0.00028158118948340416, 'timing/model': 0.4178076246753335, 'learning_rate/qwen_vl_interface': 7.185493925243271e-06, 'learning_rate/action_model': 7.185493925243271e-05, 'epoch': 0.56})
+[2026-07-28 13:03:20,082][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.03491400554776192, 'train/grad_norm_pre_clip': 1.5296125411987305, 'timing/data': 0.0003446531482040882, 'timing/model': 0.4165614959783852, 'learning_rate/qwen_vl_interface': 7.181999054610053e-06, 'learning_rate/action_model': 7.181999054610052e-05, 'epoch': 0.56})
+[2026-07-28 13:03:26,000][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.02352140098810196, 'train/grad_norm_pre_clip': 0.8383126258850098, 'timing/data': 0.0003260299563407898, 'timing/model': 0.41579431807622313, 'learning_rate/qwen_vl_interface': 7.178502930323622e-06, 'learning_rate/action_model': 7.178502930323622e-05, 'epoch': 0.56})
+[2026-07-28 13:03:31,956][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.044016651809215546, 'train/grad_norm_pre_clip': 1.5635334253311157, 'timing/data': 0.00026386789977550507, 'timing/model': 0.416185628157109, 'learning_rate/qwen_vl_interface': 7.175005554652577e-06, 'learning_rate/action_model': 7.175005554652578e-05, 'epoch': 0.56})
+[2026-07-28 13:03:37,861][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.021008946001529694, 'train/grad_norm_pre_clip': 0.5972880721092224, 'timing/data': 0.0002811141312122345, 'timing/model': 0.4175880323164165, 'learning_rate/qwen_vl_interface': 7.171506929866328e-06, 'learning_rate/action_model': 7.171506929866328e-05, 'epoch': 0.56})
+[2026-07-28 13:03:44,000][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.035978466272354126, 'train/grad_norm_pre_clip': 0.8761951327323914, 'timing/data': 0.0002837618812918663, 'timing/model': 0.41336334217339754, 'learning_rate/qwen_vl_interface': 7.168007058235095e-06, 'learning_rate/action_model': 7.168007058235095e-05, 'epoch': 0.56})
+[2026-07-28 13:03:50,180][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.020167194306850433, 'train/grad_norm_pre_clip': 0.8570349812507629, 'timing/data': 0.00028298888355493546, 'timing/model': 0.6665126881562173, 'learning_rate/qwen_vl_interface': 7.164505942029905e-06, 'learning_rate/action_model': 7.164505942029904e-05, 'epoch': 0.56})
+[2026-07-28 13:03:56,046][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.03457809239625931, 'train/grad_norm_pre_clip': 1.6166075468063354, 'timing/data': 0.0002633100375533104, 'timing/model': 0.4130834760144353, 'learning_rate/qwen_vl_interface': 7.161003583522594e-06, 'learning_rate/action_model': 7.161003583522593e-05, 'epoch': 0.56})
+[2026-07-28 13:04:02,168][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.04391235113143921, 'train/grad_norm_pre_clip': 0.8145190477371216, 'timing/data': 0.00033448170870542526, 'timing/model': 0.4148354330100119, 'learning_rate/qwen_vl_interface': 7.157499984985804e-06, 'learning_rate/action_model': 7.157499984985803e-05, 'epoch': 0.56})
+[2026-07-28 13:04:08,105][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.06571075320243835, 'train/grad_norm_pre_clip': 3.474679470062256, 'timing/data': 0.00026392005383968353, 'timing/model': 0.4146673930808902, 'learning_rate/qwen_vl_interface': 7.153995148692985e-06, 'learning_rate/action_model': 7.153995148692985e-05, 'epoch': 0.56})
+[2026-07-28 13:04:13,992][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.03681580722332001, 'train/grad_norm_pre_clip': 1.1479215621948242, 'timing/data': 0.0002657659351825714, 'timing/model': 0.41541768331080675, 'learning_rate/qwen_vl_interface': 7.150489076918383e-06, 'learning_rate/action_model': 7.150489076918384e-05, 'epoch': 0.56})
+[2026-07-28 13:04:20,108][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.02329391986131668, 'train/grad_norm_pre_clip': 1.04865300655365, 'timing/data': 0.000356034841388464, 'timing/model': 0.4126767721027136, 'learning_rate/qwen_vl_interface': 7.1469817719370535e-06, 'learning_rate/action_model': 7.146981771937053e-05, 'epoch': 0.56})
+[2026-07-28 13:04:26,258][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.014297508634626865, 'train/grad_norm_pre_clip': 0.6659245491027832, 'timing/data': 0.00026219524443149567, 'timing/model': 0.6488088080659509, 'learning_rate/qwen_vl_interface': 7.1434732360248505e-06, 'learning_rate/action_model': 7.14347323602485e-05, 'epoch': 0.56})
+[2026-07-28 13:04:32,189][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.058176666498184204, 'train/grad_norm_pre_clip': 2.788658380508423, 'timing/data': 0.0002652909606695175, 'timing/model': 0.42985552782192826, 'learning_rate/qwen_vl_interface': 7.139963471458418e-06, 'learning_rate/action_model': 7.139963471458418e-05, 'epoch': 0.56})
+[2026-07-28 13:04:38,309][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.040476005524396896, 'train/grad_norm_pre_clip': 1.74996817111969, 'timing/data': 0.0005794609896838665, 'timing/model': 0.4145014835521579, 'learning_rate/qwen_vl_interface': 7.136452480515208e-06, 'learning_rate/action_model': 7.136452480515208e-05, 'epoch': 0.56})
+[2026-07-28 13:04:44,271][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.02579076588153839, 'train/grad_norm_pre_clip': 1.592244267463684, 'timing/data': 0.00033744098618626595, 'timing/model': 0.4178557158447802, 'learning_rate/qwen_vl_interface': 7.132940265473466e-06, 'learning_rate/action_model': 7.132940265473466e-05, 'epoch': 0.56})
+[2026-07-28 13:04:50,473][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.02260640822350979, 'train/grad_norm_pre_clip': 0.8431066870689392, 'timing/data': 0.0002802349627017975, 'timing/model': 0.41660191072151065, 'learning_rate/qwen_vl_interface': 7.129426828612227e-06, 'learning_rate/action_model': 7.129426828612227e-05, 'epoch': 0.56})
+[2026-07-28 13:04:56,446][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.04896925389766693, 'train/grad_norm_pre_clip': 0.9753277897834778, 'timing/data': 0.00027460744604468346, 'timing/model': 0.4173494935967028, 'learning_rate/qwen_vl_interface': 7.125912172211326e-06, 'learning_rate/action_model': 7.125912172211325e-05, 'epoch': 0.56})
+[2026-07-28 13:05:02,599][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.03042668290436268, 'train/grad_norm_pre_clip': 2.671189785003662, 'timing/data': 0.00034771813079714775, 'timing/model': 0.6526834350079298, 'learning_rate/qwen_vl_interface': 7.122396298551381e-06, 'learning_rate/action_model': 7.12239629855138e-05, 'epoch': 0.56})
+[2026-07-28 13:05:08,521][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.05055547133088112, 'train/grad_norm_pre_clip': 1.5538980960845947, 'timing/data': 0.0002657040022313595, 'timing/model': 0.4145831800997257, 'learning_rate/qwen_vl_interface': 7.118879209913806e-06, 'learning_rate/action_model': 7.118879209913805e-05, 'epoch': 0.56})
+[2026-07-28 13:05:14,695][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.04195673391222954, 'train/grad_norm_pre_clip': 1.071667194366455, 'timing/data': 0.00033029308542609215, 'timing/model': 0.4151027072221041, 'learning_rate/qwen_vl_interface': 7.115360908580803e-06, 'learning_rate/action_model': 7.115360908580802e-05, 'epoch': 0.56})
+[2026-07-28 13:05:20,636][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.025596696883440018, 'train/grad_norm_pre_clip': 1.2671159505844116, 'timing/data': 0.00037319399416446686, 'timing/model': 0.41811332292854786, 'learning_rate/qwen_vl_interface': 7.1118413968353575e-06, 'learning_rate/action_model': 7.111841396835357e-05, 'epoch': 0.56})
+[2026-07-28 13:05:26,584][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.01809804141521454, 'train/grad_norm_pre_clip': 0.9757769107818604, 'timing/data': 0.0003632279112935066, 'timing/model': 0.4187711561098695, 'learning_rate/qwen_vl_interface': 7.108320676961242e-06, 'learning_rate/action_model': 7.108320676961242e-05, 'epoch': 0.56})
+[2026-07-28 13:05:32,532][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.057189058512449265, 'train/grad_norm_pre_clip': 1.6889941692352295, 'timing/data': 0.0002670278772711754, 'timing/model': 0.42306837486103177, 'learning_rate/qwen_vl_interface': 7.104798751243018e-06, 'learning_rate/action_model': 7.104798751243019e-05, 'epoch': 0.56})
+[2026-07-28 13:05:38,695][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.026927415281534195, 'train/grad_norm_pre_clip': 0.9327250123023987, 'timing/data': 0.00027211708948016167, 'timing/model': 0.4156972486525774, 'learning_rate/qwen_vl_interface': 7.1012756219660195e-06, 'learning_rate/action_model': 7.101275621966019e-05, 'epoch': 0.56})
+[2026-07-28 13:05:44,971][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.05232759192585945, 'train/grad_norm_pre_clip': 1.5806987285614014, 'timing/data': 0.00034268200397491455, 'timing/model': 0.4159237463027239, 'learning_rate/qwen_vl_interface': 7.09775129141637e-06, 'learning_rate/action_model': 7.097751291416369e-05, 'epoch': 0.56})
+[2026-07-28 13:05:50,905][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.035268157720565796, 'train/grad_norm_pre_clip': 1.3044910430908203, 'timing/data': 0.00025497283786535263, 'timing/model': 0.415473953820765, 'learning_rate/qwen_vl_interface': 7.094225761880966e-06, 'learning_rate/action_model': 7.094225761880967e-05, 'epoch': 0.56})
+[2026-07-28 13:05:56,872][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.05699121579527855, 'train/grad_norm_pre_clip': 1.1018513441085815, 'timing/data': 0.0003380649723112583, 'timing/model': 0.4327397686429322, 'learning_rate/qwen_vl_interface': 7.0906990356474895e-06, 'learning_rate/action_model': 7.090699035647489e-05, 'epoch': 0.56})
+[2026-07-28 13:06:03,129][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.04225516319274902, 'train/grad_norm_pre_clip': 1.1300171613693237, 'timing/data': 0.0003540520556271076, 'timing/model': 0.4209077921696007, 'learning_rate/qwen_vl_interface': 7.087171115004393e-06, 'learning_rate/action_model': 7.087171115004393e-05, 'epoch': 0.56})
+[2026-07-28 13:06:09,072][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.022399157285690308, 'train/grad_norm_pre_clip': 0.8751769661903381, 'timing/data': 0.00028923898935317993, 'timing/model': 0.4178690081462264, 'learning_rate/qwen_vl_interface': 7.083642002240905e-06, 'learning_rate/action_model': 7.083642002240905e-05, 'epoch': 0.56})
+[2026-07-28 13:06:15,041][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.03001362830400467, 'train/grad_norm_pre_clip': 1.1822974681854248, 'timing/data': 0.0002917340025305748, 'timing/model': 0.4163473309017718, 'learning_rate/qwen_vl_interface': 7.080111699647031e-06, 'learning_rate/action_model': 7.080111699647031e-05, 'epoch': 0.56})
+[2026-07-28 13:06:21,480][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.02808312326669693, 'train/grad_norm_pre_clip': 1.1990927457809448, 'timing/data': 0.00032634520903229713, 'timing/model': 0.6610920722596347, 'learning_rate/qwen_vl_interface': 7.076580209513542e-06, 'learning_rate/action_model': 7.076580209513542e-05, 'epoch': 0.56})
+[2026-07-28 13:06:27,462][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.0418979749083519, 'train/grad_norm_pre_clip': 1.7393531799316406, 'timing/data': 0.0002959230914711952, 'timing/model': 0.4191764867864549, 'learning_rate/qwen_vl_interface': 7.073047534131987e-06, 'learning_rate/action_model': 7.073047534131986e-05, 'epoch': 0.56})
+[2026-07-28 13:06:33,444][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.0342845618724823, 'train/grad_norm_pre_clip': 1.5303329229354858, 'timing/data': 0.00028823502361774445, 'timing/model': 0.41833546571433544, 'learning_rate/qwen_vl_interface': 7.069513675794679e-06, 'learning_rate/action_model': 7.069513675794679e-05, 'epoch': 0.56})
+[2026-07-28 13:06:39,607][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.022007178515195847, 'train/grad_norm_pre_clip': 1.0382330417633057, 'timing/data': 0.00027598440647125244, 'timing/model': 0.5952491029165685, 'learning_rate/qwen_vl_interface': 7.065978636794704e-06, 'learning_rate/action_model': 7.065978636794704e-05, 'epoch': 0.56})
+[2026-07-28 13:06:45,562][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.03147020563483238, 'train/grad_norm_pre_clip': 1.4140355587005615, 'timing/data': 0.00034035276621580124, 'timing/model': 0.41861474979668856, 'learning_rate/qwen_vl_interface': 7.062442419425904e-06, 'learning_rate/action_model': 7.062442419425903e-05, 'epoch': 0.56})
+[2026-07-28 13:06:51,750][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.040233153849840164, 'train/grad_norm_pre_clip': 1.1405255794525146, 'timing/data': 0.00033810408785939217, 'timing/model': 0.4178438098169863, 'learning_rate/qwen_vl_interface': 7.058905025982895e-06, 'learning_rate/action_model': 7.058905025982894e-05, 'epoch': 0.56})
+[2026-07-28 13:06:57,745][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.03496621549129486, 'train/grad_norm_pre_clip': 1.6735436916351318, 'timing/data': 0.0002869502641260624, 'timing/model': 0.43819832103326917, 'learning_rate/qwen_vl_interface': 7.055366458761056e-06, 'learning_rate/action_model': 7.055366458761056e-05, 'epoch': 0.56})
+[2026-07-28 13:07:03,969][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.05243729054927826, 'train/grad_norm_pre_clip': 2.16850209236145, 'timing/data': 0.00027156900614500046, 'timing/model': 0.41638178285211325, 'learning_rate/qwen_vl_interface': 7.051826720056522e-06, 'learning_rate/action_model': 7.051826720056521e-05, 'epoch': 0.56})
+[2026-07-28 13:07:09,943][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.008716471493244171, 'train/grad_norm_pre_clip': 0.4693898856639862, 'timing/data': 0.000351825263351202, 'timing/model': 0.4275693451054394, 'learning_rate/qwen_vl_interface': 7.04828581216619e-06, 'learning_rate/action_model': 7.04828581216619e-05, 'epoch': 0.56})
+[2026-07-28 13:07:16,124][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.026463832706212997, 'train/grad_norm_pre_clip': 0.8411045074462891, 'timing/data': 0.00031131133437156677, 'timing/model': 0.4202620112337172, 'learning_rate/qwen_vl_interface': 7.04474373738772e-06, 'learning_rate/action_model': 7.04474373738772e-05, 'epoch': 0.56})
+[2026-07-28 13:07:22,057][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.0671391412615776, 'train/grad_norm_pre_clip': 2.257763385772705, 'timing/data': 0.0003863316960632801, 'timing/model': 0.41844787122681737, 'learning_rate/qwen_vl_interface': 7.041200498019523e-06, 'learning_rate/action_model': 7.041200498019523e-05, 'epoch': 0.56})
+[2026-07-28 13:07:28,193][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.11351527273654938, 'train/grad_norm_pre_clip': 3.086589813232422, 'timing/data': 0.0003010411746799946, 'timing/model': 0.653379525989294, 'learning_rate/qwen_vl_interface': 7.037656096360771e-06, 'learning_rate/action_model': 7.037656096360771e-05, 'epoch': 0.56})
+[2026-07-28 13:07:34,353][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.029157642275094986, 'train/grad_norm_pre_clip': 1.3127855062484741, 'timing/data': 0.0003467989154160023, 'timing/model': 0.6543739372864366, 'learning_rate/qwen_vl_interface': 7.0341105347113844e-06, 'learning_rate/action_model': 7.034110534711384e-05, 'epoch': 0.56})
+[2026-07-28 13:07:40,262][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.020070895552635193, 'train/grad_norm_pre_clip': 0.8338385820388794, 'timing/data': 0.00041664671152830124, 'timing/model': 0.4250225997529924, 'learning_rate/qwen_vl_interface': 7.030563815372042e-06, 'learning_rate/action_model': 7.030563815372042e-05, 'epoch': 0.56})
+[2026-07-28 13:07:45,975][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.05192384123802185, 'train/grad_norm_pre_clip': 1.339397668838501, 'timing/data': 0.0002895970828831196, 'timing/model': 0.41772739915177226, 'learning_rate/qwen_vl_interface': 7.027015940644173e-06, 'learning_rate/action_model': 7.027015940644173e-05, 'epoch': 0.56})
+[2026-07-28 13:07:51,940][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.04080253094434738, 'train/grad_norm_pre_clip': 1.9667567014694214, 'timing/data': 0.00028500473126769066, 'timing/model': 0.419513578992337, 'learning_rate/qwen_vl_interface': 7.02346691282995e-06, 'learning_rate/action_model': 7.02346691282995e-05, 'epoch': 0.56})
+[2026-07-28 13:07:57,628][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.07325848937034607, 'train/grad_norm_pre_clip': 1.6669983863830566, 'timing/data': 0.00028942711651325226, 'timing/model': 0.43051035795360804, 'learning_rate/qwen_vl_interface': 7.019916734232304e-06, 'learning_rate/action_model': 7.019916734232303e-05, 'epoch': 0.56})
+[2026-07-28 13:08:04,037][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.040832675993442535, 'train/grad_norm_pre_clip': 1.4001905918121338, 'timing/data': 0.0002788151614367962, 'timing/model': 0.6546532888896763, 'learning_rate/qwen_vl_interface': 7.0163654071549035e-06, 'learning_rate/action_model': 7.016365407154904e-05, 'epoch': 0.57})
+[2026-07-28 13:08:09,978][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.023887690156698227, 'train/grad_norm_pre_clip': 0.7468525767326355, 'timing/data': 0.00025893980637192726, 'timing/model': 0.4156905268318951, 'learning_rate/qwen_vl_interface': 7.012812933902167e-06, 'learning_rate/action_model': 7.012812933902167e-05, 'epoch': 0.57})
+[2026-07-28 13:08:15,902][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.07190042734146118, 'train/grad_norm_pre_clip': 2.277371644973755, 'timing/data': 0.0002493979409337044, 'timing/model': 0.41777298925444484, 'learning_rate/qwen_vl_interface': 7.009259316779258e-06, 'learning_rate/action_model': 7.009259316779257e-05, 'epoch': 0.57})
+[2026-07-28 13:08:21,840][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.05677851289510727, 'train/grad_norm_pre_clip': 1.198508381843567, 'timing/data': 0.0002958723343908787, 'timing/model': 0.4184805117547512, 'learning_rate/qwen_vl_interface': 7.005704558092078e-06, 'learning_rate/action_model': 7.005704558092078e-05, 'epoch': 0.57})
+[2026-07-28 13:08:28,034][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.031533803790807724, 'train/grad_norm_pre_clip': 1.0318126678466797, 'timing/data': 0.00032615987583994865, 'timing/model': 0.4155962010845542, 'learning_rate/qwen_vl_interface': 7.002148660147271e-06, 'learning_rate/action_model': 7.002148660147271e-05, 'epoch': 0.57})
+[2026-07-28 13:08:33,931][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.09200479835271835, 'train/grad_norm_pre_clip': 2.201538562774658, 'timing/data': 0.0002724197693169117, 'timing/model': 0.41670687263831496, 'learning_rate/qwen_vl_interface': 6.9985916252522225e-06, 'learning_rate/action_model': 6.998591625252222e-05, 'epoch': 0.57})
+[2026-07-28 13:08:40,075][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.04839954152703285, 'train/grad_norm_pre_clip': 1.5999823808670044, 'timing/data': 0.00028155604377388954, 'timing/model': 0.6715103117749095, 'learning_rate/qwen_vl_interface': 6.99503345571505e-06, 'learning_rate/action_model': 6.99503345571505e-05, 'epoch': 0.57})
+[2026-07-28 13:08:46,699][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.06241928040981293, 'train/grad_norm_pre_clip': 1.8091084957122803, 'timing/data': 0.00036988407373428345, 'timing/model': 0.6593422307632864, 'learning_rate/qwen_vl_interface': 6.991474153844614e-06, 'learning_rate/action_model': 6.991474153844613e-05, 'epoch': 0.57})
+[2026-07-28 13:08:52,705][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.03798263892531395, 'train/grad_norm_pre_clip': 1.3767880201339722, 'timing/data': 0.00027506984770298004, 'timing/model': 0.41702546272426844, 'learning_rate/qwen_vl_interface': 6.98791372195051e-06, 'learning_rate/action_model': 6.98791372195051e-05, 'epoch': 0.57})
+[2026-07-28 13:08:58,629][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.051493387669324875, 'train/grad_norm_pre_clip': 2.3927767276763916, 'timing/data': 0.0003164107911288738, 'timing/model': 0.41741616278886795, 'learning_rate/qwen_vl_interface': 6.984352162343056e-06, 'learning_rate/action_model': 6.984352162343057e-05, 'epoch': 0.57})
+[2026-07-28 13:09:04,538][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.031599096953868866, 'train/grad_norm_pre_clip': 0.8270891308784485, 'timing/data': 0.0004112771712243557, 'timing/model': 0.41991619300097227, 'learning_rate/qwen_vl_interface': 6.980789477333316e-06, 'learning_rate/action_model': 6.980789477333316e-05, 'epoch': 0.57})
+[2026-07-28 13:09:10,906][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.04944555461406708, 'train/grad_norm_pre_clip': 0.7940734624862671, 'timing/data': 0.0002617998979985714, 'timing/model': 0.41566034220159054, 'learning_rate/qwen_vl_interface': 6.977225669233074e-06, 'learning_rate/action_model': 6.977225669233074e-05, 'epoch': 0.57})
+[2026-07-28 13:09:16,858][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.04420284926891327, 'train/grad_norm_pre_clip': 1.0077505111694336, 'timing/data': 0.00028677377849817276, 'timing/model': 0.41755201714113355, 'learning_rate/qwen_vl_interface': 6.973660740354848e-06, 'learning_rate/action_model': 6.973660740354848e-05, 'epoch': 0.57})
+[2026-07-28 13:09:22,823][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.0417947918176651, 'train/grad_norm_pre_clip': 1.1116461753845215, 'timing/data': 0.0002748030237853527, 'timing/model': 0.43084227992221713, 'learning_rate/qwen_vl_interface': 6.970094693011881e-06, 'learning_rate/action_model': 6.97009469301188e-05, 'epoch': 0.57})
+[2026-07-28 13:09:29,029][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.07303348183631897, 'train/grad_norm_pre_clip': 2.1042041778564453, 'timing/data': 0.00031585386022925377, 'timing/model': 0.41633667703717947, 'learning_rate/qwen_vl_interface': 6.966527529518142e-06, 'learning_rate/action_model': 6.966527529518142e-05, 'epoch': 0.57})
+[2026-07-28 13:09:35,223][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.039059218019247055, 'train/grad_norm_pre_clip': 1.7805912494659424, 'timing/data': 0.00026615289971232414, 'timing/model': 0.4400937962345779, 'learning_rate/qwen_vl_interface': 6.962959252188325e-06, 'learning_rate/action_model': 6.962959252188325e-05, 'epoch': 0.57})
+[2026-07-28 13:09:41,147][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.030706916004419327, 'train/grad_norm_pre_clip': 0.6596999168395996, 'timing/data': 0.00026170071214437485, 'timing/model': 0.41584453033283353, 'learning_rate/qwen_vl_interface': 6.9593898633378476e-06, 'learning_rate/action_model': 6.959389863337847e-05, 'epoch': 0.57})
+[2026-07-28 13:09:47,341][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.057957082986831665, 'train/grad_norm_pre_clip': 2.347388505935669, 'timing/data': 0.00035132793709635735, 'timing/model': 0.5865750396624207, 'learning_rate/qwen_vl_interface': 6.955819365282846e-06, 'learning_rate/action_model': 6.955819365282845e-05, 'epoch': 0.57})
+[2026-07-28 13:09:53,265][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.02355942130088806, 'train/grad_norm_pre_clip': 1.0884840488433838, 'timing/data': 0.0003354051150381565, 'timing/model': 0.4278913461603224, 'learning_rate/qwen_vl_interface': 6.952247760340176e-06, 'learning_rate/action_model': 6.952247760340176e-05, 'epoch': 0.57})
+[2026-07-28 13:09:59,246][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.03957483172416687, 'train/grad_norm_pre_clip': 0.775706946849823, 'timing/data': 0.0002966071479022503, 'timing/model': 0.4156707883812487, 'learning_rate/qwen_vl_interface': 6.948675050827418e-06, 'learning_rate/action_model': 6.948675050827419e-05, 'epoch': 0.57})
+[2026-07-28 13:10:05,480][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.04210663586854935, 'train/grad_norm_pre_clip': 1.0227423906326294, 'timing/data': 0.000499403104186058, 'timing/model': 0.6801694552414119, 'learning_rate/qwen_vl_interface': 6.945101239062862e-06, 'learning_rate/action_model': 6.945101239062863e-05, 'epoch': 0.57})
+[2026-07-28 13:10:11,601][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.03262452036142349, 'train/grad_norm_pre_clip': 1.082465410232544, 'timing/data': 0.00027679000049829483, 'timing/model': 0.41427928395569324, 'learning_rate/qwen_vl_interface': 6.941526327365513e-06, 'learning_rate/action_model': 6.941526327365513e-05, 'epoch': 0.57})
+[2026-07-28 13:10:17,493][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.04490111768245697, 'train/grad_norm_pre_clip': 1.3605211973190308, 'timing/data': 0.0003036521375179291, 'timing/model': 0.4169748811982572, 'learning_rate/qwen_vl_interface': 6.937950318055098e-06, 'learning_rate/action_model': 6.937950318055098e-05, 'epoch': 0.57})
+[2026-07-28 13:10:23,379][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.06667529791593552, 'train/grad_norm_pre_clip': 2.1919829845428467, 'timing/data': 0.0003554951399564743, 'timing/model': 0.4167590062133968, 'learning_rate/qwen_vl_interface': 6.934373213452045e-06, 'learning_rate/action_model': 6.934373213452044e-05, 'epoch': 0.57})
+[2026-07-28 13:10:29,295][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.02659234032034874, 'train/grad_norm_pre_clip': 0.7350310683250427, 'timing/data': 0.00026394613087177277, 'timing/model': 0.41579149290919304, 'learning_rate/qwen_vl_interface': 6.930795015877502e-06, 'learning_rate/action_model': 6.930795015877502e-05, 'epoch': 0.57})
+[2026-07-28 13:10:35,436][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.06911695748567581, 'train/grad_norm_pre_clip': 2.636300563812256, 'timing/data': 0.00032585905864834785, 'timing/model': 0.4157517538405955, 'learning_rate/qwen_vl_interface': 6.927215727653321e-06, 'learning_rate/action_model': 6.927215727653321e-05, 'epoch': 0.57})
+[2026-07-28 13:10:41,352][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.061589404940605164, 'train/grad_norm_pre_clip': 2.0259413719177246, 'timing/data': 0.0002911561168730259, 'timing/model': 0.41558274207636714, 'learning_rate/qwen_vl_interface': 6.9236353511020625e-06, 'learning_rate/action_model': 6.923635351102063e-05, 'epoch': 0.57})
+[2026-07-28 13:10:47,515][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.05750184506177902, 'train/grad_norm_pre_clip': 1.5204994678497314, 'timing/data': 0.0002650660462677479, 'timing/model': 0.4154427908360958, 'learning_rate/qwen_vl_interface': 6.920053888546995e-06, 'learning_rate/action_model': 6.920053888546994e-05, 'epoch': 0.57})
+[2026-07-28 13:10:53,463][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.04228150099515915, 'train/grad_norm_pre_clip': 1.5361669063568115, 'timing/data': 0.00029060011729598045, 'timing/model': 0.41503974376246333, 'learning_rate/qwen_vl_interface': 6.916471342312093e-06, 'learning_rate/action_model': 6.916471342312093e-05, 'epoch': 0.57})
+[2026-07-28 13:10:59,349][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.06166781485080719, 'train/grad_norm_pre_clip': 1.9956276416778564, 'timing/data': 0.00037452997639775276, 'timing/model': 0.41661406913772225, 'learning_rate/qwen_vl_interface': 6.9128877147220274e-06, 'learning_rate/action_model': 6.912887714722027e-05, 'epoch': 0.57})
+[2026-07-28 13:11:05,502][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.04369742423295975, 'train/grad_norm_pre_clip': 0.911221444606781, 'timing/data': 0.00028939591720700264, 'timing/model': 0.6506718746386468, 'learning_rate/qwen_vl_interface': 6.909303008102179e-06, 'learning_rate/action_model': 6.909303008102179e-05, 'epoch': 0.57})
+[2026-07-28 13:11:11,611][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.030334101989865303, 'train/grad_norm_pre_clip': 0.7580643892288208, 'timing/data': 0.0003408328630030155, 'timing/model': 0.6510691707953811, 'learning_rate/qwen_vl_interface': 6.905717224778623e-06, 'learning_rate/action_model': 6.905717224778622e-05, 'epoch': 0.57})
+[2026-07-28 13:11:17,536][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.04323437809944153, 'train/grad_norm_pre_clip': 0.8321219086647034, 'timing/data': 0.00029162270948290825, 'timing/model': 0.41582473181188107, 'learning_rate/qwen_vl_interface': 6.902130367078135e-06, 'learning_rate/action_model': 6.902130367078134e-05, 'epoch': 0.57})
+[2026-07-28 13:11:23,645][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.04460010305047035, 'train/grad_norm_pre_clip': 1.6826841831207275, 'timing/data': 0.00036774808540940285, 'timing/model': 0.43203722266480327, 'learning_rate/qwen_vl_interface': 6.8985424373281895e-06, 'learning_rate/action_model': 6.89854243732819e-05, 'epoch': 0.57})
+[2026-07-28 13:11:29,348][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.08271165192127228, 'train/grad_norm_pre_clip': 1.409282922744751, 'timing/data': 0.0003624819219112396, 'timing/model': 0.41689228918403387, 'learning_rate/qwen_vl_interface': 6.894953437856954e-06, 'learning_rate/action_model': 6.894953437856953e-05, 'epoch': 0.57})
+[2026-07-28 13:11:35,509][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.05230307951569557, 'train/grad_norm_pre_clip': 1.588301658630371, 'timing/data': 0.00033211009576916695, 'timing/model': 0.4131953199394047, 'learning_rate/qwen_vl_interface': 6.891363370993292e-06, 'learning_rate/action_model': 6.891363370993293e-05, 'epoch': 0.57})
+[2026-07-28 13:11:41,435][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.03669404610991478, 'train/grad_norm_pre_clip': 0.5028387308120728, 'timing/data': 0.0003513018600642681, 'timing/model': 0.4165854169987142, 'learning_rate/qwen_vl_interface': 6.88777223906676e-06, 'learning_rate/action_model': 6.88777223906676e-05, 'epoch': 0.57})
+[2026-07-28 13:11:47,314][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.06535212695598602, 'train/grad_norm_pre_clip': 2.134727716445923, 'timing/data': 0.00034258514642715454, 'timing/model': 0.4168513799086213, 'learning_rate/qwen_vl_interface': 6.8841800444076e-06, 'learning_rate/action_model': 6.8841800444076e-05, 'epoch': 0.57})
+[2026-07-28 13:11:53,455][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.04943748563528061, 'train/grad_norm_pre_clip': 2.716616630554199, 'timing/data': 0.00027447985485196114, 'timing/model': 0.416346977930516, 'learning_rate/qwen_vl_interface': 6.8805867893467525e-06, 'learning_rate/action_model': 6.880586789346752e-05, 'epoch': 0.57})
+[2026-07-28 13:11:59,606][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.03408212959766388, 'train/grad_norm_pre_clip': 1.7854247093200684, 'timing/data': 0.0003527319058775902, 'timing/model': 0.6555259991437197, 'learning_rate/qwen_vl_interface': 6.87699247621584e-06, 'learning_rate/action_model': 6.87699247621584e-05, 'epoch': 0.57})
+[2026-07-28 13:12:05,543][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.024889405816793442, 'train/grad_norm_pre_clip': 1.3117910623550415, 'timing/data': 0.0003564748913049698, 'timing/model': 0.41713959304615855, 'learning_rate/qwen_vl_interface': 6.873397107347176e-06, 'learning_rate/action_model': 6.873397107347176e-05, 'epoch': 0.57})
+[2026-07-28 13:12:11,482][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.03881291672587395, 'train/grad_norm_pre_clip': 1.9477274417877197, 'timing/data': 0.00033665401861071587, 'timing/model': 0.416196851991117, 'learning_rate/qwen_vl_interface': 6.869800685073752e-06, 'learning_rate/action_model': 6.869800685073752e-05, 'epoch': 0.57})
+[2026-07-28 13:12:17,692][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.03642464056611061, 'train/grad_norm_pre_clip': 0.8605462908744812, 'timing/data': 0.0002914709039032459, 'timing/model': 0.6723847081884742, 'learning_rate/qwen_vl_interface': 6.86620321172925e-06, 'learning_rate/action_model': 6.86620321172925e-05, 'epoch': 0.57})
+[2026-07-28 13:12:23,833][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.06162205711007118, 'train/grad_norm_pre_clip': 1.535878300666809, 'timing/data': 0.0005708611570298672, 'timing/model': 0.41844553081318736, 'learning_rate/qwen_vl_interface': 6.862604689648029e-06, 'learning_rate/action_model': 6.862604689648029e-05, 'epoch': 0.57})
+[2026-07-28 13:12:29,766][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.0330272912979126, 'train/grad_norm_pre_clip': 1.1629904508590698, 'timing/data': 0.00026003364473581314, 'timing/model': 0.42502647964283824, 'learning_rate/qwen_vl_interface': 6.859005121165135e-06, 'learning_rate/action_model': 6.859005121165135e-05, 'epoch': 0.57})
+[2026-07-28 13:12:35,707][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.06502515822649002, 'train/grad_norm_pre_clip': 1.602792501449585, 'timing/data': 0.00036201300099492073, 'timing/model': 0.4156756498850882, 'learning_rate/qwen_vl_interface': 6.855404508616285e-06, 'learning_rate/action_model': 6.855404508616285e-05, 'epoch': 0.57})
+[2026-07-28 13:12:41,647][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.04021438956260681, 'train/grad_norm_pre_clip': 1.427675485610962, 'timing/data': 0.0003330768086016178, 'timing/model': 0.4287788150832057, 'learning_rate/qwen_vl_interface': 6.851802854337881e-06, 'learning_rate/action_model': 6.851802854337881e-05, 'epoch': 0.57})
+[2026-07-28 13:12:47,790][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.04678063094615936, 'train/grad_norm_pre_clip': 1.9740986824035645, 'timing/data': 0.0005168979987502098, 'timing/model': 0.4132514647208154, 'learning_rate/qwen_vl_interface': 6.848200160666993e-06, 'learning_rate/action_model': 6.848200160666993e-05, 'epoch': 0.57})
+[2026-07-28 13:12:53,905][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.07108642905950546, 'train/grad_norm_pre_clip': 1.9021493196487427, 'timing/data': 0.00035370374098420143, 'timing/model': 0.41404862701892853, 'learning_rate/qwen_vl_interface': 6.844596429941374e-06, 'learning_rate/action_model': 6.844596429941373e-05, 'epoch': 0.57})
+[2026-07-28 13:12:59,850][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.051478542387485504, 'train/grad_norm_pre_clip': 0.763376772403717, 'timing/data': 0.0003788759931921959, 'timing/model': 0.42096418095752597, 'learning_rate/qwen_vl_interface': 6.8409916644994425e-06, 'learning_rate/action_model': 6.840991664499442e-05, 'epoch': 0.57})
+[2026-07-28 13:13:05,990][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.03994815796613693, 'train/grad_norm_pre_clip': 0.8088338971138, 'timing/data': 0.00035627372562885284, 'timing/model': 0.4124882980249822, 'learning_rate/qwen_vl_interface': 6.837385866680294e-06, 'learning_rate/action_model': 6.837385866680293e-05, 'epoch': 0.57})
+[2026-07-28 13:13:11,897][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.04520568996667862, 'train/grad_norm_pre_clip': 1.2762820720672607, 'timing/data': 0.00036962423473596573, 'timing/model': 0.41621688613668084, 'learning_rate/qwen_vl_interface': 6.833779038823691e-06, 'learning_rate/action_model': 6.833779038823691e-05, 'epoch': 0.57})
+[2026-07-28 13:13:18,025][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.023133305832743645, 'train/grad_norm_pre_clip': 1.2101434469223022, 'timing/data': 0.00044097891077399254, 'timing/model': 0.41699378564953804, 'learning_rate/qwen_vl_interface': 6.830171183270063e-06, 'learning_rate/action_model': 6.830171183270064e-05, 'epoch': 0.57})
+[2026-07-28 13:13:23,924][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.04423295706510544, 'train/grad_norm_pre_clip': 0.9065119028091431, 'timing/data': 0.00026845838874578476, 'timing/model': 0.4152774722315371, 'learning_rate/qwen_vl_interface': 6.826562302360511e-06, 'learning_rate/action_model': 6.82656230236051e-05, 'epoch': 0.57})
+[2026-07-28 13:13:30,164][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.03081788867712021, 'train/grad_norm_pre_clip': 0.9770260453224182, 'timing/data': 0.0003670528531074524, 'timing/model': 0.6959568737074733, 'learning_rate/qwen_vl_interface': 6.822952398436797e-06, 'learning_rate/action_model': 6.822952398436797e-05, 'epoch': 0.58})
+[2026-07-28 13:13:36,302][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.045271605253219604, 'train/grad_norm_pre_clip': 1.336603045463562, 'timing/data': 0.00034751929342746735, 'timing/model': 0.4138355008326471, 'learning_rate/qwen_vl_interface': 6.81934147384135e-06, 'learning_rate/action_model': 6.81934147384135e-05, 'epoch': 0.58})
+[2026-07-28 13:13:42,227][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.06807990372180939, 'train/grad_norm_pre_clip': 1.890146017074585, 'timing/data': 0.0002661189064383507, 'timing/model': 0.4153009820729494, 'learning_rate/qwen_vl_interface': 6.815729530917257e-06, 'learning_rate/action_model': 6.815729530917256e-05, 'epoch': 0.58})
+[2026-07-28 13:13:48,128][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.0356733612716198, 'train/grad_norm_pre_clip': 2.58198618888855, 'timing/data': 0.00027955323457717896, 'timing/model': 0.4187822570092976, 'learning_rate/qwen_vl_interface': 6.812116572008271e-06, 'learning_rate/action_model': 6.812116572008271e-05, 'epoch': 0.58})
+[2026-07-28 13:13:54,428][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.041458114981651306, 'train/grad_norm_pre_clip': 1.4769431352615356, 'timing/data': 0.00026693195104599, 'timing/model': 0.5795294530689716, 'learning_rate/qwen_vl_interface': 6.808502599458803e-06, 'learning_rate/action_model': 6.808502599458803e-05, 'epoch': 0.58})
+[2026-07-28 13:14:00,289][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.07799798995256424, 'train/grad_norm_pre_clip': 2.948768138885498, 'timing/data': 0.0002808067947626114, 'timing/model': 0.41429965663701296, 'learning_rate/qwen_vl_interface': 6.8048876156139154e-06, 'learning_rate/action_model': 6.804887615613915e-05, 'epoch': 0.58})
+[2026-07-28 13:14:06,155][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.037085361778736115, 'train/grad_norm_pre_clip': 1.6335586309432983, 'timing/data': 0.0002719201147556305, 'timing/model': 0.4154714113101363, 'learning_rate/qwen_vl_interface': 6.801271622819337e-06, 'learning_rate/action_model': 6.801271622819336e-05, 'epoch': 0.58})
+[2026-07-28 13:14:12,020][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.02307868003845215, 'train/grad_norm_pre_clip': 0.9580554366111755, 'timing/data': 0.00027795787900686264, 'timing/model': 0.41348367696627975, 'learning_rate/qwen_vl_interface': 6.797654623421443e-06, 'learning_rate/action_model': 6.797654623421444e-05, 'epoch': 0.58})
+[2026-07-28 13:14:18,125][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.03892384469509125, 'train/grad_norm_pre_clip': 1.092489242553711, 'timing/data': 0.00026248302310705185, 'timing/model': 0.4142619767226279, 'learning_rate/qwen_vl_interface': 6.794036619767269e-06, 'learning_rate/action_model': 6.79403661976727e-05, 'epoch': 0.58})
+[2026-07-28 13:14:23,759][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.034336719661951065, 'train/grad_norm_pre_clip': 1.083484172821045, 'timing/data': 0.0002880990505218506, 'timing/model': 0.4148290199227631, 'learning_rate/qwen_vl_interface': 6.790417614204494e-06, 'learning_rate/action_model': 6.790417614204494e-05, 'epoch': 0.58})
+[2026-07-28 13:14:29,630][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.030791886150836945, 'train/grad_norm_pre_clip': 1.223279595375061, 'timing/data': 0.00028525805100798607, 'timing/model': 0.420027872081846, 'learning_rate/qwen_vl_interface': 6.786797609081453e-06, 'learning_rate/action_model': 6.786797609081453e-05, 'epoch': 0.58})
+[2026-07-28 13:14:35,752][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.017645573243498802, 'train/grad_norm_pre_clip': 0.8532872796058655, 'timing/data': 0.0002713296562433243, 'timing/model': 0.412892596796155, 'learning_rate/qwen_vl_interface': 6.783176606747129e-06, 'learning_rate/action_model': 6.78317660674713e-05, 'epoch': 0.58})
+[2026-07-28 13:14:41,963][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.050324734300374985, 'train/grad_norm_pre_clip': 1.0652318000793457, 'timing/data': 0.0002804589457809925, 'timing/model': 0.6547255543991923, 'learning_rate/qwen_vl_interface': 6.77955460955115e-06, 'learning_rate/action_model': 6.77955460955115e-05, 'epoch': 0.58})
+[2026-07-28 13:14:47,875][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.055116426199674606, 'train/grad_norm_pre_clip': 1.5299737453460693, 'timing/data': 0.0003452901728451252, 'timing/model': 0.4140228349715471, 'learning_rate/qwen_vl_interface': 6.775931619843791e-06, 'learning_rate/action_model': 6.77593161984379e-05, 'epoch': 0.58})
+[2026-07-28 13:14:53,772][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.033392589539289474, 'train/grad_norm_pre_clip': 1.0781437158584595, 'timing/data': 0.00032949866726994514, 'timing/model': 0.42332705995067954, 'learning_rate/qwen_vl_interface': 6.7723076399759715e-06, 'learning_rate/action_model': 6.772307639975972e-05, 'epoch': 0.58})
+[2026-07-28 13:14:59,923][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.03680052608251572, 'train/grad_norm_pre_clip': 1.5237621068954468, 'timing/data': 0.00037706829607486725, 'timing/model': 0.6593090794049203, 'learning_rate/qwen_vl_interface': 6.7686826722992516e-06, 'learning_rate/action_model': 6.768682672299252e-05, 'epoch': 0.58})
+[2026-07-28 13:15:05,794][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.046645570546388626, 'train/grad_norm_pre_clip': 1.420194149017334, 'timing/data': 0.00033749593421816826, 'timing/model': 0.41757391300052404, 'learning_rate/qwen_vl_interface': 6.765056719165833e-06, 'learning_rate/action_model': 6.765056719165834e-05, 'epoch': 0.58})
+[2026-07-28 13:15:11,716][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.02096414752304554, 'train/grad_norm_pre_clip': 0.6648088097572327, 'timing/data': 0.00036644982174038887, 'timing/model': 0.41887163603678346, 'learning_rate/qwen_vl_interface': 6.761429782928558e-06, 'learning_rate/action_model': 6.761429782928557e-05, 'epoch': 0.58})
+[2026-07-28 13:15:17,815][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.0580587275326252, 'train/grad_norm_pre_clip': 1.183681607246399, 'timing/data': 0.0002982872538268566, 'timing/model': 0.41712764371186495, 'learning_rate/qwen_vl_interface': 6.757801865940908e-06, 'learning_rate/action_model': 6.757801865940908e-05, 'epoch': 0.58})
+[2026-07-28 13:15:23,946][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.08061109483242035, 'train/grad_norm_pre_clip': 3.042003870010376, 'timing/data': 0.000278327614068985, 'timing/model': 0.4140298571437597, 'learning_rate/qwen_vl_interface': 6.754172970556996e-06, 'learning_rate/action_model': 6.754172970556996e-05, 'epoch': 0.58})
+[2026-07-28 13:15:29,885][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.031216993927955627, 'train/grad_norm_pre_clip': 0.8637501001358032, 'timing/data': 0.0009079487062990665, 'timing/model': 0.4207386840134859, 'learning_rate/qwen_vl_interface': 6.750543099131573e-06, 'learning_rate/action_model': 6.750543099131573e-05, 'epoch': 0.58})
+[2026-07-28 13:15:35,790][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.02603866159915924, 'train/grad_norm_pre_clip': 0.7712253928184509, 'timing/data': 0.00035750214010477066, 'timing/model': 0.41908295592293143, 'learning_rate/qwen_vl_interface': 6.746912254020025e-06, 'learning_rate/action_model': 6.746912254020025e-05, 'epoch': 0.58})
+[2026-07-28 13:15:41,909][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.03394494205713272, 'train/grad_norm_pre_clip': 1.4435349702835083, 'timing/data': 0.0003266846761107445, 'timing/model': 0.4164627566933632, 'learning_rate/qwen_vl_interface': 6.743280437578366e-06, 'learning_rate/action_model': 6.743280437578366e-05, 'epoch': 0.58})
+[2026-07-28 13:15:48,071][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.06116916611790657, 'train/grad_norm_pre_clip': 1.1534829139709473, 'timing/data': 0.0002591768279671669, 'timing/model': 0.41437210608273745, 'learning_rate/qwen_vl_interface': 6.7396476521632456e-06, 'learning_rate/action_model': 6.739647652163246e-05, 'epoch': 0.58})
+[2026-07-28 13:15:53,984][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.046619512140750885, 'train/grad_norm_pre_clip': 1.183348536491394, 'timing/data': 0.0002780868671834469, 'timing/model': 0.41656863782554865, 'learning_rate/qwen_vl_interface': 6.736013900131934e-06, 'learning_rate/action_model': 6.736013900131934e-05, 'epoch': 0.58})
+[2026-07-28 13:16:00,160][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.03133459761738777, 'train/grad_norm_pre_clip': 0.6802840828895569, 'timing/data': 0.00028919707983732224, 'timing/model': 0.6494490406475961, 'learning_rate/qwen_vl_interface': 6.732379183842339e-06, 'learning_rate/action_model': 6.73237918384234e-05, 'epoch': 0.58})
+[2026-07-28 13:16:06,278][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.03707081452012062, 'train/grad_norm_pre_clip': 0.7963968515396118, 'timing/data': 0.0002585342153906822, 'timing/model': 0.41315885493531823, 'learning_rate/qwen_vl_interface': 6.728743505652984e-06, 'learning_rate/action_model': 6.728743505652983e-05, 'epoch': 0.58})
+[2026-07-28 13:16:12,144][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.033035725355148315, 'train/grad_norm_pre_clip': 0.8438639640808105, 'timing/data': 0.000284779816865921, 'timing/model': 0.4167713620699942, 'learning_rate/qwen_vl_interface': 6.725106867923023e-06, 'learning_rate/action_model': 6.725106867923023e-05, 'epoch': 0.58})
+[2026-07-28 13:16:18,013][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.047900162637233734, 'train/grad_norm_pre_clip': 1.155057668685913, 'timing/data': 0.00027617765590548515, 'timing/model': 0.4164413958787918, 'learning_rate/qwen_vl_interface': 6.721469273012229e-06, 'learning_rate/action_model': 6.721469273012229e-05, 'epoch': 0.58})
+[2026-07-28 13:16:24,102][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.035995662212371826, 'train/grad_norm_pre_clip': 1.17490553855896, 'timing/data': 0.00033815018832683563, 'timing/model': 0.413194774184376, 'learning_rate/qwen_vl_interface': 6.7178307232810015e-06, 'learning_rate/action_model': 6.717830723281e-05, 'epoch': 0.58})
+[2026-07-28 13:16:30,218][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.024479562416672707, 'train/grad_norm_pre_clip': 1.1991987228393555, 'timing/data': 0.00027132779359817505, 'timing/model': 0.6568435947410762, 'learning_rate/qwen_vl_interface': 6.714191221090352e-06, 'learning_rate/action_model': 6.714191221090352e-05, 'epoch': 0.58})
+[2026-07-28 13:16:36,070][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.040092747658491135, 'train/grad_norm_pre_clip': 2.566818952560425, 'timing/data': 0.00028765201568603516, 'timing/model': 0.4166580019518733, 'learning_rate/qwen_vl_interface': 6.710550768801916e-06, 'learning_rate/action_model': 6.710550768801915e-05, 'epoch': 0.58})
+[2026-07-28 13:16:42,185][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.06575778871774673, 'train/grad_norm_pre_clip': 1.3757836818695068, 'timing/data': 0.00025617703795433044, 'timing/model': 0.41618572594597936, 'learning_rate/qwen_vl_interface': 6.706909368777944e-06, 'learning_rate/action_model': 6.706909368777944e-05, 'epoch': 0.58})
+[2026-07-28 13:16:48,060][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.0370536707341671, 'train/grad_norm_pre_clip': 1.1805429458618164, 'timing/data': 0.0002769189886748791, 'timing/model': 0.41446012491360307, 'learning_rate/qwen_vl_interface': 6.703267023381299e-06, 'learning_rate/action_model': 6.7032670233813e-05, 'epoch': 0.58})
+[2026-07-28 13:16:54,165][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.07909702509641647, 'train/grad_norm_pre_clip': 2.7683699131011963, 'timing/data': 0.00024079345166683197, 'timing/model': 0.4142164345830679, 'learning_rate/qwen_vl_interface': 6.69962373497546e-06, 'learning_rate/action_model': 6.69962373497546e-05, 'epoch': 0.58})
+[2026-07-28 13:17:00,286][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.0477634072303772, 'train/grad_norm_pre_clip': 1.2616395950317383, 'timing/data': 0.0003025759942829609, 'timing/model': 0.5831682593561709, 'learning_rate/qwen_vl_interface': 6.69597950592452e-06, 'learning_rate/action_model': 6.69597950592452e-05, 'epoch': 0.58})
+[2026-07-28 13:17:06,166][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.09901026636362076, 'train/grad_norm_pre_clip': 2.841531991958618, 'timing/data': 0.0003292430192232132, 'timing/model': 0.4151284289546311, 'learning_rate/qwen_vl_interface': 6.692334338593175e-06, 'learning_rate/action_model': 6.692334338593175e-05, 'epoch': 0.58})
+[2026-07-28 13:17:12,065][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.061886489391326904, 'train/grad_norm_pre_clip': 2.025002956390381, 'timing/data': 0.0003616153262555599, 'timing/model': 0.4161158590577543, 'learning_rate/qwen_vl_interface': 6.688688235346736e-06, 'learning_rate/action_model': 6.688688235346736e-05, 'epoch': 0.58})
+[2026-07-28 13:17:18,172][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.044845644384622574, 'train/grad_norm_pre_clip': 1.4442044496536255, 'timing/data': 0.000255086924880743, 'timing/model': 0.4124960689805448, 'learning_rate/qwen_vl_interface': 6.685041198551125e-06, 'learning_rate/action_model': 6.685041198551124e-05, 'epoch': 0.58})
+[2026-07-28 13:17:24,254][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.028581012040376663, 'train/grad_norm_pre_clip': 0.779329776763916, 'timing/data': 0.00033235084265470505, 'timing/model': 0.4140626569278538, 'learning_rate/qwen_vl_interface': 6.681393230572857e-06, 'learning_rate/action_model': 6.681393230572856e-05, 'epoch': 0.58})
+[2026-07-28 13:17:30,154][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.032661035656929016, 'train/grad_norm_pre_clip': 0.8882501721382141, 'timing/data': 0.00034296978265047073, 'timing/model': 0.41497933212667704, 'learning_rate/qwen_vl_interface': 6.677744333779063e-06, 'learning_rate/action_model': 6.677744333779062e-05, 'epoch': 0.58})
+[2026-07-28 13:17:36,294][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.03364386782050133, 'train/grad_norm_pre_clip': 0.8227373361587524, 'timing/data': 0.0003448519855737686, 'timing/model': 0.41510307788848877, 'learning_rate/qwen_vl_interface': 6.674094510537471e-06, 'learning_rate/action_model': 6.674094510537471e-05, 'epoch': 0.58})
+[2026-07-28 13:17:42,167][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.0353315994143486, 'train/grad_norm_pre_clip': 0.6527990102767944, 'timing/data': 0.00036111893132328987, 'timing/model': 0.4158337223343551, 'learning_rate/qwen_vl_interface': 6.670443763216412e-06, 'learning_rate/action_model': 6.670443763216411e-05, 'epoch': 0.58})
+[2026-07-28 13:17:48,327][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.07585781812667847, 'train/grad_norm_pre_clip': 3.0708093643188477, 'timing/data': 0.0003424929454922676, 'timing/model': 0.6404702281579375, 'learning_rate/qwen_vl_interface': 6.666792094184817e-06, 'learning_rate/action_model': 6.666792094184818e-05, 'epoch': 0.58})
+[2026-07-28 13:17:54,277][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.04099143669009209, 'train/grad_norm_pre_clip': 1.621382474899292, 'timing/data': 0.0003394479863345623, 'timing/model': 0.41639669379219413, 'learning_rate/qwen_vl_interface': 6.663139505812215e-06, 'learning_rate/action_model': 6.663139505812215e-05, 'epoch': 0.58})
+[2026-07-28 13:18:00,375][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.035057954490184784, 'train/grad_norm_pre_clip': 0.7418152689933777, 'timing/data': 0.00029027974233031273, 'timing/model': 0.6531476289965212, 'learning_rate/qwen_vl_interface': 6.659486000468728e-06, 'learning_rate/action_model': 6.659486000468728e-05, 'epoch': 0.58})
+[2026-07-28 13:18:06,252][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.059103067964315414, 'train/grad_norm_pre_clip': 0.7290005087852478, 'timing/data': 0.00028479471802711487, 'timing/model': 0.41416330402716994, 'learning_rate/qwen_vl_interface': 6.6558315805250805e-06, 'learning_rate/action_model': 6.65583158052508e-05, 'epoch': 0.58})
+[2026-07-28 13:18:12,165][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.045460477471351624, 'train/grad_norm_pre_clip': 1.0579829216003418, 'timing/data': 0.00031124893575906754, 'timing/model': 0.41474514175206423, 'learning_rate/qwen_vl_interface': 6.652176248352581e-06, 'learning_rate/action_model': 6.65217624835258e-05, 'epoch': 0.58})
+[2026-07-28 13:18:18,015][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.04476296901702881, 'train/grad_norm_pre_clip': 0.9510773420333862, 'timing/data': 0.0002802829258143902, 'timing/model': 0.41593433218076825, 'learning_rate/qwen_vl_interface': 6.64852000632314e-06, 'learning_rate/action_model': 6.648520006323139e-05, 'epoch': 0.58})
+[2026-07-28 13:18:24,087][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.0664510577917099, 'train/grad_norm_pre_clip': 1.4855921268463135, 'timing/data': 0.0002803271636366844, 'timing/model': 0.4144049738533795, 'learning_rate/qwen_vl_interface': 6.644862856809247e-06, 'learning_rate/action_model': 6.644862856809247e-05, 'epoch': 0.58})
+[2026-07-28 13:18:30,036][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.04178505018353462, 'train/grad_norm_pre_clip': 1.6561880111694336, 'timing/data': 0.00028670206665992737, 'timing/model': 0.42219438729807734, 'learning_rate/qwen_vl_interface': 6.64120480218399e-06, 'learning_rate/action_model': 6.64120480218399e-05, 'epoch': 0.58})
+[2026-07-28 13:18:35,978][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.019277920946478844, 'train/grad_norm_pre_clip': 0.699720561504364, 'timing/data': 0.0003507849760353565, 'timing/model': 0.4488310110755265, 'learning_rate/qwen_vl_interface': 6.637545844821043e-06, 'learning_rate/action_model': 6.637545844821043e-05, 'epoch': 0.58})
+[2026-07-28 13:18:41,877][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.03360630199313164, 'train/grad_norm_pre_clip': 0.8153571486473083, 'timing/data': 0.00027607381343841553, 'timing/model': 0.4156093238852918, 'learning_rate/qwen_vl_interface': 6.633885987094661e-06, 'learning_rate/action_model': 6.633885987094661e-05, 'epoch': 0.58})
+[2026-07-28 13:18:48,000][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.024060958996415138, 'train/grad_norm_pre_clip': 0.5897172689437866, 'timing/data': 0.00026603415608406067, 'timing/model': 0.41338981688022614, 'learning_rate/qwen_vl_interface': 6.630225231379687e-06, 'learning_rate/action_model': 6.630225231379687e-05, 'epoch': 0.58})
+[2026-07-28 13:18:53,899][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.03645768016576767, 'train/grad_norm_pre_clip': 0.8162497878074646, 'timing/data': 0.00029484881088137627, 'timing/model': 0.4243914606049657, 'learning_rate/qwen_vl_interface': 6.626563580051546e-06, 'learning_rate/action_model': 6.626563580051546e-05, 'epoch': 0.59})
+[2026-07-28 13:18:59,790][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.05535554140806198, 'train/grad_norm_pre_clip': 1.0759358406066895, 'timing/data': 0.00029457593336701393, 'timing/model': 0.4153684279881418, 'learning_rate/qwen_vl_interface': 6.622901035486241e-06, 'learning_rate/action_model': 6.622901035486241e-05, 'epoch': 0.59})
+[2026-07-28 13:19:06,134][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.007540919352322817, 'train/grad_norm_pre_clip': 0.2612625062465668, 'timing/data': 0.000340132974088192, 'timing/model': 0.6387148350477219, 'learning_rate/qwen_vl_interface': 6.61923760006036e-06, 'learning_rate/action_model': 6.61923760006036e-05, 'epoch': 0.59})
+[2026-07-28 13:19:11,782][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.05157378315925598, 'train/grad_norm_pre_clip': 0.7814986109733582, 'timing/data': 0.0003564273938536644, 'timing/model': 0.4166276128962636, 'learning_rate/qwen_vl_interface': 6.615573276151068e-06, 'learning_rate/action_model': 6.615573276151068e-05, 'epoch': 0.59})
+[2026-07-28 13:19:17,463][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.03218764811754227, 'train/grad_norm_pre_clip': 1.4109467267990112, 'timing/data': 0.0002808179706335068, 'timing/model': 0.4140049940906465, 'learning_rate/qwen_vl_interface': 6.6119080661361e-06, 'learning_rate/action_model': 6.6119080661361e-05, 'epoch': 0.59})
+[2026-07-28 13:19:23,219][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.05526316910982132, 'train/grad_norm_pre_clip': 0.7966216802597046, 'timing/data': 0.0006363377906382084, 'timing/model': 0.4470860240980983, 'learning_rate/qwen_vl_interface': 6.608241972393774e-06, 'learning_rate/action_model': 6.608241972393774e-05, 'epoch': 0.59})
+[2026-07-28 13:19:29,534][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.029013406485319138, 'train/grad_norm_pre_clip': 0.7546945810317993, 'timing/data': 0.000256112776696682, 'timing/model': 0.6441418998874724, 'learning_rate/qwen_vl_interface': 6.60457499730298e-06, 'learning_rate/action_model': 6.60457499730298e-05, 'epoch': 0.59})
+[2026-07-28 13:19:35,370][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.034478768706321716, 'train/grad_norm_pre_clip': 2.1370911598205566, 'timing/data': 0.00035500992089509964, 'timing/model': 0.4147929330356419, 'learning_rate/qwen_vl_interface': 6.600907143243176e-06, 'learning_rate/action_model': 6.600907143243176e-05, 'epoch': 0.59})
+[2026-07-28 13:19:41,215][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.030701834708452225, 'train/grad_norm_pre_clip': 0.7789117097854614, 'timing/data': 0.0002853008918464184, 'timing/model': 0.41492342576384544, 'learning_rate/qwen_vl_interface': 6.5972384125943945e-06, 'learning_rate/action_model': 6.597238412594394e-05, 'epoch': 0.59})
+[2026-07-28 13:19:47,324][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.02778349071741104, 'train/grad_norm_pre_clip': 1.4821830987930298, 'timing/data': 0.0003317040391266346, 'timing/model': 0.41535049583762884, 'learning_rate/qwen_vl_interface': 6.593568807737233e-06, 'learning_rate/action_model': 6.593568807737233e-05, 'epoch': 0.59})
+[2026-07-28 13:19:53,188][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.03578661382198334, 'train/grad_norm_pre_clip': 1.4910129308700562, 'timing/data': 0.0003941608592867851, 'timing/model': 0.41644257865846157, 'learning_rate/qwen_vl_interface': 6.5898983310528584e-06, 'learning_rate/action_model': 6.589898331052858e-05, 'epoch': 0.59})
+[2026-07-28 13:19:59,337][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.0289931520819664, 'train/grad_norm_pre_clip': 1.9311697483062744, 'timing/data': 0.00027150101959705353, 'timing/model': 0.4137384817004204, 'learning_rate/qwen_vl_interface': 6.586226984923004e-06, 'learning_rate/action_model': 6.586226984923004e-05, 'epoch': 0.59})
+[2026-07-28 13:20:05,217][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.03386078029870987, 'train/grad_norm_pre_clip': 1.228364109992981, 'timing/data': 0.0002848636358976364, 'timing/model': 0.4163609310053289, 'learning_rate/qwen_vl_interface': 6.582554771729967e-06, 'learning_rate/action_model': 6.582554771729966e-05, 'epoch': 0.59})
+[2026-07-28 13:20:11,488][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.0725216418504715, 'train/grad_norm_pre_clip': 1.2905393838882446, 'timing/data': 0.00035209907218813896, 'timing/model': 0.4148714770562947, 'learning_rate/qwen_vl_interface': 6.578881693856605e-06, 'learning_rate/action_model': 6.578881693856604e-05, 'epoch': 0.59})
+[2026-07-28 13:20:17,607][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.044610343873500824, 'train/grad_norm_pre_clip': 1.1272389888763428, 'timing/data': 0.0003295689821243286, 'timing/model': 0.414367298129946, 'learning_rate/qwen_vl_interface': 6.5752077536863404e-06, 'learning_rate/action_model': 6.57520775368634e-05, 'epoch': 0.59})
+[2026-07-28 13:20:23,490][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.055510953068733215, 'train/grad_norm_pre_clip': 1.29134202003479, 'timing/data': 0.0003461441956460476, 'timing/model': 0.4144142880104482, 'learning_rate/qwen_vl_interface': 6.571532953603151e-06, 'learning_rate/action_model': 6.571532953603151e-05, 'epoch': 0.59})
+[2026-07-28 13:20:29,400][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.03311803191900253, 'train/grad_norm_pre_clip': 0.7480611205101013, 'timing/data': 0.00028346898034214973, 'timing/model': 0.41417642310261726, 'learning_rate/qwen_vl_interface': 6.567857295991577e-06, 'learning_rate/action_model': 6.567857295991576e-05, 'epoch': 0.59})
+[2026-07-28 13:20:35,488][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.04813941568136215, 'train/grad_norm_pre_clip': 0.9887763261795044, 'timing/data': 0.0002740798518061638, 'timing/model': 0.4121500956825912, 'learning_rate/qwen_vl_interface': 6.56418078323671e-06, 'learning_rate/action_model': 6.56418078323671e-05, 'epoch': 0.59})
+[2026-07-28 13:20:41,625][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.01810920611023903, 'train/grad_norm_pre_clip': 0.4925168454647064, 'timing/data': 0.0003511141985654831, 'timing/model': 0.6593289412558079, 'learning_rate/qwen_vl_interface': 6.5605034177242e-06, 'learning_rate/action_model': 6.5605034177242e-05, 'epoch': 0.59})
+[2026-07-28 13:20:47,503][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.02401912584900856, 'train/grad_norm_pre_clip': 0.6077352166175842, 'timing/data': 0.0002881190739572048, 'timing/model': 0.42097195982933044, 'learning_rate/qwen_vl_interface': 6.556825201840252e-06, 'learning_rate/action_model': 6.556825201840252e-05, 'epoch': 0.59})
+[2026-07-28 13:20:53,407][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.06297615170478821, 'train/grad_norm_pre_clip': 1.7751346826553345, 'timing/data': 0.0007001119665801525, 'timing/model': 0.4215477551333606, 'learning_rate/qwen_vl_interface': 6.553146137971618e-06, 'learning_rate/action_model': 6.553146137971618e-05, 'epoch': 0.59})
+[2026-07-28 13:20:59,277][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.020798467099666595, 'train/grad_norm_pre_clip': 0.5663374662399292, 'timing/data': 0.00033122021704912186, 'timing/model': 0.415279861073941, 'learning_rate/qwen_vl_interface': 6.549466228505603e-06, 'learning_rate/action_model': 6.549466228505603e-05, 'epoch': 0.59})
+[2026-07-28 13:21:05,410][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.02423073723912239, 'train/grad_norm_pre_clip': 1.279475212097168, 'timing/data': 0.00029665883630514145, 'timing/model': 0.6437591179274023, 'learning_rate/qwen_vl_interface': 6.545785475830061e-06, 'learning_rate/action_model': 6.54578547583006e-05, 'epoch': 0.59})
+[2026-07-28 13:21:11,536][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.027148514986038208, 'train/grad_norm_pre_clip': 1.4397554397583008, 'timing/data': 0.00026555173099040985, 'timing/model': 0.649104034062475, 'learning_rate/qwen_vl_interface': 6.54210388233339e-06, 'learning_rate/action_model': 6.542103882333391e-05, 'epoch': 0.59})
+[2026-07-28 13:21:17,437][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.02870389260351658, 'train/grad_norm_pre_clip': 1.8681343793869019, 'timing/data': 0.0002807960845530033, 'timing/model': 0.4137772135436535, 'learning_rate/qwen_vl_interface': 6.538421450404538e-06, 'learning_rate/action_model': 6.538421450404538e-05, 'epoch': 0.59})
+[2026-07-28 13:21:23,329][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.052107781171798706, 'train/grad_norm_pre_clip': 0.9350643754005432, 'timing/data': 0.00034894607961177826, 'timing/model': 0.4162529679015279, 'learning_rate/qwen_vl_interface': 6.534738182432999e-06, 'learning_rate/action_model': 6.534738182432999e-05, 'epoch': 0.59})
+[2026-07-28 13:21:29,193][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.07861106842756271, 'train/grad_norm_pre_clip': 1.7655396461486816, 'timing/data': 0.000292294193059206, 'timing/model': 0.4158964450471103, 'learning_rate/qwen_vl_interface': 6.531054080808798e-06, 'learning_rate/action_model': 6.531054080808798e-05, 'epoch': 0.59})
+[2026-07-28 13:21:35,276][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.038776081055402756, 'train/grad_norm_pre_clip': 0.7603821754455566, 'timing/data': 0.000270073302090168, 'timing/model': 0.41433145478367805, 'learning_rate/qwen_vl_interface': 6.5273691479225124e-06, 'learning_rate/action_model': 6.527369147922511e-05, 'epoch': 0.59})
+[2026-07-28 13:21:41,134][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.05300642549991608, 'train/grad_norm_pre_clip': 1.0386958122253418, 'timing/data': 0.0002980087883770466, 'timing/model': 0.41698123002424836, 'learning_rate/qwen_vl_interface': 6.523683386165257e-06, 'learning_rate/action_model': 6.523683386165256e-05, 'epoch': 0.59})
+[2026-07-28 13:21:47,261][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.03664065897464752, 'train/grad_norm_pre_clip': 0.9195969700813293, 'timing/data': 0.00034037092700600624, 'timing/model': 0.6577859367243946, 'learning_rate/qwen_vl_interface': 6.519996797928677e-06, 'learning_rate/action_model': 6.519996797928677e-05, 'epoch': 0.59})
+[2026-07-28 13:21:53,401][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.04966563358902931, 'train/grad_norm_pre_clip': 1.3613675832748413, 'timing/data': 0.0002770526334643364, 'timing/model': 0.4136763601563871, 'learning_rate/qwen_vl_interface': 6.5163093856049664e-06, 'learning_rate/action_model': 6.516309385604967e-05, 'epoch': 0.59})
+[2026-07-28 13:21:59,224][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.029200851917266846, 'train/grad_norm_pre_clip': 1.036403775215149, 'timing/data': 0.00027814414352178574, 'timing/model': 0.4208608539775014, 'learning_rate/qwen_vl_interface': 6.512621151586842e-06, 'learning_rate/action_model': 6.512621151586842e-05, 'epoch': 0.59})
+[2026-07-28 13:22:05,018][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.047797806560993195, 'train/grad_norm_pre_clip': 0.939068078994751, 'timing/data': 0.00032904185354709625, 'timing/model': 0.41617129324004054, 'learning_rate/qwen_vl_interface': 6.508932098267561e-06, 'learning_rate/action_model': 6.50893209826756e-05, 'epoch': 0.59})
+[2026-07-28 13:22:10,773][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.02574087679386139, 'train/grad_norm_pre_clip': 1.2150709629058838, 'timing/data': 0.0003370591439306736, 'timing/model': 0.42943698121234775, 'learning_rate/qwen_vl_interface': 6.505242228040908e-06, 'learning_rate/action_model': 6.505242228040907e-05, 'epoch': 0.59})
+[2026-07-28 13:22:17,300][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.022883683443069458, 'train/grad_norm_pre_clip': 1.0289362668991089, 'timing/data': 0.000260369386523962, 'timing/model': 0.7726021111011505, 'learning_rate/qwen_vl_interface': 6.501551543301203e-06, 'learning_rate/action_model': 6.501551543301203e-05, 'epoch': 0.59})
+[2026-07-28 13:22:23,472][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.04994729906320572, 'train/grad_norm_pre_clip': 1.4942418336868286, 'timing/data': 0.00026132911443710327, 'timing/model': 0.42518738796934485, 'learning_rate/qwen_vl_interface': 6.497860046443288e-06, 'learning_rate/action_model': 6.497860046443288e-05, 'epoch': 0.59})
+[2026-07-28 13:22:29,627][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.06683947145938873, 'train/grad_norm_pre_clip': 2.084437131881714, 'timing/data': 0.0002703983336687088, 'timing/model': 0.4260646188631654, 'learning_rate/qwen_vl_interface': 6.494167739862538e-06, 'learning_rate/action_model': 6.494167739862538e-05, 'epoch': 0.59})
+[2026-07-28 13:22:35,720][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.08407354354858398, 'train/grad_norm_pre_clip': 2.9389808177948, 'timing/data': 0.0002876291982829571, 'timing/model': 0.41537689743563533, 'learning_rate/qwen_vl_interface': 6.490474625954849e-06, 'learning_rate/action_model': 6.490474625954849e-05, 'epoch': 0.59})
+[2026-07-28 13:22:41,922][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.04142363741993904, 'train/grad_norm_pre_clip': 1.8131307363510132, 'timing/data': 0.0002564508467912674, 'timing/model': 0.42477651266381145, 'learning_rate/qwen_vl_interface': 6.486780707116642e-06, 'learning_rate/action_model': 6.486780707116642e-05, 'epoch': 0.59})
+[2026-07-28 13:22:47,934][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.023477518931031227, 'train/grad_norm_pre_clip': 0.8944239616394043, 'timing/data': 0.0003579570911824703, 'timing/model': 0.41897039068862796, 'learning_rate/qwen_vl_interface': 6.483085985744864e-06, 'learning_rate/action_model': 6.483085985744863e-05, 'epoch': 0.59})
+[2026-07-28 13:22:54,105][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.02409033291041851, 'train/grad_norm_pre_clip': 1.2519121170043945, 'timing/data': 0.00031820405274629593, 'timing/model': 0.414507573004812, 'learning_rate/qwen_vl_interface': 6.479390464236974e-06, 'learning_rate/action_model': 6.479390464236974e-05, 'epoch': 0.59})
+[2026-07-28 13:23:00,047][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.11551795899868011, 'train/grad_norm_pre_clip': 2.237989902496338, 'timing/data': 0.00028505316004157066, 'timing/model': 0.41575092915445566, 'learning_rate/qwen_vl_interface': 6.475694144990961e-06, 'learning_rate/action_model': 6.475694144990961e-05, 'epoch': 0.59})
+[2026-07-28 13:23:05,943][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.028612717986106873, 'train/grad_norm_pre_clip': 1.0705456733703613, 'timing/data': 0.0002903090789914131, 'timing/model': 0.41391376918181777, 'learning_rate/qwen_vl_interface': 6.471997030405326e-06, 'learning_rate/action_model': 6.471997030405325e-05, 'epoch': 0.59})
+[2026-07-28 13:23:12,166][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.03791102021932602, 'train/grad_norm_pre_clip': 1.1070359945297241, 'timing/data': 0.00042551569640636444, 'timing/model': 0.6642878316342831, 'learning_rate/qwen_vl_interface': 6.468299122879082e-06, 'learning_rate/action_model': 6.468299122879081e-05, 'epoch': 0.59})
+[2026-07-28 13:23:18,340][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.04463814198970795, 'train/grad_norm_pre_clip': 1.0489282608032227, 'timing/data': 0.0002704109065234661, 'timing/model': 0.42250752774998546, 'learning_rate/qwen_vl_interface': 6.464600424811765e-06, 'learning_rate/action_model': 6.464600424811764e-05, 'epoch': 0.59})
+[2026-07-28 13:23:24,205][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.042715754359960556, 'train/grad_norm_pre_clip': 1.5166635513305664, 'timing/data': 0.0002732807770371437, 'timing/model': 0.41985591081902385, 'learning_rate/qwen_vl_interface': 6.460900938603418e-06, 'learning_rate/action_model': 6.460900938603418e-05, 'epoch': 0.59})
+[2026-07-28 13:23:30,065][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.02092374488711357, 'train/grad_norm_pre_clip': 0.6142362952232361, 'timing/data': 0.0002748146653175354, 'timing/model': 0.4139202879741788, 'learning_rate/qwen_vl_interface': 6.457200666654601e-06, 'learning_rate/action_model': 6.457200666654601e-05, 'epoch': 0.59})
+[2026-07-28 13:23:36,165][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.025686275213956833, 'train/grad_norm_pre_clip': 0.7002941966056824, 'timing/data': 0.00034472299739718437, 'timing/model': 0.4138815961778164, 'learning_rate/qwen_vl_interface': 6.453499611366376e-06, 'learning_rate/action_model': 6.453499611366376e-05, 'epoch': 0.59})
+[2026-07-28 13:23:42,014][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.03551940619945526, 'train/grad_norm_pre_clip': 1.367030143737793, 'timing/data': 0.000292058102786541, 'timing/model': 0.41648297384381294, 'learning_rate/qwen_vl_interface': 6.449797775140321e-06, 'learning_rate/action_model': 6.44979777514032e-05, 'epoch': 0.59})
+[2026-07-28 13:23:47,888][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.03262278810143471, 'train/grad_norm_pre_clip': 0.9668104648590088, 'timing/data': 0.0003521186299622059, 'timing/model': 0.4163408321328461, 'learning_rate/qwen_vl_interface': 6.446095160378514e-06, 'learning_rate/action_model': 6.446095160378514e-05, 'epoch': 0.59})
+[2026-07-28 13:23:54,177][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.06860106438398361, 'train/grad_norm_pre_clip': 1.7708009481430054, 'timing/data': 0.00028302520513534546, 'timing/model': 0.5903433510102332, 'learning_rate/qwen_vl_interface': 6.442391769483544e-06, 'learning_rate/action_model': 6.442391769483544e-05, 'epoch': 0.59})
+[2026-07-28 13:24:00,027][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.06366530805826187, 'train/grad_norm_pre_clip': 1.7060439586639404, 'timing/data': 0.000369243323802948, 'timing/model': 0.41544661298394203, 'learning_rate/qwen_vl_interface': 6.4386876048585e-06, 'learning_rate/action_model': 6.4386876048585e-05, 'epoch': 0.59})
+[2026-07-28 13:24:05,715][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.025617945939302444, 'train/grad_norm_pre_clip': 0.8133237361907959, 'timing/data': 0.0003493828698992729, 'timing/model': 0.43215495301410556, 'learning_rate/qwen_vl_interface': 6.434982668906975e-06, 'learning_rate/action_model': 6.434982668906974e-05, 'epoch': 0.59})
+[2026-07-28 13:24:11,413][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.028449829667806625, 'train/grad_norm_pre_clip': 1.475398063659668, 'timing/data': 0.00028412695974111557, 'timing/model': 0.4142821920104325, 'learning_rate/qwen_vl_interface': 6.431276964033062e-06, 'learning_rate/action_model': 6.431276964033062e-05, 'epoch': 0.59})
+[2026-07-28 13:24:17,542][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.02194470912218094, 'train/grad_norm_pre_clip': 1.3013360500335693, 'timing/data': 0.0002920404076576233, 'timing/model': 0.6545999380759895, 'learning_rate/qwen_vl_interface': 6.427570492641351e-06, 'learning_rate/action_model': 6.427570492641351e-05, 'epoch': 0.59})
+[2026-07-28 13:24:23,671][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.08378323912620544, 'train/grad_norm_pre_clip': 1.6866345405578613, 'timing/data': 0.00025887880474328995, 'timing/model': 0.4248781939968467, 'learning_rate/qwen_vl_interface': 6.423863257136931e-06, 'learning_rate/action_model': 6.42386325713693e-05, 'epoch': 0.6})
+[2026-07-28 13:24:29,551][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.029949843883514404, 'train/grad_norm_pre_clip': 0.6817092895507812, 'timing/data': 0.0003504771739244461, 'timing/model': 0.41543053183704615, 'learning_rate/qwen_vl_interface': 6.42015525992539e-06, 'learning_rate/action_model': 6.42015525992539e-05, 'epoch': 0.6})
+[2026-07-28 13:24:35,539][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.03350403159856796, 'train/grad_norm_pre_clip': 0.9981123208999634, 'timing/data': 0.00039182882755994797, 'timing/model': 0.42731195595115423, 'learning_rate/qwen_vl_interface': 6.416446503412803e-06, 'learning_rate/action_model': 6.416446503412803e-05, 'epoch': 0.6})
+[2026-07-28 13:24:41,761][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.029377590864896774, 'train/grad_norm_pre_clip': 1.3154771327972412, 'timing/data': 0.0002637789584696293, 'timing/model': 0.413119925186038, 'learning_rate/qwen_vl_interface': 6.412736990005746e-06, 'learning_rate/action_model': 6.412736990005746e-05, 'epoch': 0.6})
+[2026-07-28 13:24:47,793][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.020971687510609627, 'train/grad_norm_pre_clip': 0.7796088457107544, 'timing/data': 0.000291312113404274, 'timing/model': 0.41563045512884855, 'learning_rate/qwen_vl_interface': 6.409026722111279e-06, 'learning_rate/action_model': 6.409026722111278e-05, 'epoch': 0.6})
+[2026-07-28 13:24:54,118][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.04045889526605606, 'train/grad_norm_pre_clip': 1.7060081958770752, 'timing/data': 0.0003538280725479126, 'timing/model': 0.64765217108652, 'learning_rate/qwen_vl_interface': 6.405315702136955e-06, 'learning_rate/action_model': 6.405315702136954e-05, 'epoch': 0.6})
+[2026-07-28 13:24:59,995][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.027886521071195602, 'train/grad_norm_pre_clip': 0.6114420890808105, 'timing/data': 0.0003240196965634823, 'timing/model': 0.42098339181393385, 'learning_rate/qwen_vl_interface': 6.401603932490815e-06, 'learning_rate/action_model': 6.401603932490815e-05, 'epoch': 0.6})
+[2026-07-28 13:25:06,110][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.01943974941968918, 'train/grad_norm_pre_clip': 0.45470091700553894, 'timing/data': 0.00028022006154060364, 'timing/model': 0.6552232182584703, 'learning_rate/qwen_vl_interface': 6.3978914155813865e-06, 'learning_rate/action_model': 6.397891415581386e-05, 'epoch': 0.6})
+[2026-07-28 13:25:12,007][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.03239354118704796, 'train/grad_norm_pre_clip': 1.2997711896896362, 'timing/data': 0.00038383621722459793, 'timing/model': 0.4175563999451697, 'learning_rate/qwen_vl_interface': 6.394178153817682e-06, 'learning_rate/action_model': 6.394178153817682e-05, 'epoch': 0.6})
+[2026-07-28 13:25:17,914][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.045020535588264465, 'train/grad_norm_pre_clip': 1.364525318145752, 'timing/data': 0.00036770710721611977, 'timing/model': 0.41326637007296085, 'learning_rate/qwen_vl_interface': 6.390464149609193e-06, 'learning_rate/action_model': 6.390464149609193e-05, 'epoch': 0.6})
+[2026-07-28 13:25:24,337][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3250
+[2026-07-28 13:25:29,292][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3250
+[2026-07-28 13:32:47,520][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3250
+[2026-07-28 13:32:47,521][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.06920134276151657, 'train/grad_norm_pre_clip': 2.086993932723999, 'eval/action_loss/seconds': 4.950759737286717, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.04505721107125282, 'eval/latency_3/loss': 0.04505721107125282, 'eval/flappy/loss': 0.04505721107125282, 'eval/flappy/latency_3/loss': 0.04505721107125282, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 347.12000408545134, 'rl_games_eval/mid_train/mean_length': 2815.25, 'rl_games_eval/mid_train/std_reward': 150.30974728200624, 'rl_games_eval/mid_train/std_length': 1210.177171946323, 'rl_games_eval/mid_train/macro_mean_reward': 347.12000408545134, 'rl_games_eval/mid_train/macro_mean_length': 2815.25, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 347.12000408545134, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2815.25, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 150.30974728200624, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1210.177171946323, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 347.12000408545134, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00027588289231061935, 'timing/model': 0.4125402350910008, 'learning_rate/qwen_vl_interface': 6.386749405365901e-06, 'learning_rate/action_model': 6.3867494053659e-05, 'epoch': 0.6})
+[2026-07-28 13:32:53,582][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.06815734505653381, 'train/grad_norm_pre_clip': 2.0149877071380615, 'timing/data': 0.00029362691566348076, 'timing/model': 0.4292196216993034, 'learning_rate/qwen_vl_interface': 6.3830339234982596e-06, 'learning_rate/action_model': 6.38303392349826e-05, 'epoch': 0.6})
+[2026-07-28 13:32:59,507][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.05978246033191681, 'train/grad_norm_pre_clip': 1.9137344360351562, 'timing/data': 0.0002987422049045563, 'timing/model': 0.4154845359735191, 'learning_rate/qwen_vl_interface': 6.3793177064172075e-06, 'learning_rate/action_model': 6.379317706417207e-05, 'epoch': 0.6})
+[2026-07-28 13:33:05,628][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.03155070170760155, 'train/grad_norm_pre_clip': 1.4606022834777832, 'timing/data': 0.00030608614906668663, 'timing/model': 0.6638483731076121, 'learning_rate/qwen_vl_interface': 6.375600756534157e-06, 'learning_rate/action_model': 6.375600756534157e-05, 'epoch': 0.6})
+[2026-07-28 13:33:11,729][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.04236958175897598, 'train/grad_norm_pre_clip': 1.921411156654358, 'timing/data': 0.00033102789893746376, 'timing/model': 0.41380520118400455, 'learning_rate/qwen_vl_interface': 6.371883076260996e-06, 'learning_rate/action_model': 6.371883076260997e-05, 'epoch': 0.6})
+[2026-07-28 13:33:17,647][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.03915870189666748, 'train/grad_norm_pre_clip': 1.505751609802246, 'timing/data': 0.0002995012328028679, 'timing/model': 0.4230732279829681, 'learning_rate/qwen_vl_interface': 6.368164668010087e-06, 'learning_rate/action_model': 6.368164668010087e-05, 'epoch': 0.6})
+[2026-07-28 13:33:23,557][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.04294347018003464, 'train/grad_norm_pre_clip': 2.0014257431030273, 'timing/data': 0.0002910890616476536, 'timing/model': 0.41488457517698407, 'learning_rate/qwen_vl_interface': 6.364445534194263e-06, 'learning_rate/action_model': 6.364445534194263e-05, 'epoch': 0.6})
+[2026-07-28 13:33:29,507][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.02710592746734619, 'train/grad_norm_pre_clip': 1.0461786985397339, 'timing/data': 0.0003248020075261593, 'timing/model': 0.4161747177131474, 'learning_rate/qwen_vl_interface': 6.360725677226834e-06, 'learning_rate/action_model': 6.360725677226834e-05, 'epoch': 0.6})
+[2026-07-28 13:33:35,668][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.02907322719693184, 'train/grad_norm_pre_clip': 0.6670939326286316, 'timing/data': 0.0002765296958386898, 'timing/model': 0.6466889390721917, 'learning_rate/qwen_vl_interface': 6.357005099521569e-06, 'learning_rate/action_model': 6.357005099521569e-05, 'epoch': 0.6})
+[2026-07-28 13:33:41,584][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.03961435705423355, 'train/grad_norm_pre_clip': 2.1472558975219727, 'timing/data': 0.0005795699544250965, 'timing/model': 0.4134733830578625, 'learning_rate/qwen_vl_interface': 6.353283803492717e-06, 'learning_rate/action_model': 6.353283803492717e-05, 'epoch': 0.6})
+[2026-07-28 13:33:47,495][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.02975437231361866, 'train/grad_norm_pre_clip': 1.1542153358459473, 'timing/data': 0.00029002316296100616, 'timing/model': 0.4294403321109712, 'learning_rate/qwen_vl_interface': 6.349561791554981e-06, 'learning_rate/action_model': 6.349561791554981e-05, 'epoch': 0.6})
+[2026-07-28 13:33:53,635][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.043806180357933044, 'train/grad_norm_pre_clip': 1.5620026588439941, 'timing/data': 0.00028573302552103996, 'timing/model': 0.6420001373626292, 'learning_rate/qwen_vl_interface': 6.345839066123536e-06, 'learning_rate/action_model': 6.345839066123535e-05, 'epoch': 0.6})
+[2026-07-28 13:33:59,540][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.052208442240953445, 'train/grad_norm_pre_clip': 1.660311222076416, 'timing/data': 0.0002974090166389942, 'timing/model': 0.4153285133652389, 'learning_rate/qwen_vl_interface': 6.342115629614019e-06, 'learning_rate/action_model': 6.342115629614019e-05, 'epoch': 0.6})
+[2026-07-28 13:34:05,445][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.023179760202765465, 'train/grad_norm_pre_clip': 0.691561758518219, 'timing/data': 0.0003445791080594063, 'timing/model': 0.4184504640288651, 'learning_rate/qwen_vl_interface': 6.338391484442527e-06, 'learning_rate/action_model': 6.338391484442526e-05, 'epoch': 0.6})
+[2026-07-28 13:34:11,571][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.026749012991786003, 'train/grad_norm_pre_clip': 0.8873118162155151, 'timing/data': 0.00034667830914258957, 'timing/model': 0.6571829300373793, 'learning_rate/qwen_vl_interface': 6.334666633025616e-06, 'learning_rate/action_model': 6.334666633025615e-05, 'epoch': 0.6})
+[2026-07-28 13:34:17,457][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.0328361876308918, 'train/grad_norm_pre_clip': 0.9995361566543579, 'timing/data': 0.0003383723087608814, 'timing/model': 0.41275919135659933, 'learning_rate/qwen_vl_interface': 6.3309410777803025e-06, 'learning_rate/action_model': 6.330941077780302e-05, 'epoch': 0.6})
+[2026-07-28 13:34:23,581][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.030453171581029892, 'train/grad_norm_pre_clip': 1.2543590068817139, 'timing/data': 0.0002717110328376293, 'timing/model': 0.41448735864833, 'learning_rate/qwen_vl_interface': 6.327214821124063e-06, 'learning_rate/action_model': 6.327214821124062e-05, 'epoch': 0.6})
+[2026-07-28 13:34:29,519][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.030489560216665268, 'train/grad_norm_pre_clip': 1.4180082082748413, 'timing/data': 0.0002704048529267311, 'timing/model': 0.41580452397465706, 'learning_rate/qwen_vl_interface': 6.323487865474818e-06, 'learning_rate/action_model': 6.323487865474817e-05, 'epoch': 0.6})
+[2026-07-28 13:34:35,415][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.015188182704150677, 'train/grad_norm_pre_clip': 0.9075674414634705, 'timing/data': 0.00033696508035063744, 'timing/model': 0.4153317338787019, 'learning_rate/qwen_vl_interface': 6.319760213250952e-06, 'learning_rate/action_model': 6.319760213250951e-05, 'epoch': 0.6})
+[2026-07-28 13:34:41,578][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.04849427938461304, 'train/grad_norm_pre_clip': 1.5491970777511597, 'timing/data': 0.0002803499810397625, 'timing/model': 0.4143321989104152, 'learning_rate/qwen_vl_interface': 6.316031866871297e-06, 'learning_rate/action_model': 6.316031866871297e-05, 'epoch': 0.6})
+[2026-07-28 13:34:47,504][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.0303964763879776, 'train/grad_norm_pre_clip': 0.8710349798202515, 'timing/data': 0.0003668367862701416, 'timing/model': 0.42891675885766745, 'learning_rate/qwen_vl_interface': 6.31230282875514e-06, 'learning_rate/action_model': 6.31230282875514e-05, 'epoch': 0.6})
+[2026-07-28 13:34:53,699][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.02457069605588913, 'train/grad_norm_pre_clip': 0.9111555814743042, 'timing/data': 0.00028172973543405533, 'timing/model': 0.6414803881198168, 'learning_rate/qwen_vl_interface': 6.308573101322209e-06, 'learning_rate/action_model': 6.308573101322209e-05, 'epoch': 0.6})
+[2026-07-28 13:34:59,679][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.05299828201532364, 'train/grad_norm_pre_clip': 1.0661287307739258, 'timing/data': 0.00028847530484199524, 'timing/model': 0.4154368028976023, 'learning_rate/qwen_vl_interface': 6.304842686992685e-06, 'learning_rate/action_model': 6.304842686992684e-05, 'epoch': 0.6})
+[2026-07-28 13:35:05,791][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.0502980537712574, 'train/grad_norm_pre_clip': 1.2424577474594116, 'timing/data': 0.00026484904810786247, 'timing/model': 0.4139066068455577, 'learning_rate/qwen_vl_interface': 6.301111588187194e-06, 'learning_rate/action_model': 6.301111588187194e-05, 'epoch': 0.6})
+[2026-07-28 13:35:11,678][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.02610817179083824, 'train/grad_norm_pre_clip': 0.9035791158676147, 'timing/data': 0.0002833418548107147, 'timing/model': 0.41450196225196123, 'learning_rate/qwen_vl_interface': 6.297379807326806e-06, 'learning_rate/action_model': 6.297379807326806e-05, 'epoch': 0.6})
+[2026-07-28 13:35:17,540][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.02763817459344864, 'train/grad_norm_pre_clip': 1.4052883386611938, 'timing/data': 0.0002713417634367943, 'timing/model': 0.4172621536999941, 'learning_rate/qwen_vl_interface': 6.293647346833033e-06, 'learning_rate/action_model': 6.293647346833032e-05, 'epoch': 0.6})
+[2026-07-28 13:35:23,676][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.04217841476202011, 'train/grad_norm_pre_clip': 0.9401031732559204, 'timing/data': 0.00025790976360440254, 'timing/model': 0.41616067523136735, 'learning_rate/qwen_vl_interface': 6.289914209127826e-06, 'learning_rate/action_model': 6.289914209127825e-05, 'epoch': 0.6})
+[2026-07-28 13:35:29,593][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.04067782312631607, 'train/grad_norm_pre_clip': 1.6381652355194092, 'timing/data': 0.00028524594381451607, 'timing/model': 0.41641375981271267, 'learning_rate/qwen_vl_interface': 6.286180396633579e-06, 'learning_rate/action_model': 6.286180396633579e-05, 'epoch': 0.6})
+[2026-07-28 13:35:35,751][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.029125621542334557, 'train/grad_norm_pre_clip': 0.8848812580108643, 'timing/data': 0.0003620712086558342, 'timing/model': 0.6612897580489516, 'learning_rate/qwen_vl_interface': 6.282445911773122e-06, 'learning_rate/action_model': 6.282445911773121e-05, 'epoch': 0.6})
+[2026-07-28 13:35:41,866][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.015465679578483105, 'train/grad_norm_pre_clip': 0.5190316438674927, 'timing/data': 0.000345367006957531, 'timing/model': 0.41342787304893136, 'learning_rate/qwen_vl_interface': 6.27871075696972e-06, 'learning_rate/action_model': 6.27871075696972e-05, 'epoch': 0.6})
+[2026-07-28 13:35:47,749][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.04119756072759628, 'train/grad_norm_pre_clip': 1.0815770626068115, 'timing/data': 0.0003473539836704731, 'timing/model': 0.41416999604552984, 'learning_rate/qwen_vl_interface': 6.27497493464708e-06, 'learning_rate/action_model': 6.27497493464708e-05, 'epoch': 0.6})
+[2026-07-28 13:35:53,641][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.04208323732018471, 'train/grad_norm_pre_clip': 1.2175114154815674, 'timing/data': 0.0002785148099064827, 'timing/model': 0.41535363625735044, 'learning_rate/qwen_vl_interface': 6.271238447229332e-06, 'learning_rate/action_model': 6.271238447229332e-05, 'epoch': 0.6})
+[2026-07-28 13:35:59,745][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.04058172181248665, 'train/grad_norm_pre_clip': 1.5728815793991089, 'timing/data': 0.0003051459789276123, 'timing/model': 0.5821423409506679, 'learning_rate/qwen_vl_interface': 6.267501297141042e-06, 'learning_rate/action_model': 6.267501297141042e-05, 'epoch': 0.6})
+[2026-07-28 13:36:05,667][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.051656439900398254, 'train/grad_norm_pre_clip': 0.7874777913093567, 'timing/data': 0.00037648435682058334, 'timing/model': 0.41620402224361897, 'learning_rate/qwen_vl_interface': 6.263763486807208e-06, 'learning_rate/action_model': 6.263763486807208e-05, 'epoch': 0.6})
+[2026-07-28 13:36:11,589][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.034305084496736526, 'train/grad_norm_pre_clip': 0.7888147234916687, 'timing/data': 0.00028386712074279785, 'timing/model': 0.41827390482649207, 'learning_rate/qwen_vl_interface': 6.260025018653253e-06, 'learning_rate/action_model': 6.260025018653253e-05, 'epoch': 0.6})
+[2026-07-28 13:36:17,737][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.05388621985912323, 'train/grad_norm_pre_clip': 1.2423895597457886, 'timing/data': 0.00028156209737062454, 'timing/model': 0.41294980980455875, 'learning_rate/qwen_vl_interface': 6.2562858951050294e-06, 'learning_rate/action_model': 6.256285895105029e-05, 'epoch': 0.6})
+[2026-07-28 13:36:23,910][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.03611842915415764, 'train/grad_norm_pre_clip': 1.2728277444839478, 'timing/data': 0.0005875052884221077, 'timing/model': 0.4148020097054541, 'learning_rate/qwen_vl_interface': 6.252546118588814e-06, 'learning_rate/action_model': 6.252546118588813e-05, 'epoch': 0.6})
+[2026-07-28 13:36:29,817][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.060643114149570465, 'train/grad_norm_pre_clip': 1.5151394605636597, 'timing/data': 0.00027085887268185616, 'timing/model': 0.4203815539367497, 'learning_rate/qwen_vl_interface': 6.248805691531306e-06, 'learning_rate/action_model': 6.248805691531307e-05, 'epoch': 0.6})
+[2026-07-28 13:36:35,739][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.04617879539728165, 'train/grad_norm_pre_clip': 0.9876746535301208, 'timing/data': 0.00029376987367868423, 'timing/model': 0.41850742185488343, 'learning_rate/qwen_vl_interface': 6.245064616359628e-06, 'learning_rate/action_model': 6.245064616359628e-05, 'epoch': 0.6})
+[2026-07-28 13:36:41,873][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.06016124412417412, 'train/grad_norm_pre_clip': 2.4399755001068115, 'timing/data': 0.00039091380313038826, 'timing/model': 0.6670737648382783, 'learning_rate/qwen_vl_interface': 6.241322895501325e-06, 'learning_rate/action_model': 6.241322895501325e-05, 'epoch': 0.6})
+[2026-07-28 13:36:48,021][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.043655380606651306, 'train/grad_norm_pre_clip': 2.241117000579834, 'timing/data': 0.00025962945073843, 'timing/model': 0.6620573974214494, 'learning_rate/qwen_vl_interface': 6.2375805313843564e-06, 'learning_rate/action_model': 6.237580531384356e-05, 'epoch': 0.6})
+[2026-07-28 13:36:54,008][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.028557289391756058, 'train/grad_norm_pre_clip': 0.91157066822052, 'timing/data': 0.0003408808261156082, 'timing/model': 0.42663602298125625, 'learning_rate/qwen_vl_interface': 6.233837526437106e-06, 'learning_rate/action_model': 6.233837526437105e-05, 'epoch': 0.6})
+[2026-07-28 13:36:59,953][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.04953031241893768, 'train/grad_norm_pre_clip': 2.099672317504883, 'timing/data': 0.0003729867748916149, 'timing/model': 0.42277424316853285, 'learning_rate/qwen_vl_interface': 6.230093883088364e-06, 'learning_rate/action_model': 6.230093883088364e-05, 'epoch': 0.6})
+[2026-07-28 13:37:06,190][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.044135864824056625, 'train/grad_norm_pre_clip': 1.3511862754821777, 'timing/data': 0.000338762067258358, 'timing/model': 0.42268294002860785, 'learning_rate/qwen_vl_interface': 6.226349603767341e-06, 'learning_rate/action_model': 6.22634960376734e-05, 'epoch': 0.6})
+[2026-07-28 13:37:12,115][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.04139316827058792, 'train/grad_norm_pre_clip': 1.8903827667236328, 'timing/data': 0.0002854638732969761, 'timing/model': 0.420557270757854, 'learning_rate/qwen_vl_interface': 6.2226046909036616e-06, 'learning_rate/action_model': 6.222604690903662e-05, 'epoch': 0.61})
+[2026-07-28 13:37:18,052][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.041234150528907776, 'train/grad_norm_pre_clip': 1.9508765935897827, 'timing/data': 0.0003664642572402954, 'timing/model': 0.4375937799923122, 'learning_rate/qwen_vl_interface': 6.218859146927358e-06, 'learning_rate/action_model': 6.218859146927358e-05, 'epoch': 0.61})
+[2026-07-28 13:37:24,205][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.03414638340473175, 'train/grad_norm_pre_clip': 0.8660195469856262, 'timing/data': 0.00034310203045606613, 'timing/model': 0.4173809550702572, 'learning_rate/qwen_vl_interface': 6.215112974268871e-06, 'learning_rate/action_model': 6.21511297426887e-05, 'epoch': 0.61})
+[2026-07-28 13:37:30,345][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.037103213369846344, 'train/grad_norm_pre_clip': 0.7343777418136597, 'timing/data': 0.00025761500000953674, 'timing/model': 0.4146859310567379, 'learning_rate/qwen_vl_interface': 6.211366175359055e-06, 'learning_rate/action_model': 6.211366175359054e-05, 'epoch': 0.61})
+[2026-07-28 13:37:36,237][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.023889310657978058, 'train/grad_norm_pre_clip': 0.6922809481620789, 'timing/data': 0.00034526921808719635, 'timing/model': 0.41399409295991063, 'learning_rate/qwen_vl_interface': 6.207618752629164e-06, 'learning_rate/action_model': 6.207618752629163e-05, 'epoch': 0.61})
+[2026-07-28 13:37:42,139][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.04210478067398071, 'train/grad_norm_pre_clip': 1.22763991355896, 'timing/data': 0.00027843331918120384, 'timing/model': 0.42217565700411797, 'learning_rate/qwen_vl_interface': 6.203870708510862e-06, 'learning_rate/action_model': 6.203870708510862e-05, 'epoch': 0.61})
+[2026-07-28 13:37:48,523][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.0324748270213604, 'train/grad_norm_pre_clip': 0.9508435726165771, 'timing/data': 0.0002703568898141384, 'timing/model': 0.6519862306304276, 'learning_rate/qwen_vl_interface': 6.200122045436212e-06, 'learning_rate/action_model': 6.200122045436211e-05, 'epoch': 0.61})
+[2026-07-28 13:37:54,423][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.025834836065769196, 'train/grad_norm_pre_clip': 0.7748063206672668, 'timing/data': 0.00035715289413928986, 'timing/model': 0.41606588708236814, 'learning_rate/qwen_vl_interface': 6.196372765837682e-06, 'learning_rate/action_model': 6.196372765837682e-05, 'epoch': 0.61})
+[2026-07-28 13:38:00,309][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.025449713692069054, 'train/grad_norm_pre_clip': 0.8090121150016785, 'timing/data': 0.00027622003108263016, 'timing/model': 0.4149982170201838, 'learning_rate/qwen_vl_interface': 6.19262287214814e-06, 'learning_rate/action_model': 6.19262287214814e-05, 'epoch': 0.61})
+[2026-07-28 13:38:06,428][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.040384165942668915, 'train/grad_norm_pre_clip': 1.3964719772338867, 'timing/data': 0.00028596632182598114, 'timing/model': 0.6455724681727588, 'learning_rate/qwen_vl_interface': 6.18887236680085e-06, 'learning_rate/action_model': 6.18887236680085e-05, 'epoch': 0.61})
+[2026-07-28 13:38:12,347][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.026838330551981926, 'train/grad_norm_pre_clip': 0.839843213558197, 'timing/data': 0.00028406083583831787, 'timing/model': 0.41987243527546525, 'learning_rate/qwen_vl_interface': 6.185121252229475e-06, 'learning_rate/action_model': 6.185121252229474e-05, 'epoch': 0.61})
+[2026-07-28 13:38:18,468][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.027233392000198364, 'train/grad_norm_pre_clip': 1.204961895942688, 'timing/data': 0.0003748321905732155, 'timing/model': 0.6532340412959456, 'learning_rate/qwen_vl_interface': 6.181369530868071e-06, 'learning_rate/action_model': 6.181369530868071e-05, 'epoch': 0.61})
+[2026-07-28 13:38:24,409][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.06345714628696442, 'train/grad_norm_pre_clip': 2.5759780406951904, 'timing/data': 0.0002569272182881832, 'timing/model': 0.41174840508028865, 'learning_rate/qwen_vl_interface': 6.17761720515109e-06, 'learning_rate/action_model': 6.17761720515109e-05, 'epoch': 0.61})
+[2026-07-28 13:38:30,583][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.09711035341024399, 'train/grad_norm_pre_clip': 2.1308560371398926, 'timing/data': 0.000269819051027298, 'timing/model': 0.41294032614678144, 'learning_rate/qwen_vl_interface': 6.173864277513376e-06, 'learning_rate/action_model': 6.173864277513376e-05, 'epoch': 0.61})
+[2026-07-28 13:38:36,242][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.013618113473057747, 'train/grad_norm_pre_clip': 0.6812916994094849, 'timing/data': 0.0002812780439853668, 'timing/model': 0.41485003288835287, 'learning_rate/qwen_vl_interface': 6.170110750390163e-06, 'learning_rate/action_model': 6.170110750390163e-05, 'epoch': 0.61})
+[2026-07-28 13:38:42,143][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.028551213443279266, 'train/grad_norm_pre_clip': 1.0986757278442383, 'timing/data': 0.0003179740160703659, 'timing/model': 0.41367876110598445, 'learning_rate/qwen_vl_interface': 6.1663566262170725e-06, 'learning_rate/action_model': 6.166356626217072e-05, 'epoch': 0.61})
+[2026-07-28 13:38:48,039][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.028526876121759415, 'train/grad_norm_pre_clip': 1.3106030225753784, 'timing/data': 0.0003467029891908169, 'timing/model': 0.41512974025681615, 'learning_rate/qwen_vl_interface': 6.162601907430116e-06, 'learning_rate/action_model': 6.162601907430116e-05, 'epoch': 0.61})
+[2026-07-28 13:38:54,229][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.022868042811751366, 'train/grad_norm_pre_clip': 0.7648959755897522, 'timing/data': 0.0002712951973080635, 'timing/model': 0.6647669491358101, 'learning_rate/qwen_vl_interface': 6.15884659646569e-06, 'learning_rate/action_model': 6.15884659646569e-05, 'epoch': 0.61})
+[2026-07-28 13:39:00,119][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.028146781027317047, 'train/grad_norm_pre_clip': 1.3284066915512085, 'timing/data': 0.00026090024039149284, 'timing/model': 0.41573110269382596, 'learning_rate/qwen_vl_interface': 6.155090695760574e-06, 'learning_rate/action_model': 6.155090695760573e-05, 'epoch': 0.61})
+[2026-07-28 13:39:05,784][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.046371471136808395, 'train/grad_norm_pre_clip': 0.8068974018096924, 'timing/data': 0.0008490378968417645, 'timing/model': 0.4157944438047707, 'learning_rate/qwen_vl_interface': 6.151334207751933e-06, 'learning_rate/action_model': 6.151334207751932e-05, 'epoch': 0.61})
+[2026-07-28 13:39:11,644][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.01962359808385372, 'train/grad_norm_pre_clip': 0.5988948941230774, 'timing/data': 0.00034778984263539314, 'timing/model': 0.415567415766418, 'learning_rate/qwen_vl_interface': 6.1475771348773115e-06, 'learning_rate/action_model': 6.147577134877312e-05, 'epoch': 0.61})
+[2026-07-28 13:39:17,573][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.022492948919534683, 'train/grad_norm_pre_clip': 1.3270376920700073, 'timing/data': 0.00036926427856087685, 'timing/model': 0.4179655071347952, 'learning_rate/qwen_vl_interface': 6.14381947957463e-06, 'learning_rate/action_model': 6.14381947957463e-05, 'epoch': 0.61})
+[2026-07-28 13:39:23,691][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.044168345630168915, 'train/grad_norm_pre_clip': 1.3021208047866821, 'timing/data': 0.0003573428839445114, 'timing/model': 0.4157977048307657, 'learning_rate/qwen_vl_interface': 6.140061244282194e-06, 'learning_rate/action_model': 6.140061244282194e-05, 'epoch': 0.61})
+[2026-07-28 13:39:29,807][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.02594383992254734, 'train/grad_norm_pre_clip': 0.8104795813560486, 'timing/data': 0.0002810833975672722, 'timing/model': 0.4154856209643185, 'learning_rate/qwen_vl_interface': 6.136302431438677e-06, 'learning_rate/action_model': 6.136302431438677e-05, 'epoch': 0.61})
+[2026-07-28 13:39:35,709][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.0686565488576889, 'train/grad_norm_pre_clip': 2.6435093879699707, 'timing/data': 0.00027476297691464424, 'timing/model': 0.416244403924793, 'learning_rate/qwen_vl_interface': 6.132543043483137e-06, 'learning_rate/action_model': 6.132543043483136e-05, 'epoch': 0.61})
+[2026-07-28 13:39:41,601][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.04458408057689667, 'train/grad_norm_pre_clip': 3.271099328994751, 'timing/data': 0.0003144913353025913, 'timing/model': 0.41712272819131613, 'learning_rate/qwen_vl_interface': 6.1287830828549975e-06, 'learning_rate/action_model': 6.128783082854997e-05, 'epoch': 0.61})
+[2026-07-28 13:39:47,822][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.030013762414455414, 'train/grad_norm_pre_clip': 1.3812792301177979, 'timing/data': 0.0002932320348918438, 'timing/model': 0.4141740743070841, 'learning_rate/qwen_vl_interface': 6.125022551994055e-06, 'learning_rate/action_model': 6.125022551994056e-05, 'epoch': 0.61})
+[2026-07-28 13:39:54,061][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.09896882623434067, 'train/grad_norm_pre_clip': 2.5182976722717285, 'timing/data': 0.0002846098504960537, 'timing/model': 0.41472153179347515, 'learning_rate/qwen_vl_interface': 6.121261453340481e-06, 'learning_rate/action_model': 6.12126145334048e-05, 'epoch': 0.61})
+[2026-07-28 13:39:59,992][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.05437048524618149, 'train/grad_norm_pre_clip': 1.8433009386062622, 'timing/data': 0.0002686460502445698, 'timing/model': 0.4150366773828864, 'learning_rate/qwen_vl_interface': 6.117499789334806e-06, 'learning_rate/action_model': 6.117499789334807e-05, 'epoch': 0.61})
+[2026-07-28 13:40:05,953][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.042485933750867844, 'train/grad_norm_pre_clip': 2.3769915103912354, 'timing/data': 0.00035223690792918205, 'timing/model': 0.4158306112512946, 'learning_rate/qwen_vl_interface': 6.113737562417939e-06, 'learning_rate/action_model': 6.113737562417938e-05, 'epoch': 0.61})
+[2026-07-28 13:40:12,086][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.030892567709088326, 'train/grad_norm_pre_clip': 0.9416208267211914, 'timing/data': 0.0002612420357763767, 'timing/model': 0.6582003249786794, 'learning_rate/qwen_vl_interface': 6.109974775031144e-06, 'learning_rate/action_model': 6.109974775031144e-05, 'epoch': 0.61})
+[2026-07-28 13:40:18,239][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.03351914882659912, 'train/grad_norm_pre_clip': 1.1157474517822266, 'timing/data': 0.0002843886613845825, 'timing/model': 0.6557220509275794, 'learning_rate/qwen_vl_interface': 6.106211429616054e-06, 'learning_rate/action_model': 6.106211429616053e-05, 'epoch': 0.61})
+[2026-07-28 13:40:24,240][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.02092842385172844, 'train/grad_norm_pre_clip': 0.9375960826873779, 'timing/data': 0.0002597901038825512, 'timing/model': 0.41474678879603744, 'learning_rate/qwen_vl_interface': 6.102447528614662e-06, 'learning_rate/action_model': 6.102447528614662e-05, 'epoch': 0.61})
+[2026-07-28 13:40:30,146][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.033159609884023666, 'train/grad_norm_pre_clip': 1.3792301416397095, 'timing/data': 0.00028602685779333115, 'timing/model': 0.41391072887927294, 'learning_rate/qwen_vl_interface': 6.098683074469325e-06, 'learning_rate/action_model': 6.0986830744693246e-05, 'epoch': 0.61})
+[2026-07-28 13:40:36,041][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.06025698781013489, 'train/grad_norm_pre_clip': 1.9194170236587524, 'timing/data': 0.00028726691380143166, 'timing/model': 0.416039383970201, 'learning_rate/qwen_vl_interface': 6.094918069622754e-06, 'learning_rate/action_model': 6.0949180696227534e-05, 'epoch': 0.61})
+[2026-07-28 13:40:42,138][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.03395818918943405, 'train/grad_norm_pre_clip': 0.817135214805603, 'timing/data': 0.0003326199948787689, 'timing/model': 0.41513719176873565, 'learning_rate/qwen_vl_interface': 6.091152516518022e-06, 'learning_rate/action_model': 6.0911525165180216e-05, 'epoch': 0.61})
+[2026-07-28 13:40:48,048][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.021931761875748634, 'train/grad_norm_pre_clip': 0.5999876856803894, 'timing/data': 0.0002646828070282936, 'timing/model': 0.41430786484852433, 'learning_rate/qwen_vl_interface': 6.087386417598554e-06, 'learning_rate/action_model': 6.0873864175985537e-05, 'epoch': 0.61})
+[2026-07-28 13:40:53,932][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.04250013455748558, 'train/grad_norm_pre_clip': 2.349393844604492, 'timing/data': 0.0002841521054506302, 'timing/model': 0.41444841399788857, 'learning_rate/qwen_vl_interface': 6.083619775308129e-06, 'learning_rate/action_model': 6.083619775308129e-05, 'epoch': 0.61})
+[2026-07-28 13:41:00,154][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.05549392104148865, 'train/grad_norm_pre_clip': 1.6726322174072266, 'timing/data': 0.00025049690157175064, 'timing/model': 0.4133581230416894, 'learning_rate/qwen_vl_interface': 6.079852592090885e-06, 'learning_rate/action_model': 6.079852592090884e-05, 'epoch': 0.61})
+[2026-07-28 13:41:06,083][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.03497153893113136, 'train/grad_norm_pre_clip': 1.207531213760376, 'timing/data': 0.0002701752819120884, 'timing/model': 0.41449096892029047, 'learning_rate/qwen_vl_interface': 6.076084870391303e-06, 'learning_rate/action_model': 6.076084870391302e-05, 'epoch': 0.61})
+[2026-07-28 13:41:12,189][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.04257760941982269, 'train/grad_norm_pre_clip': 1.139710783958435, 'timing/data': 0.0003536790609359741, 'timing/model': 0.657909631729126, 'learning_rate/qwen_vl_interface': 6.072316612654215e-06, 'learning_rate/action_model': 6.072316612654215e-05, 'epoch': 0.61})
+[2026-07-28 13:41:18,097][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.02726449817419052, 'train/grad_norm_pre_clip': 1.3928723335266113, 'timing/data': 0.00028152018785476685, 'timing/model': 0.43251531897112727, 'learning_rate/qwen_vl_interface': 6.068547821324808e-06, 'learning_rate/action_model': 6.068547821324807e-05, 'epoch': 0.61})
+[2026-07-28 13:41:24,349][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.01843465305864811, 'train/grad_norm_pre_clip': 0.8307210803031921, 'timing/data': 0.0002839593216776848, 'timing/model': 0.654097044840455, 'learning_rate/qwen_vl_interface': 6.064778498848605e-06, 'learning_rate/action_model': 6.064778498848604e-05, 'epoch': 0.61})
+[2026-07-28 13:41:30,234][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.03519831597805023, 'train/grad_norm_pre_clip': 1.1633504629135132, 'timing/data': 0.0003017028793692589, 'timing/model': 0.41487943287938833, 'learning_rate/qwen_vl_interface': 6.061008647671479e-06, 'learning_rate/action_model': 6.061008647671479e-05, 'epoch': 0.61})
+[2026-07-28 13:41:36,088][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.04375804588198662, 'train/grad_norm_pre_clip': 1.144805669784546, 'timing/data': 0.0003572702407836914, 'timing/model': 0.4152867761440575, 'learning_rate/qwen_vl_interface': 6.05723827023965e-06, 'learning_rate/action_model': 6.057238270239649e-05, 'epoch': 0.61})
+[2026-07-28 13:41:42,238][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.03679352253675461, 'train/grad_norm_pre_clip': 1.187863826751709, 'timing/data': 0.0002958569675683975, 'timing/model': 0.4151768363080919, 'learning_rate/qwen_vl_interface': 6.053467368999667e-06, 'learning_rate/action_model': 6.0534673689996665e-05, 'epoch': 0.61})
+[2026-07-28 13:41:48,399][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.03846439719200134, 'train/grad_norm_pre_clip': 1.643810749053955, 'timing/data': 0.0007510152645409107, 'timing/model': 0.41466790018603206, 'learning_rate/qwen_vl_interface': 6.049695946398433e-06, 'learning_rate/action_model': 6.049695946398433e-05, 'epoch': 0.61})
+[2026-07-28 13:41:54,323][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.045067161321640015, 'train/grad_norm_pre_clip': 1.1565238237380981, 'timing/data': 0.00026686815544962883, 'timing/model': 0.415643772110343, 'learning_rate/qwen_vl_interface': 6.045924004883184e-06, 'learning_rate/action_model': 6.045924004883183e-05, 'epoch': 0.61})
+[2026-07-28 13:42:00,421][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.01796102523803711, 'train/grad_norm_pre_clip': 1.1882295608520508, 'timing/data': 0.0002887547016143799, 'timing/model': 0.41471376875415444, 'learning_rate/qwen_vl_interface': 6.042151546901487e-06, 'learning_rate/action_model': 6.042151546901487e-05, 'epoch': 0.61})
+[2026-07-28 13:42:06,309][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.03184967860579491, 'train/grad_norm_pre_clip': 1.25493323802948, 'timing/data': 0.0003838632255792618, 'timing/model': 0.4166245670057833, 'learning_rate/qwen_vl_interface': 6.038378574901252e-06, 'learning_rate/action_model': 6.0383785749012514e-05, 'epoch': 0.61})
+[2026-07-28 13:42:12,405][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.05397633835673332, 'train/grad_norm_pre_clip': 1.4417015314102173, 'timing/data': 0.0002842438407242298, 'timing/model': 0.41347510274499655, 'learning_rate/qwen_vl_interface': 6.0346050913307195e-06, 'learning_rate/action_model': 6.0346050913307195e-05, 'epoch': 0.61})
+[2026-07-28 13:42:18,474][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.022932730615139008, 'train/grad_norm_pre_clip': 0.8692978620529175, 'timing/data': 0.0002629500813782215, 'timing/model': 0.5914550782181323, 'learning_rate/qwen_vl_interface': 6.03083109863846e-06, 'learning_rate/action_model': 6.03083109863846e-05, 'epoch': 0.61})
+[2026-07-28 15:13:47,864][starVLA.model.framework.share_tools][INFO] - [*] [apply_config_compat] normalised config from version_id=0.21 to '0.21'
+[2026-07-28 15:13:47,865][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 15:13:47,866][starVLA.training.train_starvla][INFO] - VLA Training :: Warming Up
+[2026-07-28 15:13:48,092][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 15:13:48,092][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 15:13:48,092][huggingface_hub._login][WARNING] - Note: Environment variable`HF_TOKEN` is set and is the current active token independently from the token you've just configured.
+[2026-07-28 15:13:48,093][starVLA.training.rl_games.auth][INFO] - Logged into Hugging Face Hub
+[2026-07-28 15:13:48,691][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 15:13:48,706][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 15:13:48,712][starVLA.training.rl_games.auth][INFO] - Logged into Weights & Biases
+[2026-07-28 15:13:48,728][starVLA.training.train_starvla][INFO] - ✅ Configuration wrapped for access tracking
+[2026-07-28 15:13:51,230][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 15:13:51,231][starVLA.training.train_starvla][INFO] - Creating VLA Dataset with Mixture `flappy_train__bridge`
+[2026-07-28 15:13:52,192][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 15:13:52,193][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 15:13:52,193][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 15:13:52,193][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 15:13:52,194][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 15:13:52,194][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 15:13:52,194][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 15:13:52,195][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 15:13:52,195][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 15:13:52,195][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 15:13:52,195][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 15:13:52,195][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 15:13:52,195][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 15:13:52,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 15:13:52,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 15:13:52,197][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 15:13:52,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 15:13:52,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 15:13:52,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 15:13:52,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 15:13:52,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 15:13:52,198][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 15:13:52,199][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 15:13:52,199][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 15:13:52,199][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 15:13:52,200][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 15:13:52,200][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 15:13:52,200][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 15:13:52,201][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 15:13:52,201][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 15:13:52,201][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 15:13:52,259][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Behavior (data_config): ['BEHAVIOR_challenge']
+[2026-07-28 15:13:52,260][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from DOMINO: ['robotwin']
+[2026-07-28 15:13:52,260][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from DOMINO (data_config): []
+[2026-07-28 15:13:52,260][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from DOMINO (data_config): ['domino', 'domino_clean', 'domino_random', 'domino_cotrain']
+[2026-07-28 15:13:52,261][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Franka: ['custom_robot_config', 'demo_sim_franka_delta_joints', 'SO101']
+[2026-07-28 15:13:52,261][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Franka (data_config): []
+[2026-07-28 15:13:52,261][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Franka (data_config): ['custom_dataset', 'custom_dataset_2', 'demo_sim_pick_place', 'SO101_pick']
+[2026-07-28 15:13:52,261][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from LIBERO: ['libero_franka']
+[2026-07-28 15:13:52,261][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from LIBERO (data_config): []
+[2026-07-28 15:13:52,261][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from LIBERO (data_config): ['libero_all', 'libero_goal', 'multi_robot']
+[2026-07-28 15:13:52,262][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from RoboChallenge_table30v2: ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 15:13:52,262][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from RoboChallenge_table30v2 (data_config): ['ur5_robochallenge', 'arx5_robochallenge', 'dosw1_robochallenge']
+[2026-07-28 15:13:52,262][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from RoboChallenge_table30v2 (data_config): ['robochallenge_table30v2_shred_paper', 'robochallenge_table30v2_ur5_all', 'robochallenge_table30v2_arx5_all', 'robochallenge_table30v2_dosw1_all']
+[2026-07-28 15:13:52,263][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_365: ['panda_omron_robocasa365']
+[2026-07-28 15:13:52,263][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_365 (data_config): []
+[2026-07-28 15:13:52,264][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_365 (data_config): ['robocasa365_open_drawer_target_human', 'robocasa365_atomic_target_human_all', 'robocasa365_composite_target_human_all', 'robocasa365_target_human_all']
+[2026-07-28 15:13:52,264][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robocasa_tabletop: ['fourier_gr1_arms_waist']
+[2026-07-28 15:13:52,264][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robocasa_tabletop (data_config): []
+[2026-07-28 15:13:52,264][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robocasa_tabletop (data_config): ['fourier_gr1_unified_1000']
+[2026-07-28 15:13:52,265][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from Robotwin: ['robotwin', 'robotwin50', 'arx_x5']
+[2026-07-28 15:13:52,265][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from Robotwin (data_config): []
+[2026-07-28 15:13:52,265][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from Robotwin (data_config): ['robotwin_all', 'robotwin_all_50', 'robotwin', 'robotwin_task1', 'robotwin_task2', 'arx_x5']
+[2026-07-28 15:13:52,266][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from SimplerEnv: ['oxe_droid', 'oxe_bridge', 'oxe_rt1']
+[2026-07-28 15:13:52,266][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from SimplerEnv (data_config): []
+[2026-07-28 15:13:52,266][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from SimplerEnv (data_config): ['bridge', 'bridge_rt_1']
+[2026-07-28 15:13:52,267][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from VLA-Arena: ['vla_arena_franka']
+[2026-07-28 15:13:52,267][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from VLA-Arena (data_config): []
+[2026-07-28 15:13:52,267][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from VLA-Arena (data_config): ['vla_arena_L0_S', 'vla_arena_L0_M', 'vla_arena_L0_L']
+[2026-07-28 15:13:52,267][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded data_config from rl_games: ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 15:13:52,267][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded embodiment_tags from rl_games (data_config): ['rl_games_flappy', 'rl_games_demon_attack', 'rl_games_deadly_corridor']
+[2026-07-28 15:13:52,267][starVLA.dataloader.gr00t_lerobot.registry][INFO] - [registry] Loaded mixtures from rl_games (data_config): ['flappy_train', 'flappy_train__bridge', 'flappy_mixed_latency_train', 'flappy_mixed_latency_train__bridge', 'demon_attack_train', 'demon_attack_train__bridge', 'demon_attack_mixed_latency_train', 'demon_attack_mixed_latency_train__bridge', 'deadly_corridor_train', 'deadly_corridor_train__bridge', 'deadly_corridor_mixed_latency_train', 'deadly_corridor_mixed_latency_train__bridge']
+[2026-07-28 15:13:53,532][starVLA.dataloader][INFO] - Created train VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 15:13:53,535][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 15:13:53,542][starVLA.training.train_starvla][INFO] - Creating VLA Eval Dataset with Mixture `flappy_train__bridge__val`
+[2026-07-28 15:13:53,558][starVLA.dataloader][INFO] - Created eval VLA dataloader with num_workers=8, persistent_workers=True, cpu_only_workers=True
+[2026-07-28 15:13:53,565][starVLA.training.train_starvla][INFO] - LR Group qwen_vl_interface: lr=1e-05, num_params=713
+[2026-07-28 15:13:53,565][starVLA.training.train_starvla][INFO] - LR Group action_model: lr=0.0001, num_params=16
+[2026-07-28 15:13:54,951][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state, steps: 3000
+[2026-07-28 15:13:54,963][starVLA.training.train_starvla][INFO] - 📝 Full config saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/config.full.yaml
+[2026-07-28 15:13:54,976][starVLA.training.train_starvla][INFO] - 📊 Accessed config snapshot saved at /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/config.yaml
+[2026-07-28 15:13:54,976][starVLA.training.train_starvla][INFO] - Will resume full training state from checkpoint: /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state, steps: 3000
+[2026-07-28 15:13:54,981][accelerate.accelerator][INFO] - Since you passed both train and evaluation dataloader, `is_train_batch_min` (here True will decide the `train_batch_size` (8).
+[2026-07-28 15:14:30,768][accelerate.accelerator][INFO] - Loading states from /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state
+[2026-07-28 15:14:30,768][accelerate.accelerator][INFO] - Loading DeepSpeed Model and Optimizer
+[2026-07-28 15:15:03,581][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer loaded from input dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3000_state/pytorch_model
+[2026-07-28 15:15:03,582][accelerate.checkpointing][INFO] - All model weights loaded successfully
+[2026-07-28 15:15:03,582][accelerate.checkpointing][INFO] - All optimizer states loaded successfully
+[2026-07-28 15:15:03,582][accelerate.checkpointing][INFO] - All scheduler states loaded successfully
+[2026-07-28 15:15:03,582][accelerate.checkpointing][INFO] - All dataloader sampler states loaded successfully
+[2026-07-28 15:15:03,590][accelerate.checkpointing][INFO] - All random states loaded successfully
+[2026-07-28 15:15:03,592][accelerate.accelerator][INFO] - Loading in 0 custom states
+[2026-07-28 15:15:06,565][starVLA.training.train_starvla][INFO] - ***** Training Configuration *****
+[2026-07-28 15:15:06,565][starVLA.training.train_starvla][INFO] -   Total optimization steps = 4000
+[2026-07-28 15:15:06,565][starVLA.training.train_starvla][INFO] -   Per device batch size = 8
+[2026-07-28 15:15:06,566][starVLA.training.train_starvla][INFO] -   Gradient accumulation steps = 16
+[2026-07-28 15:15:06,566][starVLA.training.train_starvla][INFO] -   Total batch size = 256
+[2026-07-28 15:16:24,616][starVLA.training.train_starvla][INFO] - Step 3001, Loss: {'train/loss': 0.027354031801223755, 'train/grad_norm_pre_clip': 1.0200598239898682, 'timing/data': 0.0003129360266029835, 'timing/model': 0.6100591910071671, 'learning_rate/qwen_vl_interface': 1.0000000000000001e-07, 'learning_rate/action_model': 1.0000000000000002e-06, 'epoch': 0.55})
+[2026-07-28 15:16:29,027][starVLA.training.train_starvla][INFO] - Step 3002, Loss: {'train/loss': 0.08026114106178284, 'train/grad_norm_pre_clip': 1.1515222787857056, 'timing/data': 0.000331277959048748, 'timing/model': 0.41270648082718253, 'learning_rate/qwen_vl_interface': 2.0000000000000002e-07, 'learning_rate/action_model': 2.0000000000000003e-06, 'epoch': 0.55})
+[2026-07-28 15:16:33,265][starVLA.training.train_starvla][INFO] - Step 3003, Loss: {'train/loss': 0.03763314336538315, 'train/grad_norm_pre_clip': 1.3158577680587769, 'timing/data': 0.0003901529125869274, 'timing/model': 0.35155627597123384, 'learning_rate/qwen_vl_interface': 3.0000000000000004e-07, 'learning_rate/action_model': 3e-06, 'epoch': 0.55})
+[2026-07-28 15:16:37,705][starVLA.training.train_starvla][INFO] - Step 3004, Loss: {'train/loss': 0.032343558967113495, 'train/grad_norm_pre_clip': 1.6218138933181763, 'timing/data': 0.0007729819044470787, 'timing/model': 0.3413374163210392, 'learning_rate/qwen_vl_interface': 4.0000000000000003e-07, 'learning_rate/action_model': 4.000000000000001e-06, 'epoch': 0.55})
+[2026-07-28 15:16:41,850][starVLA.training.train_starvla][INFO] - Step 3005, Loss: {'train/loss': 0.026793399825692177, 'train/grad_norm_pre_clip': 1.2283258438110352, 'timing/data': 0.0006349286995828152, 'timing/model': 0.34489795099943876, 'learning_rate/qwen_vl_interface': 5.000000000000001e-07, 'learning_rate/action_model': 5e-06, 'epoch': 0.55})
+[2026-07-28 15:16:46,180][starVLA.training.train_starvla][INFO] - Step 3006, Loss: {'train/loss': 0.06427288055419922, 'train/grad_norm_pre_clip': 1.5148197412490845, 'timing/data': 0.00033980002626776695, 'timing/model': 0.3414873220026493, 'learning_rate/qwen_vl_interface': 6.000000000000001e-07, 'learning_rate/action_model': 6e-06, 'epoch': 0.55})
+[2026-07-28 15:16:50,571][starVLA.training.train_starvla][INFO] - Step 3007, Loss: {'train/loss': 0.018978692591190338, 'train/grad_norm_pre_clip': 0.8080756664276123, 'timing/data': 0.00039857299998402596, 'timing/model': 0.3522010571323335, 'learning_rate/qwen_vl_interface': 7.000000000000001e-07, 'learning_rate/action_model': 7.000000000000001e-06, 'epoch': 0.55})
+[2026-07-28 15:16:54,713][starVLA.training.train_starvla][INFO] - Step 3008, Loss: {'train/loss': 0.028400951996445656, 'train/grad_norm_pre_clip': 1.208213210105896, 'timing/data': 0.00042782723903656006, 'timing/model': 0.334040658082813, 'learning_rate/qwen_vl_interface': 8.000000000000001e-07, 'learning_rate/action_model': 8.000000000000001e-06, 'epoch': 0.55})
+[2026-07-28 15:16:59,156][starVLA.training.train_starvla][INFO] - Step 3009, Loss: {'train/loss': 0.031314607709646225, 'train/grad_norm_pre_clip': 1.343248724937439, 'timing/data': 0.0003643641248345375, 'timing/model': 0.33691520197317004, 'learning_rate/qwen_vl_interface': 9.000000000000001e-07, 'learning_rate/action_model': 9e-06, 'epoch': 0.55})
+[2026-07-28 15:17:03,500][starVLA.training.train_starvla][INFO] - Step 3010, Loss: {'train/loss': 0.026314163580536842, 'train/grad_norm_pre_clip': 0.8472521901130676, 'timing/data': 0.00027750618755817413, 'timing/model': 0.3738831989467144, 'learning_rate/qwen_vl_interface': 1.0000000000000002e-06, 'learning_rate/action_model': 1e-05, 'epoch': 0.55})
+[2026-07-28 15:17:07,684][starVLA.training.train_starvla][INFO] - Step 3011, Loss: {'train/loss': 0.058411434292793274, 'train/grad_norm_pre_clip': 1.5825448036193848, 'timing/data': 0.0003256942145526409, 'timing/model': 0.3345052972435951, 'learning_rate/qwen_vl_interface': 1.1e-06, 'learning_rate/action_model': 1.1000000000000001e-05, 'epoch': 0.55})
+[2026-07-28 15:17:11,888][starVLA.training.train_starvla][INFO] - Step 3012, Loss: {'train/loss': 0.05829497054219246, 'train/grad_norm_pre_clip': 1.357184886932373, 'timing/data': 0.0003746119327843189, 'timing/model': 0.35214931471273303, 'learning_rate/qwen_vl_interface': 1.2000000000000002e-06, 'learning_rate/action_model': 1.2e-05, 'epoch': 0.55})
+[2026-07-28 15:17:16,271][starVLA.training.train_starvla][INFO] - Step 3013, Loss: {'train/loss': 0.02521582879126072, 'train/grad_norm_pre_clip': 1.1032695770263672, 'timing/data': 0.0004220111295580864, 'timing/model': 0.37592414394021034, 'learning_rate/qwen_vl_interface': 1.3e-06, 'learning_rate/action_model': 1.3000000000000001e-05, 'epoch': 0.55})
+[2026-07-28 15:17:20,592][starVLA.training.train_starvla][INFO] - Step 3014, Loss: {'train/loss': 0.06988812237977982, 'train/grad_norm_pre_clip': 2.078523874282837, 'timing/data': 0.00031043728813529015, 'timing/model': 0.34950121166184545, 'learning_rate/qwen_vl_interface': 1.4000000000000001e-06, 'learning_rate/action_model': 1.4000000000000001e-05, 'epoch': 0.55})
+[2026-07-28 15:17:24,808][starVLA.training.train_starvla][INFO] - Step 3015, Loss: {'train/loss': 0.05782516300678253, 'train/grad_norm_pre_clip': 1.5845973491668701, 'timing/data': 0.0003577698953449726, 'timing/model': 0.3527017901651561, 'learning_rate/qwen_vl_interface': 1.5e-06, 'learning_rate/action_model': 1.5e-05, 'epoch': 0.55})
+[2026-07-28 15:17:29,130][starVLA.training.train_starvla][INFO] - Step 3016, Loss: {'train/loss': 0.0412653312087059, 'train/grad_norm_pre_clip': 2.5316078662872314, 'timing/data': 0.0003677443601191044, 'timing/model': 0.3479950069449842, 'learning_rate/qwen_vl_interface': 1.6000000000000001e-06, 'learning_rate/action_model': 1.6000000000000003e-05, 'epoch': 0.55})
+[2026-07-28 15:17:33,346][starVLA.training.train_starvla][INFO] - Step 3017, Loss: {'train/loss': 0.05204744637012482, 'train/grad_norm_pre_clip': 1.0967355966567993, 'timing/data': 0.0003700028173625469, 'timing/model': 0.33783615240827203, 'learning_rate/qwen_vl_interface': 1.7000000000000002e-06, 'learning_rate/action_model': 1.7000000000000003e-05, 'epoch': 0.55})
+[2026-07-28 15:17:37,610][starVLA.training.train_starvla][INFO] - Step 3018, Loss: {'train/loss': 0.04392148181796074, 'train/grad_norm_pre_clip': 1.42672860622406, 'timing/data': 0.00042787007987499237, 'timing/model': 0.3560413820669055, 'learning_rate/qwen_vl_interface': 1.8000000000000001e-06, 'learning_rate/action_model': 1.8e-05, 'epoch': 0.55})
+[2026-07-28 15:17:41,859][starVLA.training.train_starvla][INFO] - Step 3019, Loss: {'train/loss': 0.03751590847969055, 'train/grad_norm_pre_clip': 1.0397140979766846, 'timing/data': 0.0003158990293741226, 'timing/model': 0.36375615699216723, 'learning_rate/qwen_vl_interface': 1.9000000000000002e-06, 'learning_rate/action_model': 1.9e-05, 'epoch': 0.55})
+[2026-07-28 15:17:46,059][starVLA.training.train_starvla][INFO] - Step 3020, Loss: {'train/loss': 0.057919394224882126, 'train/grad_norm_pre_clip': 2.619784355163574, 'timing/data': 0.00032801879569888115, 'timing/model': 0.336427157279104, 'learning_rate/qwen_vl_interface': 2.0000000000000003e-06, 'learning_rate/action_model': 2e-05, 'epoch': 0.55})
+[2026-07-28 15:17:50,180][starVLA.training.train_starvla][INFO] - Step 3021, Loss: {'train/loss': 0.07129130512475967, 'train/grad_norm_pre_clip': 2.156181812286377, 'timing/data': 0.0005025286227464676, 'timing/model': 0.3346488969400525, 'learning_rate/qwen_vl_interface': 2.1000000000000002e-06, 'learning_rate/action_model': 2.1e-05, 'epoch': 0.55})
+[2026-07-28 15:17:54,509][starVLA.training.train_starvla][INFO] - Step 3022, Loss: {'train/loss': 0.03452330082654953, 'train/grad_norm_pre_clip': 1.1791822910308838, 'timing/data': 0.0005774619057774544, 'timing/model': 0.35868629813194275, 'learning_rate/qwen_vl_interface': 2.2e-06, 'learning_rate/action_model': 2.2000000000000003e-05, 'epoch': 0.56})
+[2026-07-28 15:17:58,635][starVLA.training.train_starvla][INFO] - Step 3023, Loss: {'train/loss': 0.030919181182980537, 'train/grad_norm_pre_clip': 1.441738486289978, 'timing/data': 0.0003381469286978245, 'timing/model': 0.333779523614794, 'learning_rate/qwen_vl_interface': 2.3000000000000004e-06, 'learning_rate/action_model': 2.3000000000000003e-05, 'epoch': 0.56})
+[2026-07-28 15:18:02,785][starVLA.training.train_starvla][INFO] - Step 3024, Loss: {'train/loss': 0.028802700340747833, 'train/grad_norm_pre_clip': 1.4507861137390137, 'timing/data': 0.00039181066676974297, 'timing/model': 0.3510556509718299, 'learning_rate/qwen_vl_interface': 2.4000000000000003e-06, 'learning_rate/action_model': 2.4e-05, 'epoch': 0.56})
+[2026-07-28 15:18:06,896][starVLA.training.train_starvla][INFO] - Step 3025, Loss: {'train/loss': 0.052901893854141235, 'train/grad_norm_pre_clip': 1.4530274868011475, 'timing/data': 0.0003400677815079689, 'timing/model': 0.3324860529974103, 'learning_rate/qwen_vl_interface': 2.5e-06, 'learning_rate/action_model': 2.5e-05, 'epoch': 0.56})
+[2026-07-28 15:18:11,194][starVLA.training.train_starvla][INFO] - Step 3026, Loss: {'train/loss': 0.0323052741587162, 'train/grad_norm_pre_clip': 0.8407903909683228, 'timing/data': 0.0003605671226978302, 'timing/model': 0.34355568001046777, 'learning_rate/qwen_vl_interface': 2.6e-06, 'learning_rate/action_model': 2.6000000000000002e-05, 'epoch': 0.56})
+[2026-07-28 15:18:15,459][starVLA.training.train_starvla][INFO] - Step 3027, Loss: {'train/loss': 0.0338553786277771, 'train/grad_norm_pre_clip': 0.7617139220237732, 'timing/data': 0.0003079008311033249, 'timing/model': 0.3462179312482476, 'learning_rate/qwen_vl_interface': 2.7000000000000004e-06, 'learning_rate/action_model': 2.7000000000000002e-05, 'epoch': 0.56})
+[2026-07-28 15:18:19,724][starVLA.training.train_starvla][INFO] - Step 3028, Loss: {'train/loss': 0.04734804853796959, 'train/grad_norm_pre_clip': 1.8120291233062744, 'timing/data': 0.00034020887687802315, 'timing/model': 0.3469559117220342, 'learning_rate/qwen_vl_interface': 2.8000000000000003e-06, 'learning_rate/action_model': 2.8000000000000003e-05, 'epoch': 0.56})
+[2026-07-28 15:18:24,094][starVLA.training.train_starvla][INFO] - Step 3029, Loss: {'train/loss': 0.029442520812153816, 'train/grad_norm_pre_clip': 1.4383562803268433, 'timing/data': 0.0003790217451751232, 'timing/model': 0.3340941551141441, 'learning_rate/qwen_vl_interface': 2.9e-06, 'learning_rate/action_model': 2.9e-05, 'epoch': 0.56})
+[2026-07-28 15:18:28,235][starVLA.training.train_starvla][INFO] - Step 3030, Loss: {'train/loss': 0.03726359084248543, 'train/grad_norm_pre_clip': 0.9908667802810669, 'timing/data': 0.00036756694316864014, 'timing/model': 0.3353257020935416, 'learning_rate/qwen_vl_interface': 3e-06, 'learning_rate/action_model': 3e-05, 'epoch': 0.56})
+[2026-07-28 15:18:32,398][starVLA.training.train_starvla][INFO] - Step 3031, Loss: {'train/loss': 0.016094882041215897, 'train/grad_norm_pre_clip': 0.6450479030609131, 'timing/data': 0.000764352735131979, 'timing/model': 0.33312449511140585, 'learning_rate/qwen_vl_interface': 3.1000000000000004e-06, 'learning_rate/action_model': 3.1e-05, 'epoch': 0.56})
+[2026-07-28 15:18:36,734][starVLA.training.train_starvla][INFO] - Step 3032, Loss: {'train/loss': 0.04248952493071556, 'train/grad_norm_pre_clip': 1.0244473218917847, 'timing/data': 0.0003142543137073517, 'timing/model': 0.3865903811529279, 'learning_rate/qwen_vl_interface': 3.2000000000000003e-06, 'learning_rate/action_model': 3.2000000000000005e-05, 'epoch': 0.56})
+[2026-07-28 15:18:40,943][starVLA.training.train_starvla][INFO] - Step 3033, Loss: {'train/loss': 0.056465260684490204, 'train/grad_norm_pre_clip': 1.2794922590255737, 'timing/data': 0.00039536599069833755, 'timing/model': 0.3404720830731094, 'learning_rate/qwen_vl_interface': 3.3000000000000006e-06, 'learning_rate/action_model': 3.3e-05, 'epoch': 0.56})
+[2026-07-28 15:18:45,950][starVLA.training.train_starvla][INFO] - Step 3034, Loss: {'train/loss': 0.0348941795527935, 'train/grad_norm_pre_clip': 1.005871295928955, 'timing/data': 0.0002863421104848385, 'timing/model': 0.3471774165518582, 'learning_rate/qwen_vl_interface': 3.4000000000000005e-06, 'learning_rate/action_model': 3.4000000000000007e-05, 'epoch': 0.56})
+[2026-07-28 15:18:50,209][starVLA.training.train_starvla][INFO] - Step 3035, Loss: {'train/loss': 0.028444644063711166, 'train/grad_norm_pre_clip': 0.7775392532348633, 'timing/data': 0.0006116582080721855, 'timing/model': 0.342525377869606, 'learning_rate/qwen_vl_interface': 3.5e-06, 'learning_rate/action_model': 3.5e-05, 'epoch': 0.56})
+[2026-07-28 15:18:54,565][starVLA.training.train_starvla][INFO] - Step 3036, Loss: {'train/loss': 0.040358997881412506, 'train/grad_norm_pre_clip': 1.0257660150527954, 'timing/data': 0.0003142249770462513, 'timing/model': 0.347604607231915, 'learning_rate/qwen_vl_interface': 3.6000000000000003e-06, 'learning_rate/action_model': 3.6e-05, 'epoch': 0.56})
+[2026-07-28 15:18:58,855][starVLA.training.train_starvla][INFO] - Step 3037, Loss: {'train/loss': 0.021793905645608902, 'train/grad_norm_pre_clip': 0.4345345199108124, 'timing/data': 0.00034038908779621124, 'timing/model': 0.3376971338875592, 'learning_rate/qwen_vl_interface': 3.7e-06, 'learning_rate/action_model': 3.7e-05, 'epoch': 0.56})
+[2026-07-28 15:19:03,347][starVLA.training.train_starvla][INFO] - Step 3038, Loss: {'train/loss': 0.025867627933621407, 'train/grad_norm_pre_clip': 0.519112765789032, 'timing/data': 0.0002870219759643078, 'timing/model': 0.3390890220180154, 'learning_rate/qwen_vl_interface': 3.8000000000000005e-06, 'learning_rate/action_model': 3.8e-05, 'epoch': 0.56})
+[2026-07-28 15:19:07,691][starVLA.training.train_starvla][INFO] - Step 3039, Loss: {'train/loss': 0.060467444360256195, 'train/grad_norm_pre_clip': 1.5916129350662231, 'timing/data': 0.000450722873210907, 'timing/model': 0.3385511660017073, 'learning_rate/qwen_vl_interface': 3.900000000000001e-06, 'learning_rate/action_model': 3.9000000000000006e-05, 'epoch': 0.56})
+[2026-07-28 15:19:11,983][starVLA.training.train_starvla][INFO] - Step 3040, Loss: {'train/loss': 0.029414862394332886, 'train/grad_norm_pre_clip': 1.081102967262268, 'timing/data': 0.00039943819865584373, 'timing/model': 0.345393936149776, 'learning_rate/qwen_vl_interface': 4.000000000000001e-06, 'learning_rate/action_model': 4e-05, 'epoch': 0.56})
+[2026-07-28 15:19:16,283][starVLA.training.train_starvla][INFO] - Step 3041, Loss: {'train/loss': 0.02414080500602722, 'train/grad_norm_pre_clip': 0.7043944001197815, 'timing/data': 0.00034516118466854095, 'timing/model': 0.3604784165509045, 'learning_rate/qwen_vl_interface': 4.1e-06, 'learning_rate/action_model': 4.1e-05, 'epoch': 0.56})
+[2026-07-28 15:19:20,656][starVLA.training.train_starvla][INFO] - Step 3042, Loss: {'train/loss': 0.047828011214733124, 'train/grad_norm_pre_clip': 1.983094334602356, 'timing/data': 0.00041649723425507545, 'timing/model': 0.34353716718032956, 'learning_rate/qwen_vl_interface': 4.2000000000000004e-06, 'learning_rate/action_model': 4.2e-05, 'epoch': 0.56})
+[2026-07-28 15:19:24,849][starVLA.training.train_starvla][INFO] - Step 3043, Loss: {'train/loss': 0.032000426203012466, 'train/grad_norm_pre_clip': 1.0962737798690796, 'timing/data': 0.0004164399579167366, 'timing/model': 0.35191742004826665, 'learning_rate/qwen_vl_interface': 4.3e-06, 'learning_rate/action_model': 4.3e-05, 'epoch': 0.56})
+[2026-07-28 15:19:29,033][starVLA.training.train_starvla][INFO] - Step 3044, Loss: {'train/loss': 0.028508242219686508, 'train/grad_norm_pre_clip': 1.6498405933380127, 'timing/data': 0.0003852169029414654, 'timing/model': 0.33606366999447346, 'learning_rate/qwen_vl_interface': 4.4e-06, 'learning_rate/action_model': 4.4000000000000006e-05, 'epoch': 0.56})
+[2026-07-28 15:19:33,291][starVLA.training.train_starvla][INFO] - Step 3045, Loss: {'train/loss': 0.025088181719183922, 'train/grad_norm_pre_clip': 0.696202278137207, 'timing/data': 0.0002994062379002571, 'timing/model': 0.34417030587792397, 'learning_rate/qwen_vl_interface': 4.5e-06, 'learning_rate/action_model': 4.5e-05, 'epoch': 0.56})
+[2026-07-28 15:19:37,583][starVLA.training.train_starvla][INFO] - Step 3046, Loss: {'train/loss': 0.05503307655453682, 'train/grad_norm_pre_clip': 4.254855632781982, 'timing/data': 0.0003128969110548496, 'timing/model': 0.3432704503647983, 'learning_rate/qwen_vl_interface': 4.600000000000001e-06, 'learning_rate/action_model': 4.600000000000001e-05, 'epoch': 0.56})
+[2026-07-28 15:19:41,966][starVLA.training.train_starvla][INFO] - Step 3047, Loss: {'train/loss': 0.086920827627182, 'train/grad_norm_pre_clip': 3.7824361324310303, 'timing/data': 0.00032182689756155014, 'timing/model': 0.33916673297062516, 'learning_rate/qwen_vl_interface': 4.7e-06, 'learning_rate/action_model': 4.7e-05, 'epoch': 0.56})
+[2026-07-28 15:19:46,154][starVLA.training.train_starvla][INFO] - Step 3048, Loss: {'train/loss': 0.04838739335536957, 'train/grad_norm_pre_clip': 2.5867977142333984, 'timing/data': 0.00035886093974113464, 'timing/model': 0.336977431550622, 'learning_rate/qwen_vl_interface': 4.800000000000001e-06, 'learning_rate/action_model': 4.8e-05, 'epoch': 0.56})
+[2026-07-28 15:19:50,368][starVLA.training.train_starvla][INFO] - Step 3049, Loss: {'train/loss': 0.0445413738489151, 'train/grad_norm_pre_clip': 1.6059083938598633, 'timing/data': 0.0003461088053882122, 'timing/model': 0.34508083993569016, 'learning_rate/qwen_vl_interface': 4.9000000000000005e-06, 'learning_rate/action_model': 4.9e-05, 'epoch': 0.56})
+[2026-07-28 15:19:54,648][starVLA.training.train_starvla][INFO] - Step 3050, Loss: {'train/loss': 0.021782152354717255, 'train/grad_norm_pre_clip': 0.6297441720962524, 'timing/data': 0.00038514239713549614, 'timing/model': 0.33805404230952263, 'learning_rate/qwen_vl_interface': 5e-06, 'learning_rate/action_model': 5e-05, 'epoch': 0.56})
+[2026-07-28 15:19:58,842][starVLA.training.train_starvla][INFO] - Step 3051, Loss: {'train/loss': 0.07134328037500381, 'train/grad_norm_pre_clip': 1.801712155342102, 'timing/data': 0.00032684020698070526, 'timing/model': 0.34764633467420936, 'learning_rate/qwen_vl_interface': 5.1e-06, 'learning_rate/action_model': 5.1000000000000006e-05, 'epoch': 0.56})
+[2026-07-28 15:20:03,109][starVLA.training.train_starvla][INFO] - Step 3052, Loss: {'train/loss': 0.02903807908296585, 'train/grad_norm_pre_clip': 1.5447789430618286, 'timing/data': 0.00031385989859700203, 'timing/model': 0.3636553562246263, 'learning_rate/qwen_vl_interface': 5.2e-06, 'learning_rate/action_model': 5.2000000000000004e-05, 'epoch': 0.56})
+[2026-07-28 15:20:07,313][starVLA.training.train_starvla][INFO] - Step 3053, Loss: {'train/loss': 0.027826158329844475, 'train/grad_norm_pre_clip': 1.5676311254501343, 'timing/data': 0.0003304285928606987, 'timing/model': 0.3398698028177023, 'learning_rate/qwen_vl_interface': 5.300000000000001e-06, 'learning_rate/action_model': 5.300000000000001e-05, 'epoch': 0.56})
+[2026-07-28 15:20:11,581][starVLA.training.train_starvla][INFO] - Step 3054, Loss: {'train/loss': 0.06722244620323181, 'train/grad_norm_pre_clip': 2.5683951377868652, 'timing/data': 0.0003231121227145195, 'timing/model': 0.3382542380131781, 'learning_rate/qwen_vl_interface': 5.400000000000001e-06, 'learning_rate/action_model': 5.4000000000000005e-05, 'epoch': 0.56})
+[2026-07-28 15:20:15,851][starVLA.training.train_starvla][INFO] - Step 3055, Loss: {'train/loss': 0.025067631155252457, 'train/grad_norm_pre_clip': 0.8602940440177917, 'timing/data': 0.00037629902362823486, 'timing/model': 0.34541563456878066, 'learning_rate/qwen_vl_interface': 5.500000000000001e-06, 'learning_rate/action_model': 5.500000000000001e-05, 'epoch': 0.56})
+[2026-07-28 15:20:20,202][starVLA.training.train_starvla][INFO] - Step 3056, Loss: {'train/loss': 0.03269997611641884, 'train/grad_norm_pre_clip': 1.1554909944534302, 'timing/data': 0.00040057767182588577, 'timing/model': 0.37500928016379476, 'learning_rate/qwen_vl_interface': 5.600000000000001e-06, 'learning_rate/action_model': 5.6000000000000006e-05, 'epoch': 0.56})
+[2026-07-28 15:20:24,522][starVLA.training.train_starvla][INFO] - Step 3057, Loss: {'train/loss': 0.01997119002044201, 'train/grad_norm_pre_clip': 0.4563736617565155, 'timing/data': 0.0005058967508375645, 'timing/model': 0.33490284578874707, 'learning_rate/qwen_vl_interface': 5.7e-06, 'learning_rate/action_model': 5.6999999999999996e-05, 'epoch': 0.56})
+[2026-07-28 15:20:28,687][starVLA.training.train_starvla][INFO] - Step 3058, Loss: {'train/loss': 0.051553621888160706, 'train/grad_norm_pre_clip': 1.3644936084747314, 'timing/data': 0.0008218288421630859, 'timing/model': 0.3358917818404734, 'learning_rate/qwen_vl_interface': 5.8e-06, 'learning_rate/action_model': 5.8e-05, 'epoch': 0.56})
+[2026-07-28 15:20:32,821][starVLA.training.train_starvla][INFO] - Step 3059, Loss: {'train/loss': 0.04734151437878609, 'train/grad_norm_pre_clip': 1.5860204696655273, 'timing/data': 0.0003062766045331955, 'timing/model': 0.33850510139018297, 'learning_rate/qwen_vl_interface': 5.9e-06, 'learning_rate/action_model': 5.9e-05, 'epoch': 0.56})
+[2026-07-28 15:20:37,238][starVLA.training.train_starvla][INFO] - Step 3060, Loss: {'train/loss': 0.022308746352791786, 'train/grad_norm_pre_clip': 0.674200177192688, 'timing/data': 0.00029994500800967216, 'timing/model': 0.34263376519083977, 'learning_rate/qwen_vl_interface': 6e-06, 'learning_rate/action_model': 6e-05, 'epoch': 0.56})
+[2026-07-28 15:20:41,433][starVLA.training.train_starvla][INFO] - Step 3061, Loss: {'train/loss': 0.03877938538789749, 'train/grad_norm_pre_clip': 1.159766674041748, 'timing/data': 0.0004215710796415806, 'timing/model': 0.33489344408735633, 'learning_rate/qwen_vl_interface': 6.1e-06, 'learning_rate/action_model': 6.1e-05, 'epoch': 0.56})
+[2026-07-28 15:20:45,617][starVLA.training.train_starvla][INFO] - Step 3062, Loss: {'train/loss': 0.019792869687080383, 'train/grad_norm_pre_clip': 0.6841102838516235, 'timing/data': 0.00038019707426428795, 'timing/model': 0.3371962383389473, 'learning_rate/qwen_vl_interface': 6.200000000000001e-06, 'learning_rate/action_model': 6.2e-05, 'epoch': 0.56})
+[2026-07-28 15:20:49,866][starVLA.training.train_starvla][INFO] - Step 3063, Loss: {'train/loss': 0.015130916610360146, 'train/grad_norm_pre_clip': 0.8510429859161377, 'timing/data': 0.0003678719513118267, 'timing/model': 0.33950555929914117, 'learning_rate/qwen_vl_interface': 6.300000000000001e-06, 'learning_rate/action_model': 6.3e-05, 'epoch': 0.56})
+[2026-07-28 15:20:54,049][starVLA.training.train_starvla][INFO] - Step 3064, Loss: {'train/loss': 0.03025868907570839, 'train/grad_norm_pre_clip': 1.0737171173095703, 'timing/data': 0.0004257829859852791, 'timing/model': 0.3451946503482759, 'learning_rate/qwen_vl_interface': 6.4000000000000006e-06, 'learning_rate/action_model': 6.400000000000001e-05, 'epoch': 0.56})
+[2026-07-28 15:20:58,171][starVLA.training.train_starvla][INFO] - Step 3065, Loss: {'train/loss': 0.033837996423244476, 'train/grad_norm_pre_clip': 1.1940416097640991, 'timing/data': 0.00030662817880511284, 'timing/model': 0.3350918600335717, 'learning_rate/qwen_vl_interface': 6.5000000000000004e-06, 'learning_rate/action_model': 6.500000000000001e-05, 'epoch': 0.56})
+[2026-07-28 15:21:02,332][starVLA.training.train_starvla][INFO] - Step 3066, Loss: {'train/loss': 0.03205585479736328, 'train/grad_norm_pre_clip': 1.5817885398864746, 'timing/data': 0.00037419330328702927, 'timing/model': 0.33627650374546647, 'learning_rate/qwen_vl_interface': 6.600000000000001e-06, 'learning_rate/action_model': 6.6e-05, 'epoch': 0.56})
+[2026-07-28 15:21:06,509][starVLA.training.train_starvla][INFO] - Step 3067, Loss: {'train/loss': 0.01874621957540512, 'train/grad_norm_pre_clip': 1.1039444208145142, 'timing/data': 0.00039552198722958565, 'timing/model': 0.3463640338741243, 'learning_rate/qwen_vl_interface': 6.700000000000001e-06, 'learning_rate/action_model': 6.7e-05, 'epoch': 0.56})
+[2026-07-28 15:21:10,651][starVLA.training.train_starvla][INFO] - Step 3068, Loss: {'train/loss': 0.0234049241989851, 'train/grad_norm_pre_clip': 1.805471420288086, 'timing/data': 0.0003581368364393711, 'timing/model': 0.33611229807138443, 'learning_rate/qwen_vl_interface': 6.800000000000001e-06, 'learning_rate/action_model': 6.800000000000001e-05, 'epoch': 0.56})
+[2026-07-28 15:21:14,779][starVLA.training.train_starvla][INFO] - Step 3069, Loss: {'train/loss': 0.018260546028614044, 'train/grad_norm_pre_clip': 0.9364317655563354, 'timing/data': 0.0003593210130929947, 'timing/model': 0.34338451735675335, 'learning_rate/qwen_vl_interface': 6.9e-06, 'learning_rate/action_model': 6.9e-05, 'epoch': 0.56})
+[2026-07-28 15:21:19,168][starVLA.training.train_starvla][INFO] - Step 3070, Loss: {'train/loss': 0.06320022791624069, 'train/grad_norm_pre_clip': 0.9044634103775024, 'timing/data': 0.00030381977558135986, 'timing/model': 0.35791044728830457, 'learning_rate/qwen_vl_interface': 7e-06, 'learning_rate/action_model': 7e-05, 'epoch': 0.56})
+[2026-07-28 15:21:23,411][starVLA.training.train_starvla][INFO] - Step 3071, Loss: {'train/loss': 0.03584523871541023, 'train/grad_norm_pre_clip': 1.0930997133255005, 'timing/data': 0.00032105622813105583, 'timing/model': 0.3392961681820452, 'learning_rate/qwen_vl_interface': 7.100000000000001e-06, 'learning_rate/action_model': 7.1e-05, 'epoch': 0.56})
+[2026-07-28 15:21:27,637][starVLA.training.train_starvla][INFO] - Step 3072, Loss: {'train/loss': 0.08169412612915039, 'train/grad_norm_pre_clip': 1.8245482444763184, 'timing/data': 0.0002982630394399166, 'timing/model': 0.3465609438717365, 'learning_rate/qwen_vl_interface': 7.2000000000000005e-06, 'learning_rate/action_model': 7.2e-05, 'epoch': 0.56})
+[2026-07-28 15:21:31,927][starVLA.training.train_starvla][INFO] - Step 3073, Loss: {'train/loss': 0.06952392309904099, 'train/grad_norm_pre_clip': 2.5543437004089355, 'timing/data': 0.00037899985909461975, 'timing/model': 0.338721658103168, 'learning_rate/qwen_vl_interface': 7.3e-06, 'learning_rate/action_model': 7.3e-05, 'epoch': 0.56})
+[2026-07-28 15:21:36,208][starVLA.training.train_starvla][INFO] - Step 3074, Loss: {'train/loss': 0.03787313401699066, 'train/grad_norm_pre_clip': 1.2639057636260986, 'timing/data': 0.0004491237923502922, 'timing/model': 0.3421670510433614, 'learning_rate/qwen_vl_interface': 7.4e-06, 'learning_rate/action_model': 7.4e-05, 'epoch': 0.56})
+[2026-07-28 15:21:40,399][starVLA.training.train_starvla][INFO] - Step 3075, Loss: {'train/loss': 0.0218880046159029, 'train/grad_norm_pre_clip': 0.7489914894104004, 'timing/data': 0.000354803167283535, 'timing/model': 0.3393992097117007, 'learning_rate/qwen_vl_interface': 7.500000000000001e-06, 'learning_rate/action_model': 7.500000000000001e-05, 'epoch': 0.56})
+[2026-07-28 15:21:44,682][starVLA.training.train_starvla][INFO] - Step 3076, Loss: {'train/loss': 0.06474698334932327, 'train/grad_norm_pre_clip': 1.9351202249526978, 'timing/data': 0.0004939609207212925, 'timing/model': 0.36619484424591064, 'learning_rate/qwen_vl_interface': 7.600000000000001e-06, 'learning_rate/action_model': 7.6e-05, 'epoch': 0.56})
+[2026-07-28 15:21:49,136][starVLA.training.train_starvla][INFO] - Step 3077, Loss: {'train/loss': 0.020231284201145172, 'train/grad_norm_pre_clip': 1.1050522327423096, 'timing/data': 0.0003365287557244301, 'timing/model': 0.3394660307094455, 'learning_rate/qwen_vl_interface': 7.7e-06, 'learning_rate/action_model': 7.7e-05, 'epoch': 0.57})
+[2026-07-28 15:21:53,546][starVLA.training.train_starvla][INFO] - Step 3078, Loss: {'train/loss': 0.0348181426525116, 'train/grad_norm_pre_clip': 1.7354037761688232, 'timing/data': 0.00048337364569306374, 'timing/model': 0.37439217092469335, 'learning_rate/qwen_vl_interface': 7.800000000000002e-06, 'learning_rate/action_model': 7.800000000000001e-05, 'epoch': 0.57})
+[2026-07-28 15:21:57,839][starVLA.training.train_starvla][INFO] - Step 3079, Loss: {'train/loss': 0.020443949848413467, 'train/grad_norm_pre_clip': 0.893182635307312, 'timing/data': 0.0005364380776882172, 'timing/model': 0.37395901093259454, 'learning_rate/qwen_vl_interface': 7.9e-06, 'learning_rate/action_model': 7.900000000000001e-05, 'epoch': 0.57})
+[2026-07-28 15:22:02,139][starVLA.training.train_starvla][INFO] - Step 3080, Loss: {'train/loss': 0.06547756493091583, 'train/grad_norm_pre_clip': 1.5968369245529175, 'timing/data': 0.0005252971313893795, 'timing/model': 0.3404181329533458, 'learning_rate/qwen_vl_interface': 8.000000000000001e-06, 'learning_rate/action_model': 8e-05, 'epoch': 0.57})
+[2026-07-28 15:22:06,425][starVLA.training.train_starvla][INFO] - Step 3081, Loss: {'train/loss': 0.05222329497337341, 'train/grad_norm_pre_clip': 2.806328058242798, 'timing/data': 0.00028929905965924263, 'timing/model': 0.350184952840209, 'learning_rate/qwen_vl_interface': 8.1e-06, 'learning_rate/action_model': 8.1e-05, 'epoch': 0.57})
+[2026-07-28 15:22:11,019][starVLA.training.train_starvla][INFO] - Step 3082, Loss: {'train/loss': 0.026749255135655403, 'train/grad_norm_pre_clip': 1.2637665271759033, 'timing/data': 0.000356829259544611, 'timing/model': 0.33368168817833066, 'learning_rate/qwen_vl_interface': 8.2e-06, 'learning_rate/action_model': 8.2e-05, 'epoch': 0.57})
+[2026-07-28 15:22:15,215][starVLA.training.train_starvla][INFO] - Step 3083, Loss: {'train/loss': 0.03868292644619942, 'train/grad_norm_pre_clip': 1.8298014402389526, 'timing/data': 0.0003421558067202568, 'timing/model': 0.34322603419423103, 'learning_rate/qwen_vl_interface': 8.3e-06, 'learning_rate/action_model': 8.3e-05, 'epoch': 0.57})
+[2026-07-28 15:22:19,543][starVLA.training.train_starvla][INFO] - Step 3084, Loss: {'train/loss': 0.029240265488624573, 'train/grad_norm_pre_clip': 2.586127996444702, 'timing/data': 0.00036423373967409134, 'timing/model': 0.3452918007969856, 'learning_rate/qwen_vl_interface': 8.400000000000001e-06, 'learning_rate/action_model': 8.4e-05, 'epoch': 0.57})
+[2026-07-28 15:22:23,885][starVLA.training.train_starvla][INFO] - Step 3085, Loss: {'train/loss': 0.04452816769480705, 'train/grad_norm_pre_clip': 0.7539384365081787, 'timing/data': 0.0008237473666667938, 'timing/model': 0.3468630649149418, 'learning_rate/qwen_vl_interface': 8.5e-06, 'learning_rate/action_model': 8.5e-05, 'epoch': 0.57})
+[2026-07-28 15:22:28,119][starVLA.training.train_starvla][INFO] - Step 3086, Loss: {'train/loss': 0.02144785039126873, 'train/grad_norm_pre_clip': 0.6858577132225037, 'timing/data': 0.0004452192224562168, 'timing/model': 0.3488330701366067, 'learning_rate/qwen_vl_interface': 8.6e-06, 'learning_rate/action_model': 8.6e-05, 'epoch': 0.57})
+[2026-07-28 15:22:32,335][starVLA.training.train_starvla][INFO] - Step 3087, Loss: {'train/loss': 0.04311906918883324, 'train/grad_norm_pre_clip': 0.9970955848693848, 'timing/data': 0.0003118719905614853, 'timing/model': 0.3492181529290974, 'learning_rate/qwen_vl_interface': 8.700000000000001e-06, 'learning_rate/action_model': 8.7e-05, 'epoch': 0.57})
+[2026-07-28 15:22:36,621][starVLA.training.train_starvla][INFO] - Step 3088, Loss: {'train/loss': 0.0433049350976944, 'train/grad_norm_pre_clip': 2.8160228729248047, 'timing/data': 0.00031790509819984436, 'timing/model': 0.3942159400321543, 'learning_rate/qwen_vl_interface': 8.8e-06, 'learning_rate/action_model': 8.800000000000001e-05, 'epoch': 0.57})
+[2026-07-28 15:22:40,875][starVLA.training.train_starvla][INFO] - Step 3089, Loss: {'train/loss': 0.027430880814790726, 'train/grad_norm_pre_clip': 1.0132670402526855, 'timing/data': 0.0005244133062660694, 'timing/model': 0.338839263189584, 'learning_rate/qwen_vl_interface': 8.900000000000001e-06, 'learning_rate/action_model': 8.900000000000001e-05, 'epoch': 0.57})
+[2026-07-28 15:22:45,122][starVLA.training.train_starvla][INFO] - Step 3090, Loss: {'train/loss': 0.053057655692100525, 'train/grad_norm_pre_clip': 1.2110061645507812, 'timing/data': 0.00033807381987571716, 'timing/model': 0.34756715362891555, 'learning_rate/qwen_vl_interface': 9e-06, 'learning_rate/action_model': 9e-05, 'epoch': 0.57})
+[2026-07-28 15:22:49,296][starVLA.training.train_starvla][INFO] - Step 3091, Loss: {'train/loss': 0.04801104962825775, 'train/grad_norm_pre_clip': 0.7120378017425537, 'timing/data': 0.0003569037653505802, 'timing/model': 0.3350078440271318, 'learning_rate/qwen_vl_interface': 9.100000000000001e-06, 'learning_rate/action_model': 9.1e-05, 'epoch': 0.57})
+[2026-07-28 15:22:53,433][starVLA.training.train_starvla][INFO] - Step 3092, Loss: {'train/loss': 0.04690832272171974, 'train/grad_norm_pre_clip': 1.448601245880127, 'timing/data': 0.00033308984711766243, 'timing/model': 0.3342999620363116, 'learning_rate/qwen_vl_interface': 9.200000000000002e-06, 'learning_rate/action_model': 9.200000000000001e-05, 'epoch': 0.57})
+[2026-07-28 15:22:57,556][starVLA.training.train_starvla][INFO] - Step 3093, Loss: {'train/loss': 0.06983183324337006, 'train/grad_norm_pre_clip': 3.120332956314087, 'timing/data': 0.0004379628226161003, 'timing/model': 0.3435773686505854, 'learning_rate/qwen_vl_interface': 9.3e-06, 'learning_rate/action_model': 9.300000000000001e-05, 'epoch': 0.57})
+[2026-07-28 15:23:01,781][starVLA.training.train_starvla][INFO] - Step 3094, Loss: {'train/loss': 0.05249064415693283, 'train/grad_norm_pre_clip': 1.4094599485397339, 'timing/data': 0.0003885049372911453, 'timing/model': 0.33455704478546977, 'learning_rate/qwen_vl_interface': 9.4e-06, 'learning_rate/action_model': 9.4e-05, 'epoch': 0.57})
+[2026-07-28 15:23:06,063][starVLA.training.train_starvla][INFO] - Step 3095, Loss: {'train/loss': 0.037332966923713684, 'train/grad_norm_pre_clip': 0.8873981237411499, 'timing/data': 0.000774323008954525, 'timing/model': 0.3552469378337264, 'learning_rate/qwen_vl_interface': 9.5e-06, 'learning_rate/action_model': 9.5e-05, 'epoch': 0.57})
+[2026-07-28 15:23:10,290][starVLA.training.train_starvla][INFO] - Step 3096, Loss: {'train/loss': 0.04287060722708702, 'train/grad_norm_pre_clip': 1.3158025741577148, 'timing/data': 0.0004755551926791668, 'timing/model': 0.36989566730335355, 'learning_rate/qwen_vl_interface': 9.600000000000001e-06, 'learning_rate/action_model': 9.6e-05, 'epoch': 0.57})
+[2026-07-28 15:23:14,596][starVLA.training.train_starvla][INFO] - Step 3097, Loss: {'train/loss': 0.038893017917871475, 'train/grad_norm_pre_clip': 0.9750807881355286, 'timing/data': 0.000508100725710392, 'timing/model': 0.3406482837162912, 'learning_rate/qwen_vl_interface': 9.7e-06, 'learning_rate/action_model': 9.7e-05, 'epoch': 0.57})
+[2026-07-28 15:23:18,727][starVLA.training.train_starvla][INFO] - Step 3098, Loss: {'train/loss': 0.053397297859191895, 'train/grad_norm_pre_clip': 2.243291139602661, 'timing/data': 0.0003178049810230732, 'timing/model': 0.33890028623864055, 'learning_rate/qwen_vl_interface': 9.800000000000001e-06, 'learning_rate/action_model': 9.8e-05, 'epoch': 0.57})
+[2026-07-28 15:23:22,826][starVLA.training.train_starvla][INFO] - Step 3099, Loss: {'train/loss': 0.06302051246166229, 'train/grad_norm_pre_clip': 1.3280154466629028, 'timing/data': 0.00030366936698555946, 'timing/model': 0.33366368198767304, 'learning_rate/qwen_vl_interface': 9.9e-06, 'learning_rate/action_model': 9.900000000000001e-05, 'epoch': 0.57})
+[2026-07-28 15:23:26,934][starVLA.training.train_starvla][INFO] - Step 3100, Loss: {'train/loss': 0.02198510244488716, 'train/grad_norm_pre_clip': 0.5147536993026733, 'timing/data': 0.0003006746992468834, 'timing/model': 0.3345660581253469, 'learning_rate/qwen_vl_interface': 1e-05, 'learning_rate/action_model': 0.0001, 'epoch': 0.57})
+[2026-07-28 15:23:31,043][starVLA.training.train_starvla][INFO] - Step 3101, Loss: {'train/loss': 0.03602513670921326, 'train/grad_norm_pre_clip': 1.7605825662612915, 'timing/data': 0.0003724321722984314, 'timing/model': 0.34034581296145916, 'learning_rate/qwen_vl_interface': 9.999998458888285e-06, 'learning_rate/action_model': 9.999998458888285e-05, 'epoch': 0.57})
+[2026-07-28 15:23:35,123][starVLA.training.train_starvla][INFO] - Step 3102, Loss: {'train/loss': 0.08438272774219513, 'train/grad_norm_pre_clip': 2.663592576980591, 'timing/data': 0.0003800629638135433, 'timing/model': 0.33275496074929833, 'learning_rate/qwen_vl_interface': 9.999993835554141e-06, 'learning_rate/action_model': 9.999993835554141e-05, 'epoch': 0.57})
+[2026-07-28 15:23:39,242][starVLA.training.train_starvla][INFO] - Step 3103, Loss: {'train/loss': 0.05110183358192444, 'train/grad_norm_pre_clip': 1.2215375900268555, 'timing/data': 0.0005580452270805836, 'timing/model': 0.33420631010085344, 'learning_rate/qwen_vl_interface': 9.999986130000566e-06, 'learning_rate/action_model': 9.999986130000566e-05, 'epoch': 0.57})
+[2026-07-28 15:23:43,674][starVLA.training.train_starvla][INFO] - Step 3104, Loss: {'train/loss': 0.03333107382059097, 'train/grad_norm_pre_clip': 1.2325721979141235, 'timing/data': 0.00032299989834427834, 'timing/model': 0.387612157035619, 'learning_rate/qwen_vl_interface': 9.99997534223256e-06, 'learning_rate/action_model': 9.99997534223256e-05, 'epoch': 0.57})
+[2026-07-28 15:23:48,100][starVLA.training.train_starvla][INFO] - Step 3105, Loss: {'train/loss': 0.04698307067155838, 'train/grad_norm_pre_clip': 1.1890767812728882, 'timing/data': 0.00035655079409480095, 'timing/model': 0.3669933369383216, 'learning_rate/qwen_vl_interface': 9.999961472257127e-06, 'learning_rate/action_model': 9.999961472257127e-05, 'epoch': 0.57})
+[2026-07-28 15:23:52,187][starVLA.training.train_starvla][INFO] - Step 3106, Loss: {'train/loss': 0.04213358461856842, 'train/grad_norm_pre_clip': 1.3189496994018555, 'timing/data': 0.0003138650208711624, 'timing/model': 0.33256288012489676, 'learning_rate/qwen_vl_interface': 9.999944520083262e-06, 'learning_rate/action_model': 9.999944520083261e-05, 'epoch': 0.57})
+[2026-07-28 15:23:56,342][starVLA.training.train_starvla][INFO] - Step 3107, Loss: {'train/loss': 0.04051333665847778, 'train/grad_norm_pre_clip': 0.8787927627563477, 'timing/data': 0.00029981089755892754, 'timing/model': 0.35379939107224345, 'learning_rate/qwen_vl_interface': 9.999924485721968e-06, 'learning_rate/action_model': 9.999924485721968e-05, 'epoch': 0.57})
+[2026-07-28 15:24:00,449][starVLA.training.train_starvla][INFO] - Step 3108, Loss: {'train/loss': 0.042731624096632004, 'train/grad_norm_pre_clip': 1.795779824256897, 'timing/data': 0.00032906606793403625, 'timing/model': 0.3347387299872935, 'learning_rate/qwen_vl_interface': 9.999901369186244e-06, 'learning_rate/action_model': 9.999901369186244e-05, 'epoch': 0.57})
+[2026-07-28 15:24:04,583][starVLA.training.train_starvla][INFO] - Step 3109, Loss: {'train/loss': 0.031956400722265244, 'train/grad_norm_pre_clip': 1.4291460514068604, 'timing/data': 0.000318372156471014, 'timing/model': 0.33260604878887534, 'learning_rate/qwen_vl_interface': 9.999875170491092e-06, 'learning_rate/action_model': 9.999875170491091e-05, 'epoch': 0.57})
+[2026-07-28 15:24:08,778][starVLA.training.train_starvla][INFO] - Step 3110, Loss: {'train/loss': 0.05991104245185852, 'train/grad_norm_pre_clip': 2.194110870361328, 'timing/data': 0.0003390982747077942, 'timing/model': 0.34930822998285294, 'learning_rate/qwen_vl_interface': 9.999845889653509e-06, 'learning_rate/action_model': 9.999845889653508e-05, 'epoch': 0.57})
+[2026-07-28 15:24:12,947][starVLA.training.train_starvla][INFO] - Step 3111, Loss: {'train/loss': 0.04812587797641754, 'train/grad_norm_pre_clip': 2.4022433757781982, 'timing/data': 0.00031213369220495224, 'timing/model': 0.33452270925045013, 'learning_rate/qwen_vl_interface': 9.999813526692496e-06, 'learning_rate/action_model': 9.999813526692495e-05, 'epoch': 0.57})
+[2026-07-28 15:24:17,035][starVLA.training.train_starvla][INFO] - Step 3112, Loss: {'train/loss': 0.016385823488235474, 'train/grad_norm_pre_clip': 0.7757654190063477, 'timing/data': 0.0008324338123202324, 'timing/model': 0.3325006030499935, 'learning_rate/qwen_vl_interface': 9.999778081629055e-06, 'learning_rate/action_model': 9.999778081629054e-05, 'epoch': 0.57})
+[2026-07-28 15:24:21,137][starVLA.training.train_starvla][INFO] - Step 3113, Loss: {'train/loss': 0.0185086652636528, 'train/grad_norm_pre_clip': 0.8391175270080566, 'timing/data': 0.00030493317171931267, 'timing/model': 0.33290168084204197, 'learning_rate/qwen_vl_interface': 9.999739554486184e-06, 'learning_rate/action_model': 9.999739554486182e-05, 'epoch': 0.57})
+[2026-07-28 15:24:25,300][starVLA.training.train_starvla][INFO] - Step 3114, Loss: {'train/loss': 0.027154527604579926, 'train/grad_norm_pre_clip': 1.242692470550537, 'timing/data': 0.0004002000205218792, 'timing/model': 0.33239431492984295, 'learning_rate/qwen_vl_interface': 9.999697945288881e-06, 'learning_rate/action_model': 9.999697945288881e-05, 'epoch': 0.57})
+[2026-07-28 15:24:29,414][starVLA.training.train_starvla][INFO] - Step 3115, Loss: {'train/loss': 0.04383620247244835, 'train/grad_norm_pre_clip': 1.6052191257476807, 'timing/data': 0.00040413998067379, 'timing/model': 0.33991785533726215, 'learning_rate/qwen_vl_interface': 9.999653254064153e-06, 'learning_rate/action_model': 9.999653254064152e-05, 'epoch': 0.57})
+[2026-07-28 15:24:33,532][starVLA.training.train_starvla][INFO] - Step 3116, Loss: {'train/loss': 0.036978788673877716, 'train/grad_norm_pre_clip': 1.0396101474761963, 'timing/data': 0.0002710479311645031, 'timing/model': 0.3332807463593781, 'learning_rate/qwen_vl_interface': 9.99960548084099e-06, 'learning_rate/action_model': 9.99960548084099e-05, 'epoch': 0.57})
+[2026-07-28 15:24:37,664][starVLA.training.train_starvla][INFO] - Step 3117, Loss: {'train/loss': 0.06819118559360504, 'train/grad_norm_pre_clip': 1.9125735759735107, 'timing/data': 0.00046522216871380806, 'timing/model': 0.3406263370998204, 'learning_rate/qwen_vl_interface': 9.999554625650398e-06, 'learning_rate/action_model': 9.999554625650399e-05, 'epoch': 0.57})
+[2026-07-28 15:24:41,841][starVLA.training.train_starvla][INFO] - Step 3118, Loss: {'train/loss': 0.0496157705783844, 'train/grad_norm_pre_clip': 1.1111820936203003, 'timing/data': 0.0003019440919160843, 'timing/model': 0.35558698419481516, 'learning_rate/qwen_vl_interface': 9.999500688525376e-06, 'learning_rate/action_model': 9.999500688525375e-05, 'epoch': 0.57})
+[2026-07-28 15:24:46,056][starVLA.training.train_starvla][INFO] - Step 3119, Loss: {'train/loss': 0.03375307470560074, 'train/grad_norm_pre_clip': 1.1924364566802979, 'timing/data': 0.0003041527234017849, 'timing/model': 0.35577381774783134, 'learning_rate/qwen_vl_interface': 9.99944366950092e-06, 'learning_rate/action_model': 9.999443669500919e-05, 'epoch': 0.57})
+[2026-07-28 15:24:50,204][starVLA.training.train_starvla][INFO] - Step 3120, Loss: {'train/loss': 0.07494527101516724, 'train/grad_norm_pre_clip': 5.1653733253479, 'timing/data': 0.0003297249786555767, 'timing/model': 0.3339762953110039, 'learning_rate/qwen_vl_interface': 9.99938356861403e-06, 'learning_rate/action_model': 9.99938356861403e-05, 'epoch': 0.57})
+[2026-07-28 15:24:54,661][starVLA.training.train_starvla][INFO] - Step 3121, Loss: {'train/loss': 0.07064235210418701, 'train/grad_norm_pre_clip': 2.2627315521240234, 'timing/data': 0.00030493782833218575, 'timing/model': 0.3318837550468743, 'learning_rate/qwen_vl_interface': 9.999320385903709e-06, 'learning_rate/action_model': 9.999320385903708e-05, 'epoch': 0.57})
+[2026-07-28 15:24:58,736][starVLA.training.train_starvla][INFO] - Step 3122, Loss: {'train/loss': 0.053537994623184204, 'train/grad_norm_pre_clip': 3.2395951747894287, 'timing/data': 0.00037217000499367714, 'timing/model': 0.33189714699983597, 'learning_rate/qwen_vl_interface': 9.999254121410948e-06, 'learning_rate/action_model': 9.999254121410949e-05, 'epoch': 0.57})
+[2026-07-28 15:25:02,830][starVLA.training.train_starvla][INFO] - Step 3123, Loss: {'train/loss': 0.08907774835824966, 'train/grad_norm_pre_clip': 2.744050979614258, 'timing/data': 0.0003043282777070999, 'timing/model': 0.34834294114261866, 'learning_rate/qwen_vl_interface': 9.999184775178751e-06, 'learning_rate/action_model': 9.999184775178751e-05, 'epoch': 0.57})
+[2026-07-28 15:25:06,987][starVLA.training.train_starvla][INFO] - Step 3124, Loss: {'train/loss': 0.05514860898256302, 'train/grad_norm_pre_clip': 1.707561731338501, 'timing/data': 0.000468467827886343, 'timing/model': 0.34057488292455673, 'learning_rate/qwen_vl_interface': 9.999112347252117e-06, 'learning_rate/action_model': 9.999112347252116e-05, 'epoch': 0.57})
+[2026-07-28 15:25:11,096][starVLA.training.train_starvla][INFO] - Step 3125, Loss: {'train/loss': 0.03923140466213226, 'train/grad_norm_pre_clip': 1.161824107170105, 'timing/data': 0.0003141108900308609, 'timing/model': 0.3329597353003919, 'learning_rate/qwen_vl_interface': 9.99903683767804e-06, 'learning_rate/action_model': 9.99903683767804e-05, 'epoch': 0.57})
+[2026-07-28 15:25:15,190][starVLA.training.train_starvla][INFO] - Step 3126, Loss: {'train/loss': 0.03539682924747467, 'train/grad_norm_pre_clip': 1.3613207340240479, 'timing/data': 0.000359503086656332, 'timing/model': 0.3337124898098409, 'learning_rate/qwen_vl_interface': 9.998958246505518e-06, 'learning_rate/action_model': 9.998958246505516e-05, 'epoch': 0.57})
+[2026-07-28 15:25:19,316][starVLA.training.train_starvla][INFO] - Step 3127, Loss: {'train/loss': 0.024364590644836426, 'train/grad_norm_pre_clip': 0.6728657484054565, 'timing/data': 0.0002926350571215153, 'timing/model': 0.335505039896816, 'learning_rate/qwen_vl_interface': 9.998876573785547e-06, 'learning_rate/action_model': 9.998876573785547e-05, 'epoch': 0.57})
+[2026-07-28 15:25:23,588][starVLA.training.train_starvla][INFO] - Step 3128, Loss: {'train/loss': 0.08914297819137573, 'train/grad_norm_pre_clip': 1.5511741638183594, 'timing/data': 0.0002813623286783695, 'timing/model': 0.36232199100777507, 'learning_rate/qwen_vl_interface': 9.998791819571126e-06, 'learning_rate/action_model': 9.998791819571127e-05, 'epoch': 0.57})
+[2026-07-28 15:25:27,809][starVLA.training.train_starvla][INFO] - Step 3129, Loss: {'train/loss': 0.05856182426214218, 'train/grad_norm_pre_clip': 1.4736137390136719, 'timing/data': 0.0003146640956401825, 'timing/model': 0.3362895702011883, 'learning_rate/qwen_vl_interface': 9.998703983917252e-06, 'learning_rate/action_model': 9.998703983917252e-05, 'epoch': 0.57})
+[2026-07-28 15:25:32,222][starVLA.training.train_starvla][INFO] - Step 3130, Loss: {'train/loss': 0.0380622036755085, 'train/grad_norm_pre_clip': 1.356866478919983, 'timing/data': 0.00045274198055267334, 'timing/model': 0.33705911599099636, 'learning_rate/qwen_vl_interface': 9.998613066880919e-06, 'learning_rate/action_model': 9.998613066880918e-05, 'epoch': 0.57})
+[2026-07-28 15:25:36,373][starVLA.training.train_starvla][INFO] - Step 3131, Loss: {'train/loss': 0.03341064602136612, 'train/grad_norm_pre_clip': 0.7254997491836548, 'timing/data': 0.00029110535979270935, 'timing/model': 0.33322714688256383, 'learning_rate/qwen_vl_interface': 9.99851906852112e-06, 'learning_rate/action_model': 9.998519068521119e-05, 'epoch': 0.58})
+[2026-07-28 15:25:40,487][starVLA.training.train_starvla][INFO] - Step 3132, Loss: {'train/loss': 0.061509907245635986, 'train/grad_norm_pre_clip': 0.828105092048645, 'timing/data': 0.00035654613748192787, 'timing/model': 0.3423210163600743, 'learning_rate/qwen_vl_interface': 9.998421988898851e-06, 'learning_rate/action_model': 9.998421988898851e-05, 'epoch': 0.58})
+[2026-07-28 15:25:44,758][starVLA.training.train_starvla][INFO] - Step 3133, Loss: {'train/loss': 0.03496735915541649, 'train/grad_norm_pre_clip': 0.8689619302749634, 'timing/data': 0.00031177792698144913, 'timing/model': 0.34752711839973927, 'learning_rate/qwen_vl_interface': 9.998321828077107e-06, 'learning_rate/action_model': 9.998321828077106e-05, 'epoch': 0.58})
+[2026-07-28 15:25:48,911][starVLA.training.train_starvla][INFO] - Step 3134, Loss: {'train/loss': 0.03145371377468109, 'train/grad_norm_pre_clip': 0.5754076242446899, 'timing/data': 0.00039542512968182564, 'timing/model': 0.3404051763936877, 'learning_rate/qwen_vl_interface': 9.99821858612088e-06, 'learning_rate/action_model': 9.99821858612088e-05, 'epoch': 0.58})
+[2026-07-28 15:25:53,091][starVLA.training.train_starvla][INFO] - Step 3135, Loss: {'train/loss': 0.03622494637966156, 'train/grad_norm_pre_clip': 0.8537060022354126, 'timing/data': 0.0002978099510073662, 'timing/model': 0.3671746118925512, 'learning_rate/qwen_vl_interface': 9.998112263097164e-06, 'learning_rate/action_model': 9.998112263097165e-05, 'epoch': 0.58})
+[2026-07-28 15:25:57,260][starVLA.training.train_starvla][INFO] - Step 3136, Loss: {'train/loss': 0.033117979764938354, 'train/grad_norm_pre_clip': 1.1268200874328613, 'timing/data': 0.00033970782533288, 'timing/model': 0.34875616896897554, 'learning_rate/qwen_vl_interface': 9.99800285907495e-06, 'learning_rate/action_model': 9.99800285907495e-05, 'epoch': 0.58})
+[2026-07-28 15:26:01,372][starVLA.training.train_starvla][INFO] - Step 3137, Loss: {'train/loss': 0.030634352937340736, 'train/grad_norm_pre_clip': 1.1821513175964355, 'timing/data': 0.0003765248693525791, 'timing/model': 0.3358806879259646, 'learning_rate/qwen_vl_interface': 9.997890374125228e-06, 'learning_rate/action_model': 9.997890374125227e-05, 'epoch': 0.58})
+[2026-07-28 15:26:05,524][starVLA.training.train_starvla][INFO] - Step 3138, Loss: {'train/loss': 0.0694127306342125, 'train/grad_norm_pre_clip': 1.4120488166809082, 'timing/data': 0.0002997498959302902, 'timing/model': 0.333384629804641, 'learning_rate/qwen_vl_interface': 9.997774808320992e-06, 'learning_rate/action_model': 9.997774808320992e-05, 'epoch': 0.58})
+[2026-07-28 15:26:09,643][starVLA.training.train_starvla][INFO] - Step 3139, Loss: {'train/loss': 0.02773672342300415, 'train/grad_norm_pre_clip': 1.2883751392364502, 'timing/data': 0.0005148560740053654, 'timing/model': 0.33368012495338917, 'learning_rate/qwen_vl_interface': 9.997656161737227e-06, 'learning_rate/action_model': 9.997656161737226e-05, 'epoch': 0.58})
+[2026-07-28 15:26:13,793][starVLA.training.train_starvla][INFO] - Step 3140, Loss: {'train/loss': 0.041203536093235016, 'train/grad_norm_pre_clip': 1.1640520095825195, 'timing/data': 0.00031578680500388145, 'timing/model': 0.3343651620671153, 'learning_rate/qwen_vl_interface': 9.997534434450922e-06, 'learning_rate/action_model': 9.997534434450921e-05, 'epoch': 0.58})
+[2026-07-28 15:26:17,934][starVLA.training.train_starvla][INFO] - Step 3141, Loss: {'train/loss': 0.05359676480293274, 'train/grad_norm_pre_clip': 1.6498228311538696, 'timing/data': 0.00035294797271490097, 'timing/model': 0.347945773974061, 'learning_rate/qwen_vl_interface': 9.997409626541065e-06, 'learning_rate/action_model': 9.997409626541066e-05, 'epoch': 0.58})
+[2026-07-28 15:26:22,097][starVLA.training.train_starvla][INFO] - Step 3142, Loss: {'train/loss': 0.04131341725587845, 'train/grad_norm_pre_clip': 1.0976526737213135, 'timing/data': 0.00030356505885720253, 'timing/model': 0.33232853561639786, 'learning_rate/qwen_vl_interface': 9.997281738088647e-06, 'learning_rate/action_model': 9.997281738088646e-05, 'epoch': 0.58})
+[2026-07-28 15:26:26,211][starVLA.training.train_starvla][INFO] - Step 3143, Loss: {'train/loss': 0.02565000392496586, 'train/grad_norm_pre_clip': 1.120396614074707, 'timing/data': 0.000370705034583807, 'timing/model': 0.33236114913597703, 'learning_rate/qwen_vl_interface': 9.997150769176648e-06, 'learning_rate/action_model': 9.997150769176647e-05, 'epoch': 0.58})
+[2026-07-28 15:26:30,296][starVLA.training.train_starvla][INFO] - Step 3144, Loss: {'train/loss': 0.03542115539312363, 'train/grad_norm_pre_clip': 1.497970461845398, 'timing/data': 0.00030364328995347023, 'timing/model': 0.33337582694366574, 'learning_rate/qwen_vl_interface': 9.997016719890052e-06, 'learning_rate/action_model': 9.997016719890052e-05, 'epoch': 0.58})
+[2026-07-28 15:26:34,460][starVLA.training.train_starvla][INFO] - Step 3145, Loss: {'train/loss': 0.051057249307632446, 'train/grad_norm_pre_clip': 2.1739554405212402, 'timing/data': 0.0003121369518339634, 'timing/model': 0.33495461707934737, 'learning_rate/qwen_vl_interface': 9.996879590315847e-06, 'learning_rate/action_model': 9.996879590315846e-05, 'epoch': 0.58})
+[2026-07-28 15:26:38,549][starVLA.training.train_starvla][INFO] - Step 3146, Loss: {'train/loss': 0.1326911747455597, 'train/grad_norm_pre_clip': 3.2600390911102295, 'timing/data': 0.00032411515712738037, 'timing/model': 0.33890596497803926, 'learning_rate/qwen_vl_interface': 9.99673938054301e-06, 'learning_rate/action_model': 9.996739380543009e-05, 'epoch': 0.58})
+[2026-07-28 15:26:42,648][starVLA.training.train_starvla][INFO] - Step 3147, Loss: {'train/loss': 0.07047390937805176, 'train/grad_norm_pre_clip': 0.9411023259162903, 'timing/data': 0.0003149723634123802, 'timing/model': 0.3344252179376781, 'learning_rate/qwen_vl_interface': 9.996596090662525e-06, 'learning_rate/action_model': 9.996596090662525e-05, 'epoch': 0.58})
+[2026-07-28 15:26:46,785][starVLA.training.train_starvla][INFO] - Step 3148, Loss: {'train/loss': 0.016818590462207794, 'train/grad_norm_pre_clip': 1.104689359664917, 'timing/data': 0.0003214878961443901, 'timing/model': 0.34407690865918994, 'learning_rate/qwen_vl_interface': 9.996449720767369e-06, 'learning_rate/action_model': 9.996449720767368e-05, 'epoch': 0.58})
+[2026-07-28 15:26:51,012][starVLA.training.train_starvla][INFO] - Step 3149, Loss: {'train/loss': 0.026420675218105316, 'train/grad_norm_pre_clip': 0.4812138080596924, 'timing/data': 0.00035034725442528725, 'timing/model': 0.37374305399134755, 'learning_rate/qwen_vl_interface': 9.99630027095252e-06, 'learning_rate/action_model': 9.99630027095252e-05, 'epoch': 0.58})
+[2026-07-28 15:26:55,191][starVLA.training.train_starvla][INFO] - Step 3150, Loss: {'train/loss': 0.03405708819627762, 'train/grad_norm_pre_clip': 0.8518646955490112, 'timing/data': 0.0004486590623855591, 'timing/model': 0.33396244794130325, 'learning_rate/qwen_vl_interface': 9.996147741314956e-06, 'learning_rate/action_model': 9.996147741314957e-05, 'epoch': 0.58})
+[2026-07-28 15:26:59,304][starVLA.training.train_starvla][INFO] - Step 3151, Loss: {'train/loss': 0.03740936145186424, 'train/grad_norm_pre_clip': 1.219409465789795, 'timing/data': 0.0005253860726952553, 'timing/model': 0.33578302804380655, 'learning_rate/qwen_vl_interface': 9.99599213195365e-06, 'learning_rate/action_model': 9.99599213195365e-05, 'epoch': 0.58})
+[2026-07-28 15:27:03,394][starVLA.training.train_starvla][INFO] - Step 3152, Loss: {'train/loss': 0.036197323352098465, 'train/grad_norm_pre_clip': 1.4793519973754883, 'timing/data': 0.00041321665048599243, 'timing/model': 0.334899066016078, 'learning_rate/qwen_vl_interface': 9.995833442969578e-06, 'learning_rate/action_model': 9.995833442969578e-05, 'epoch': 0.58})
+[2026-07-28 15:27:07,514][starVLA.training.train_starvla][INFO] - Step 3153, Loss: {'train/loss': 0.029843691736459732, 'train/grad_norm_pre_clip': 1.2168736457824707, 'timing/data': 0.0003171307034790516, 'timing/model': 0.33378452295437455, 'learning_rate/qwen_vl_interface': 9.995671674465709e-06, 'learning_rate/action_model': 9.995671674465709e-05, 'epoch': 0.58})
+[2026-07-28 15:27:11,623][starVLA.training.train_starvla][INFO] - Step 3154, Loss: {'train/loss': 0.03162805363535881, 'train/grad_norm_pre_clip': 0.9608531594276428, 'timing/data': 0.0003636288456618786, 'timing/model': 0.3349074781872332, 'learning_rate/qwen_vl_interface': 9.99550682654701e-06, 'learning_rate/action_model': 9.995506826547011e-05, 'epoch': 0.58})
+[2026-07-28 15:27:15,724][starVLA.training.train_starvla][INFO] - Step 3155, Loss: {'train/loss': 0.06332488358020782, 'train/grad_norm_pre_clip': 1.6158968210220337, 'timing/data': 0.00030858302488923073, 'timing/model': 0.3344303569756448, 'learning_rate/qwen_vl_interface': 9.995338899320457e-06, 'learning_rate/action_model': 9.995338899320456e-05, 'epoch': 0.58})
+[2026-07-28 15:27:19,814][starVLA.training.train_starvla][INFO] - Step 3156, Loss: {'train/loss': 0.05470278114080429, 'train/grad_norm_pre_clip': 1.2266358137130737, 'timing/data': 0.0002938881516456604, 'timing/model': 0.33310448192059994, 'learning_rate/qwen_vl_interface': 9.995167892895011e-06, 'learning_rate/action_model': 9.99516789289501e-05, 'epoch': 0.58})
+[2026-07-28 15:27:23,894][starVLA.training.train_starvla][INFO] - Step 3157, Loss: {'train/loss': 0.05253874510526657, 'train/grad_norm_pre_clip': 2.129335880279541, 'timing/data': 0.00035072164610028267, 'timing/model': 0.33405721886083484, 'learning_rate/qwen_vl_interface': 9.994993807381635e-06, 'learning_rate/action_model': 9.994993807381635e-05, 'epoch': 0.58})
+[2026-07-28 15:27:27,970][starVLA.training.train_starvla][INFO] - Step 3158, Loss: {'train/loss': 0.0748971551656723, 'train/grad_norm_pre_clip': 3.1922523975372314, 'timing/data': 0.0004633180797100067, 'timing/model': 0.3332121754065156, 'learning_rate/qwen_vl_interface': 9.994816642893294e-06, 'learning_rate/action_model': 9.994816642893293e-05, 'epoch': 0.58})
+[2026-07-28 15:27:32,045][starVLA.training.train_starvla][INFO] - Step 3159, Loss: {'train/loss': 0.029344432055950165, 'train/grad_norm_pre_clip': 0.941624641418457, 'timing/data': 0.00033415062353014946, 'timing/model': 0.33180242870002985, 'learning_rate/qwen_vl_interface': 9.994636399544948e-06, 'learning_rate/action_model': 9.994636399544947e-05, 'epoch': 0.58})
+[2026-07-28 15:27:36,154][starVLA.training.train_starvla][INFO] - Step 3160, Loss: {'train/loss': 0.07293085753917694, 'train/grad_norm_pre_clip': 2.1881701946258545, 'timing/data': 0.0005129789933562279, 'timing/model': 0.3328486126847565, 'learning_rate/qwen_vl_interface': 9.994453077453552e-06, 'learning_rate/action_model': 9.994453077453552e-05, 'epoch': 0.58})
+[2026-07-28 15:27:40,284][starVLA.training.train_starvla][INFO] - Step 3161, Loss: {'train/loss': 0.041232120245695114, 'train/grad_norm_pre_clip': 0.8112317323684692, 'timing/data': 0.0004602670669555664, 'timing/model': 0.3329161312431097, 'learning_rate/qwen_vl_interface': 9.994266676738065e-06, 'learning_rate/action_model': 9.994266676738065e-05, 'epoch': 0.58})
+[2026-07-28 15:27:44,379][starVLA.training.train_starvla][INFO] - Step 3162, Loss: {'train/loss': 0.034050408750772476, 'train/grad_norm_pre_clip': 0.9242420792579651, 'timing/data': 0.000296515878289938, 'timing/model': 0.33341630967333913, 'learning_rate/qwen_vl_interface': 9.99407719751944e-06, 'learning_rate/action_model': 9.994077197519439e-05, 'epoch': 0.58})
+[2026-07-28 15:27:48,490][starVLA.training.train_starvla][INFO] - Step 3163, Loss: {'train/loss': 0.03620595484972, 'train/grad_norm_pre_clip': 1.1320182085037231, 'timing/data': 0.00037471530959010124, 'timing/model': 0.33410273119807243, 'learning_rate/qwen_vl_interface': 9.993884639920626e-06, 'learning_rate/action_model': 9.993884639920626e-05, 'epoch': 0.58})
+[2026-07-28 15:27:52,621][starVLA.training.train_starvla][INFO] - Step 3164, Loss: {'train/loss': 0.03489980846643448, 'train/grad_norm_pre_clip': 0.8412628769874573, 'timing/data': 0.00029737409204244614, 'timing/model': 0.34733459120616317, 'learning_rate/qwen_vl_interface': 9.993689004066573e-06, 'learning_rate/action_model': 9.993689004066573e-05, 'epoch': 0.58})
+[2026-07-28 15:27:56,709][starVLA.training.train_starvla][INFO] - Step 3165, Loss: {'train/loss': 0.037810057401657104, 'train/grad_norm_pre_clip': 1.128080129623413, 'timing/data': 0.0003465781919658184, 'timing/model': 0.33260991610586643, 'learning_rate/qwen_vl_interface': 9.993490290084225e-06, 'learning_rate/action_model': 9.993490290084225e-05, 'epoch': 0.58})
+[2026-07-28 15:28:01,090][starVLA.training.train_starvla][INFO] - Step 3166, Loss: {'train/loss': 0.03672308847308159, 'train/grad_norm_pre_clip': 0.695111870765686, 'timing/data': 0.00038263481110334396, 'timing/model': 0.3339108806103468, 'learning_rate/qwen_vl_interface': 9.993288498102531e-06, 'learning_rate/action_model': 9.993288498102532e-05, 'epoch': 0.58})
+[2026-07-28 15:28:05,264][starVLA.training.train_starvla][INFO] - Step 3167, Loss: {'train/loss': 0.06601421535015106, 'train/grad_norm_pre_clip': 1.6843891143798828, 'timing/data': 0.0003572646528482437, 'timing/model': 0.33307516016066074, 'learning_rate/qwen_vl_interface': 9.993083628252424e-06, 'learning_rate/action_model': 9.993083628252424e-05, 'epoch': 0.58})
+[2026-07-28 15:28:09,390][starVLA.training.train_starvla][INFO] - Step 3168, Loss: {'train/loss': 0.05462165176868439, 'train/grad_norm_pre_clip': 1.7995784282684326, 'timing/data': 0.00050400011241436, 'timing/model': 0.3341948059387505, 'learning_rate/qwen_vl_interface': 9.992875680666849e-06, 'learning_rate/action_model': 9.992875680666849e-05, 'epoch': 0.58})
+[2026-07-28 15:28:13,533][starVLA.training.train_starvla][INFO] - Step 3169, Loss: {'train/loss': 0.042569588869810104, 'train/grad_norm_pre_clip': 0.9985770583152771, 'timing/data': 0.00036011915653944016, 'timing/model': 0.34983448777347803, 'learning_rate/qwen_vl_interface': 9.992664655480736e-06, 'learning_rate/action_model': 9.992664655480735e-05, 'epoch': 0.58})
+[2026-07-28 15:28:17,826][starVLA.training.train_starvla][INFO] - Step 3170, Loss: {'train/loss': 0.05275975167751312, 'train/grad_norm_pre_clip': 1.2004245519638062, 'timing/data': 0.00030079297721385956, 'timing/model': 0.35049965884536505, 'learning_rate/qwen_vl_interface': 9.992450552831018e-06, 'learning_rate/action_model': 9.992450552831017e-05, 'epoch': 0.58})
+[2026-07-28 15:28:21,928][starVLA.training.train_starvla][INFO] - Step 3171, Loss: {'train/loss': 0.024481846019625664, 'train/grad_norm_pre_clip': 1.276468276977539, 'timing/data': 0.0003206329420208931, 'timing/model': 0.3325371127575636, 'learning_rate/qwen_vl_interface': 9.992233372856623e-06, 'learning_rate/action_model': 9.992233372856624e-05, 'epoch': 0.58})
+[2026-07-28 15:28:26,015][starVLA.training.train_starvla][INFO] - Step 3172, Loss: {'train/loss': 0.06881983578205109, 'train/grad_norm_pre_clip': 1.1553715467453003, 'timing/data': 0.0003013717941939831, 'timing/model': 0.3486900422722101, 'learning_rate/qwen_vl_interface': 9.99201311569848e-06, 'learning_rate/action_model': 9.99201311569848e-05, 'epoch': 0.58})
+[2026-07-28 15:28:30,170][starVLA.training.train_starvla][INFO] - Step 3173, Loss: {'train/loss': 0.060396283864974976, 'train/grad_norm_pre_clip': 1.1356687545776367, 'timing/data': 0.00030242884531617165, 'timing/model': 0.332742135040462, 'learning_rate/qwen_vl_interface': 9.99178978149951e-06, 'learning_rate/action_model': 9.99178978149951e-05, 'epoch': 0.58})
+[2026-07-28 15:28:34,356][starVLA.training.train_starvla][INFO] - Step 3174, Loss: {'train/loss': 0.015180767513811588, 'train/grad_norm_pre_clip': 1.0211576223373413, 'timing/data': 0.00033972831442952156, 'timing/model': 0.33319320203736424, 'learning_rate/qwen_vl_interface': 9.99156337040463e-06, 'learning_rate/action_model': 9.991563370404629e-05, 'epoch': 0.58})
+[2026-07-28 15:28:38,742][starVLA.training.train_starvla][INFO] - Step 3175, Loss: {'train/loss': 0.034960437566041946, 'train/grad_norm_pre_clip': 1.443284034729004, 'timing/data': 0.0003066752105951309, 'timing/model': 0.3390326569788158, 'learning_rate/qwen_vl_interface': 9.991333882560759e-06, 'learning_rate/action_model': 9.991333882560759e-05, 'epoch': 0.58})
+[2026-07-28 15:28:42,872][starVLA.training.train_starvla][INFO] - Step 3176, Loss: {'train/loss': 0.079631507396698, 'train/grad_norm_pre_clip': 1.4178462028503418, 'timing/data': 0.00035273609682917595, 'timing/model': 0.34076455421745777, 'learning_rate/qwen_vl_interface': 9.991101318116806e-06, 'learning_rate/action_model': 9.991101318116806e-05, 'epoch': 0.58})
+[2026-07-28 15:28:47,004][starVLA.training.train_starvla][INFO] - Step 3177, Loss: {'train/loss': 0.10227344930171967, 'train/grad_norm_pre_clip': 2.2860472202301025, 'timing/data': 0.00034887809306383133, 'timing/model': 0.3339282772503793, 'learning_rate/qwen_vl_interface': 9.990865677223684e-06, 'learning_rate/action_model': 9.990865677223684e-05, 'epoch': 0.58})
+[2026-07-28 15:28:51,100][starVLA.training.train_starvla][INFO] - Step 3178, Loss: {'train/loss': 0.06315312534570694, 'train/grad_norm_pre_clip': 1.2459205389022827, 'timing/data': 0.000313508789986372, 'timing/model': 0.33381758304312825, 'learning_rate/qwen_vl_interface': 9.990626960034291e-06, 'learning_rate/action_model': 9.990626960034291e-05, 'epoch': 0.58})
+[2026-07-28 15:28:55,193][starVLA.training.train_starvla][INFO] - Step 3179, Loss: {'train/loss': 0.04097443446516991, 'train/grad_norm_pre_clip': 1.1364458799362183, 'timing/data': 0.0003664162941277027, 'timing/model': 0.33407924603670835, 'learning_rate/qwen_vl_interface': 9.990385166703534e-06, 'learning_rate/action_model': 9.990385166703533e-05, 'epoch': 0.58})
+[2026-07-28 15:28:59,308][starVLA.training.train_starvla][INFO] - Step 3180, Loss: {'train/loss': 0.03808323293924332, 'train/grad_norm_pre_clip': 1.0600491762161255, 'timing/data': 0.00045356014743447304, 'timing/model': 0.3432592200115323, 'learning_rate/qwen_vl_interface': 9.990140297388306e-06, 'learning_rate/action_model': 9.990140297388307e-05, 'epoch': 0.58})
+[2026-07-28 15:29:03,587][starVLA.training.train_starvla][INFO] - Step 3181, Loss: {'train/loss': 0.04344508796930313, 'train/grad_norm_pre_clip': 1.0596753358840942, 'timing/data': 0.0004914328455924988, 'timing/model': 0.3379022809676826, 'learning_rate/qwen_vl_interface': 9.989892352247505e-06, 'learning_rate/action_model': 9.989892352247504e-05, 'epoch': 0.58})
+[2026-07-28 15:29:07,702][starVLA.training.train_starvla][INFO] - Step 3182, Loss: {'train/loss': 0.04143820330500603, 'train/grad_norm_pre_clip': 1.1985903978347778, 'timing/data': 0.00029642507433891296, 'timing/model': 0.34143046708777547, 'learning_rate/qwen_vl_interface': 9.989641331442014e-06, 'learning_rate/action_model': 9.989641331442013e-05, 'epoch': 0.58})
+[2026-07-28 15:29:11,850][starVLA.training.train_starvla][INFO] - Step 3183, Loss: {'train/loss': 0.07332898676395416, 'train/grad_norm_pre_clip': 3.481680154800415, 'timing/data': 0.0003748089075088501, 'timing/model': 0.34642975218594074, 'learning_rate/qwen_vl_interface': 9.989387235134722e-06, 'learning_rate/action_model': 9.989387235134722e-05, 'epoch': 0.58})
+[2026-07-28 15:29:15,976][starVLA.training.train_starvla][INFO] - Step 3184, Loss: {'train/loss': 0.026935188099741936, 'train/grad_norm_pre_clip': 0.9246460199356079, 'timing/data': 0.00037947995588183403, 'timing/model': 0.34442723728716373, 'learning_rate/qwen_vl_interface': 9.98913006349051e-06, 'learning_rate/action_model': 9.989130063490508e-05, 'epoch': 0.58})
+[2026-07-28 15:29:20,070][starVLA.training.train_starvla][INFO] - Step 3185, Loss: {'train/loss': 0.03803978115320206, 'train/grad_norm_pre_clip': 0.9603959918022156, 'timing/data': 0.000279740896075964, 'timing/model': 0.33195427898317575, 'learning_rate/qwen_vl_interface': 9.988869816676247e-06, 'learning_rate/action_model': 9.988869816676246e-05, 'epoch': 0.59})
+[2026-07-28 15:29:24,171][starVLA.training.train_starvla][INFO] - Step 3186, Loss: {'train/loss': 0.016758620738983154, 'train/grad_norm_pre_clip': 0.6829037070274353, 'timing/data': 0.0002873069606721401, 'timing/model': 0.33316962188109756, 'learning_rate/qwen_vl_interface': 9.988606494860811e-06, 'learning_rate/action_model': 9.988606494860812e-05, 'epoch': 0.59})
+[2026-07-28 15:29:28,291][starVLA.training.train_starvla][INFO] - Step 3187, Loss: {'train/loss': 0.051640450954437256, 'train/grad_norm_pre_clip': 0.9248554706573486, 'timing/data': 0.00028333812952041626, 'timing/model': 0.3489466910250485, 'learning_rate/qwen_vl_interface': 9.98834009821507e-06, 'learning_rate/action_model': 9.988340098215069e-05, 'epoch': 0.59})
+[2026-07-28 15:29:32,489][starVLA.training.train_starvla][INFO] - Step 3188, Loss: {'train/loss': 0.03598395735025406, 'train/grad_norm_pre_clip': 1.1871976852416992, 'timing/data': 0.00028643570840358734, 'timing/model': 0.3329767258837819, 'learning_rate/qwen_vl_interface': 9.98807062691188e-06, 'learning_rate/action_model': 9.988070626911879e-05, 'epoch': 0.59})
+[2026-07-28 15:29:36,821][starVLA.training.train_starvla][INFO] - Step 3189, Loss: {'train/loss': 0.036445144563913345, 'train/grad_norm_pre_clip': 1.609288215637207, 'timing/data': 0.0003002360463142395, 'timing/model': 0.3330541639588773, 'learning_rate/qwen_vl_interface': 9.987798081126099e-06, 'learning_rate/action_model': 9.987798081126099e-05, 'epoch': 0.59})
+[2026-07-28 15:29:40,927][starVLA.training.train_starvla][INFO] - Step 3190, Loss: {'train/loss': 0.046169131994247437, 'train/grad_norm_pre_clip': 1.2870911359786987, 'timing/data': 0.0003790389746427536, 'timing/model': 0.3322051651775837, 'learning_rate/qwen_vl_interface': 9.987522461034584e-06, 'learning_rate/action_model': 9.987522461034584e-05, 'epoch': 0.59})
+[2026-07-28 15:29:45,128][starVLA.training.train_starvla][INFO] - Step 3191, Loss: {'train/loss': 0.01689794659614563, 'train/grad_norm_pre_clip': 1.0030272006988525, 'timing/data': 0.0003570658154785633, 'timing/model': 0.45199050195515156, 'learning_rate/qwen_vl_interface': 9.987243766816175e-06, 'learning_rate/action_model': 9.987243766816176e-05, 'epoch': 0.59})
+[2026-07-28 15:29:49,308][starVLA.training.train_starvla][INFO] - Step 3192, Loss: {'train/loss': 0.03033854067325592, 'train/grad_norm_pre_clip': 1.302370548248291, 'timing/data': 0.0003849058412015438, 'timing/model': 0.33297367999330163, 'learning_rate/qwen_vl_interface': 9.98696199865172e-06, 'learning_rate/action_model': 9.98696199865172e-05, 'epoch': 0.59})
+[2026-07-28 15:29:53,444][starVLA.training.train_starvla][INFO] - Step 3193, Loss: {'train/loss': 0.07505346834659576, 'train/grad_norm_pre_clip': 3.6386048793792725, 'timing/data': 0.0007666032761335373, 'timing/model': 0.33402008889243007, 'learning_rate/qwen_vl_interface': 9.986677156724054e-06, 'learning_rate/action_model': 9.986677156724054e-05, 'epoch': 0.59})
+[2026-07-28 15:29:57,576][starVLA.training.train_starvla][INFO] - Step 3194, Loss: {'train/loss': 0.049962736666202545, 'train/grad_norm_pre_clip': 2.02109694480896, 'timing/data': 0.0003101453185081482, 'timing/model': 0.33219681307673454, 'learning_rate/qwen_vl_interface': 9.986389241218003e-06, 'learning_rate/action_model': 9.986389241218003e-05, 'epoch': 0.59})
+[2026-07-28 15:30:01,715][starVLA.training.train_starvla][INFO] - Step 3195, Loss: {'train/loss': 0.09159942716360092, 'train/grad_norm_pre_clip': 2.8378424644470215, 'timing/data': 0.00030392128974199295, 'timing/model': 0.3320946916937828, 'learning_rate/qwen_vl_interface': 9.986098252320397e-06, 'learning_rate/action_model': 9.986098252320397e-05, 'epoch': 0.59})
+[2026-07-28 15:30:05,853][starVLA.training.train_starvla][INFO] - Step 3196, Loss: {'train/loss': 0.03307632356882095, 'train/grad_norm_pre_clip': 2.0462870597839355, 'timing/data': 0.00030716415494680405, 'timing/model': 0.33279430819675326, 'learning_rate/qwen_vl_interface': 9.985804190220054e-06, 'learning_rate/action_model': 9.985804190220053e-05, 'epoch': 0.59})
+[2026-07-28 15:30:10,004][starVLA.training.train_starvla][INFO] - Step 3197, Loss: {'train/loss': 0.05690282955765724, 'train/grad_norm_pre_clip': 1.991949439048767, 'timing/data': 0.0002968539483845234, 'timing/model': 0.3338293517008424, 'learning_rate/qwen_vl_interface': 9.985507055107788e-06, 'learning_rate/action_model': 9.985507055107788e-05, 'epoch': 0.59})
+[2026-07-28 15:30:14,190][starVLA.training.train_starvla][INFO] - Step 3198, Loss: {'train/loss': 0.05741669237613678, 'train/grad_norm_pre_clip': 2.344473361968994, 'timing/data': 0.00030087120831012726, 'timing/model': 0.33277616184204817, 'learning_rate/qwen_vl_interface': 9.985206847176406e-06, 'learning_rate/action_model': 9.985206847176406e-05, 'epoch': 0.59})
+[2026-07-28 15:30:18,393][starVLA.training.train_starvla][INFO] - Step 3199, Loss: {'train/loss': 0.03227616101503372, 'train/grad_norm_pre_clip': 1.7006245851516724, 'timing/data': 0.00033141300082206726, 'timing/model': 0.33225973322987556, 'learning_rate/qwen_vl_interface': 9.984903566620711e-06, 'learning_rate/action_model': 9.984903566620712e-05, 'epoch': 0.59})
+[2026-07-28 15:30:22,503][starVLA.training.train_starvla][INFO] - Step 3200, Loss: {'train/loss': 0.05009389668703079, 'train/grad_norm_pre_clip': 1.4937077760696411, 'timing/data': 0.00048605306074023247, 'timing/model': 0.33679624926298857, 'learning_rate/qwen_vl_interface': 9.984597213637497e-06, 'learning_rate/action_model': 9.984597213637497e-05, 'epoch': 0.59})
+[2026-07-28 15:30:26,641][starVLA.training.train_starvla][INFO] - Step 3201, Loss: {'train/loss': 0.06030276417732239, 'train/grad_norm_pre_clip': 3.1374707221984863, 'timing/data': 0.0004804232157766819, 'timing/model': 0.3346671489998698, 'learning_rate/qwen_vl_interface': 9.984287788425555e-06, 'learning_rate/action_model': 9.984287788425555e-05, 'epoch': 0.59})
+[2026-07-28 15:30:30,747][starVLA.training.train_starvla][INFO] - Step 3202, Loss: {'train/loss': 0.042104873806238174, 'train/grad_norm_pre_clip': 3.0538244247436523, 'timing/data': 0.0004781740717589855, 'timing/model': 0.3353203763253987, 'learning_rate/qwen_vl_interface': 9.983975291185666e-06, 'learning_rate/action_model': 9.983975291185667e-05, 'epoch': 0.59})
+[2026-07-28 15:30:34,935][starVLA.training.train_starvla][INFO] - Step 3203, Loss: {'train/loss': 0.06503540277481079, 'train/grad_norm_pre_clip': 2.6791179180145264, 'timing/data': 0.0002977699041366577, 'timing/model': 0.33116645412519574, 'learning_rate/qwen_vl_interface': 9.983659722120607e-06, 'learning_rate/action_model': 9.983659722120608e-05, 'epoch': 0.59})
+[2026-07-28 15:30:39,040][starVLA.training.train_starvla][INFO] - Step 3204, Loss: {'train/loss': 0.02464986778795719, 'train/grad_norm_pre_clip': 0.6766848564147949, 'timing/data': 0.00030299462378025055, 'timing/model': 0.33223973400890827, 'learning_rate/qwen_vl_interface': 9.983341081435145e-06, 'learning_rate/action_model': 9.983341081435145e-05, 'epoch': 0.59})
+[2026-07-28 15:30:43,143][starVLA.training.train_starvla][INFO] - Step 3205, Loss: {'train/loss': 0.03361520171165466, 'train/grad_norm_pre_clip': 0.9308140277862549, 'timing/data': 0.0006186328828334808, 'timing/model': 0.3357400759123266, 'learning_rate/qwen_vl_interface': 9.983019369336048e-06, 'learning_rate/action_model': 9.983019369336048e-05, 'epoch': 0.59})
+[2026-07-28 15:30:47,261][starVLA.training.train_starvla][INFO] - Step 3206, Loss: {'train/loss': 0.07058476656675339, 'train/grad_norm_pre_clip': 2.1869382858276367, 'timing/data': 0.00033910805359482765, 'timing/model': 0.3372936216183007, 'learning_rate/qwen_vl_interface': 9.982694586032066e-06, 'learning_rate/action_model': 9.982694586032065e-05, 'epoch': 0.59})
+[2026-07-28 15:30:51,349][starVLA.training.train_starvla][INFO] - Step 3207, Loss: {'train/loss': 0.06499400734901428, 'train/grad_norm_pre_clip': 1.2650370597839355, 'timing/data': 0.00034399470314383507, 'timing/model': 0.334745641797781, 'learning_rate/qwen_vl_interface': 9.982366731733948e-06, 'learning_rate/action_model': 9.982366731733947e-05, 'epoch': 0.59})
+[2026-07-28 15:30:55,435][starVLA.training.train_starvla][INFO] - Step 3208, Loss: {'train/loss': 0.04783634841442108, 'train/grad_norm_pre_clip': 1.1154342889785767, 'timing/data': 0.00034773023799061775, 'timing/model': 0.3332597860135138, 'learning_rate/qwen_vl_interface': 9.982035806654437e-06, 'learning_rate/action_model': 9.982035806654437e-05, 'epoch': 0.59})
+[2026-07-28 15:30:59,831][starVLA.training.train_starvla][INFO] - Step 3209, Loss: {'train/loss': 0.051473088562488556, 'train/grad_norm_pre_clip': 1.884453535079956, 'timing/data': 0.00045171426609158516, 'timing/model': 0.33291321294382215, 'learning_rate/qwen_vl_interface': 9.981701811008267e-06, 'learning_rate/action_model': 9.981701811008267e-05, 'epoch': 0.59})
+[2026-07-28 15:31:03,963][starVLA.training.train_starvla][INFO] - Step 3210, Loss: {'train/loss': 0.026748651638627052, 'train/grad_norm_pre_clip': 0.9397324323654175, 'timing/data': 0.00030105607584118843, 'timing/model': 0.3374447822570801, 'learning_rate/qwen_vl_interface': 9.981364745012163e-06, 'learning_rate/action_model': 9.981364745012163e-05, 'epoch': 0.59})
+[2026-07-28 15:31:08,066][starVLA.training.train_starvla][INFO] - Step 3211, Loss: {'train/loss': 0.07003889977931976, 'train/grad_norm_pre_clip': 1.5117032527923584, 'timing/data': 0.000380645040422678, 'timing/model': 0.3416496771387756, 'learning_rate/qwen_vl_interface': 9.981024608884842e-06, 'learning_rate/action_model': 9.981024608884843e-05, 'epoch': 0.59})
+[2026-07-28 15:31:12,183][starVLA.training.train_starvla][INFO] - Step 3212, Loss: {'train/loss': 0.060291312634944916, 'train/grad_norm_pre_clip': 0.9431646466255188, 'timing/data': 0.00038276705890893936, 'timing/model': 0.33778786612674594, 'learning_rate/qwen_vl_interface': 9.980681402847017e-06, 'learning_rate/action_model': 9.980681402847016e-05, 'epoch': 0.59})
+[2026-07-28 15:31:16,483][starVLA.training.train_starvla][INFO] - Step 3213, Loss: {'train/loss': 0.037816718220710754, 'train/grad_norm_pre_clip': 1.161812424659729, 'timing/data': 0.00034738704562187195, 'timing/model': 0.33353572711348534, 'learning_rate/qwen_vl_interface': 9.980335127121391e-06, 'learning_rate/action_model': 9.98033512712139e-05, 'epoch': 0.59})
+[2026-07-28 15:31:20,597][starVLA.training.train_starvla][INFO] - Step 3214, Loss: {'train/loss': 0.04720362275838852, 'train/grad_norm_pre_clip': 2.033906936645508, 'timing/data': 0.00032400013878941536, 'timing/model': 0.33961535384878516, 'learning_rate/qwen_vl_interface': 9.979985781932658e-06, 'learning_rate/action_model': 9.979985781932657e-05, 'epoch': 0.59})
+[2026-07-28 15:31:24,699][starVLA.training.train_starvla][INFO] - Step 3215, Loss: {'train/loss': 0.022642185911536217, 'train/grad_norm_pre_clip': 0.544370174407959, 'timing/data': 0.0003034481778740883, 'timing/model': 0.33553112437948585, 'learning_rate/qwen_vl_interface': 9.979633367507501e-06, 'learning_rate/action_model': 9.979633367507502e-05, 'epoch': 0.59})
+[2026-07-28 15:31:28,804][starVLA.training.train_starvla][INFO] - Step 3216, Loss: {'train/loss': 0.04324553161859512, 'train/grad_norm_pre_clip': 1.3562909364700317, 'timing/data': 0.00028744013980031013, 'timing/model': 0.3340573678724468, 'learning_rate/qwen_vl_interface': 9.979277884074604e-06, 'learning_rate/action_model': 9.979277884074604e-05, 'epoch': 0.59})
+[2026-07-28 15:31:32,892][starVLA.training.train_starvla][INFO] - Step 3217, Loss: {'train/loss': 0.037959031760692596, 'train/grad_norm_pre_clip': 1.8017408847808838, 'timing/data': 0.0002981778234243393, 'timing/model': 0.3352168509736657, 'learning_rate/qwen_vl_interface': 9.978919331864631e-06, 'learning_rate/action_model': 9.97891933186463e-05, 'epoch': 0.59})
+[2026-07-28 15:31:37,018][starVLA.training.train_starvla][INFO] - Step 3218, Loss: {'train/loss': 0.046127475798130035, 'train/grad_norm_pre_clip': 1.1877853870391846, 'timing/data': 0.0003251209855079651, 'timing/model': 0.33248657127842307, 'learning_rate/qwen_vl_interface': 9.978557711110245e-06, 'learning_rate/action_model': 9.978557711110244e-05, 'epoch': 0.59})
+[2026-07-28 15:31:41,103][starVLA.training.train_starvla][INFO] - Step 3219, Loss: {'train/loss': 0.023970462381839752, 'train/grad_norm_pre_clip': 0.4717332422733307, 'timing/data': 0.00029633333906531334, 'timing/model': 0.3334049950353801, 'learning_rate/qwen_vl_interface': 9.978193022046097e-06, 'learning_rate/action_model': 9.978193022046097e-05, 'epoch': 0.59})
+[2026-07-28 15:31:45,192][starVLA.training.train_starvla][INFO] - Step 3220, Loss: {'train/loss': 0.05668775737285614, 'train/grad_norm_pre_clip': 1.4686347246170044, 'timing/data': 0.0007512890733778477, 'timing/model': 0.3332586991600692, 'learning_rate/qwen_vl_interface': 9.977825264908832e-06, 'learning_rate/action_model': 9.977825264908832e-05, 'epoch': 0.59})
+[2026-07-28 15:31:49,581][starVLA.training.train_starvla][INFO] - Step 3221, Loss: {'train/loss': 0.045547965914011, 'train/grad_norm_pre_clip': 1.4720796346664429, 'timing/data': 0.00035548675805330276, 'timing/model': 0.3331249179318547, 'learning_rate/qwen_vl_interface': 9.97745443993708e-06, 'learning_rate/action_model': 9.97745443993708e-05, 'epoch': 0.59})
+[2026-07-28 15:31:53,668][starVLA.training.train_starvla][INFO] - Step 3222, Loss: {'train/loss': 0.015387384220957756, 'train/grad_norm_pre_clip': 0.5986495018005371, 'timing/data': 0.0005446411669254303, 'timing/model': 0.3321351897902787, 'learning_rate/qwen_vl_interface': 9.977080547371468e-06, 'learning_rate/action_model': 9.977080547371468e-05, 'epoch': 0.59})
+[2026-07-28 15:31:57,751][starVLA.training.train_starvla][INFO] - Step 3223, Loss: {'train/loss': 0.018174629658460617, 'train/grad_norm_pre_clip': 1.0916776657104492, 'timing/data': 0.0003120601177215576, 'timing/model': 0.3349146521650255, 'learning_rate/qwen_vl_interface': 9.97670358745461e-06, 'learning_rate/action_model': 9.97670358745461e-05, 'epoch': 0.59})
+[2026-07-28 15:32:01,968][starVLA.training.train_starvla][INFO] - Step 3224, Loss: {'train/loss': 0.058591075241565704, 'train/grad_norm_pre_clip': 2.6935062408447266, 'timing/data': 0.0003388170152902603, 'timing/model': 0.3337309737689793, 'learning_rate/qwen_vl_interface': 9.976323560431112e-06, 'learning_rate/action_model': 9.976323560431111e-05, 'epoch': 0.59})
+[2026-07-28 15:32:06,074][starVLA.training.train_starvla][INFO] - Step 3225, Loss: {'train/loss': 0.019590897485613823, 'train/grad_norm_pre_clip': 1.5108546018600464, 'timing/data': 0.0003101709298789501, 'timing/model': 0.3361002239398658, 'learning_rate/qwen_vl_interface': 9.975940466547567e-06, 'learning_rate/action_model': 9.975940466547567e-05, 'epoch': 0.59})
+[2026-07-28 15:32:10,168][starVLA.training.train_starvla][INFO] - Step 3226, Loss: {'train/loss': 0.05699312686920166, 'train/grad_norm_pre_clip': 2.471686601638794, 'timing/data': 0.0003442317247390747, 'timing/model': 0.33296544663608074, 'learning_rate/qwen_vl_interface': 9.975554306052563e-06, 'learning_rate/action_model': 9.975554306052562e-05, 'epoch': 0.59})
+[2026-07-28 15:32:14,271][starVLA.training.train_starvla][INFO] - Step 3227, Loss: {'train/loss': 0.024842452257871628, 'train/grad_norm_pre_clip': 1.2427306175231934, 'timing/data': 0.00030247773975133896, 'timing/model': 0.3424541843123734, 'learning_rate/qwen_vl_interface': 9.975165079196674e-06, 'learning_rate/action_model': 9.975165079196673e-05, 'epoch': 0.59})
+[2026-07-28 15:32:18,354][starVLA.training.train_starvla][INFO] - Step 3228, Loss: {'train/loss': 0.04411645978689194, 'train/grad_norm_pre_clip': 1.4872618913650513, 'timing/data': 0.00031464220955967903, 'timing/model': 0.33324888767674565, 'learning_rate/qwen_vl_interface': 9.974772786232466e-06, 'learning_rate/action_model': 9.974772786232466e-05, 'epoch': 0.59})
+[2026-07-28 15:32:22,479][starVLA.training.train_starvla][INFO] - Step 3229, Loss: {'train/loss': 0.04355659335851669, 'train/grad_norm_pre_clip': 1.7071945667266846, 'timing/data': 0.000544142909348011, 'timing/model': 0.33500041300430894, 'learning_rate/qwen_vl_interface': 9.974377427414493e-06, 'learning_rate/action_model': 9.974377427414493e-05, 'epoch': 0.59})
+[2026-07-28 15:32:26,708][starVLA.training.train_starvla][INFO] - Step 3230, Loss: {'train/loss': 0.020090464502573013, 'train/grad_norm_pre_clip': 1.10256826877594, 'timing/data': 0.00031291181221604347, 'timing/model': 0.34433177672326565, 'learning_rate/qwen_vl_interface': 9.973979002999299e-06, 'learning_rate/action_model': 9.973979002999299e-05, 'epoch': 0.59})
+[2026-07-28 15:32:30,985][starVLA.training.train_starvla][INFO] - Step 3231, Loss: {'train/loss': 0.060745421797037125, 'train/grad_norm_pre_clip': 2.8018152713775635, 'timing/data': 0.0002741948701441288, 'timing/model': 0.33257319312542677, 'learning_rate/qwen_vl_interface': 9.973577513245417e-06, 'learning_rate/action_model': 9.973577513245417e-05, 'epoch': 0.59})
+[2026-07-28 15:32:35,214][starVLA.training.train_starvla][INFO] - Step 3232, Loss: {'train/loss': 0.037794098258018494, 'train/grad_norm_pre_clip': 1.481407642364502, 'timing/data': 0.00028828810900449753, 'timing/model': 0.33319599088281393, 'learning_rate/qwen_vl_interface': 9.973172958413372e-06, 'learning_rate/action_model': 9.97317295841337e-05, 'epoch': 0.59})
+[2026-07-28 15:32:39,421][starVLA.training.train_starvla][INFO] - Step 3233, Loss: {'train/loss': 0.031941063702106476, 'train/grad_norm_pre_clip': 0.6817057132720947, 'timing/data': 0.00036312080919742584, 'timing/model': 0.3339705793187022, 'learning_rate/qwen_vl_interface': 9.97276533876567e-06, 'learning_rate/action_model': 9.97276533876567e-05, 'epoch': 0.59})
+[2026-07-28 15:32:43,618][starVLA.training.train_starvla][INFO] - Step 3234, Loss: {'train/loss': 0.045167747884988785, 'train/grad_norm_pre_clip': 1.6300393342971802, 'timing/data': 0.00037889275699853897, 'timing/model': 0.33231381606310606, 'learning_rate/qwen_vl_interface': 9.972354654566816e-06, 'learning_rate/action_model': 9.972354654566816e-05, 'epoch': 0.59})
+[2026-07-28 15:32:47,717][starVLA.training.train_starvla][INFO] - Step 3235, Loss: {'train/loss': 0.05202488601207733, 'train/grad_norm_pre_clip': 1.8700608015060425, 'timing/data': 0.00045776693150401115, 'timing/model': 0.333229070995003, 'learning_rate/qwen_vl_interface': 9.971940906083295e-06, 'learning_rate/action_model': 9.971940906083295e-05, 'epoch': 0.59})
+[2026-07-28 15:32:51,889][starVLA.training.train_starvla][INFO] - Step 3236, Loss: {'train/loss': 0.029064243659377098, 'train/grad_norm_pre_clip': 0.6615384817123413, 'timing/data': 0.00037270598113536835, 'timing/model': 0.3356745750643313, 'learning_rate/qwen_vl_interface': 9.971524093583586e-06, 'learning_rate/action_model': 9.971524093583586e-05, 'epoch': 0.59})
+[2026-07-28 15:32:56,055][starVLA.training.train_starvla][INFO] - Step 3237, Loss: {'train/loss': 0.06943285465240479, 'train/grad_norm_pre_clip': 2.27160906791687, 'timing/data': 0.0003543747588992119, 'timing/model': 0.33283646404743195, 'learning_rate/qwen_vl_interface': 9.971104217338154e-06, 'learning_rate/action_model': 9.971104217338153e-05, 'epoch': 0.59})
+[2026-07-28 15:33:00,138][starVLA.training.train_starvla][INFO] - Step 3238, Loss: {'train/loss': 0.030889760702848434, 'train/grad_norm_pre_clip': 1.2133324146270752, 'timing/data': 0.0003046290948987007, 'timing/model': 0.3337133442983031, 'learning_rate/qwen_vl_interface': 9.970681277619452e-06, 'learning_rate/action_model': 9.970681277619452e-05, 'epoch': 0.59})
+[2026-07-28 15:33:04,223][starVLA.training.train_starvla][INFO] - Step 3239, Loss: {'train/loss': 0.03733063116669655, 'train/grad_norm_pre_clip': 1.3120512962341309, 'timing/data': 0.0003430210053920746, 'timing/model': 0.3322394760325551, 'learning_rate/qwen_vl_interface': 9.97025527470192e-06, 'learning_rate/action_model': 9.970255274701919e-05, 'epoch': 0.59})
+[2026-07-28 15:33:08,314][starVLA.training.train_starvla][INFO] - Step 3240, Loss: {'train/loss': 0.03841503709554672, 'train/grad_norm_pre_clip': 0.81699138879776, 'timing/data': 0.0003895307891070843, 'timing/model': 0.3342922627925873, 'learning_rate/qwen_vl_interface': 9.969826208861987e-06, 'learning_rate/action_model': 9.969826208861987e-05, 'epoch': 0.6})
+[2026-07-28 15:33:12,440][starVLA.training.train_starvla][INFO] - Step 3241, Loss: {'train/loss': 0.09273958206176758, 'train/grad_norm_pre_clip': 3.658566474914551, 'timing/data': 0.000304302666336298, 'timing/model': 0.334548267070204, 'learning_rate/qwen_vl_interface': 9.96939408037807e-06, 'learning_rate/action_model': 9.96939408037807e-05, 'epoch': 0.6})
+[2026-07-28 15:33:16,534][starVLA.training.train_starvla][INFO] - Step 3242, Loss: {'train/loss': 0.026206515729427338, 'train/grad_norm_pre_clip': 0.7007870674133301, 'timing/data': 0.00036186305806040764, 'timing/model': 0.3344878298230469, 'learning_rate/qwen_vl_interface': 9.968958889530573e-06, 'learning_rate/action_model': 9.968958889530573e-05, 'epoch': 0.6})
+[2026-07-28 15:33:20,666][starVLA.training.train_starvla][INFO] - Step 3243, Loss: {'train/loss': 0.05096883326768875, 'train/grad_norm_pre_clip': 1.244310975074768, 'timing/data': 0.000493606086820364, 'timing/model': 0.33361203921958804, 'learning_rate/qwen_vl_interface': 9.968520636601883e-06, 'learning_rate/action_model': 9.968520636601883e-05, 'epoch': 0.6})
+[2026-07-28 15:33:24,782][starVLA.training.train_starvla][INFO] - Step 3244, Loss: {'train/loss': 0.03417648747563362, 'train/grad_norm_pre_clip': 1.1271108388900757, 'timing/data': 0.000294386874884367, 'timing/model': 0.34304401092231274, 'learning_rate/qwen_vl_interface': 9.968079321876382e-06, 'learning_rate/action_model': 9.968079321876382e-05, 'epoch': 0.6})
+[2026-07-28 15:33:28,912][starVLA.training.train_starvla][INFO] - Step 3245, Loss: {'train/loss': 0.04027639329433441, 'train/grad_norm_pre_clip': 1.3399852514266968, 'timing/data': 0.00033429916948080063, 'timing/model': 0.33417190983891487, 'learning_rate/qwen_vl_interface': 9.967634945640432e-06, 'learning_rate/action_model': 9.967634945640431e-05, 'epoch': 0.6})
+[2026-07-28 15:33:33,068][starVLA.training.train_starvla][INFO] - Step 3246, Loss: {'train/loss': 0.07612371444702148, 'train/grad_norm_pre_clip': 1.493475317955017, 'timing/data': 0.00035246508195996284, 'timing/model': 0.3327938332222402, 'learning_rate/qwen_vl_interface': 9.967187508182385e-06, 'learning_rate/action_model': 9.967187508182385e-05, 'epoch': 0.6})
+[2026-07-28 15:33:37,211][starVLA.training.train_starvla][INFO] - Step 3247, Loss: {'train/loss': 0.037054091691970825, 'train/grad_norm_pre_clip': 0.8031986355781555, 'timing/data': 0.0008725319057703018, 'timing/model': 0.33395821461454034, 'learning_rate/qwen_vl_interface': 9.966737009792577e-06, 'learning_rate/action_model': 9.966737009792577e-05, 'epoch': 0.6})
+[2026-07-28 15:33:41,377][starVLA.training.train_starvla][INFO] - Step 3248, Loss: {'train/loss': 0.056618623435497284, 'train/grad_norm_pre_clip': 2.337273597717285, 'timing/data': 0.0002771611325442791, 'timing/model': 0.3537114281207323, 'learning_rate/qwen_vl_interface': 9.966283450763333e-06, 'learning_rate/action_model': 9.966283450763333e-05, 'epoch': 0.6})
+[2026-07-28 15:33:45,513][starVLA.training.train_starvla][INFO] - Step 3249, Loss: {'train/loss': 0.11067633330821991, 'train/grad_norm_pre_clip': 2.413154125213623, 'timing/data': 0.00029766419902443886, 'timing/model': 0.332651627715677, 'learning_rate/qwen_vl_interface': 9.965826831388962e-06, 'learning_rate/action_model': 9.965826831388961e-05, 'epoch': 0.6})
+[2026-07-28 15:33:49,626][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3250
+[2026-07-28 15:35:02,379][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3250
+[2026-07-28 15:42:53,765][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3250
+[2026-07-28 15:42:53,766][starVLA.training.train_starvla][INFO] - Step 3250, Loss: {'train/loss': 0.04516266658902168, 'train/grad_norm_pre_clip': 0.7871258854866028, 'eval/action_loss/seconds': 72.75030690804124, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.03849292919039726, 'eval/latency_3/loss': 0.03849292919039726, 'eval/flappy/loss': 0.03849292919039726, 'eval/flappy/latency_3/loss': 0.03849292919039726, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 398.7300046890974, 'rl_games_eval/mid_train/mean_length': 3231.45, 'rl_games_eval/mid_train/std_reward': 83.87884276974052, 'rl_games_eval/mid_train/std_length': 674.9740346857797, 'rl_games_eval/mid_train/macro_mean_reward': 398.7300046890974, 'rl_games_eval/mid_train/macro_mean_length': 3231.45, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 398.7300046890974, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 3231.45, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 83.87884276974052, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 674.9740346857797, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 398.7300046890974, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0002833106555044651, 'timing/model': 0.33442204678431153, 'learning_rate/qwen_vl_interface': 9.965367151965757e-06, 'learning_rate/action_model': 9.965367151965756e-05, 'epoch': 0.6})
+[2026-07-28 15:43:00,001][starVLA.training.train_starvla][INFO] - Step 3251, Loss: {'train/loss': 0.026804102584719658, 'train/grad_norm_pre_clip': 0.7699863910675049, 'timing/data': 0.00029164692386984825, 'timing/model': 0.4244580599479377, 'learning_rate/qwen_vl_interface': 9.964904412792002e-06, 'learning_rate/action_model': 9.964904412792002e-05, 'epoch': 0.6})
+[2026-07-28 15:43:06,002][starVLA.training.train_starvla][INFO] - Step 3252, Loss: {'train/loss': 0.03796039894223213, 'train/grad_norm_pre_clip': 1.191038966178894, 'timing/data': 0.00033478857949376106, 'timing/model': 0.4213465084321797, 'learning_rate/qwen_vl_interface': 9.964438614167964e-06, 'learning_rate/action_model': 9.964438614167964e-05, 'epoch': 0.6})
+[2026-07-28 15:43:12,329][starVLA.training.train_starvla][INFO] - Step 3253, Loss: {'train/loss': 0.024412851780653, 'train/grad_norm_pre_clip': 1.110327959060669, 'timing/data': 0.00026301387697458267, 'timing/model': 0.42791877686977386, 'learning_rate/qwen_vl_interface': 9.963969756395892e-06, 'learning_rate/action_model': 9.96396975639589e-05, 'epoch': 0.6})
+[2026-07-28 15:43:18,614][starVLA.training.train_starvla][INFO] - Step 3254, Loss: {'train/loss': 0.03333892673254013, 'train/grad_norm_pre_clip': 1.146895170211792, 'timing/data': 0.0002748807892203331, 'timing/model': 0.6704255333170295, 'learning_rate/qwen_vl_interface': 9.963497839780025e-06, 'learning_rate/action_model': 9.963497839780024e-05, 'epoch': 0.6})
+[2026-07-28 15:43:24,599][starVLA.training.train_starvla][INFO] - Step 3255, Loss: {'train/loss': 0.03712726756930351, 'train/grad_norm_pre_clip': 1.8452041149139404, 'timing/data': 0.0003254879266023636, 'timing/model': 0.4260450410656631, 'learning_rate/qwen_vl_interface': 9.963022864626582e-06, 'learning_rate/action_model': 9.963022864626581e-05, 'epoch': 0.6})
+[2026-07-28 15:43:30,856][starVLA.training.train_starvla][INFO] - Step 3256, Loss: {'train/loss': 0.06164021044969559, 'train/grad_norm_pre_clip': 1.5002723932266235, 'timing/data': 0.00027457624673843384, 'timing/model': 0.6531712128780782, 'learning_rate/qwen_vl_interface': 9.96254483124377e-06, 'learning_rate/action_model': 9.96254483124377e-05, 'epoch': 0.6})
+[2026-07-28 15:43:36,957][starVLA.training.train_starvla][INFO] - Step 3257, Loss: {'train/loss': 0.05229981988668442, 'train/grad_norm_pre_clip': 1.1542744636535645, 'timing/data': 0.0002872222103178501, 'timing/model': 0.42737739719450474, 'learning_rate/qwen_vl_interface': 9.962063739941783e-06, 'learning_rate/action_model': 9.962063739941783e-05, 'epoch': 0.6})
+[2026-07-28 15:43:43,202][starVLA.training.train_starvla][INFO] - Step 3258, Loss: {'train/loss': 0.043552421033382416, 'train/grad_norm_pre_clip': 2.405134677886963, 'timing/data': 0.00030620209872722626, 'timing/model': 0.4249072363600135, 'learning_rate/qwen_vl_interface': 9.961579591032791e-06, 'learning_rate/action_model': 9.96157959103279e-05, 'epoch': 0.6})
+[2026-07-28 15:43:49,297][starVLA.training.train_starvla][INFO] - Step 3259, Loss: {'train/loss': 0.02862001582980156, 'train/grad_norm_pre_clip': 0.9496915936470032, 'timing/data': 0.0002659992314875126, 'timing/model': 0.4301180182956159, 'learning_rate/qwen_vl_interface': 9.961092384830956e-06, 'learning_rate/action_model': 9.961092384830956e-05, 'epoch': 0.6})
+[2026-07-28 15:43:55,376][starVLA.training.train_starvla][INFO] - Step 3260, Loss: {'train/loss': 0.0803302749991417, 'train/grad_norm_pre_clip': 3.098501443862915, 'timing/data': 0.00039725610986351967, 'timing/model': 0.42917621321976185, 'learning_rate/qwen_vl_interface': 9.96060212165242e-06, 'learning_rate/action_model': 9.96060212165242e-05, 'epoch': 0.6})
+[2026-07-28 15:44:01,622][starVLA.training.train_starvla][INFO] - Step 3261, Loss: {'train/loss': 0.02840026654303074, 'train/grad_norm_pre_clip': 1.4715094566345215, 'timing/data': 0.00025375187397003174, 'timing/model': 0.6512664295732975, 'learning_rate/qwen_vl_interface': 9.96010880181531e-06, 'learning_rate/action_model': 9.960108801815309e-05, 'epoch': 0.6})
+[2026-07-28 15:44:07,655][starVLA.training.train_starvla][INFO] - Step 3262, Loss: {'train/loss': 0.06213614344596863, 'train/grad_norm_pre_clip': 1.5195306539535522, 'timing/data': 0.0002517509274184704, 'timing/model': 0.42390365712344646, 'learning_rate/qwen_vl_interface': 9.959612425639736e-06, 'learning_rate/action_model': 9.959612425639735e-05, 'epoch': 0.6})
+[2026-07-28 15:44:13,663][starVLA.training.train_starvla][INFO] - Step 3263, Loss: {'train/loss': 0.0226697139441967, 'train/grad_norm_pre_clip': 1.3599170446395874, 'timing/data': 0.0003313487395644188, 'timing/model': 0.42286442779004574, 'learning_rate/qwen_vl_interface': 9.95911299344779e-06, 'learning_rate/action_model': 9.959112993447789e-05, 'epoch': 0.6})
+[2026-07-28 15:44:19,900][starVLA.training.train_starvla][INFO] - Step 3264, Loss: {'train/loss': 0.0402042493224144, 'train/grad_norm_pre_clip': 1.4594788551330566, 'timing/data': 0.00030461978167295456, 'timing/model': 0.6703972178511322, 'learning_rate/qwen_vl_interface': 9.958610505563546e-06, 'learning_rate/action_model': 9.958610505563546e-05, 'epoch': 0.6})
+[2026-07-28 15:44:25,949][starVLA.training.train_starvla][INFO] - Step 3265, Loss: {'train/loss': 0.06418730318546295, 'train/grad_norm_pre_clip': 1.6323094367980957, 'timing/data': 0.00027884775772690773, 'timing/model': 0.4370192512869835, 'learning_rate/qwen_vl_interface': 9.95810496231307e-06, 'learning_rate/action_model': 9.95810496231307e-05, 'epoch': 0.6})
+[2026-07-28 15:44:32,061][starVLA.training.train_starvla][INFO] - Step 3266, Loss: {'train/loss': 0.018362483009696007, 'train/grad_norm_pre_clip': 0.6281144022941589, 'timing/data': 0.00038362620398402214, 'timing/model': 0.4312651939690113, 'learning_rate/qwen_vl_interface': 9.957596364024396e-06, 'learning_rate/action_model': 9.957596364024396e-05, 'epoch': 0.6})
+[2026-07-28 15:44:38,340][starVLA.training.train_starvla][INFO] - Step 3267, Loss: {'train/loss': 0.06729291379451752, 'train/grad_norm_pre_clip': 1.4595365524291992, 'timing/data': 0.0003551412373781204, 'timing/model': 0.6725585274398327, 'learning_rate/qwen_vl_interface': 9.957084711027553e-06, 'learning_rate/action_model': 9.957084711027552e-05, 'epoch': 0.6})
+[2026-07-28 15:44:44,337][starVLA.training.train_starvla][INFO] - Step 3268, Loss: {'train/loss': 0.024730198085308075, 'train/grad_norm_pre_clip': 0.7277894020080566, 'timing/data': 0.0002588629722595215, 'timing/model': 0.42287945095449686, 'learning_rate/qwen_vl_interface': 9.956570003654544e-06, 'learning_rate/action_model': 9.956570003654544e-05, 'epoch': 0.6})
+[2026-07-28 15:44:50,328][starVLA.training.train_starvla][INFO] - Step 3269, Loss: {'train/loss': 0.07399237155914307, 'train/grad_norm_pre_clip': 2.9806010723114014, 'timing/data': 0.00031099608168005943, 'timing/model': 0.4215953089296818, 'learning_rate/qwen_vl_interface': 9.956052242239358e-06, 'learning_rate/action_model': 9.956052242239358e-05, 'epoch': 0.6})
+[2026-07-28 15:44:56,321][starVLA.training.train_starvla][INFO] - Step 3270, Loss: {'train/loss': 0.05409298092126846, 'train/grad_norm_pre_clip': 2.0144872665405273, 'timing/data': 0.00033211521804332733, 'timing/model': 0.4210198661312461, 'learning_rate/qwen_vl_interface': 9.955531427117964e-06, 'learning_rate/action_model': 9.955531427117964e-05, 'epoch': 0.6})
+[2026-07-28 15:45:02,571][starVLA.training.train_starvla][INFO] - Step 3271, Loss: {'train/loss': 0.028598470613360405, 'train/grad_norm_pre_clip': 2.261032819747925, 'timing/data': 0.00028954586014151573, 'timing/model': 0.4230910949409008, 'learning_rate/qwen_vl_interface': 9.955007558628316e-06, 'learning_rate/action_model': 9.955007558628316e-05, 'epoch': 0.6})
+[2026-07-28 15:45:08,891][starVLA.training.train_starvla][INFO] - Step 3272, Loss: {'train/loss': 0.05324436351656914, 'train/grad_norm_pre_clip': 2.5399885177612305, 'timing/data': 0.00026534078642725945, 'timing/model': 0.4224128071218729, 'learning_rate/qwen_vl_interface': 9.954480637110346e-06, 'learning_rate/action_model': 9.954480637110344e-05, 'epoch': 0.6})
+[2026-07-28 15:45:14,895][starVLA.training.train_starvla][INFO] - Step 3273, Loss: {'train/loss': 0.06705747544765472, 'train/grad_norm_pre_clip': 1.5821398496627808, 'timing/data': 0.0006111077964305878, 'timing/model': 0.4216240248642862, 'learning_rate/qwen_vl_interface': 9.953950662905963e-06, 'learning_rate/action_model': 9.953950662905964e-05, 'epoch': 0.6})
+[2026-07-28 15:45:21,077][starVLA.training.train_starvla][INFO] - Step 3274, Loss: {'train/loss': 0.03278307616710663, 'train/grad_norm_pre_clip': 0.8862614631652832, 'timing/data': 0.0005223439075052738, 'timing/model': 0.6102098259143531, 'learning_rate/qwen_vl_interface': 9.953417636359067e-06, 'learning_rate/action_model': 9.953417636359067e-05, 'epoch': 0.6})
+[2026-07-28 15:45:27,090][starVLA.training.train_starvla][INFO] - Step 3275, Loss: {'train/loss': 0.030525919049978256, 'train/grad_norm_pre_clip': 1.3528432846069336, 'timing/data': 0.0003001452423632145, 'timing/model': 0.41906946897506714, 'learning_rate/qwen_vl_interface': 9.95288155781553e-06, 'learning_rate/action_model': 9.95288155781553e-05, 'epoch': 0.6})
+[2026-07-28 15:45:33,350][starVLA.training.train_starvla][INFO] - Step 3276, Loss: {'train/loss': 0.04287395626306534, 'train/grad_norm_pre_clip': 1.0669612884521484, 'timing/data': 0.00025841686874628067, 'timing/model': 0.4326199349015951, 'learning_rate/qwen_vl_interface': 9.952342427623209e-06, 'learning_rate/action_model': 9.952342427623208e-05, 'epoch': 0.6})
+[2026-07-28 15:45:39,374][starVLA.training.train_starvla][INFO] - Step 3277, Loss: {'train/loss': 0.05197301506996155, 'train/grad_norm_pre_clip': 1.969988226890564, 'timing/data': 0.000355797354131937, 'timing/model': 0.4232913888990879, 'learning_rate/qwen_vl_interface': 9.95180024613194e-06, 'learning_rate/action_model': 9.951800246131939e-05, 'epoch': 0.6})
+[2026-07-28 15:45:45,628][starVLA.training.train_starvla][INFO] - Step 3278, Loss: {'train/loss': 0.03861560672521591, 'train/grad_norm_pre_clip': 1.4137957096099854, 'timing/data': 0.0002907402813434601, 'timing/model': 0.43762995628640056, 'learning_rate/qwen_vl_interface': 9.951255013693537e-06, 'learning_rate/action_model': 9.951255013693536e-05, 'epoch': 0.6})
+[2026-07-28 15:45:51,694][starVLA.training.train_starvla][INFO] - Step 3279, Loss: {'train/loss': 0.03618403524160385, 'train/grad_norm_pre_clip': 1.5871134996414185, 'timing/data': 0.0002852887846529484, 'timing/model': 0.42289694398641586, 'learning_rate/qwen_vl_interface': 9.950706730661797e-06, 'learning_rate/action_model': 9.950706730661797e-05, 'epoch': 0.6})
+[2026-07-28 15:45:57,807][starVLA.training.train_starvla][INFO] - Step 3280, Loss: {'train/loss': 0.023235293105244637, 'train/grad_norm_pre_clip': 1.0043708086013794, 'timing/data': 0.0003349659964442253, 'timing/model': 0.4426186429336667, 'learning_rate/qwen_vl_interface': 9.950155397392494e-06, 'learning_rate/action_model': 9.950155397392493e-05, 'epoch': 0.6})
+[2026-07-28 15:46:04,099][starVLA.training.train_starvla][INFO] - Step 3281, Loss: {'train/loss': 0.03776603192090988, 'train/grad_norm_pre_clip': 1.4001953601837158, 'timing/data': 0.00026747677475214005, 'timing/model': 0.6696043838746846, 'learning_rate/qwen_vl_interface': 9.949601014243382e-06, 'learning_rate/action_model': 9.949601014243382e-05, 'epoch': 0.6})
+[2026-07-28 15:46:10,128][starVLA.training.train_starvla][INFO] - Step 3282, Loss: {'train/loss': 0.02925814501941204, 'train/grad_norm_pre_clip': 0.9584765434265137, 'timing/data': 0.0002881670370697975, 'timing/model': 0.43632453214377165, 'learning_rate/qwen_vl_interface': 9.949043581574196e-06, 'learning_rate/action_model': 9.949043581574196e-05, 'epoch': 0.6})
+[2026-07-28 15:46:16,123][starVLA.training.train_starvla][INFO] - Step 3283, Loss: {'train/loss': 0.05260653793811798, 'train/grad_norm_pre_clip': 0.9583831429481506, 'timing/data': 0.00029662298038601875, 'timing/model': 0.4314849767833948, 'learning_rate/qwen_vl_interface': 9.948483099746645e-06, 'learning_rate/action_model': 9.948483099746645e-05, 'epoch': 0.6})
+[2026-07-28 15:46:22,150][starVLA.training.train_starvla][INFO] - Step 3284, Loss: {'train/loss': 0.054818231612443924, 'train/grad_norm_pre_clip': 0.9783283472061157, 'timing/data': 0.00040168408304452896, 'timing/model': 0.4222185588441789, 'learning_rate/qwen_vl_interface': 9.947919569124423e-06, 'learning_rate/action_model': 9.947919569124422e-05, 'epoch': 0.6})
+[2026-07-28 15:46:28,166][starVLA.training.train_starvla][INFO] - Step 3285, Loss: {'train/loss': 0.08274263143539429, 'train/grad_norm_pre_clip': 1.9015648365020752, 'timing/data': 0.00039075687527656555, 'timing/model': 0.4333917610347271, 'learning_rate/qwen_vl_interface': 9.947352990073196e-06, 'learning_rate/action_model': 9.947352990073195e-05, 'epoch': 0.6})
+[2026-07-28 15:46:34,434][starVLA.training.train_starvla][INFO] - Step 3286, Loss: {'train/loss': 0.04509073123335838, 'train/grad_norm_pre_clip': 1.6742531061172485, 'timing/data': 0.0003382437862455845, 'timing/model': 0.4280875981785357, 'learning_rate/qwen_vl_interface': 9.946783362960612e-06, 'learning_rate/action_model': 9.946783362960612e-05, 'epoch': 0.6})
+[2026-07-28 15:46:40,454][starVLA.training.train_starvla][INFO] - Step 3287, Loss: {'train/loss': 0.018494442105293274, 'train/grad_norm_pre_clip': 0.8597998023033142, 'timing/data': 0.0003522648476064205, 'timing/model': 0.42327150283381343, 'learning_rate/qwen_vl_interface': 9.946210688156297e-06, 'learning_rate/action_model': 9.946210688156297e-05, 'epoch': 0.6})
+[2026-07-28 15:46:46,713][starVLA.training.train_starvla][INFO] - Step 3288, Loss: {'train/loss': 0.05704883486032486, 'train/grad_norm_pre_clip': 1.4423459768295288, 'timing/data': 0.0002863309346139431, 'timing/model': 0.421941549051553, 'learning_rate/qwen_vl_interface': 9.945634966031852e-06, 'learning_rate/action_model': 9.945634966031851e-05, 'epoch': 0.6})
+[2026-07-28 15:46:52,760][starVLA.training.train_starvla][INFO] - Step 3289, Loss: {'train/loss': 0.059614088386297226, 'train/grad_norm_pre_clip': 2.286968946456909, 'timing/data': 0.00031828926876187325, 'timing/model': 0.42116132099181414, 'learning_rate/qwen_vl_interface': 9.945056196960858e-06, 'learning_rate/action_model': 9.945056196960858e-05, 'epoch': 0.6})
+[2026-07-28 15:46:58,766][starVLA.training.train_starvla][INFO] - Step 3290, Loss: {'train/loss': 0.02460930123925209, 'train/grad_norm_pre_clip': 1.155144453048706, 'timing/data': 0.0003717946819961071, 'timing/model': 0.42073530703783035, 'learning_rate/qwen_vl_interface': 9.944474381318871e-06, 'learning_rate/action_model': 9.94447438131887e-05, 'epoch': 0.6})
+[2026-07-28 15:47:05,085][starVLA.training.train_starvla][INFO] - Step 3291, Loss: {'train/loss': 0.03037748858332634, 'train/grad_norm_pre_clip': 1.2606239318847656, 'timing/data': 0.0003061862662434578, 'timing/model': 0.661213044077158, 'learning_rate/qwen_vl_interface': 9.943889519483424e-06, 'learning_rate/action_model': 9.943889519483424e-05, 'epoch': 0.6})
+[2026-07-28 15:47:11,418][starVLA.training.train_starvla][INFO] - Step 3292, Loss: {'train/loss': 0.04184790700674057, 'train/grad_norm_pre_clip': 1.430362343788147, 'timing/data': 0.000395218376070261, 'timing/model': 0.6610980182886124, 'learning_rate/qwen_vl_interface': 9.94330161183403e-06, 'learning_rate/action_model': 9.94330161183403e-05, 'epoch': 0.6})
+[2026-07-28 15:47:17,464][starVLA.training.train_starvla][INFO] - Step 3293, Loss: {'train/loss': 0.05607617273926735, 'train/grad_norm_pre_clip': 1.530781626701355, 'timing/data': 0.00035028625279664993, 'timing/model': 0.4300347329117358, 'learning_rate/qwen_vl_interface': 9.942710658752175e-06, 'learning_rate/action_model': 9.942710658752174e-05, 'epoch': 0.6})
+[2026-07-28 15:47:23,483][starVLA.training.train_starvla][INFO] - Step 3294, Loss: {'train/loss': 0.038129523396492004, 'train/grad_norm_pre_clip': 1.3617894649505615, 'timing/data': 0.00038183387368917465, 'timing/model': 0.42960804607719183, 'learning_rate/qwen_vl_interface': 9.942116660621319e-06, 'learning_rate/action_model': 9.94211666062132e-05, 'epoch': 0.61})
+[2026-07-28 15:47:29,517][starVLA.training.train_starvla][INFO] - Step 3295, Loss: {'train/loss': 0.02913731336593628, 'train/grad_norm_pre_clip': 0.7480789422988892, 'timing/data': 0.00048552779480814934, 'timing/model': 0.42933887988328934, 'learning_rate/qwen_vl_interface': 9.941519617826906e-06, 'learning_rate/action_model': 9.941519617826906e-05, 'epoch': 0.61})
+[2026-07-28 15:47:36,073][starVLA.training.train_starvla][INFO] - Step 3296, Loss: {'train/loss': 0.030590392649173737, 'train/grad_norm_pre_clip': 1.0807600021362305, 'timing/data': 0.0002977694384753704, 'timing/model': 0.4242498790845275, 'learning_rate/qwen_vl_interface': 9.940919530756347e-06, 'learning_rate/action_model': 9.940919530756347e-05, 'epoch': 0.61})
+[2026-07-28 15:47:42,107][starVLA.training.train_starvla][INFO] - Step 3297, Loss: {'train/loss': 0.0610455758869648, 'train/grad_norm_pre_clip': 2.6301608085632324, 'timing/data': 0.0003169858828186989, 'timing/model': 0.42369120800867677, 'learning_rate/qwen_vl_interface': 9.940316399799032e-06, 'learning_rate/action_model': 9.940316399799033e-05, 'epoch': 0.61})
+[2026-07-28 15:47:48,150][starVLA.training.train_starvla][INFO] - Step 3298, Loss: {'train/loss': 0.030289597809314728, 'train/grad_norm_pre_clip': 1.0500380992889404, 'timing/data': 0.00028094789013266563, 'timing/model': 0.42305256612598896, 'learning_rate/qwen_vl_interface': 9.939710225346329e-06, 'learning_rate/action_model': 9.939710225346329e-05, 'epoch': 0.61})
+[2026-07-28 15:47:54,399][starVLA.training.train_starvla][INFO] - Step 3299, Loss: {'train/loss': 0.05009932070970535, 'train/grad_norm_pre_clip': 1.7884557247161865, 'timing/data': 0.0002822261303663254, 'timing/model': 0.42403932102024555, 'learning_rate/qwen_vl_interface': 9.939101007791575e-06, 'learning_rate/action_model': 9.939101007791575e-05, 'epoch': 0.61})
+[2026-07-28 15:48:00,725][starVLA.training.train_starvla][INFO] - Step 3300, Loss: {'train/loss': 0.034476861357688904, 'train/grad_norm_pre_clip': 1.5300898551940918, 'timing/data': 0.00025541801005601883, 'timing/model': 0.42422923306003213, 'learning_rate/qwen_vl_interface': 9.938488747530087e-06, 'learning_rate/action_model': 9.938488747530086e-05, 'epoch': 0.61})
+[2026-07-28 15:48:06,792][starVLA.training.train_starvla][INFO] - Step 3301, Loss: {'train/loss': 0.036893125623464584, 'train/grad_norm_pre_clip': 1.932090163230896, 'timing/data': 0.0004347129724919796, 'timing/model': 0.42515808809548616, 'learning_rate/qwen_vl_interface': 9.937873444959151e-06, 'learning_rate/action_model': 9.937873444959151e-05, 'epoch': 0.61})
+[2026-07-28 15:48:13,248][starVLA.training.train_starvla][INFO] - Step 3302, Loss: {'train/loss': 0.05741257965564728, 'train/grad_norm_pre_clip': 1.3598570823669434, 'timing/data': 0.00035736383870244026, 'timing/model': 0.7107620849274099, 'learning_rate/qwen_vl_interface': 9.937255100478036e-06, 'learning_rate/action_model': 9.937255100478036e-05, 'epoch': 0.61})
+[2026-07-28 15:48:19,749][starVLA.training.train_starvla][INFO] - Step 3303, Loss: {'train/loss': 0.10047048330307007, 'train/grad_norm_pre_clip': 3.496880054473877, 'timing/data': 0.0003226231783628464, 'timing/model': 0.6770422169938684, 'learning_rate/qwen_vl_interface': 9.93663371448797e-06, 'learning_rate/action_model': 9.93663371448797e-05, 'epoch': 0.61})
+[2026-07-28 15:48:25,803][starVLA.training.train_starvla][INFO] - Step 3304, Loss: {'train/loss': 0.04936403036117554, 'train/grad_norm_pre_clip': 1.5322508811950684, 'timing/data': 0.0002919039689004421, 'timing/model': 0.42977253813296556, 'learning_rate/qwen_vl_interface': 9.936009287392173e-06, 'learning_rate/action_model': 9.936009287392171e-05, 'epoch': 0.61})
+[2026-07-28 15:48:31,849][starVLA.training.train_starvla][INFO] - Step 3305, Loss: {'train/loss': 0.021273624151945114, 'train/grad_norm_pre_clip': 0.8002073764801025, 'timing/data': 0.0003374856896698475, 'timing/model': 0.4241094649769366, 'learning_rate/qwen_vl_interface': 9.935381819595822e-06, 'learning_rate/action_model': 9.935381819595822e-05, 'epoch': 0.61})
+[2026-07-28 15:48:38,149][starVLA.training.train_starvla][INFO] - Step 3306, Loss: {'train/loss': 0.03628756105899811, 'train/grad_norm_pre_clip': 1.3195074796676636, 'timing/data': 0.00025553395971655846, 'timing/model': 0.4241481120698154, 'learning_rate/qwen_vl_interface': 9.934751311506078e-06, 'learning_rate/action_model': 9.934751311506077e-05, 'epoch': 0.61})
+[2026-07-28 15:48:44,263][starVLA.training.train_starvla][INFO] - Step 3307, Loss: {'train/loss': 0.07845384627580643, 'train/grad_norm_pre_clip': 1.2999273538589478, 'timing/data': 0.0002676020376384258, 'timing/model': 0.43075632583349943, 'learning_rate/qwen_vl_interface': 9.93411776353207e-06, 'learning_rate/action_model': 9.934117763532069e-05, 'epoch': 0.61})
+[2026-07-28 15:48:50,584][starVLA.training.train_starvla][INFO] - Step 3308, Loss: {'train/loss': 0.024012885987758636, 'train/grad_norm_pre_clip': 0.8479453921318054, 'timing/data': 0.000274518970400095, 'timing/model': 0.6544429031200707, 'learning_rate/qwen_vl_interface': 9.933481176084899e-06, 'learning_rate/action_model': 9.933481176084899e-05, 'epoch': 0.61})
+[2026-07-28 15:48:56,633][starVLA.training.train_starvla][INFO] - Step 3309, Loss: {'train/loss': 0.027737539261579514, 'train/grad_norm_pre_clip': 0.743699848651886, 'timing/data': 0.0002497527748346329, 'timing/model': 0.42671339493244886, 'learning_rate/qwen_vl_interface': 9.932841549577642e-06, 'learning_rate/action_model': 9.932841549577641e-05, 'epoch': 0.61})
+[2026-07-28 15:49:02,939][starVLA.training.train_starvla][INFO] - Step 3310, Loss: {'train/loss': 0.035238392651081085, 'train/grad_norm_pre_clip': 1.0299811363220215, 'timing/data': 0.0002651461400091648, 'timing/model': 0.6843138509429991, 'learning_rate/qwen_vl_interface': 9.932198884425343e-06, 'learning_rate/action_model': 9.932198884425342e-05, 'epoch': 0.61})
+[2026-07-28 15:49:09,104][starVLA.training.train_starvla][INFO] - Step 3311, Loss: {'train/loss': 0.054367970675230026, 'train/grad_norm_pre_clip': 1.397581696510315, 'timing/data': 0.0003333771601319313, 'timing/model': 0.43063888186588883, 'learning_rate/qwen_vl_interface': 9.931553181045022e-06, 'learning_rate/action_model': 9.931553181045022e-05, 'epoch': 0.61})
+[2026-07-28 15:49:15,137][starVLA.training.train_starvla][INFO] - Step 3312, Loss: {'train/loss': 0.01703527756035328, 'train/grad_norm_pre_clip': 0.7823500037193298, 'timing/data': 0.0003358498215675354, 'timing/model': 0.423248543869704, 'learning_rate/qwen_vl_interface': 9.930904439855667e-06, 'learning_rate/action_model': 9.930904439855667e-05, 'epoch': 0.61})
+[2026-07-28 15:49:21,224][starVLA.training.train_starvla][INFO] - Step 3313, Loss: {'train/loss': 0.03676677122712135, 'train/grad_norm_pre_clip': 1.1145102977752686, 'timing/data': 0.0003112279810011387, 'timing/model': 0.4326567021198571, 'learning_rate/qwen_vl_interface': 9.930252661278245e-06, 'learning_rate/action_model': 9.930252661278244e-05, 'epoch': 0.61})
+[2026-07-28 15:49:27,513][starVLA.training.train_starvla][INFO] - Step 3314, Loss: {'train/loss': 0.024997934699058533, 'train/grad_norm_pre_clip': 0.8634566068649292, 'timing/data': 0.00032425159588456154, 'timing/model': 0.6774540427140892, 'learning_rate/qwen_vl_interface': 9.92959784573568e-06, 'learning_rate/action_model': 9.92959784573568e-05, 'epoch': 0.61})
+[2026-07-28 15:49:33,502][starVLA.training.train_starvla][INFO] - Step 3315, Loss: {'train/loss': 0.04879150539636612, 'train/grad_norm_pre_clip': 2.3991920948028564, 'timing/data': 0.000273875892162323, 'timing/model': 0.4212483558803797, 'learning_rate/qwen_vl_interface': 9.92893999365288e-06, 'learning_rate/action_model': 9.92893999365288e-05, 'epoch': 0.61})
+[2026-07-28 15:49:39,815][starVLA.training.train_starvla][INFO] - Step 3316, Loss: {'train/loss': 0.030846713110804558, 'train/grad_norm_pre_clip': 1.4388221502304077, 'timing/data': 0.00028046080842614174, 'timing/model': 0.421205744612962, 'learning_rate/qwen_vl_interface': 9.928279105456717e-06, 'learning_rate/action_model': 9.928279105456718e-05, 'epoch': 0.61})
+[2026-07-28 15:49:45,864][starVLA.training.train_starvla][INFO] - Step 3317, Loss: {'train/loss': 0.05601366236805916, 'train/grad_norm_pre_clip': 2.4710583686828613, 'timing/data': 0.00030904030427336693, 'timing/model': 0.41966074611991644, 'learning_rate/qwen_vl_interface': 9.927615181576034e-06, 'learning_rate/action_model': 9.927615181576034e-05, 'epoch': 0.61})
+[2026-07-28 15:49:51,864][starVLA.training.train_starvla][INFO] - Step 3318, Loss: {'train/loss': 0.03202832490205765, 'train/grad_norm_pre_clip': 1.5360352993011475, 'timing/data': 0.00031481683254241943, 'timing/model': 0.4212801386602223, 'learning_rate/qwen_vl_interface': 9.926948222441644e-06, 'learning_rate/action_model': 9.926948222441643e-05, 'epoch': 0.61})
+[2026-07-28 15:49:58,101][starVLA.training.train_starvla][INFO] - Step 3319, Loss: {'train/loss': 0.011287668719887733, 'train/grad_norm_pre_clip': 0.5056926012039185, 'timing/data': 0.0002587810158729553, 'timing/model': 0.42206071875989437, 'learning_rate/qwen_vl_interface': 9.926278228486328e-06, 'learning_rate/action_model': 9.926278228486327e-05, 'epoch': 0.61})
+[2026-07-28 15:50:04,313][starVLA.training.train_starvla][INFO] - Step 3320, Loss: {'train/loss': 0.060850419104099274, 'train/grad_norm_pre_clip': 1.8525718450546265, 'timing/data': 0.00024882610887289047, 'timing/model': 0.4297303422354162, 'learning_rate/qwen_vl_interface': 9.92560520014484e-06, 'learning_rate/action_model': 9.92560520014484e-05, 'epoch': 0.61})
+[2026-07-28 15:50:10,491][starVLA.training.train_starvla][INFO] - Step 3321, Loss: {'train/loss': 0.035927869379520416, 'train/grad_norm_pre_clip': 1.3312528133392334, 'timing/data': 0.0003837011754512787, 'timing/model': 0.421320500317961, 'learning_rate/qwen_vl_interface': 9.924929137853899e-06, 'learning_rate/action_model': 9.924929137853898e-05, 'epoch': 0.61})
+[2026-07-28 15:50:16,507][starVLA.training.train_starvla][INFO] - Step 3322, Loss: {'train/loss': 0.092984139919281, 'train/grad_norm_pre_clip': 1.9692732095718384, 'timing/data': 0.0003113849088549614, 'timing/model': 0.42130648205056787, 'learning_rate/qwen_vl_interface': 9.924250042052197e-06, 'learning_rate/action_model': 9.924250042052197e-05, 'epoch': 0.61})
+[2026-07-28 15:50:22,983][starVLA.training.train_starvla][INFO] - Step 3323, Loss: {'train/loss': 0.044799596071243286, 'train/grad_norm_pre_clip': 1.2446891069412231, 'timing/data': 0.00026233214884996414, 'timing/model': 0.6635382450185716, 'learning_rate/qwen_vl_interface': 9.92356791318039e-06, 'learning_rate/action_model': 9.92356791318039e-05, 'epoch': 0.61})
+[2026-07-28 15:50:29,075][starVLA.training.train_starvla][INFO] - Step 3324, Loss: {'train/loss': 0.05233525484800339, 'train/grad_norm_pre_clip': 1.2389519214630127, 'timing/data': 0.00031024497002363205, 'timing/model': 0.4215958430431783, 'learning_rate/qwen_vl_interface': 9.922882751681104e-06, 'learning_rate/action_model': 9.922882751681104e-05, 'epoch': 0.61})
+[2026-07-28 15:50:34,820][starVLA.training.train_starvla][INFO] - Step 3325, Loss: {'train/loss': 0.02781415358185768, 'train/grad_norm_pre_clip': 0.619519829750061, 'timing/data': 0.000328658614307642, 'timing/model': 0.42213994916528463, 'learning_rate/qwen_vl_interface': 9.922194557998932e-06, 'learning_rate/action_model': 9.922194557998933e-05, 'epoch': 0.61})
+[2026-07-28 15:50:40,643][starVLA.training.train_starvla][INFO] - Step 3326, Loss: {'train/loss': 0.026319444179534912, 'train/grad_norm_pre_clip': 0.8618413209915161, 'timing/data': 0.0003549172542989254, 'timing/model': 0.4211429599672556, 'learning_rate/qwen_vl_interface': 9.921503332580439e-06, 'learning_rate/action_model': 9.921503332580439e-05, 'epoch': 0.61})
+[2026-07-28 15:50:46,999][starVLA.training.train_starvla][INFO] - Step 3327, Loss: {'train/loss': 0.05832584202289581, 'train/grad_norm_pre_clip': 1.388375997543335, 'timing/data': 0.00031306222081184387, 'timing/model': 0.42733416613191366, 'learning_rate/qwen_vl_interface': 9.92080907587415e-06, 'learning_rate/action_model': 9.92080907587415e-05, 'epoch': 0.61})
+[2026-07-28 15:50:53,251][starVLA.training.train_starvla][INFO] - Step 3328, Loss: {'train/loss': 0.04202158376574516, 'train/grad_norm_pre_clip': 1.0172961950302124, 'timing/data': 0.00045329006388783455, 'timing/model': 0.6829967801459134, 'learning_rate/qwen_vl_interface': 9.92011178833056e-06, 'learning_rate/action_model': 9.92011178833056e-05, 'epoch': 0.61})
+[2026-07-28 15:50:59,238][starVLA.training.train_starvla][INFO] - Step 3329, Loss: {'train/loss': 0.035005852580070496, 'train/grad_norm_pre_clip': 0.9893499612808228, 'timing/data': 0.0003123432397842407, 'timing/model': 0.42018600506708026, 'learning_rate/qwen_vl_interface': 9.919411470402136e-06, 'learning_rate/action_model': 9.919411470402136e-05, 'epoch': 0.61})
+[2026-07-28 15:51:05,483][starVLA.training.train_starvla][INFO] - Step 3330, Loss: {'train/loss': 0.04392237216234207, 'train/grad_norm_pre_clip': 0.7825603485107422, 'timing/data': 0.00031460216268897057, 'timing/model': 0.4213838470168412, 'learning_rate/qwen_vl_interface': 9.918708122543305e-06, 'learning_rate/action_model': 9.918708122543305e-05, 'epoch': 0.61})
+[2026-07-28 15:51:11,469][starVLA.training.train_starvla][INFO] - Step 3331, Loss: {'train/loss': 0.04383876919746399, 'train/grad_norm_pre_clip': 2.1315228939056396, 'timing/data': 0.00029703136533498764, 'timing/model': 0.42143750516697764, 'learning_rate/qwen_vl_interface': 9.918001745210458e-06, 'learning_rate/action_model': 9.918001745210456e-05, 'epoch': 0.61})
+[2026-07-28 15:51:17,461][starVLA.training.train_starvla][INFO] - Step 3332, Loss: {'train/loss': 0.07164354622364044, 'train/grad_norm_pre_clip': 1.7803778648376465, 'timing/data': 0.00027427589520812035, 'timing/model': 0.4205895313061774, 'learning_rate/qwen_vl_interface': 9.917292338861958e-06, 'learning_rate/action_model': 9.917292338861957e-05, 'epoch': 0.61})
+[2026-07-28 15:51:23,657][starVLA.training.train_starvla][INFO] - Step 3333, Loss: {'train/loss': 0.04207036271691322, 'train/grad_norm_pre_clip': 1.6285011768341064, 'timing/data': 0.000425481703132391, 'timing/model': 0.6263404139317572, 'learning_rate/qwen_vl_interface': 9.916579903958134e-06, 'learning_rate/action_model': 9.916579903958133e-05, 'epoch': 0.61})
+[2026-07-28 15:51:29,624][starVLA.training.train_starvla][INFO] - Step 3334, Loss: {'train/loss': 0.032941918820142746, 'train/grad_norm_pre_clip': 1.1752358675003052, 'timing/data': 0.0003100298345088959, 'timing/model': 0.42114025028422475, 'learning_rate/qwen_vl_interface': 9.915864440961272e-06, 'learning_rate/action_model': 9.915864440961272e-05, 'epoch': 0.61})
+[2026-07-28 15:51:35,828][starVLA.training.train_starvla][INFO] - Step 3335, Loss: {'train/loss': 0.054612837731838226, 'train/grad_norm_pre_clip': 1.2539772987365723, 'timing/data': 0.0002527679316699505, 'timing/model': 0.41980882501229644, 'learning_rate/qwen_vl_interface': 9.915145950335631e-06, 'learning_rate/action_model': 9.915145950335632e-05, 'epoch': 0.61})
+[2026-07-28 15:51:41,785][starVLA.training.train_starvla][INFO] - Step 3336, Loss: {'train/loss': 0.037475522607564926, 'train/grad_norm_pre_clip': 0.9484372735023499, 'timing/data': 0.00031743571162223816, 'timing/model': 0.41985767893493176, 'learning_rate/qwen_vl_interface': 9.914424432547434e-06, 'learning_rate/action_model': 9.914424432547433e-05, 'epoch': 0.61})
+[2026-07-28 15:51:47,984][starVLA.training.train_starvla][INFO] - Step 3337, Loss: {'train/loss': 0.06694228947162628, 'train/grad_norm_pre_clip': 1.9726847410202026, 'timing/data': 0.0002963058650493622, 'timing/model': 0.4203293491154909, 'learning_rate/qwen_vl_interface': 9.913699888064862e-06, 'learning_rate/action_model': 9.913699888064861e-05, 'epoch': 0.61})
+[2026-07-28 15:51:53,950][starVLA.training.train_starvla][INFO] - Step 3338, Loss: {'train/loss': 0.06497912853956223, 'train/grad_norm_pre_clip': 0.8378243446350098, 'timing/data': 0.0002747410908341408, 'timing/model': 0.42734365398064256, 'learning_rate/qwen_vl_interface': 9.912972317358066e-06, 'learning_rate/action_model': 9.912972317358066e-05, 'epoch': 0.61})
+[2026-07-28 15:52:00,174][starVLA.training.train_starvla][INFO] - Step 3339, Loss: {'train/loss': 0.05535627901554108, 'train/grad_norm_pre_clip': 3.934781312942505, 'timing/data': 0.00030065886676311493, 'timing/model': 0.662946445401758, 'learning_rate/qwen_vl_interface': 9.912241720899157e-06, 'learning_rate/action_model': 9.912241720899157e-05, 'epoch': 0.61})
+[2026-07-28 15:52:06,140][starVLA.training.train_starvla][INFO] - Step 3340, Loss: {'train/loss': 0.018184855580329895, 'train/grad_norm_pre_clip': 0.48106083273887634, 'timing/data': 0.00033363187685608864, 'timing/model': 0.41992341727018356, 'learning_rate/qwen_vl_interface': 9.911508099162215e-06, 'learning_rate/action_model': 9.911508099162214e-05, 'epoch': 0.61})
+[2026-07-28 15:52:12,332][starVLA.training.train_starvla][INFO] - Step 3341, Loss: {'train/loss': 0.04089290276169777, 'train/grad_norm_pre_clip': 1.4667912721633911, 'timing/data': 0.00030860211700201035, 'timing/model': 0.42079449025914073, 'learning_rate/qwen_vl_interface': 9.910771452623276e-06, 'learning_rate/action_model': 9.910771452623275e-05, 'epoch': 0.61})
+[2026-07-28 15:52:18,311][starVLA.training.train_starvla][INFO] - Step 3342, Loss: {'train/loss': 0.09920559078454971, 'train/grad_norm_pre_clip': 2.6074116230010986, 'timing/data': 0.00034848088398575783, 'timing/model': 0.4186391858384013, 'learning_rate/qwen_vl_interface': 9.910031781760342e-06, 'learning_rate/action_model': 9.910031781760342e-05, 'epoch': 0.61})
+[2026-07-28 15:52:24,520][starVLA.training.train_starvla][INFO] - Step 3343, Loss: {'train/loss': 0.023191552609205246, 'train/grad_norm_pre_clip': 0.6736221313476562, 'timing/data': 0.0002888469025492668, 'timing/model': 0.41992900893092155, 'learning_rate/qwen_vl_interface': 9.909289087053378e-06, 'learning_rate/action_model': 9.909289087053378e-05, 'epoch': 0.61})
+[2026-07-28 15:52:30,533][starVLA.training.train_starvla][INFO] - Step 3344, Loss: {'train/loss': 0.03435672074556351, 'train/grad_norm_pre_clip': 1.547250509262085, 'timing/data': 0.0004015550948679447, 'timing/model': 0.42066430719569325, 'learning_rate/qwen_vl_interface': 9.90854336898431e-06, 'learning_rate/action_model': 9.90854336898431e-05, 'epoch': 0.61})
+[2026-07-28 15:52:36,512][starVLA.training.train_starvla][INFO] - Step 3345, Loss: {'train/loss': 0.034763604402542114, 'train/grad_norm_pre_clip': 1.0546241998672485, 'timing/data': 0.00045084208250045776, 'timing/model': 0.4195731049403548, 'learning_rate/qwen_vl_interface': 9.907794628037028e-06, 'learning_rate/action_model': 9.907794628037028e-05, 'epoch': 0.61})
+[2026-07-28 15:52:42,772][starVLA.training.train_starvla][INFO] - Step 3346, Loss: {'train/loss': 0.04010511934757233, 'train/grad_norm_pre_clip': 1.767924189567566, 'timing/data': 0.0002870536409318447, 'timing/model': 0.683234769385308, 'learning_rate/qwen_vl_interface': 9.90704286469738e-06, 'learning_rate/action_model': 9.90704286469738e-05, 'epoch': 0.61})
+[2026-07-28 15:52:48,784][starVLA.training.train_starvla][INFO] - Step 3347, Loss: {'train/loss': 0.060624636709690094, 'train/grad_norm_pre_clip': 1.2830811738967896, 'timing/data': 0.00028986111283302307, 'timing/model': 0.419873827137053, 'learning_rate/qwen_vl_interface': 9.90628807945318e-06, 'learning_rate/action_model': 9.90628807945318e-05, 'epoch': 0.61})
+[2026-07-28 15:52:54,773][starVLA.training.train_starvla][INFO] - Step 3348, Loss: {'train/loss': 0.04925695061683655, 'train/grad_norm_pre_clip': 2.1264374256134033, 'timing/data': 0.0002762996591627598, 'timing/model': 0.42096775490790606, 'learning_rate/qwen_vl_interface': 9.905530272794195e-06, 'learning_rate/action_model': 9.905530272794195e-05, 'epoch': 0.61})
+[2026-07-28 15:53:00,498][starVLA.training.train_starvla][INFO] - Step 3349, Loss: {'train/loss': 0.028569022193551064, 'train/grad_norm_pre_clip': 1.5833219289779663, 'timing/data': 0.0003616507165133953, 'timing/model': 0.4204365028999746, 'learning_rate/qwen_vl_interface': 9.904769445212163e-06, 'learning_rate/action_model': 9.904769445212162e-05, 'epoch': 0.62})
+[2026-07-28 15:53:06,687][starVLA.training.train_starvla][INFO] - Step 3350, Loss: {'train/loss': 0.0382419228553772, 'train/grad_norm_pre_clip': 1.0770772695541382, 'timing/data': 0.00030296389013528824, 'timing/model': 0.6583265750668943, 'learning_rate/qwen_vl_interface': 9.904005597200772e-06, 'learning_rate/action_model': 9.904005597200772e-05, 'epoch': 0.62})
+[2026-07-28 15:53:12,838][starVLA.training.train_starvla][INFO] - Step 3351, Loss: {'train/loss': 0.014642568305134773, 'train/grad_norm_pre_clip': 0.7678760290145874, 'timing/data': 0.00025346409529447556, 'timing/model': 0.4197403169237077, 'learning_rate/qwen_vl_interface': 9.903238729255678e-06, 'learning_rate/action_model': 9.903238729255677e-05, 'epoch': 0.62})
+[2026-07-28 15:53:18,796][starVLA.training.train_starvla][INFO] - Step 3352, Loss: {'train/loss': 0.03862562030553818, 'train/grad_norm_pre_clip': 0.9097667336463928, 'timing/data': 0.00026253610849380493, 'timing/model': 0.42104699881747365, 'learning_rate/qwen_vl_interface': 9.902468841874492e-06, 'learning_rate/action_model': 9.902468841874492e-05, 'epoch': 0.62})
+[2026-07-28 15:53:24,547][starVLA.training.train_starvla][INFO] - Step 3353, Loss: {'train/loss': 0.04956740140914917, 'train/grad_norm_pre_clip': 1.2538601160049438, 'timing/data': 0.0003638830967247486, 'timing/model': 0.4199654436670244, 'learning_rate/qwen_vl_interface': 9.901695935556784e-06, 'learning_rate/action_model': 9.901695935556783e-05, 'epoch': 0.62})
+[2026-07-28 15:53:30,777][starVLA.training.train_starvla][INFO] - Step 3354, Loss: {'train/loss': 0.021946493536233902, 'train/grad_norm_pre_clip': 0.6239603161811829, 'timing/data': 0.0002589849755167961, 'timing/model': 0.42109988490119576, 'learning_rate/qwen_vl_interface': 9.900920010804088e-06, 'learning_rate/action_model': 9.900920010804088e-05, 'epoch': 0.62})
+[2026-07-28 15:53:37,035][starVLA.training.train_starvla][INFO] - Step 3355, Loss: {'train/loss': 0.045938968658447266, 'train/grad_norm_pre_clip': 1.2664669752120972, 'timing/data': 0.0006133690476417542, 'timing/model': 0.6580444630235434, 'learning_rate/qwen_vl_interface': 9.900141068119891e-06, 'learning_rate/action_model': 9.900141068119891e-05, 'epoch': 0.62})
+[2026-07-28 15:53:43,088][starVLA.training.train_starvla][INFO] - Step 3356, Loss: {'train/loss': 0.012231487780809402, 'train/grad_norm_pre_clip': 0.3739846646785736, 'timing/data': 0.00025459006428718567, 'timing/model': 0.4190089860931039, 'learning_rate/qwen_vl_interface': 9.89935910800964e-06, 'learning_rate/action_model': 9.89935910800964e-05, 'epoch': 0.62})
+[2026-07-28 15:53:49,069][starVLA.training.train_starvla][INFO] - Step 3357, Loss: {'train/loss': 0.04645863547921181, 'train/grad_norm_pre_clip': 1.518812894821167, 'timing/data': 0.0003705378621816635, 'timing/model': 0.421517726033926, 'learning_rate/qwen_vl_interface': 9.898574130980743e-06, 'learning_rate/action_model': 9.898574130980742e-05, 'epoch': 0.62})
+[2026-07-28 15:53:55,310][starVLA.training.train_starvla][INFO] - Step 3358, Loss: {'train/loss': 0.03380068019032478, 'train/grad_norm_pre_clip': 1.0320428609848022, 'timing/data': 0.00037025613710284233, 'timing/model': 0.6624845340847969, 'learning_rate/qwen_vl_interface': 9.89778613754256e-06, 'learning_rate/action_model': 9.89778613754256e-05, 'epoch': 0.62})
+[2026-07-28 15:54:01,340][starVLA.training.train_starvla][INFO] - Step 3359, Loss: {'train/loss': 0.03181367367506027, 'train/grad_norm_pre_clip': 1.4039173126220703, 'timing/data': 0.00026245182380080223, 'timing/model': 0.42084517516195774, 'learning_rate/qwen_vl_interface': 9.896995128206412e-06, 'learning_rate/action_model': 9.896995128206411e-05, 'epoch': 0.62})
+[2026-07-28 15:54:07,358][starVLA.training.train_starvla][INFO] - Step 3360, Loss: {'train/loss': 0.08626638352870941, 'train/grad_norm_pre_clip': 3.7488255500793457, 'timing/data': 0.00030144909396767616, 'timing/model': 0.4213429018855095, 'learning_rate/qwen_vl_interface': 9.896201103485577e-06, 'learning_rate/action_model': 9.896201103485577e-05, 'epoch': 0.62})
+[2026-07-28 15:54:13,614][starVLA.training.train_starvla][INFO] - Step 3361, Loss: {'train/loss': 0.032956331968307495, 'train/grad_norm_pre_clip': 1.9268765449523926, 'timing/data': 0.00029043108224868774, 'timing/model': 0.6704644849523902, 'learning_rate/qwen_vl_interface': 9.89540406389529e-06, 'learning_rate/action_model': 9.89540406389529e-05, 'epoch': 0.62})
+[2026-07-28 15:54:19,652][starVLA.training.train_starvla][INFO] - Step 3362, Loss: {'train/loss': 0.05802832543849945, 'train/grad_norm_pre_clip': 1.5665806531906128, 'timing/data': 0.0002585509791970253, 'timing/model': 0.42064922070130706, 'learning_rate/qwen_vl_interface': 9.894604009952742e-06, 'learning_rate/action_model': 9.89460400995274e-05, 'epoch': 0.62})
+[2026-07-28 15:54:25,833][starVLA.training.train_starvla][INFO] - Step 3363, Loss: {'train/loss': 0.022360051050782204, 'train/grad_norm_pre_clip': 0.9007349610328674, 'timing/data': 0.00029259780421853065, 'timing/model': 0.42058568727225065, 'learning_rate/qwen_vl_interface': 9.893800942177073e-06, 'learning_rate/action_model': 9.893800942177073e-05, 'epoch': 0.62})
+[2026-07-28 15:54:32,054][starVLA.training.train_starvla][INFO] - Step 3364, Loss: {'train/loss': 0.05033000186085701, 'train/grad_norm_pre_clip': 1.6787587404251099, 'timing/data': 0.0002543148584663868, 'timing/model': 0.4183029206469655, 'learning_rate/qwen_vl_interface': 9.892994861089393e-06, 'learning_rate/action_model': 9.892994861089393e-05, 'epoch': 0.62})
+[2026-07-28 15:54:38,036][starVLA.training.train_starvla][INFO] - Step 3365, Loss: {'train/loss': 0.04475616663694382, 'train/grad_norm_pre_clip': 1.424826741218567, 'timing/data': 0.0003140149638056755, 'timing/model': 0.4214265076443553, 'learning_rate/qwen_vl_interface': 9.892185767212756e-06, 'learning_rate/action_model': 9.892185767212756e-05, 'epoch': 0.62})
+[2026-07-28 15:54:44,293][starVLA.training.train_starvla][INFO] - Step 3366, Loss: {'train/loss': 0.026798242703080177, 'train/grad_norm_pre_clip': 1.5697847604751587, 'timing/data': 0.0002559889107942581, 'timing/model': 0.41857285192236304, 'learning_rate/qwen_vl_interface': 9.891373661072172e-06, 'learning_rate/action_model': 9.891373661072173e-05, 'epoch': 0.62})
+[2026-07-28 15:54:50,284][starVLA.training.train_starvla][INFO] - Step 3367, Loss: {'train/loss': 0.03349937126040459, 'train/grad_norm_pre_clip': 0.9222425222396851, 'timing/data': 0.00032161222770810127, 'timing/model': 0.42634404031559825, 'learning_rate/qwen_vl_interface': 9.89055854319461e-06, 'learning_rate/action_model': 9.890558543194611e-05, 'epoch': 0.62})
+[2026-07-28 15:54:56,520][starVLA.training.train_starvla][INFO] - Step 3368, Loss: {'train/loss': 0.0590677410364151, 'train/grad_norm_pre_clip': 1.695987343788147, 'timing/data': 0.00031491974368691444, 'timing/model': 0.4214750239625573, 'learning_rate/qwen_vl_interface': 9.889740414108994e-06, 'learning_rate/action_model': 9.889740414108995e-05, 'epoch': 0.62})
+[2026-07-28 15:55:02,526][starVLA.training.train_starvla][INFO] - Step 3369, Loss: {'train/loss': 0.03412003815174103, 'train/grad_norm_pre_clip': 1.364804744720459, 'timing/data': 0.000298134982585907, 'timing/model': 0.4271313939243555, 'learning_rate/qwen_vl_interface': 9.888919274346196e-06, 'learning_rate/action_model': 9.888919274346195e-05, 'epoch': 0.62})
+[2026-07-28 15:55:08,778][starVLA.training.train_starvla][INFO] - Step 3370, Loss: {'train/loss': 0.04988471418619156, 'train/grad_norm_pre_clip': 1.9632436037063599, 'timing/data': 0.00029165204614400864, 'timing/model': 0.4212368060834706, 'learning_rate/qwen_vl_interface': 9.888095124439043e-06, 'learning_rate/action_model': 9.888095124439043e-05, 'epoch': 0.62})
+[2026-07-28 15:55:15,000][starVLA.training.train_starvla][INFO] - Step 3371, Loss: {'train/loss': 0.062396809458732605, 'train/grad_norm_pre_clip': 3.768934726715088, 'timing/data': 0.00025474606081843376, 'timing/model': 0.4185645398683846, 'learning_rate/qwen_vl_interface': 9.88726796492232e-06, 'learning_rate/action_model': 9.88726796492232e-05, 'epoch': 0.62})
+[2026-07-28 15:55:20,996][starVLA.training.train_starvla][INFO] - Step 3372, Loss: {'train/loss': 0.050267767161130905, 'train/grad_norm_pre_clip': 0.9383177161216736, 'timing/data': 0.0003200201317667961, 'timing/model': 0.425569923594594, 'learning_rate/qwen_vl_interface': 9.886437796332759e-06, 'learning_rate/action_model': 9.88643779633276e-05, 'epoch': 0.62})
+[2026-07-28 15:55:27,031][starVLA.training.train_starvla][INFO] - Step 3373, Loss: {'train/loss': 0.04803389683365822, 'train/grad_norm_pre_clip': 0.9455209374427795, 'timing/data': 0.0005586310289800167, 'timing/model': 0.46017436403781176, 'learning_rate/qwen_vl_interface': 9.885604619209052e-06, 'learning_rate/action_model': 9.885604619209051e-05, 'epoch': 0.62})
+[2026-07-28 15:55:33,390][starVLA.training.train_starvla][INFO] - Step 3374, Loss: {'train/loss': 0.03321271017193794, 'train/grad_norm_pre_clip': 0.799157977104187, 'timing/data': 0.0003900197334587574, 'timing/model': 0.6217964701354504, 'learning_rate/qwen_vl_interface': 9.884768434091832e-06, 'learning_rate/action_model': 9.884768434091832e-05, 'epoch': 0.62})
+[2026-07-28 15:55:39,634][starVLA.training.train_starvla][INFO] - Step 3375, Loss: {'train/loss': 0.026907173916697502, 'train/grad_norm_pre_clip': 0.6955748200416565, 'timing/data': 0.0003459108993411064, 'timing/model': 0.6528967646881938, 'learning_rate/qwen_vl_interface': 9.883929241523698e-06, 'learning_rate/action_model': 9.883929241523697e-05, 'epoch': 0.62})
+[2026-07-28 15:55:45,635][starVLA.training.train_starvla][INFO] - Step 3376, Loss: {'train/loss': 0.03541108965873718, 'train/grad_norm_pre_clip': 0.8629978895187378, 'timing/data': 0.0003135618753731251, 'timing/model': 0.4216646198183298, 'learning_rate/qwen_vl_interface': 9.883087042049185e-06, 'learning_rate/action_model': 9.883087042049186e-05, 'epoch': 0.62})
+[2026-07-28 15:55:51,619][starVLA.training.train_starvla][INFO] - Step 3377, Loss: {'train/loss': 0.062493011355400085, 'train/grad_norm_pre_clip': 3.475309371948242, 'timing/data': 0.00034017302095890045, 'timing/model': 0.42187081836164, 'learning_rate/qwen_vl_interface': 9.882241836214794e-06, 'learning_rate/action_model': 9.882241836214794e-05, 'epoch': 0.62})
+[2026-07-28 15:55:57,834][starVLA.training.train_starvla][INFO] - Step 3378, Loss: {'train/loss': 0.0499856173992157, 'train/grad_norm_pre_clip': 1.3411335945129395, 'timing/data': 0.0003021238371729851, 'timing/model': 0.422573653049767, 'learning_rate/qwen_vl_interface': 9.881393624568965e-06, 'learning_rate/action_model': 9.881393624568965e-05, 'epoch': 0.62})
+[2026-07-28 15:56:03,832][starVLA.training.train_starvla][INFO] - Step 3379, Loss: {'train/loss': 0.048084937036037445, 'train/grad_norm_pre_clip': 2.1550662517547607, 'timing/data': 0.0003039408475160599, 'timing/model': 0.41998333716765046, 'learning_rate/qwen_vl_interface': 9.880542407662098e-06, 'learning_rate/action_model': 9.880542407662097e-05, 'epoch': 0.62})
+[2026-07-28 15:56:09,804][starVLA.training.train_starvla][INFO] - Step 3380, Loss: {'train/loss': 0.02977897971868515, 'train/grad_norm_pre_clip': 1.4738695621490479, 'timing/data': 0.00031108129769563675, 'timing/model': 0.4200322888791561, 'learning_rate/qwen_vl_interface': 9.879688186046535e-06, 'learning_rate/action_model': 9.879688186046534e-05, 'epoch': 0.62})
+[2026-07-28 15:56:15,806][starVLA.training.train_starvla][INFO] - Step 3381, Loss: {'train/loss': 0.04082707315683365, 'train/grad_norm_pre_clip': 0.673216700553894, 'timing/data': 0.00045141996815800667, 'timing/model': 0.4204615871421993, 'learning_rate/qwen_vl_interface': 9.87883096027657e-06, 'learning_rate/action_model': 9.87883096027657e-05, 'epoch': 0.62})
+[2026-07-28 15:56:22,469][starVLA.training.train_starvla][INFO] - Step 3382, Loss: {'train/loss': 0.0500059500336647, 'train/grad_norm_pre_clip': 1.354248285293579, 'timing/data': 0.0003400910645723343, 'timing/model': 0.4337108451873064, 'learning_rate/qwen_vl_interface': 9.87797073090845e-06, 'learning_rate/action_model': 9.87797073090845e-05, 'epoch': 0.62})
+[2026-07-28 15:56:28,483][starVLA.training.train_starvla][INFO] - Step 3383, Loss: {'train/loss': 0.03735845908522606, 'train/grad_norm_pre_clip': 1.256105899810791, 'timing/data': 0.00036487309262156487, 'timing/model': 0.41912259720265865, 'learning_rate/qwen_vl_interface': 9.877107498500368e-06, 'learning_rate/action_model': 9.877107498500367e-05, 'epoch': 0.62})
+[2026-07-28 15:56:34,475][starVLA.training.train_starvla][INFO] - Step 3384, Loss: {'train/loss': 0.02802756242454052, 'train/grad_norm_pre_clip': 0.655599057674408, 'timing/data': 0.00028397515416145325, 'timing/model': 0.433314124122262, 'learning_rate/qwen_vl_interface': 9.876241263612467e-06, 'learning_rate/action_model': 9.876241263612465e-05, 'epoch': 0.62})
+[2026-07-28 15:56:40,498][starVLA.training.train_starvla][INFO] - Step 3385, Loss: {'train/loss': 0.023856155574321747, 'train/grad_norm_pre_clip': 0.9120814800262451, 'timing/data': 0.0003608991391956806, 'timing/model': 0.42815218586474657, 'learning_rate/qwen_vl_interface': 9.875372026806835e-06, 'learning_rate/action_model': 9.875372026806835e-05, 'epoch': 0.62})
+[2026-07-28 15:56:46,744][starVLA.training.train_starvla][INFO] - Step 3386, Loss: {'train/loss': 0.0212185550481081, 'train/grad_norm_pre_clip': 0.5073455572128296, 'timing/data': 0.00039333896711468697, 'timing/model': 0.6757635623216629, 'learning_rate/qwen_vl_interface': 9.874499788647512e-06, 'learning_rate/action_model': 9.874499788647513e-05, 'epoch': 0.62})
+[2026-07-28 15:56:52,948][starVLA.training.train_starvla][INFO] - Step 3387, Loss: {'train/loss': 0.02390953153371811, 'train/grad_norm_pre_clip': 0.7919082045555115, 'timing/data': 0.00035089394077658653, 'timing/model': 0.66499515902251, 'learning_rate/qwen_vl_interface': 9.873624549700485e-06, 'learning_rate/action_model': 9.873624549700484e-05, 'epoch': 0.62})
+[2026-07-28 15:56:58,922][starVLA.training.train_starvla][INFO] - Step 3388, Loss: {'train/loss': 0.11834576725959778, 'train/grad_norm_pre_clip': 3.3320541381835938, 'timing/data': 0.00025656307116150856, 'timing/model': 0.4208121527917683, 'learning_rate/qwen_vl_interface': 9.872746310533683e-06, 'learning_rate/action_model': 9.872746310533684e-05, 'epoch': 0.62})
+[2026-07-28 15:57:04,916][starVLA.training.train_starvla][INFO] - Step 3389, Loss: {'train/loss': 0.030693043023347855, 'train/grad_norm_pre_clip': 0.91652911901474, 'timing/data': 0.0003131958656013012, 'timing/model': 0.4215716142207384, 'learning_rate/qwen_vl_interface': 9.871865071716991e-06, 'learning_rate/action_model': 9.87186507171699e-05, 'epoch': 0.62})
+[2026-07-28 15:57:11,158][starVLA.training.train_starvla][INFO] - Step 3390, Loss: {'train/loss': 0.0381816104054451, 'train/grad_norm_pre_clip': 2.1231789588928223, 'timing/data': 0.0002566808834671974, 'timing/model': 0.42139805713668466, 'learning_rate/qwen_vl_interface': 9.870980833822231e-06, 'learning_rate/action_model': 9.87098083382223e-05, 'epoch': 0.62})
+[2026-07-28 15:57:17,149][starVLA.training.train_starvla][INFO] - Step 3391, Loss: {'train/loss': 0.023040737956762314, 'train/grad_norm_pre_clip': 0.5244916081428528, 'timing/data': 0.0003126799128949642, 'timing/model': 0.4217159664258361, 'learning_rate/qwen_vl_interface': 9.87009359742318e-06, 'learning_rate/action_model': 9.87009359742318e-05, 'epoch': 0.62})
+[2026-07-28 15:57:23,391][starVLA.training.train_starvla][INFO] - Step 3392, Loss: {'train/loss': 0.03888676315546036, 'train/grad_norm_pre_clip': 1.7032605409622192, 'timing/data': 0.00026352889835834503, 'timing/model': 0.4222319587133825, 'learning_rate/qwen_vl_interface': 9.86920336309555e-06, 'learning_rate/action_model': 9.86920336309555e-05, 'epoch': 0.62})
+[2026-07-28 15:57:29,422][starVLA.training.train_starvla][INFO] - Step 3393, Loss: {'train/loss': 0.03766591101884842, 'train/grad_norm_pre_clip': 0.9539644718170166, 'timing/data': 0.00031047966331243515, 'timing/model': 0.4229827360250056, 'learning_rate/qwen_vl_interface': 9.86831013141701e-06, 'learning_rate/action_model': 9.86831013141701e-05, 'epoch': 0.62})
+[2026-07-28 15:57:35,672][starVLA.training.train_starvla][INFO] - Step 3394, Loss: {'train/loss': 0.019475651904940605, 'train/grad_norm_pre_clip': 0.9833922982215881, 'timing/data': 0.00025648297742009163, 'timing/model': 0.42443038057535887, 'learning_rate/qwen_vl_interface': 9.867413902967166e-06, 'learning_rate/action_model': 9.867413902967166e-05, 'epoch': 0.62})
+[2026-07-28 15:57:41,926][starVLA.training.train_starvla][INFO] - Step 3395, Loss: {'train/loss': 0.032678358256816864, 'train/grad_norm_pre_clip': 1.3952240943908691, 'timing/data': 0.0002697603777050972, 'timing/model': 0.4268808448687196, 'learning_rate/qwen_vl_interface': 9.86651467832757e-06, 'learning_rate/action_model': 9.86651467832757e-05, 'epoch': 0.62})
+[2026-07-28 15:57:48,303][starVLA.training.train_starvla][INFO] - Step 3396, Loss: {'train/loss': 0.0404496043920517, 'train/grad_norm_pre_clip': 2.2872912883758545, 'timing/data': 0.0003121793270111084, 'timing/model': 0.42199299205094576, 'learning_rate/qwen_vl_interface': 9.865612458081718e-06, 'learning_rate/action_model': 9.865612458081719e-05, 'epoch': 0.62})
+[2026-07-28 15:57:54,535][starVLA.training.train_starvla][INFO] - Step 3397, Loss: {'train/loss': 0.053422711789608, 'train/grad_norm_pre_clip': 3.591827869415283, 'timing/data': 0.00031747808679938316, 'timing/model': 0.6601891140453517, 'learning_rate/qwen_vl_interface': 9.864707242815057e-06, 'learning_rate/action_model': 9.864707242815056e-05, 'epoch': 0.62})
+[2026-07-28 15:58:00,542][starVLA.training.train_starvla][INFO] - Step 3398, Loss: {'train/loss': 0.10297201573848724, 'train/grad_norm_pre_clip': 4.0746378898620605, 'timing/data': 0.0003192359581589699, 'timing/model': 0.42336809635162354, 'learning_rate/qwen_vl_interface': 9.863799033114966e-06, 'learning_rate/action_model': 9.863799033114966e-05, 'epoch': 0.62})
+[2026-07-28 15:58:06,912][starVLA.training.train_starvla][INFO] - Step 3399, Loss: {'train/loss': 0.0070812092162668705, 'train/grad_norm_pre_clip': 0.7916351556777954, 'timing/data': 0.00030477624386548996, 'timing/model': 0.6910228291526437, 'learning_rate/qwen_vl_interface': 9.862887829570771e-06, 'learning_rate/action_model': 9.862887829570772e-05, 'epoch': 0.62})
+[2026-07-28 15:58:12,940][starVLA.training.train_starvla][INFO] - Step 3400, Loss: {'train/loss': 0.05143596976995468, 'train/grad_norm_pre_clip': 5.183933258056641, 'timing/data': 0.00026344601064920425, 'timing/model': 0.4219058146700263, 'learning_rate/qwen_vl_interface': 9.861973632773748e-06, 'learning_rate/action_model': 9.861973632773748e-05, 'epoch': 0.62})
+[2026-07-28 15:58:19,040][starVLA.training.train_starvla][INFO] - Step 3401, Loss: {'train/loss': 0.012199269607663155, 'train/grad_norm_pre_clip': 0.8980990648269653, 'timing/data': 0.0003477041609585285, 'timing/model': 0.42177491867914796, 'learning_rate/qwen_vl_interface': 9.861056443317105e-06, 'learning_rate/action_model': 9.861056443317105e-05, 'epoch': 0.62})
+[2026-07-28 15:58:25,316][starVLA.training.train_starvla][INFO] - Step 3402, Loss: {'train/loss': 0.05711214989423752, 'train/grad_norm_pre_clip': 2.2307188510894775, 'timing/data': 0.00029379967600107193, 'timing/model': 0.6699340678751469, 'learning_rate/qwen_vl_interface': 9.860136261796e-06, 'learning_rate/action_model': 9.860136261795998e-05, 'epoch': 0.62})
+[2026-07-28 15:58:31,326][starVLA.training.train_starvla][INFO] - Step 3403, Loss: {'train/loss': 0.061789482831954956, 'train/grad_norm_pre_clip': 1.3730684518814087, 'timing/data': 0.00025428785011172295, 'timing/model': 0.42113365372642875, 'learning_rate/qwen_vl_interface': 9.859213088807521e-06, 'learning_rate/action_model': 9.85921308880752e-05, 'epoch': 0.63})
+[2026-07-28 15:58:37,415][starVLA.training.train_starvla][INFO] - Step 3404, Loss: {'train/loss': 0.0655592679977417, 'train/grad_norm_pre_clip': 1.7806172370910645, 'timing/data': 0.0003169858828186989, 'timing/model': 0.4239057432860136, 'learning_rate/qwen_vl_interface': 9.858286924950714e-06, 'learning_rate/action_model': 9.858286924950713e-05, 'epoch': 0.63})
+[2026-07-28 15:58:43,626][starVLA.training.train_starvla][INFO] - Step 3405, Loss: {'train/loss': 0.04514005780220032, 'train/grad_norm_pre_clip': 1.2955931425094604, 'timing/data': 0.00025393301621079445, 'timing/model': 0.4207519688643515, 'learning_rate/qwen_vl_interface': 9.85735777082655e-06, 'learning_rate/action_model': 9.857357770826551e-05, 'epoch': 0.63})
+[2026-07-28 15:58:49,622][starVLA.training.train_starvla][INFO] - Step 3406, Loss: {'train/loss': 0.04316827654838562, 'train/grad_norm_pre_clip': 2.409153461456299, 'timing/data': 0.00031989719718694687, 'timing/model': 0.42220737179741263, 'learning_rate/qwen_vl_interface': 9.856425627037951e-06, 'learning_rate/action_model': 9.856425627037952e-05, 'epoch': 0.63})
+[2026-07-28 15:58:55,700][starVLA.training.train_starvla][INFO] - Step 3407, Loss: {'train/loss': 0.04159759357571602, 'train/grad_norm_pre_clip': 1.6409755945205688, 'timing/data': 0.0003436109982430935, 'timing/model': 0.4204512578435242, 'learning_rate/qwen_vl_interface': 9.855490494189777e-06, 'learning_rate/action_model': 9.855490494189776e-05, 'epoch': 0.63})
+[2026-07-28 15:59:01,922][starVLA.training.train_starvla][INFO] - Step 3408, Loss: {'train/loss': 0.05588522553443909, 'train/grad_norm_pre_clip': 1.3415640592575073, 'timing/data': 0.0003705318085849285, 'timing/model': 0.6558521608822048, 'learning_rate/qwen_vl_interface': 9.854552372888819e-06, 'learning_rate/action_model': 9.854552372888819e-05, 'epoch': 0.63})
+[2026-07-28 15:59:08,232][starVLA.training.train_starvla][INFO] - Step 3409, Loss: {'train/loss': 0.07611235976219177, 'train/grad_norm_pre_clip': 1.4668200016021729, 'timing/data': 0.00037207966670393944, 'timing/model': 0.4186845668591559, 'learning_rate/qwen_vl_interface': 9.853611263743817e-06, 'learning_rate/action_model': 9.853611263743817e-05, 'epoch': 0.63})
+[2026-07-28 15:59:14,214][starVLA.training.train_starvla][INFO] - Step 3410, Loss: {'train/loss': 0.051236458122730255, 'train/grad_norm_pre_clip': 1.701491355895996, 'timing/data': 0.0003608069382607937, 'timing/model': 0.43285041488707066, 'learning_rate/qwen_vl_interface': 9.85266716736545e-06, 'learning_rate/action_model': 9.852667167365449e-05, 'epoch': 0.63})
+[2026-07-28 15:59:20,333][starVLA.training.train_starvla][INFO] - Step 3411, Loss: {'train/loss': 0.03974331542849541, 'train/grad_norm_pre_clip': 0.9510672092437744, 'timing/data': 0.0003879088908433914, 'timing/model': 0.4226503507234156, 'learning_rate/qwen_vl_interface': 9.851720084366324e-06, 'learning_rate/action_model': 9.851720084366324e-05, 'epoch': 0.63})
+[2026-07-28 15:59:26,642][starVLA.training.train_starvla][INFO] - Step 3412, Loss: {'train/loss': 0.05134512484073639, 'train/grad_norm_pre_clip': 2.397771120071411, 'timing/data': 0.0002980274148285389, 'timing/model': 0.42009699484333396, 'learning_rate/qwen_vl_interface': 9.850770015360999e-06, 'learning_rate/action_model': 9.850770015360999e-05, 'epoch': 0.63})
+[2026-07-28 15:59:32,914][starVLA.training.train_starvla][INFO] - Step 3413, Loss: {'train/loss': 0.11455316096544266, 'train/grad_norm_pre_clip': 3.8513472080230713, 'timing/data': 0.00042881863191723824, 'timing/model': 0.41992001701146364, 'learning_rate/qwen_vl_interface': 9.849816960965959e-06, 'learning_rate/action_model': 9.849816960965959e-05, 'epoch': 0.63})
+[2026-07-28 15:59:38,926][starVLA.training.train_starvla][INFO] - Step 3414, Loss: {'train/loss': 0.043386444449424744, 'train/grad_norm_pre_clip': 1.884218692779541, 'timing/data': 0.00029440829530358315, 'timing/model': 0.42141204699873924, 'learning_rate/qwen_vl_interface': 9.848860921799633e-06, 'learning_rate/action_model': 9.848860921799632e-05, 'epoch': 0.63})
+[2026-07-28 15:59:45,121][starVLA.training.train_starvla][INFO] - Step 3415, Loss: {'train/loss': 0.03535591810941696, 'train/grad_norm_pre_clip': 1.4527496099472046, 'timing/data': 0.00035049207508563995, 'timing/model': 0.5990395727567375, 'learning_rate/qwen_vl_interface': 9.847901898482385e-06, 'learning_rate/action_model': 9.847901898482385e-05, 'epoch': 0.63})
+[2026-07-28 15:59:51,141][starVLA.training.train_starvla][INFO] - Step 3416, Loss: {'train/loss': 0.08379869163036346, 'train/grad_norm_pre_clip': 2.976806163787842, 'timing/data': 0.00037203729152679443, 'timing/model': 0.42237117094919086, 'learning_rate/qwen_vl_interface': 9.846939891636512e-06, 'learning_rate/action_model': 9.846939891636512e-05, 'epoch': 0.63})
+[2026-07-28 15:59:57,666][starVLA.training.train_starvla][INFO] - Step 3417, Loss: {'train/loss': 0.0783444419503212, 'train/grad_norm_pre_clip': 1.0006510019302368, 'timing/data': 0.0002460172399878502, 'timing/model': 0.4185977941378951, 'learning_rate/qwen_vl_interface': 9.845974901886253e-06, 'learning_rate/action_model': 9.845974901886252e-05, 'epoch': 0.63})
+[2026-07-28 16:00:03,869][starVLA.training.train_starvla][INFO] - Step 3418, Loss: {'train/loss': 0.08765162527561188, 'train/grad_norm_pre_clip': 2.288050413131714, 'timing/data': 0.00030756788328289986, 'timing/model': 0.4219146301038563, 'learning_rate/qwen_vl_interface': 9.845006929857777e-06, 'learning_rate/action_model': 9.845006929857776e-05, 'epoch': 0.63})
+[2026-07-28 16:00:10,134][starVLA.training.train_starvla][INFO] - Step 3419, Loss: {'train/loss': 0.034886494278907776, 'train/grad_norm_pre_clip': 0.6780771017074585, 'timing/data': 0.00029907096177339554, 'timing/model': 0.4204844259656966, 'learning_rate/qwen_vl_interface': 9.844035976179193e-06, 'learning_rate/action_model': 9.844035976179192e-05, 'epoch': 0.63})
+[2026-07-28 16:00:16,134][starVLA.training.train_starvla][INFO] - Step 3420, Loss: {'train/loss': 0.027254637330770493, 'train/grad_norm_pre_clip': 1.0901751518249512, 'timing/data': 0.0004454776644706726, 'timing/model': 0.42049429612234235, 'learning_rate/qwen_vl_interface': 9.843062041480538e-06, 'learning_rate/action_model': 9.843062041480539e-05, 'epoch': 0.63})
+[2026-07-28 16:00:22,137][starVLA.training.train_starvla][INFO] - Step 3421, Loss: {'train/loss': 0.08812329173088074, 'train/grad_norm_pre_clip': 2.2133219242095947, 'timing/data': 0.00036193476989865303, 'timing/model': 0.42076976923272014, 'learning_rate/qwen_vl_interface': 9.842085126393793e-06, 'learning_rate/action_model': 9.842085126393792e-05, 'epoch': 0.63})
+[2026-07-28 16:00:28,400][starVLA.training.train_starvla][INFO] - Step 3422, Loss: {'train/loss': 0.03597832843661308, 'train/grad_norm_pre_clip': 1.821840763092041, 'timing/data': 0.00037760287523269653, 'timing/model': 0.6636725850403309, 'learning_rate/qwen_vl_interface': 9.841105231552863e-06, 'learning_rate/action_model': 9.841105231552864e-05, 'epoch': 0.63})
+[2026-07-28 16:00:34,671][starVLA.training.train_starvla][INFO] - Step 3423, Loss: {'train/loss': 0.07239320874214172, 'train/grad_norm_pre_clip': 1.873784065246582, 'timing/data': 0.0003212480805814266, 'timing/model': 0.6608362179249525, 'learning_rate/qwen_vl_interface': 9.840122357593594e-06, 'learning_rate/action_model': 9.840122357593594e-05, 'epoch': 0.63})
+[2026-07-28 16:00:40,761][starVLA.training.train_starvla][INFO] - Step 3424, Loss: {'train/loss': 0.03238039091229439, 'train/grad_norm_pre_clip': 1.6935679912567139, 'timing/data': 0.0002577039413154125, 'timing/model': 0.4223787900991738, 'learning_rate/qwen_vl_interface': 9.839136505153762e-06, 'learning_rate/action_model': 9.839136505153762e-05, 'epoch': 0.63})
+[2026-07-28 16:00:46,873][starVLA.training.train_starvla][INFO] - Step 3425, Loss: {'train/loss': 0.03099740296602249, 'train/grad_norm_pre_clip': 0.8700361251831055, 'timing/data': 0.0003771660849452019, 'timing/model': 0.42217214964330196, 'learning_rate/qwen_vl_interface': 9.838147674873076e-06, 'learning_rate/action_model': 9.838147674873075e-05, 'epoch': 0.63})
+[2026-07-28 16:00:52,935][starVLA.training.train_starvla][INFO] - Step 3426, Loss: {'train/loss': 0.04305645823478699, 'train/grad_norm_pre_clip': 1.1424444913864136, 'timing/data': 0.0003294297493994236, 'timing/model': 0.42373455921187997, 'learning_rate/qwen_vl_interface': 9.837155867393176e-06, 'learning_rate/action_model': 9.837155867393175e-05, 'epoch': 0.63})
+[2026-07-28 16:00:58,855][starVLA.training.train_starvla][INFO] - Step 3427, Loss: {'train/loss': 0.06115979701280594, 'train/grad_norm_pre_clip': 2.079375982284546, 'timing/data': 0.0003158445470035076, 'timing/model': 0.42248476296663284, 'learning_rate/qwen_vl_interface': 9.836161083357635e-06, 'learning_rate/action_model': 9.836161083357635e-05, 'epoch': 0.63})
+[2026-07-28 16:01:04,895][starVLA.training.train_starvla][INFO] - Step 3428, Loss: {'train/loss': 0.05850527435541153, 'train/grad_norm_pre_clip': 1.606231689453125, 'timing/data': 0.0003326740115880966, 'timing/model': 0.4215520080178976, 'learning_rate/qwen_vl_interface': 9.835163323411957e-06, 'learning_rate/action_model': 9.835163323411958e-05, 'epoch': 0.63})
+[2026-07-28 16:01:11,253][starVLA.training.train_starvla][INFO] - Step 3429, Loss: {'train/loss': 0.0624706894159317, 'train/grad_norm_pre_clip': 0.953539252281189, 'timing/data': 0.00038769282400608063, 'timing/model': 0.43072311393916607, 'learning_rate/qwen_vl_interface': 9.834162588203581e-06, 'learning_rate/action_model': 9.83416258820358e-05, 'epoch': 0.63})
+[2026-07-28 16:01:17,330][starVLA.training.train_starvla][INFO] - Step 3430, Loss: {'train/loss': 0.04598807543516159, 'train/grad_norm_pre_clip': 1.5763901472091675, 'timing/data': 0.00032421573996543884, 'timing/model': 0.42933209985494614, 'learning_rate/qwen_vl_interface': 9.833158878381872e-06, 'learning_rate/action_model': 9.83315887838187e-05, 'epoch': 0.63})
+[2026-07-28 16:01:23,091][starVLA.training.train_starvla][INFO] - Step 3431, Loss: {'train/loss': 0.028157634660601616, 'train/grad_norm_pre_clip': 1.7161870002746582, 'timing/data': 0.0003003999590873718, 'timing/model': 0.4230348630808294, 'learning_rate/qwen_vl_interface': 9.832152194598123e-06, 'learning_rate/action_model': 9.832152194598122e-05, 'epoch': 0.63})
+[2026-07-28 16:01:29,109][starVLA.training.train_starvla][INFO] - Step 3432, Loss: {'train/loss': 0.02909966930747032, 'train/grad_norm_pre_clip': 0.6712996959686279, 'timing/data': 0.0002928110770881176, 'timing/model': 0.42401614086702466, 'learning_rate/qwen_vl_interface': 9.831142537505563e-06, 'learning_rate/action_model': 9.831142537505562e-05, 'epoch': 0.63})
+[2026-07-28 16:01:35,583][starVLA.training.train_starvla][INFO] - Step 3433, Loss: {'train/loss': 0.024710627272725105, 'train/grad_norm_pre_clip': 0.9358634948730469, 'timing/data': 0.0002819821238517761, 'timing/model': 0.6570946141146123, 'learning_rate/qwen_vl_interface': 9.830129907759347e-06, 'learning_rate/action_model': 9.830129907759347e-05, 'epoch': 0.63})
+[2026-07-28 16:01:41,645][starVLA.training.train_starvla][INFO] - Step 3434, Loss: {'train/loss': 0.03690619021654129, 'train/grad_norm_pre_clip': 1.1245867013931274, 'timing/data': 0.00046495115384459496, 'timing/model': 0.42304856330156326, 'learning_rate/qwen_vl_interface': 9.82911430601656e-06, 'learning_rate/action_model': 9.82911430601656e-05, 'epoch': 0.63})
+[2026-07-28 16:01:47,669][starVLA.training.train_starvla][INFO] - Step 3435, Loss: {'train/loss': 0.05345723032951355, 'train/grad_norm_pre_clip': 1.4391236305236816, 'timing/data': 0.000333959236741066, 'timing/model': 0.4221330927684903, 'learning_rate/qwen_vl_interface': 9.828095732936216e-06, 'learning_rate/action_model': 9.828095732936216e-05, 'epoch': 0.63})
+[2026-07-28 16:01:53,992][starVLA.training.train_starvla][INFO] - Step 3436, Loss: {'train/loss': 0.023132093250751495, 'train/grad_norm_pre_clip': 1.0947468280792236, 'timing/data': 0.00033056503161787987, 'timing/model': 0.6663863970898092, 'learning_rate/qwen_vl_interface': 9.827074189179254e-06, 'learning_rate/action_model': 9.827074189179253e-05, 'epoch': 0.63})
+[2026-07-28 16:02:00,440][starVLA.training.train_starvla][INFO] - Step 3437, Loss: {'train/loss': 0.08882465958595276, 'train/grad_norm_pre_clip': 1.946413516998291, 'timing/data': 0.00027918489649891853, 'timing/model': 0.4293085061945021, 'learning_rate/qwen_vl_interface': 9.82604967540854e-06, 'learning_rate/action_model': 9.82604967540854e-05, 'epoch': 0.63})
+[2026-07-28 16:02:06,766][starVLA.training.train_starvla][INFO] - Step 3438, Loss: {'train/loss': 0.049450237303972244, 'train/grad_norm_pre_clip': 0.9141019582748413, 'timing/data': 0.00038113631308078766, 'timing/model': 0.42258603079244494, 'learning_rate/qwen_vl_interface': 9.825022192288878e-06, 'learning_rate/action_model': 9.825022192288877e-05, 'epoch': 0.63})
+[2026-07-28 16:02:12,729][starVLA.training.train_starvla][INFO] - Step 3439, Loss: {'train/loss': 0.035980224609375, 'train/grad_norm_pre_clip': 1.5935570001602173, 'timing/data': 0.00029265880584716797, 'timing/model': 0.42359553929418325, 'learning_rate/qwen_vl_interface': 9.823991740486982e-06, 'learning_rate/action_model': 9.823991740486982e-05, 'epoch': 0.63})
+[2026-07-28 16:02:18,983][starVLA.training.train_starvla][INFO] - Step 3440, Loss: {'train/loss': 0.02856769971549511, 'train/grad_norm_pre_clip': 1.4945021867752075, 'timing/data': 0.0003181649371981621, 'timing/model': 0.4213914889842272, 'learning_rate/qwen_vl_interface': 9.822958320671506e-06, 'learning_rate/action_model': 9.822958320671506e-05, 'epoch': 0.63})
+[2026-07-28 16:02:25,272][starVLA.training.train_starvla][INFO] - Step 3441, Loss: {'train/loss': 0.047686617821455, 'train/grad_norm_pre_clip': 1.9251430034637451, 'timing/data': 0.00026989588513970375, 'timing/model': 0.421713475137949, 'learning_rate/qwen_vl_interface': 9.821921933513023e-06, 'learning_rate/action_model': 9.821921933513023e-05, 'epoch': 0.63})
+[2026-07-28 16:02:31,267][starVLA.training.train_starvla][INFO] - Step 3442, Loss: {'train/loss': 0.027481935918331146, 'train/grad_norm_pre_clip': 1.4312570095062256, 'timing/data': 0.0004642489366233349, 'timing/model': 0.4240877851843834, 'learning_rate/qwen_vl_interface': 9.820882579684037e-06, 'learning_rate/action_model': 9.820882579684037e-05, 'epoch': 0.63})
+[2026-07-28 16:02:37,632][starVLA.training.train_starvla][INFO] - Step 3443, Loss: {'train/loss': 0.033335261046886444, 'train/grad_norm_pre_clip': 1.5888525247573853, 'timing/data': 0.0003219796344637871, 'timing/model': 0.41939238365739584, 'learning_rate/qwen_vl_interface': 9.819840259858968e-06, 'learning_rate/action_model': 9.819840259858968e-05, 'epoch': 0.63})
+[2026-07-28 16:02:43,870][starVLA.training.train_starvla][INFO] - Step 3444, Loss: {'train/loss': 0.020021039992570877, 'train/grad_norm_pre_clip': 0.7791344523429871, 'timing/data': 0.0003904062323272228, 'timing/model': 0.6570862899534404, 'learning_rate/qwen_vl_interface': 9.818794974714167e-06, 'learning_rate/action_model': 9.818794974714168e-05, 'epoch': 0.63})
+[2026-07-28 16:02:49,875][starVLA.training.train_starvla][INFO] - Step 3445, Loss: {'train/loss': 0.033451374620199203, 'train/grad_norm_pre_clip': 4.942644119262695, 'timing/data': 0.0005044899880886078, 'timing/model': 0.42164169531315565, 'learning_rate/qwen_vl_interface': 9.817746724927912e-06, 'learning_rate/action_model': 9.817746724927912e-05, 'epoch': 0.63})
+[2026-07-28 16:02:56,044][starVLA.training.train_starvla][INFO] - Step 3446, Loss: {'train/loss': 0.08209547400474548, 'train/grad_norm_pre_clip': 3.0740561485290527, 'timing/data': 0.0005002268590033054, 'timing/model': 0.6182741690427065, 'learning_rate/qwen_vl_interface': 9.8166955111804e-06, 'learning_rate/action_model': 9.816695511180399e-05, 'epoch': 0.63})
+[2026-07-28 16:03:02,053][starVLA.training.train_starvla][INFO] - Step 3447, Loss: {'train/loss': 0.05345990136265755, 'train/grad_norm_pre_clip': 1.1797236204147339, 'timing/data': 0.0004923241212964058, 'timing/model': 0.41994153521955013, 'learning_rate/qwen_vl_interface': 9.815641334153746e-06, 'learning_rate/action_model': 9.815641334153746e-05, 'epoch': 0.63})
+[2026-07-28 16:03:08,075][starVLA.training.train_starvla][INFO] - Step 3448, Loss: {'train/loss': 0.010135753080248833, 'train/grad_norm_pre_clip': 0.5496936440467834, 'timing/data': 0.000309172086417675, 'timing/model': 0.42147355200722814, 'learning_rate/qwen_vl_interface': 9.814584194532001e-06, 'learning_rate/action_model': 9.814584194532001e-05, 'epoch': 0.63})
+[2026-07-28 16:03:14,296][starVLA.training.train_starvla][INFO] - Step 3449, Loss: {'train/loss': 0.029352452605962753, 'train/grad_norm_pre_clip': 0.9724862575531006, 'timing/data': 0.0005068057216703892, 'timing/model': 0.6569747077301145, 'learning_rate/qwen_vl_interface': 9.813524093001128e-06, 'learning_rate/action_model': 9.813524093001129e-05, 'epoch': 0.63})
+[2026-07-28 16:03:20,511][starVLA.training.train_starvla][INFO] - Step 3450, Loss: {'train/loss': 0.047685716301202774, 'train/grad_norm_pre_clip': 1.1060693264007568, 'timing/data': 0.0004216572269797325, 'timing/model': 0.4212396699003875, 'learning_rate/qwen_vl_interface': 9.81246103024902e-06, 'learning_rate/action_model': 9.81246103024902e-05, 'epoch': 0.63})
+[2026-07-28 16:03:26,510][starVLA.training.train_starvla][INFO] - Step 3451, Loss: {'train/loss': 0.05536499246954918, 'train/grad_norm_pre_clip': 1.138352870941162, 'timing/data': 0.00030364515259861946, 'timing/model': 0.4234965820796788, 'learning_rate/qwen_vl_interface': 9.811395006965481e-06, 'learning_rate/action_model': 9.81139500696548e-05, 'epoch': 0.63})
+[2026-07-28 16:03:32,500][starVLA.training.train_starvla][INFO] - Step 3452, Loss: {'train/loss': 0.08950895071029663, 'train/grad_norm_pre_clip': 1.3644115924835205, 'timing/data': 0.00048188166692852974, 'timing/model': 0.42392732203006744, 'learning_rate/qwen_vl_interface': 9.810326023842245e-06, 'learning_rate/action_model': 9.810326023842245e-05, 'epoch': 0.63})
+[2026-07-28 16:03:38,519][starVLA.training.train_starvla][INFO] - Step 3453, Loss: {'train/loss': 0.051281049847602844, 'train/grad_norm_pre_clip': 1.8823175430297852, 'timing/data': 0.00030531780794262886, 'timing/model': 0.421741443220526, 'learning_rate/qwen_vl_interface': 9.809254081572963e-06, 'learning_rate/action_model': 9.809254081572961e-05, 'epoch': 0.63})
+[2026-07-28 16:03:44,789][starVLA.training.train_starvla][INFO] - Step 3454, Loss: {'train/loss': 0.05418074131011963, 'train/grad_norm_pre_clip': 1.7252012491226196, 'timing/data': 0.00026345020160079, 'timing/model': 0.4232674897648394, 'learning_rate/qwen_vl_interface': 9.808179180853204e-06, 'learning_rate/action_model': 9.808179180853204e-05, 'epoch': 0.63})
+[2026-07-28 16:03:51,255][starVLA.training.train_starvla][INFO] - Step 3455, Loss: {'train/loss': 0.04714810103178024, 'train/grad_norm_pre_clip': 2.1772749423980713, 'timing/data': 0.0003422689624130726, 'timing/model': 0.6662945412099361, 'learning_rate/qwen_vl_interface': 9.807101322380464e-06, 'learning_rate/action_model': 9.807101322380464e-05, 'epoch': 0.63})
+[2026-07-28 16:03:57,331][starVLA.training.train_starvla][INFO] - Step 3456, Loss: {'train/loss': 0.05839020386338234, 'train/grad_norm_pre_clip': 1.386539101600647, 'timing/data': 0.0002910429611802101, 'timing/model': 0.42213197611272335, 'learning_rate/qwen_vl_interface': 9.806020506854152e-06, 'learning_rate/action_model': 9.806020506854153e-05, 'epoch': 0.63})
+[2026-07-28 16:04:03,363][starVLA.training.train_starvla][INFO] - Step 3457, Loss: {'train/loss': 0.041472241282463074, 'train/grad_norm_pre_clip': 1.4989019632339478, 'timing/data': 0.00038782181218266487, 'timing/model': 0.42426727898418903, 'learning_rate/qwen_vl_interface': 9.804936734975596e-06, 'learning_rate/action_model': 9.804936734975596e-05, 'epoch': 0.63})
+[2026-07-28 16:04:09,376][starVLA.training.train_starvla][INFO] - Step 3458, Loss: {'train/loss': 0.051335759460926056, 'train/grad_norm_pre_clip': 0.9842913150787354, 'timing/data': 0.00033974414691329, 'timing/model': 0.42836742475628853, 'learning_rate/qwen_vl_interface': 9.803850007448046e-06, 'learning_rate/action_model': 9.803850007448045e-05, 'epoch': 0.64})
+[2026-07-28 16:04:15,637][starVLA.training.train_starvla][INFO] - Step 3459, Loss: {'train/loss': 0.06840258836746216, 'train/grad_norm_pre_clip': 1.4759488105773926, 'timing/data': 0.0003085387870669365, 'timing/model': 0.6643087822012603, 'learning_rate/qwen_vl_interface': 9.802760324976666e-06, 'learning_rate/action_model': 9.802760324976667e-05, 'epoch': 0.64})
+[2026-07-28 16:04:21,860][starVLA.training.train_starvla][INFO] - Step 3460, Loss: {'train/loss': 0.014907787553966045, 'train/grad_norm_pre_clip': 0.6751064658164978, 'timing/data': 0.0002694963477551937, 'timing/model': 0.42156263068318367, 'learning_rate/qwen_vl_interface': 9.80166768826854e-06, 'learning_rate/action_model': 9.80166768826854e-05, 'epoch': 0.64})
+[2026-07-28 16:04:27,884][starVLA.training.train_starvla][INFO] - Step 3461, Loss: {'train/loss': 0.030734196305274963, 'train/grad_norm_pre_clip': 0.5736843347549438, 'timing/data': 0.00029073283076286316, 'timing/model': 0.419926137663424, 'learning_rate/qwen_vl_interface': 9.800572098032668e-06, 'learning_rate/action_model': 9.800572098032668e-05, 'epoch': 0.64})
+[2026-07-28 16:04:33,949][starVLA.training.train_starvla][INFO] - Step 3462, Loss: {'train/loss': 0.053800541907548904, 'train/grad_norm_pre_clip': 1.9073015451431274, 'timing/data': 0.0003534378483891487, 'timing/model': 0.42151402262970805, 'learning_rate/qwen_vl_interface': 9.799473554979966e-06, 'learning_rate/action_model': 9.799473554979965e-05, 'epoch': 0.64})
+[2026-07-28 16:04:39,969][starVLA.training.train_starvla][INFO] - Step 3463, Loss: {'train/loss': 0.04592776671051979, 'train/grad_norm_pre_clip': 1.375206708908081, 'timing/data': 0.00031437911093235016, 'timing/model': 0.4208777225576341, 'learning_rate/qwen_vl_interface': 9.798372059823268e-06, 'learning_rate/action_model': 9.798372059823268e-05, 'epoch': 0.64})
+[2026-07-28 16:04:46,464][starVLA.training.train_starvla][INFO] - Step 3464, Loss: {'train/loss': 0.04569605737924576, 'train/grad_norm_pre_clip': 0.7352792024612427, 'timing/data': 0.00028648413717746735, 'timing/model': 0.4216246078722179, 'learning_rate/qwen_vl_interface': 9.797267613277319e-06, 'learning_rate/action_model': 9.797267613277319e-05, 'epoch': 0.64})
+[2026-07-28 16:04:52,506][starVLA.training.train_starvla][INFO] - Step 3465, Loss: {'train/loss': 0.042718686163425446, 'train/grad_norm_pre_clip': 0.8516988754272461, 'timing/data': 0.000293998047709465, 'timing/model': 0.4219100628979504, 'learning_rate/qwen_vl_interface': 9.79616021605879e-06, 'learning_rate/action_model': 9.796160216058788e-05, 'epoch': 0.64})
+[2026-07-28 16:04:58,617][starVLA.training.train_starvla][INFO] - Step 3466, Loss: {'train/loss': 0.04427996277809143, 'train/grad_norm_pre_clip': 1.9459768533706665, 'timing/data': 0.00037131505087018013, 'timing/model': 0.454867844004184, 'learning_rate/qwen_vl_interface': 9.795049868886247e-06, 'learning_rate/action_model': 9.795049868886247e-05, 'epoch': 0.64})
+[2026-07-28 16:05:04,997][starVLA.training.train_starvla][INFO] - Step 3467, Loss: {'train/loss': 0.03747544437646866, 'train/grad_norm_pre_clip': 1.4993185997009277, 'timing/data': 0.00026397407054901123, 'timing/model': 0.4236554130911827, 'learning_rate/qwen_vl_interface': 9.793936572480192e-06, 'learning_rate/action_model': 9.793936572480192e-05, 'epoch': 0.64})
+[2026-07-28 16:05:11,061][starVLA.training.train_starvla][INFO] - Step 3468, Loss: {'train/loss': 0.02932615950703621, 'train/grad_norm_pre_clip': 1.2730203866958618, 'timing/data': 0.0003370782360434532, 'timing/model': 0.4215478179976344, 'learning_rate/qwen_vl_interface': 9.792820327563028e-06, 'learning_rate/action_model': 9.792820327563028e-05, 'epoch': 0.64})
+[2026-07-28 16:05:17,359][starVLA.training.train_starvla][INFO] - Step 3469, Loss: {'train/loss': 0.04527532681822777, 'train/grad_norm_pre_clip': 0.9891153573989868, 'timing/data': 0.0003357641398906708, 'timing/model': 0.6835866002365947, 'learning_rate/qwen_vl_interface': 9.791701134859076e-06, 'learning_rate/action_model': 9.791701134859075e-05, 'epoch': 0.64})
+[2026-07-28 16:05:23,345][starVLA.training.train_starvla][INFO] - Step 3470, Loss: {'train/loss': 0.03767320141196251, 'train/grad_norm_pre_clip': 1.0710227489471436, 'timing/data': 0.00025188783183693886, 'timing/model': 0.4199976879172027, 'learning_rate/qwen_vl_interface': 9.790578995094565e-06, 'learning_rate/action_model': 9.790578995094563e-05, 'epoch': 0.64})
+[2026-07-28 16:05:29,601][starVLA.training.train_starvla][INFO] - Step 3471, Loss: {'train/loss': 0.025788966566324234, 'train/grad_norm_pre_clip': 1.238542914390564, 'timing/data': 0.00037989066913723946, 'timing/model': 0.6658004061318934, 'learning_rate/qwen_vl_interface': 9.789453908997639e-06, 'learning_rate/action_model': 9.789453908997638e-05, 'epoch': 0.64})
+[2026-07-28 16:05:35,569][starVLA.training.train_starvla][INFO] - Step 3472, Loss: {'train/loss': 0.037617407739162445, 'train/grad_norm_pre_clip': 0.8909083008766174, 'timing/data': 0.00035084178671240807, 'timing/model': 0.4203019910492003, 'learning_rate/qwen_vl_interface': 9.788325877298355e-06, 'learning_rate/action_model': 9.788325877298355e-05, 'epoch': 0.64})
+[2026-07-28 16:05:41,563][starVLA.training.train_starvla][INFO] - Step 3473, Loss: {'train/loss': 0.030051544308662415, 'train/grad_norm_pre_clip': 1.3733792304992676, 'timing/data': 0.00031170016154646873, 'timing/model': 0.4225244247354567, 'learning_rate/qwen_vl_interface': 9.787194900728682e-06, 'learning_rate/action_model': 9.787194900728683e-05, 'epoch': 0.64})
+[2026-07-28 16:05:47,602][starVLA.training.train_starvla][INFO] - Step 3474, Loss: {'train/loss': 0.03165017068386078, 'train/grad_norm_pre_clip': 1.063460111618042, 'timing/data': 0.00036305468529462814, 'timing/model': 0.4220833037979901, 'learning_rate/qwen_vl_interface': 9.786060980022499e-06, 'learning_rate/action_model': 9.786060980022498e-05, 'epoch': 0.64})
+[2026-07-28 16:05:53,579][starVLA.training.train_starvla][INFO] - Step 3475, Loss: {'train/loss': 0.02556273713707924, 'train/grad_norm_pre_clip': 1.4096177816390991, 'timing/data': 0.0002961978316307068, 'timing/model': 0.4217185592278838, 'learning_rate/qwen_vl_interface': 9.784924115915592e-06, 'learning_rate/action_model': 9.784924115915592e-05, 'epoch': 0.64})
+[2026-07-28 16:05:59,867][starVLA.training.train_starvla][INFO] - Step 3476, Loss: {'train/loss': 0.021725986152887344, 'train/grad_norm_pre_clip': 1.877611517906189, 'timing/data': 0.00024951668456196785, 'timing/model': 0.4215891412459314, 'learning_rate/qwen_vl_interface': 9.783784309145663e-06, 'learning_rate/action_model': 9.783784309145661e-05, 'epoch': 0.64})
+[2026-07-28 16:06:06,181][starVLA.training.train_starvla][INFO] - Step 3477, Loss: {'train/loss': 0.022667644545435905, 'train/grad_norm_pre_clip': 1.5712032318115234, 'timing/data': 0.0002473066560924053, 'timing/model': 0.4210102679207921, 'learning_rate/qwen_vl_interface': 9.782641560452317e-06, 'learning_rate/action_model': 9.782641560452316e-05, 'epoch': 0.64})
+[2026-07-28 16:06:12,209][starVLA.training.train_starvla][INFO] - Step 3478, Loss: {'train/loss': 0.023643948137760162, 'train/grad_norm_pre_clip': 2.680088758468628, 'timing/data': 0.00037177838385105133, 'timing/model': 0.4208153933286667, 'learning_rate/qwen_vl_interface': 9.78149587057707e-06, 'learning_rate/action_model': 9.78149587057707e-05, 'epoch': 0.64})
+[2026-07-28 16:06:18,301][starVLA.training.train_starvla][INFO] - Step 3479, Loss: {'train/loss': 0.10659150779247284, 'train/grad_norm_pre_clip': 3.1607706546783447, 'timing/data': 0.00042522791773080826, 'timing/model': 0.42122645396739244, 'learning_rate/qwen_vl_interface': 9.780347240263354e-06, 'learning_rate/action_model': 9.780347240263353e-05, 'epoch': 0.64})
+[2026-07-28 16:06:24,551][starVLA.training.train_starvla][INFO] - Step 3480, Loss: {'train/loss': 0.05108846724033356, 'train/grad_norm_pre_clip': 1.57054603099823, 'timing/data': 0.00030004093423485756, 'timing/model': 0.658767803106457, 'learning_rate/qwen_vl_interface': 9.779195670256498e-06, 'learning_rate/action_model': 9.779195670256497e-05, 'epoch': 0.64})
+[2026-07-28 16:06:30,542][starVLA.training.train_starvla][INFO] - Step 3481, Loss: {'train/loss': 0.06587036699056625, 'train/grad_norm_pre_clip': 2.156226634979248, 'timing/data': 0.0003233780153095722, 'timing/model': 0.4213377577252686, 'learning_rate/qwen_vl_interface': 9.778041161303746e-06, 'learning_rate/action_model': 9.778041161303745e-05, 'epoch': 0.64})
+[2026-07-28 16:06:36,522][starVLA.training.train_starvla][INFO] - Step 3482, Loss: {'train/loss': 0.03589381277561188, 'train/grad_norm_pre_clip': 2.297934055328369, 'timing/data': 0.0003015040419995785, 'timing/model': 0.42225232580676675, 'learning_rate/qwen_vl_interface': 9.776883714154242e-06, 'learning_rate/action_model': 9.776883714154241e-05, 'epoch': 0.64})
+[2026-07-28 16:06:42,505][starVLA.training.train_starvla][INFO] - Step 3483, Loss: {'train/loss': 0.011425968259572983, 'train/grad_norm_pre_clip': 0.5306787490844727, 'timing/data': 0.0003715893253684044, 'timing/model': 0.4221376171335578, 'learning_rate/qwen_vl_interface': 9.775723329559044e-06, 'learning_rate/action_model': 9.775723329559044e-05, 'epoch': 0.64})
+[2026-07-28 16:06:49,038][starVLA.training.train_starvla][INFO] - Step 3484, Loss: {'train/loss': 0.030032970011234283, 'train/grad_norm_pre_clip': 1.9745179414749146, 'timing/data': 0.00026709167286753654, 'timing/model': 0.4210192454047501, 'learning_rate/qwen_vl_interface': 9.774560008271114e-06, 'learning_rate/action_model': 9.774560008271114e-05, 'epoch': 0.64})
+[2026-07-28 16:06:55,083][starVLA.training.train_starvla][INFO] - Step 3485, Loss: {'train/loss': 0.0739501565694809, 'train/grad_norm_pre_clip': 3.3893074989318848, 'timing/data': 0.0006776829250156879, 'timing/model': 0.45519459107890725, 'learning_rate/qwen_vl_interface': 9.773393751045315e-06, 'learning_rate/action_model': 9.773393751045315e-05, 'epoch': 0.64})
+[2026-07-28 16:07:01,176][starVLA.training.train_starvla][INFO] - Step 3486, Loss: {'train/loss': 0.06912720203399658, 'train/grad_norm_pre_clip': 1.8014365434646606, 'timing/data': 0.0002690879628062248, 'timing/model': 0.421411014162004, 'learning_rate/qwen_vl_interface': 9.772224558638422e-06, 'learning_rate/action_model': 9.772224558638422e-05, 'epoch': 0.64})
+[2026-07-28 16:07:07,406][starVLA.training.train_starvla][INFO] - Step 3487, Loss: {'train/loss': 0.031045418232679367, 'train/grad_norm_pre_clip': 1.4134551286697388, 'timing/data': 0.00030937138944864273, 'timing/model': 0.6335215768776834, 'learning_rate/qwen_vl_interface': 9.77105243180911e-06, 'learning_rate/action_model': 9.771052431809109e-05, 'epoch': 0.64})
+[2026-07-28 16:07:13,631][starVLA.training.train_starvla][INFO] - Step 3488, Loss: {'train/loss': 0.10150162875652313, 'train/grad_norm_pre_clip': 2.877471446990967, 'timing/data': 0.00033025583252310753, 'timing/model': 0.421180441044271, 'learning_rate/qwen_vl_interface': 9.76987737131796e-06, 'learning_rate/action_model': 9.76987737131796e-05, 'epoch': 0.64})
+[2026-07-28 16:07:19,616][starVLA.training.train_starvla][INFO] - Step 3489, Loss: {'train/loss': 0.04876559600234032, 'train/grad_norm_pre_clip': 1.4244003295898438, 'timing/data': 0.0002932730130851269, 'timing/model': 0.4220837876200676, 'learning_rate/qwen_vl_interface': 9.768699377927454e-06, 'learning_rate/action_model': 9.768699377927453e-05, 'epoch': 0.64})
+[2026-07-28 16:07:25,594][starVLA.training.train_starvla][INFO] - Step 3490, Loss: {'train/loss': 0.04245448857545853, 'train/grad_norm_pre_clip': 1.3297584056854248, 'timing/data': 0.000603153370320797, 'timing/model': 0.43469524709507823, 'learning_rate/qwen_vl_interface': 9.767518452401981e-06, 'learning_rate/action_model': 9.76751845240198e-05, 'epoch': 0.64})
+[2026-07-28 16:07:32,061][starVLA.training.train_starvla][INFO] - Step 3491, Loss: {'train/loss': 0.04006437957286835, 'train/grad_norm_pre_clip': 1.188307285308838, 'timing/data': 0.0005073752254247665, 'timing/model': 0.6638421337120235, 'learning_rate/qwen_vl_interface': 9.76633459550783e-06, 'learning_rate/action_model': 9.76633459550783e-05, 'epoch': 0.64})
+[2026-07-28 16:07:38,050][starVLA.training.train_starvla][INFO] - Step 3492, Loss: {'train/loss': 0.06389494240283966, 'train/grad_norm_pre_clip': 1.2255216836929321, 'timing/data': 0.0003049289807677269, 'timing/model': 0.4206839259713888, 'learning_rate/qwen_vl_interface': 9.765147808013193e-06, 'learning_rate/action_model': 9.765147808013192e-05, 'epoch': 0.64})
+[2026-07-28 16:07:44,033][starVLA.training.train_starvla][INFO] - Step 3493, Loss: {'train/loss': 0.054488662630319595, 'train/grad_norm_pre_clip': 1.621429443359375, 'timing/data': 0.00041141780093312263, 'timing/model': 0.4361048750579357, 'learning_rate/qwen_vl_interface': 9.763958090688165e-06, 'learning_rate/action_model': 9.763958090688164e-05, 'epoch': 0.64})
+[2026-07-28 16:07:50,039][starVLA.training.train_starvla][INFO] - Step 3494, Loss: {'train/loss': 0.05237603932619095, 'train/grad_norm_pre_clip': 0.7882751226425171, 'timing/data': 0.00028794119134545326, 'timing/model': 0.4390815212391317, 'learning_rate/qwen_vl_interface': 9.762765444304737e-06, 'learning_rate/action_model': 9.762765444304737e-05, 'epoch': 0.64})
+[2026-07-28 16:07:56,243][starVLA.training.train_starvla][INFO] - Step 3495, Loss: {'train/loss': 0.04165727645158768, 'train/grad_norm_pre_clip': 1.0424402952194214, 'timing/data': 0.00025805598124861717, 'timing/model': 0.420467397198081, 'learning_rate/qwen_vl_interface': 9.761569869636808e-06, 'learning_rate/action_model': 9.761569869636808e-05, 'epoch': 0.64})
+[2026-07-28 16:08:02,451][starVLA.training.train_starvla][INFO] - Step 3496, Loss: {'train/loss': 0.05117274075746536, 'train/grad_norm_pre_clip': 1.3463268280029297, 'timing/data': 0.0003155660815536976, 'timing/model': 0.6581997801549733, 'learning_rate/qwen_vl_interface': 9.760371367460172e-06, 'learning_rate/action_model': 9.760371367460171e-05, 'epoch': 0.64})
+[2026-07-28 16:08:08,433][starVLA.training.train_starvla][INFO] - Step 3497, Loss: {'train/loss': 0.02444813773036003, 'train/grad_norm_pre_clip': 0.6565335392951965, 'timing/data': 0.00028821825981140137, 'timing/model': 0.41926082130521536, 'learning_rate/qwen_vl_interface': 9.759169938552524e-06, 'learning_rate/action_model': 9.759169938552524e-05, 'epoch': 0.64})
+[2026-07-28 16:08:14,432][starVLA.training.train_starvla][INFO] - Step 3498, Loss: {'train/loss': 0.04384457692503929, 'train/grad_norm_pre_clip': 0.9464494585990906, 'timing/data': 0.0003542308695614338, 'timing/model': 0.4218437420204282, 'learning_rate/qwen_vl_interface': 9.757965583693458e-06, 'learning_rate/action_model': 9.757965583693457e-05, 'epoch': 0.64})
+[2026-07-28 16:08:20,394][starVLA.training.train_starvla][INFO] - Step 3499, Loss: {'train/loss': 0.03657075762748718, 'train/grad_norm_pre_clip': 0.8546340465545654, 'timing/data': 0.00029316404834389687, 'timing/model': 0.4225956089794636, 'learning_rate/qwen_vl_interface': 9.756758303664467e-06, 'learning_rate/action_model': 9.756758303664468e-05, 'epoch': 0.64})
+[2026-07-28 16:08:26,609][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3500
+[2026-07-28 16:08:31,693][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3500
+[2026-07-28 16:16:17,489][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3500
+[2026-07-28 16:16:17,491][starVLA.training.train_starvla][INFO] - Step 3500, Loss: {'train/loss': 0.0566956028342247, 'train/grad_norm_pre_clip': 1.3621015548706055, 'eval/action_loss/seconds': 5.080289393197745, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.04610523208975792, 'eval/latency_3/loss': 0.04610523208975792, 'eval/flappy/loss': 0.04610523208975792, 'eval/flappy/latency_3/loss': 0.04610523208975792, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 335.5500039495528, 'rl_games_eval/mid_train/mean_length': 2721.9, 'rl_games_eval/mid_train/std_reward': 164.0522634087507, 'rl_games_eval/mid_train/std_length': 1322.178425931992, 'rl_games_eval/mid_train/macro_mean_reward': 335.5500039495528, 'rl_games_eval/mid_train/macro_mean_length': 2721.9, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 335.5500039495528, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2721.9, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 164.0522634087507, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1322.178425931992, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 335.5500039495528, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00029271701350808144, 'timing/model': 0.6676926086656749, 'learning_rate/qwen_vl_interface': 9.755548099248942e-06, 'learning_rate/action_model': 9.75554809924894e-05, 'epoch': 0.64})
+[2026-07-28 16:16:17,493][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3500_state
+[2026-07-28 16:16:17,494][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 16:16:55,487][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3500_state/pytorch_model
+[2026-07-28 16:16:55,488][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3500_state/sampler.bin
+[2026-07-28 16:16:55,488][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3500_state/sampler_1.bin
+[2026-07-28 16:16:55,491][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_3500_state/random_states_0.pkl
+[2026-07-28 16:17:09,084][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 3500
+[2026-07-28 16:17:09,084][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 16:17:09,107][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 16:17:15,457][starVLA.training.train_starvla][INFO] - Step 3501, Loss: {'train/loss': 0.02853921428322792, 'train/grad_norm_pre_clip': 1.0801891088485718, 'timing/data': 0.0003318837843835354, 'timing/model': 0.4197031008079648, 'learning_rate/qwen_vl_interface': 9.75433497123217e-06, 'learning_rate/action_model': 9.754334971232171e-05, 'epoch': 0.64})
+[2026-07-28 16:17:21,862][starVLA.training.train_starvla][INFO] - Step 3502, Loss: {'train/loss': 0.033562757074832916, 'train/grad_norm_pre_clip': 0.8159164190292358, 'timing/data': 0.00028269505128264427, 'timing/model': 0.6580707491375506, 'learning_rate/qwen_vl_interface': 9.753118920401341e-06, 'learning_rate/action_model': 9.75311892040134e-05, 'epoch': 0.64})
+[2026-07-28 16:17:27,874][starVLA.training.train_starvla][INFO] - Step 3503, Loss: {'train/loss': 0.01571878418326378, 'train/grad_norm_pre_clip': 0.38568055629730225, 'timing/data': 0.00024877581745386124, 'timing/model': 0.4374494641087949, 'learning_rate/qwen_vl_interface': 9.751899947545534e-06, 'learning_rate/action_model': 9.751899947545533e-05, 'epoch': 0.64})
+[2026-07-28 16:17:33,934][starVLA.training.train_starvla][INFO] - Step 3504, Loss: {'train/loss': 0.029552482068538666, 'train/grad_norm_pre_clip': 1.4318088293075562, 'timing/data': 0.0002993689849972725, 'timing/model': 0.42141031799837947, 'learning_rate/qwen_vl_interface': 9.750678053455725e-06, 'learning_rate/action_model': 9.750678053455725e-05, 'epoch': 0.64})
+[2026-07-28 16:17:40,367][starVLA.training.train_starvla][INFO] - Step 3505, Loss: {'train/loss': 0.018418261781334877, 'train/grad_norm_pre_clip': 0.4635859727859497, 'timing/data': 0.0003068773075938225, 'timing/model': 0.4204447199590504, 'learning_rate/qwen_vl_interface': 9.749453238924793e-06, 'learning_rate/action_model': 9.749453238924792e-05, 'epoch': 0.64})
+[2026-07-28 16:17:46,682][starVLA.training.train_starvla][INFO] - Step 3506, Loss: {'train/loss': 0.040218498557806015, 'train/grad_norm_pre_clip': 1.9222781658172607, 'timing/data': 0.00028410181403160095, 'timing/model': 0.4208935471251607, 'learning_rate/qwen_vl_interface': 9.7482255047475e-06, 'learning_rate/action_model': 9.7482255047475e-05, 'epoch': 0.64})
+[2026-07-28 16:17:52,692][starVLA.training.train_starvla][INFO] - Step 3507, Loss: {'train/loss': 0.03211423382163048, 'train/grad_norm_pre_clip': 1.6196478605270386, 'timing/data': 0.0002899467945098877, 'timing/model': 0.42083696695044637, 'learning_rate/qwen_vl_interface': 9.746994851720518e-06, 'learning_rate/action_model': 9.746994851720517e-05, 'epoch': 0.64})
+[2026-07-28 16:17:58,906][starVLA.training.train_starvla][INFO] - Step 3508, Loss: {'train/loss': 0.023813553154468536, 'train/grad_norm_pre_clip': 1.1177114248275757, 'timing/data': 0.00029135728254914284, 'timing/model': 0.6578956539742649, 'learning_rate/qwen_vl_interface': 9.745761280642397e-06, 'learning_rate/action_model': 9.745761280642397e-05, 'epoch': 0.64})
+[2026-07-28 16:18:05,016][starVLA.training.train_starvla][INFO] - Step 3509, Loss: {'train/loss': 0.016742175444960594, 'train/grad_norm_pre_clip': 0.7415323257446289, 'timing/data': 0.00031995540484786034, 'timing/model': 0.4194790618494153, 'learning_rate/qwen_vl_interface': 9.744524792313591e-06, 'learning_rate/action_model': 9.74452479231359e-05, 'epoch': 0.64})
+[2026-07-28 16:18:11,246][starVLA.training.train_starvla][INFO] - Step 3510, Loss: {'train/loss': 0.04990473389625549, 'train/grad_norm_pre_clip': 2.5914103984832764, 'timing/data': 0.0003865109756588936, 'timing/model': 0.66403276193887, 'learning_rate/qwen_vl_interface': 9.743285387536445e-06, 'learning_rate/action_model': 9.743285387536444e-05, 'epoch': 0.64})
+[2026-07-28 16:18:17,306][starVLA.training.train_starvla][INFO] - Step 3511, Loss: {'train/loss': 0.03177119791507721, 'train/grad_norm_pre_clip': 1.5654470920562744, 'timing/data': 0.0002632690593600273, 'timing/model': 0.4190517379902303, 'learning_rate/qwen_vl_interface': 9.742043067115193e-06, 'learning_rate/action_model': 9.742043067115192e-05, 'epoch': 0.64})
+[2026-07-28 16:18:23,412][starVLA.training.train_starvla][INFO] - Step 3512, Loss: {'train/loss': 0.01227397471666336, 'train/grad_norm_pre_clip': 0.9702064394950867, 'timing/data': 0.0003401460126042366, 'timing/model': 0.42687880899757147, 'learning_rate/qwen_vl_interface': 9.740797831855965e-06, 'learning_rate/action_model': 9.740797831855965e-05, 'epoch': 0.65})
+[2026-07-28 16:18:29,745][starVLA.training.train_starvla][INFO] - Step 3513, Loss: {'train/loss': 0.09120658785104752, 'train/grad_norm_pre_clip': 2.58971905708313, 'timing/data': 0.0002623852342367172, 'timing/model': 0.42071559792384505, 'learning_rate/qwen_vl_interface': 9.73954968256678e-06, 'learning_rate/action_model': 9.73954968256678e-05, 'epoch': 0.65})
+[2026-07-28 16:18:35,762][starVLA.training.train_starvla][INFO] - Step 3514, Loss: {'train/loss': 0.030987190082669258, 'train/grad_norm_pre_clip': 1.5328949689865112, 'timing/data': 0.00030753808096051216, 'timing/model': 0.4194276938214898, 'learning_rate/qwen_vl_interface': 9.738298620057547e-06, 'learning_rate/action_model': 9.738298620057546e-05, 'epoch': 0.65})
+[2026-07-28 16:18:41,797][starVLA.training.train_starvla][INFO] - Step 3515, Loss: {'train/loss': 0.06702326238155365, 'train/grad_norm_pre_clip': 1.5044950246810913, 'timing/data': 0.0003042779862880707, 'timing/model': 0.42099295696243644, 'learning_rate/qwen_vl_interface': 9.73704464514007e-06, 'learning_rate/action_model': 9.737044645140068e-05, 'epoch': 0.65})
+[2026-07-28 16:18:48,042][starVLA.training.train_starvla][INFO] - Step 3516, Loss: {'train/loss': 0.05109572783112526, 'train/grad_norm_pre_clip': 2.256244421005249, 'timing/data': 0.0002585980109870434, 'timing/model': 0.4202651879750192, 'learning_rate/qwen_vl_interface': 9.735787758628036e-06, 'learning_rate/action_model': 9.735787758628035e-05, 'epoch': 0.65})
+[2026-07-28 16:18:54,322][starVLA.training.train_starvla][INFO] - Step 3517, Loss: {'train/loss': 0.07613195478916168, 'train/grad_norm_pre_clip': 1.43941068649292, 'timing/data': 0.00036486564204096794, 'timing/model': 0.42027496080845594, 'learning_rate/qwen_vl_interface': 9.734527961337028e-06, 'learning_rate/action_model': 9.734527961337028e-05, 'epoch': 0.65})
+[2026-07-28 16:19:00,338][starVLA.training.train_starvla][INFO] - Step 3518, Loss: {'train/loss': 0.027694815769791603, 'train/grad_norm_pre_clip': 1.293442726135254, 'timing/data': 0.0003025899641215801, 'timing/model': 0.4199910731986165, 'learning_rate/qwen_vl_interface': 9.733265254084514e-06, 'learning_rate/action_model': 9.733265254084513e-05, 'epoch': 0.65})
+[2026-07-28 16:19:06,420][starVLA.training.train_starvla][INFO] - Step 3519, Loss: {'train/loss': 0.06346505135297775, 'train/grad_norm_pre_clip': 1.7616468667984009, 'timing/data': 0.0003116209991276264, 'timing/model': 0.42456068377941847, 'learning_rate/qwen_vl_interface': 9.73199963768985e-06, 'learning_rate/action_model': 9.731999637689849e-05, 'epoch': 0.65})
+[2026-07-28 16:19:12,658][starVLA.training.train_starvla][INFO] - Step 3520, Loss: {'train/loss': 0.05260505899786949, 'train/grad_norm_pre_clip': 1.4975502490997314, 'timing/data': 0.0003146701492369175, 'timing/model': 0.6574379419907928, 'learning_rate/qwen_vl_interface': 9.73073111297428e-06, 'learning_rate/action_model': 9.73073111297428e-05, 'epoch': 0.65})
+[2026-07-28 16:19:18,660][starVLA.training.train_starvla][INFO] - Step 3521, Loss: {'train/loss': 0.05009053647518158, 'train/grad_norm_pre_clip': 1.767393708229065, 'timing/data': 0.00030689872801303864, 'timing/model': 0.42370459297671914, 'learning_rate/qwen_vl_interface': 9.72945968076094e-06, 'learning_rate/action_model': 9.729459680760939e-05, 'epoch': 0.65})
+[2026-07-28 16:19:24,910][starVLA.training.train_starvla][INFO] - Step 3522, Loss: {'train/loss': 0.02248658612370491, 'train/grad_norm_pre_clip': 0.8511002659797668, 'timing/data': 0.000299008097499609, 'timing/model': 0.660815313924104, 'learning_rate/qwen_vl_interface': 9.728185341874844e-06, 'learning_rate/action_model': 9.728185341874844e-05, 'epoch': 0.65})
+[2026-07-28 16:19:30,928][starVLA.training.train_starvla][INFO] - Step 3523, Loss: {'train/loss': 0.04578491300344467, 'train/grad_norm_pre_clip': 1.4948034286499023, 'timing/data': 0.0002725888043642044, 'timing/model': 0.4287587758153677, 'learning_rate/qwen_vl_interface': 9.7269080971429e-06, 'learning_rate/action_model': 9.726908097142899e-05, 'epoch': 0.65})
+[2026-07-28 16:19:37,298][starVLA.training.train_starvla][INFO] - Step 3524, Loss: {'train/loss': 0.09225645661354065, 'train/grad_norm_pre_clip': 2.1681220531463623, 'timing/data': 0.00026169605553150177, 'timing/model': 0.4279426313005388, 'learning_rate/qwen_vl_interface': 9.725627947393896e-06, 'learning_rate/action_model': 9.725627947393896e-05, 'epoch': 0.65})
+[2026-07-28 16:19:43,131][starVLA.training.train_starvla][INFO] - Step 3525, Loss: {'train/loss': 0.04278922080993652, 'train/grad_norm_pre_clip': 1.2726293802261353, 'timing/data': 0.00030707940459251404, 'timing/model': 0.42133397702127695, 'learning_rate/qwen_vl_interface': 9.72434489345851e-06, 'learning_rate/action_model': 9.72434489345851e-05, 'epoch': 0.65})
+[2026-07-28 16:19:49,116][starVLA.training.train_starvla][INFO] - Step 3526, Loss: {'train/loss': 0.02929161675274372, 'train/grad_norm_pre_clip': 0.5987128615379333, 'timing/data': 0.0003038090653717518, 'timing/model': 0.42529957881197333, 'learning_rate/qwen_vl_interface': 9.723058936169298e-06, 'learning_rate/action_model': 9.723058936169297e-05, 'epoch': 0.65})
+[2026-07-28 16:19:55,120][starVLA.training.train_starvla][INFO] - Step 3527, Loss: {'train/loss': 0.03671162202954292, 'train/grad_norm_pre_clip': 1.2754385471343994, 'timing/data': 0.00029398221522569656, 'timing/model': 0.4224172979593277, 'learning_rate/qwen_vl_interface': 9.721770076360707e-06, 'learning_rate/action_model': 9.721770076360706e-05, 'epoch': 0.65})
+[2026-07-28 16:20:01,361][starVLA.training.train_starvla][INFO] - Step 3528, Loss: {'train/loss': 0.02625090442597866, 'train/grad_norm_pre_clip': 0.6216320395469666, 'timing/data': 0.0002870280295610428, 'timing/model': 0.6587581671774387, 'learning_rate/qwen_vl_interface': 9.720478314869061e-06, 'learning_rate/action_model': 9.72047831486906e-05, 'epoch': 0.65})
+[2026-07-28 16:20:07,360][starVLA.training.train_starvla][INFO] - Step 3529, Loss: {'train/loss': 0.05390394479036331, 'train/grad_norm_pre_clip': 0.7448192834854126, 'timing/data': 0.00025626691058278084, 'timing/model': 0.4216806720942259, 'learning_rate/qwen_vl_interface': 9.719183652532572e-06, 'learning_rate/action_model': 9.719183652532572e-05, 'epoch': 0.65})
+[2026-07-28 16:20:13,324][starVLA.training.train_starvla][INFO] - Step 3530, Loss: {'train/loss': 0.05656629055738449, 'train/grad_norm_pre_clip': 0.7681964039802551, 'timing/data': 0.00038358569145202637, 'timing/model': 0.47603787714615464, 'learning_rate/qwen_vl_interface': 9.717886090191334e-06, 'learning_rate/action_model': 9.717886090191333e-05, 'epoch': 0.65})
+[2026-07-28 16:20:19,530][starVLA.training.train_starvla][INFO] - Step 3531, Loss: {'train/loss': 0.032655417919158936, 'train/grad_norm_pre_clip': 0.7412170171737671, 'timing/data': 0.0003878651186823845, 'timing/model': 0.4979512239806354, 'learning_rate/qwen_vl_interface': 9.716585628687317e-06, 'learning_rate/action_model': 9.716585628687316e-05, 'epoch': 0.65})
+[2026-07-28 16:20:25,523][starVLA.training.train_starvla][INFO] - Step 3532, Loss: {'train/loss': 0.043459754437208176, 'train/grad_norm_pre_clip': 1.107012391090393, 'timing/data': 0.00031953584402799606, 'timing/model': 0.4230838119983673, 'learning_rate/qwen_vl_interface': 9.715282268864379e-06, 'learning_rate/action_model': 9.715282268864378e-05, 'epoch': 0.65})
+[2026-07-28 16:20:31,823][starVLA.training.train_starvla][INFO] - Step 3533, Loss: {'train/loss': 0.041582513600587845, 'train/grad_norm_pre_clip': 1.012601375579834, 'timing/data': 0.0003125709481537342, 'timing/model': 0.6778712291270494, 'learning_rate/qwen_vl_interface': 9.713976011568256e-06, 'learning_rate/action_model': 9.713976011568256e-05, 'epoch': 0.65})
+[2026-07-28 16:20:38,248][starVLA.training.train_starvla][INFO] - Step 3534, Loss: {'train/loss': 0.023273851722478867, 'train/grad_norm_pre_clip': 1.01283597946167, 'timing/data': 0.0002623959444463253, 'timing/model': 0.4204388651996851, 'learning_rate/qwen_vl_interface': 9.712666857646565e-06, 'learning_rate/action_model': 9.712666857646564e-05, 'epoch': 0.65})
+[2026-07-28 16:20:44,273][starVLA.training.train_starvla][INFO] - Step 3535, Loss: {'train/loss': 0.04548661783337593, 'train/grad_norm_pre_clip': 1.3335886001586914, 'timing/data': 0.0003697141073644161, 'timing/model': 0.4223922099918127, 'learning_rate/qwen_vl_interface': 9.711354807948798e-06, 'learning_rate/action_model': 9.711354807948798e-05, 'epoch': 0.65})
+[2026-07-28 16:20:50,354][starVLA.training.train_starvla][INFO] - Step 3536, Loss: {'train/loss': 0.04891516640782356, 'train/grad_norm_pre_clip': 0.7334453463554382, 'timing/data': 0.00027770083397626877, 'timing/model': 0.4201095080934465, 'learning_rate/qwen_vl_interface': 9.710039863326334e-06, 'learning_rate/action_model': 9.710039863326334e-05, 'epoch': 0.65})
+[2026-07-28 16:20:56,824][starVLA.training.train_starvla][INFO] - Step 3537, Loss: {'train/loss': 0.03277844935655594, 'train/grad_norm_pre_clip': 0.9879000186920166, 'timing/data': 0.0003213486634194851, 'timing/model': 0.41922188084572554, 'learning_rate/qwen_vl_interface': 9.708722024632425e-06, 'learning_rate/action_model': 9.708722024632424e-05, 'epoch': 0.65})
+[2026-07-28 16:21:03,181][starVLA.training.train_starvla][INFO] - Step 3538, Loss: {'train/loss': 0.045784104615449905, 'train/grad_norm_pre_clip': 0.9897965788841248, 'timing/data': 0.0002570287324488163, 'timing/model': 0.4202083717100322, 'learning_rate/qwen_vl_interface': 9.7074012927222e-06, 'learning_rate/action_model': 9.7074012927222e-05, 'epoch': 0.65})
+[2026-07-28 16:21:09,205][starVLA.training.train_starvla][INFO] - Step 3539, Loss: {'train/loss': 0.0153166763484478, 'train/grad_norm_pre_clip': 1.0678660869598389, 'timing/data': 0.0003218729980289936, 'timing/model': 0.420212232042104, 'learning_rate/qwen_vl_interface': 9.706077668452671e-06, 'learning_rate/action_model': 9.706077668452671e-05, 'epoch': 0.65})
+[2026-07-28 16:21:15,177][starVLA.training.train_starvla][INFO] - Step 3540, Loss: {'train/loss': 0.0550321601331234, 'train/grad_norm_pre_clip': 1.4173177480697632, 'timing/data': 0.0003884262405335903, 'timing/model': 0.4217430353164673, 'learning_rate/qwen_vl_interface': 9.704751152682722e-06, 'learning_rate/action_model': 9.704751152682721e-05, 'epoch': 0.65})
+[2026-07-28 16:21:21,730][starVLA.training.train_starvla][INFO] - Step 3541, Loss: {'train/loss': 0.06422823667526245, 'train/grad_norm_pre_clip': 2.997465133666992, 'timing/data': 0.00025572581216692924, 'timing/model': 0.6579642561264336, 'learning_rate/qwen_vl_interface': 9.703421746273115e-06, 'learning_rate/action_model': 9.703421746273114e-05, 'epoch': 0.65})
+[2026-07-28 16:21:27,715][starVLA.training.train_starvla][INFO] - Step 3542, Loss: {'train/loss': 0.0860675647854805, 'train/grad_norm_pre_clip': 3.085832357406616, 'timing/data': 0.00028916215524077415, 'timing/model': 0.4211691217496991, 'learning_rate/qwen_vl_interface': 9.702089450086486e-06, 'learning_rate/action_model': 9.702089450086486e-05, 'epoch': 0.65})
+[2026-07-28 16:21:33,750][starVLA.training.train_starvla][INFO] - Step 3543, Loss: {'train/loss': 0.021313834935426712, 'train/grad_norm_pre_clip': 0.8325854539871216, 'timing/data': 0.00030247587710618973, 'timing/model': 0.4223032663576305, 'learning_rate/qwen_vl_interface': 9.700754264987347e-06, 'learning_rate/action_model': 9.700754264987347e-05, 'epoch': 0.65})
+[2026-07-28 16:21:40,020][starVLA.training.train_starvla][INFO] - Step 3544, Loss: {'train/loss': 0.04566388204693794, 'train/grad_norm_pre_clip': 1.3399922847747803, 'timing/data': 0.0007277242839336395, 'timing/model': 0.6702154013328254, 'learning_rate/qwen_vl_interface': 9.69941619184209e-06, 'learning_rate/action_model': 9.699416191842089e-05, 'epoch': 0.65})
+[2026-07-28 16:21:46,445][starVLA.training.train_starvla][INFO] - Step 3545, Loss: {'train/loss': 0.06936641037464142, 'train/grad_norm_pre_clip': 1.9373079538345337, 'timing/data': 0.00030408892780542374, 'timing/model': 0.6562070962972939, 'learning_rate/qwen_vl_interface': 9.69807523151897e-06, 'learning_rate/action_model': 9.698075231518969e-05, 'epoch': 0.65})
+[2026-07-28 16:21:52,418][starVLA.training.train_starvla][INFO] - Step 3546, Loss: {'train/loss': 0.030730193480849266, 'train/grad_norm_pre_clip': 0.9342383742332458, 'timing/data': 0.00026437267661094666, 'timing/model': 0.41889523481950164, 'learning_rate/qwen_vl_interface': 9.696731384888125e-06, 'learning_rate/action_model': 9.696731384888125e-05, 'epoch': 0.65})
+[2026-07-28 16:21:58,476][starVLA.training.train_starvla][INFO] - Step 3547, Loss: {'train/loss': 0.014872809872031212, 'train/grad_norm_pre_clip': 0.5299765467643738, 'timing/data': 0.0003102370537817478, 'timing/model': 0.43504054797813296, 'learning_rate/qwen_vl_interface': 9.695384652821562e-06, 'learning_rate/action_model': 9.695384652821561e-05, 'epoch': 0.65})
+[2026-07-28 16:22:04,517][starVLA.training.train_starvla][INFO] - Step 3548, Loss: {'train/loss': 0.053163863718509674, 'train/grad_norm_pre_clip': 1.248989462852478, 'timing/data': 0.0003230827860534191, 'timing/model': 0.42253714613616467, 'learning_rate/qwen_vl_interface': 9.694035036193162e-06, 'learning_rate/action_model': 9.694035036193161e-05, 'epoch': 0.65})
+[2026-07-28 16:22:10,511][starVLA.training.train_starvla][INFO] - Step 3549, Loss: {'train/loss': 0.046319957822561264, 'train/grad_norm_pre_clip': 1.2259577512741089, 'timing/data': 0.00039294827729463577, 'timing/model': 0.42079979181289673, 'learning_rate/qwen_vl_interface': 9.692682535878674e-06, 'learning_rate/action_model': 9.692682535878673e-05, 'epoch': 0.65})
+[2026-07-28 16:22:16,817][starVLA.training.train_starvla][INFO] - Step 3550, Loss: {'train/loss': 0.03549136966466904, 'train/grad_norm_pre_clip': 1.61020028591156, 'timing/data': 0.00025495467707514763, 'timing/model': 0.4183396231383085, 'learning_rate/qwen_vl_interface': 9.691327152755721e-06, 'learning_rate/action_model': 9.691327152755722e-05, 'epoch': 0.65})
+[2026-07-28 16:22:23,053][starVLA.training.train_starvla][INFO] - Step 3551, Loss: {'train/loss': 0.04742172732949257, 'train/grad_norm_pre_clip': 1.4572850465774536, 'timing/data': 0.00033139996230602264, 'timing/model': 0.4194978619925678, 'learning_rate/qwen_vl_interface': 9.6899688877038e-06, 'learning_rate/action_model': 9.689968887703798e-05, 'epoch': 0.65})
+[2026-07-28 16:22:28,866][starVLA.training.train_starvla][INFO] - Step 3552, Loss: {'train/loss': 0.03027896024286747, 'train/grad_norm_pre_clip': 1.0550636053085327, 'timing/data': 0.00039587216451764107, 'timing/model': 0.4398898067884147, 'learning_rate/qwen_vl_interface': 9.688607741604269e-06, 'learning_rate/action_model': 9.688607741604269e-05, 'epoch': 0.65})
+[2026-07-28 16:22:34,877][starVLA.training.train_starvla][INFO] - Step 3553, Loss: {'train/loss': 0.05346185714006424, 'train/grad_norm_pre_clip': 1.0569517612457275, 'timing/data': 0.0003080936148762703, 'timing/model': 0.42158573074266315, 'learning_rate/qwen_vl_interface': 9.687243715340365e-06, 'learning_rate/action_model': 9.687243715340364e-05, 'epoch': 0.65})
+[2026-07-28 16:22:41,168][starVLA.training.train_starvla][INFO] - Step 3554, Loss: {'train/loss': 0.04782072827219963, 'train/grad_norm_pre_clip': 1.388412356376648, 'timing/data': 0.0003518438898026943, 'timing/model': 0.47078903997316957, 'learning_rate/qwen_vl_interface': 9.685876809797188e-06, 'learning_rate/action_model': 9.685876809797188e-05, 'epoch': 0.65})
+[2026-07-28 16:22:47,526][starVLA.training.train_starvla][INFO] - Step 3555, Loss: {'train/loss': 0.04468914866447449, 'train/grad_norm_pre_clip': 1.3820924758911133, 'timing/data': 0.000309669878333807, 'timing/model': 0.4264940358698368, 'learning_rate/qwen_vl_interface': 9.68450702586171e-06, 'learning_rate/action_model': 9.684507025861709e-05, 'epoch': 0.65})
+[2026-07-28 16:22:53,793][starVLA.training.train_starvla][INFO] - Step 3556, Loss: {'train/loss': 0.05481617525219917, 'train/grad_norm_pre_clip': 1.7109991312026978, 'timing/data': 0.0002641887404024601, 'timing/model': 0.4201278639957309, 'learning_rate/qwen_vl_interface': 9.683134364422766e-06, 'learning_rate/action_model': 9.683134364422765e-05, 'epoch': 0.65})
+[2026-07-28 16:22:59,822][starVLA.training.train_starvla][INFO] - Step 3557, Loss: {'train/loss': 0.050224971026182175, 'train/grad_norm_pre_clip': 1.2672703266143799, 'timing/data': 0.00030175503343343735, 'timing/model': 0.44096418330445886, 'learning_rate/qwen_vl_interface': 9.681758826371064e-06, 'learning_rate/action_model': 9.681758826371063e-05, 'epoch': 0.65})
+[2026-07-28 16:23:06,056][starVLA.training.train_starvla][INFO] - Step 3558, Loss: {'train/loss': 0.048123832792043686, 'train/grad_norm_pre_clip': 1.8116532564163208, 'timing/data': 0.0003385823220014572, 'timing/model': 0.42134894616901875, 'learning_rate/qwen_vl_interface': 9.680380412599174e-06, 'learning_rate/action_model': 9.680380412599173e-05, 'epoch': 0.65})
+[2026-07-28 16:23:12,102][starVLA.training.train_starvla][INFO] - Step 3559, Loss: {'train/loss': 0.03520435094833374, 'train/grad_norm_pre_clip': 0.7522068023681641, 'timing/data': 0.00031274696812033653, 'timing/model': 0.42688939999789, 'learning_rate/qwen_vl_interface': 9.678999124001533e-06, 'learning_rate/action_model': 9.678999124001533e-05, 'epoch': 0.65})
+[2026-07-28 16:23:18,087][starVLA.training.train_starvla][INFO] - Step 3560, Loss: {'train/loss': 0.06921364367008209, 'train/grad_norm_pre_clip': 2.2212350368499756, 'timing/data': 0.0003275717608630657, 'timing/model': 0.43201280990615487, 'learning_rate/qwen_vl_interface': 9.67761496147445e-06, 'learning_rate/action_model': 9.677614961474449e-05, 'epoch': 0.65})
+[2026-07-28 16:23:24,536][starVLA.training.train_starvla][INFO] - Step 3561, Loss: {'train/loss': 0.06233571842312813, 'train/grad_norm_pre_clip': 1.8468074798583984, 'timing/data': 0.0003271494060754776, 'timing/model': 0.5981125519610941, 'learning_rate/qwen_vl_interface': 9.676227925916082e-06, 'learning_rate/action_model': 9.676227925916082e-05, 'epoch': 0.65})
+[2026-07-28 16:23:30,527][starVLA.training.train_starvla][INFO] - Step 3562, Loss: {'train/loss': 0.05393030494451523, 'train/grad_norm_pre_clip': 1.7877737283706665, 'timing/data': 0.0005744169466197491, 'timing/model': 0.4265602892264724, 'learning_rate/qwen_vl_interface': 9.674838018226472e-06, 'learning_rate/action_model': 9.674838018226471e-05, 'epoch': 0.65})
+[2026-07-28 16:23:36,514][starVLA.training.train_starvla][INFO] - Step 3563, Loss: {'train/loss': 0.054873548448085785, 'train/grad_norm_pre_clip': 1.105725646018982, 'timing/data': 0.00029249489307403564, 'timing/model': 0.4222842217423022, 'learning_rate/qwen_vl_interface': 9.67344523930751e-06, 'learning_rate/action_model': 9.673445239307511e-05, 'epoch': 0.65})
+[2026-07-28 16:23:42,499][starVLA.training.train_starvla][INFO] - Step 3564, Loss: {'train/loss': 0.036062560975551605, 'train/grad_norm_pre_clip': 1.6249715089797974, 'timing/data': 0.0002990276552736759, 'timing/model': 0.4202970056794584, 'learning_rate/qwen_vl_interface': 9.672049590062959e-06, 'learning_rate/action_model': 9.672049590062958e-05, 'epoch': 0.65})
+[2026-07-28 16:23:48,965][starVLA.training.train_starvla][INFO] - Step 3565, Loss: {'train/loss': 0.04873163253068924, 'train/grad_norm_pre_clip': 1.2858628034591675, 'timing/data': 0.0002456018701195717, 'timing/model': 0.6571627049706876, 'learning_rate/qwen_vl_interface': 9.670651071398436e-06, 'learning_rate/action_model': 9.670651071398436e-05, 'epoch': 0.65})
+[2026-07-28 16:23:54,911][starVLA.training.train_starvla][INFO] - Step 3566, Loss: {'train/loss': 0.07119148969650269, 'train/grad_norm_pre_clip': 1.2404695749282837, 'timing/data': 0.00030770013108849525, 'timing/model': 0.4212343213148415, 'learning_rate/qwen_vl_interface': 9.669249684221427e-06, 'learning_rate/action_model': 9.669249684221428e-05, 'epoch': 0.65})
+[2026-07-28 16:24:00,654][starVLA.training.train_starvla][INFO] - Step 3567, Loss: {'train/loss': 0.010129551403224468, 'train/grad_norm_pre_clip': 0.48533332347869873, 'timing/data': 0.0003225351683795452, 'timing/model': 0.42380677396431565, 'learning_rate/qwen_vl_interface': 9.66784542944128e-06, 'learning_rate/action_model': 9.66784542944128e-05, 'epoch': 0.66})
+[2026-07-28 16:24:06,466][starVLA.training.train_starvla][INFO] - Step 3568, Loss: {'train/loss': 0.047434404492378235, 'train/grad_norm_pre_clip': 0.8402222990989685, 'timing/data': 0.00041539594531059265, 'timing/model': 0.4204704840667546, 'learning_rate/qwen_vl_interface': 9.666438307969194e-06, 'learning_rate/action_model': 9.666438307969195e-05, 'epoch': 0.66})
+[2026-07-28 16:24:12,268][starVLA.training.train_starvla][INFO] - Step 3569, Loss: {'train/loss': 0.02770400419831276, 'train/grad_norm_pre_clip': 0.7948625087738037, 'timing/data': 0.0003366977907717228, 'timing/model': 0.4244981575757265, 'learning_rate/qwen_vl_interface': 9.66502832071824e-06, 'learning_rate/action_model': 9.66502832071824e-05, 'epoch': 0.66})
+[2026-07-28 16:24:18,525][starVLA.training.train_starvla][INFO] - Step 3570, Loss: {'train/loss': 0.06290216743946075, 'train/grad_norm_pre_clip': 1.7407712936401367, 'timing/data': 0.0002596448175609112, 'timing/model': 0.42077898466959596, 'learning_rate/qwen_vl_interface': 9.663615468603344e-06, 'learning_rate/action_model': 9.663615468603343e-05, 'epoch': 0.66})
+[2026-07-28 16:24:24,515][starVLA.training.train_starvla][INFO] - Step 3571, Loss: {'train/loss': 0.038039300590753555, 'train/grad_norm_pre_clip': 0.9761413931846619, 'timing/data': 0.00027840305119752884, 'timing/model': 0.419819972012192, 'learning_rate/qwen_vl_interface': 9.662199752541287e-06, 'learning_rate/action_model': 9.662199752541287e-05, 'epoch': 0.66})
+[2026-07-28 16:24:30,512][starVLA.training.train_starvla][INFO] - Step 3572, Loss: {'train/loss': 0.055245012044906616, 'train/grad_norm_pre_clip': 2.3679797649383545, 'timing/data': 0.00030098622664809227, 'timing/model': 0.42338606622070074, 'learning_rate/qwen_vl_interface': 9.660781173450711e-06, 'learning_rate/action_model': 9.660781173450712e-05, 'epoch': 0.66})
+[2026-07-28 16:24:36,761][starVLA.training.train_starvla][INFO] - Step 3573, Loss: {'train/loss': 0.04441475123167038, 'train/grad_norm_pre_clip': 1.350258708000183, 'timing/data': 0.0004031960852444172, 'timing/model': 0.4188064751215279, 'learning_rate/qwen_vl_interface': 9.659359732252121e-06, 'learning_rate/action_model': 9.659359732252121e-05, 'epoch': 0.66})
+[2026-07-28 16:24:42,727][starVLA.training.train_starvla][INFO] - Step 3574, Loss: {'train/loss': 0.03083888441324234, 'train/grad_norm_pre_clip': 1.0673795938491821, 'timing/data': 0.0003110119141638279, 'timing/model': 0.4227205030620098, 'learning_rate/qwen_vl_interface': 9.657935429867872e-06, 'learning_rate/action_model': 9.657935429867873e-05, 'epoch': 0.66})
+[2026-07-28 16:24:48,718][starVLA.training.train_starvla][INFO] - Step 3575, Loss: {'train/loss': 0.04517307132482529, 'train/grad_norm_pre_clip': 2.3690719604492188, 'timing/data': 0.00031481776386499405, 'timing/model': 0.4208273747935891, 'learning_rate/qwen_vl_interface': 9.65650826722218e-06, 'learning_rate/action_model': 9.656508267222179e-05, 'epoch': 0.66})
+[2026-07-28 16:24:55,081][starVLA.training.train_starvla][INFO] - Step 3576, Loss: {'train/loss': 0.04669848829507828, 'train/grad_norm_pre_clip': 2.026831865310669, 'timing/data': 0.0002960436977446079, 'timing/model': 0.8215343710035086, 'learning_rate/qwen_vl_interface': 9.655078245241114e-06, 'learning_rate/action_model': 9.655078245241112e-05, 'epoch': 0.66})
+[2026-07-28 16:25:00,822][starVLA.training.train_starvla][INFO] - Step 3577, Loss: {'train/loss': 0.04568719118833542, 'train/grad_norm_pre_clip': 1.2487279176712036, 'timing/data': 0.00029158638790249825, 'timing/model': 0.42388609470799565, 'learning_rate/qwen_vl_interface': 9.653645364852598e-06, 'learning_rate/action_model': 9.653645364852597e-05, 'epoch': 0.66})
+[2026-07-28 16:25:06,977][starVLA.training.train_starvla][INFO] - Step 3578, Loss: {'train/loss': 0.06690025329589844, 'train/grad_norm_pre_clip': 1.6038975715637207, 'timing/data': 0.000294731929898262, 'timing/model': 0.4203277239575982, 'learning_rate/qwen_vl_interface': 9.652209626986414e-06, 'learning_rate/action_model': 9.652209626986413e-05, 'epoch': 0.66})
+[2026-07-28 16:25:13,038][starVLA.training.train_starvla][INFO] - Step 3579, Loss: {'train/loss': 0.010136345401406288, 'train/grad_norm_pre_clip': 0.8905214071273804, 'timing/data': 0.0002988278865814209, 'timing/model': 0.42740549985319376, 'learning_rate/qwen_vl_interface': 9.6507710325742e-06, 'learning_rate/action_model': 9.6507710325742e-05, 'epoch': 0.66})
+[2026-07-28 16:25:19,229][starVLA.training.train_starvla][INFO] - Step 3580, Loss: {'train/loss': 0.017695201560854912, 'train/grad_norm_pre_clip': 1.027138113975525, 'timing/data': 0.00031001074239611626, 'timing/model': 0.4208031939342618, 'learning_rate/qwen_vl_interface': 9.649329582549439e-06, 'learning_rate/action_model': 9.649329582549439e-05, 'epoch': 0.66})
+[2026-07-28 16:25:25,440][starVLA.training.train_starvla][INFO] - Step 3581, Loss: {'train/loss': 0.02896576002240181, 'train/grad_norm_pre_clip': 1.340026617050171, 'timing/data': 0.0002599232830107212, 'timing/model': 0.42155385576188564, 'learning_rate/qwen_vl_interface': 9.647885277847476e-06, 'learning_rate/action_model': 9.647885277847476e-05, 'epoch': 0.66})
+[2026-07-28 16:25:31,411][starVLA.training.train_starvla][INFO] - Step 3582, Loss: {'train/loss': 0.030070800334215164, 'train/grad_norm_pre_clip': 2.1988461017608643, 'timing/data': 0.0002979608252644539, 'timing/model': 0.4203146961517632, 'learning_rate/qwen_vl_interface': 9.646438119405503e-06, 'learning_rate/action_model': 9.646438119405503e-05, 'epoch': 0.66})
+[2026-07-28 16:25:37,382][starVLA.training.train_starvla][INFO] - Step 3583, Loss: {'train/loss': 0.022222096100449562, 'train/grad_norm_pre_clip': 0.8239524364471436, 'timing/data': 0.00034277187660336494, 'timing/model': 0.4194520181044936, 'learning_rate/qwen_vl_interface': 9.644988108162566e-06, 'learning_rate/action_model': 9.644988108162567e-05, 'epoch': 0.66})
+[2026-07-28 16:25:43,607][starVLA.training.train_starvla][INFO] - Step 3584, Loss: {'train/loss': 0.08054214715957642, 'train/grad_norm_pre_clip': 3.630831718444824, 'timing/data': 0.0002720542252063751, 'timing/model': 0.41785477194935083, 'learning_rate/qwen_vl_interface': 9.643535245059562e-06, 'learning_rate/action_model': 9.643535245059561e-05, 'epoch': 0.66})
+[2026-07-28 16:25:49,598][starVLA.training.train_starvla][INFO] - Step 3585, Loss: {'train/loss': 0.07057136297225952, 'train/grad_norm_pre_clip': 2.4088497161865234, 'timing/data': 0.00030067889019846916, 'timing/model': 0.4208320891484618, 'learning_rate/qwen_vl_interface': 9.642079531039235e-06, 'learning_rate/action_model': 9.642079531039235e-05, 'epoch': 0.66})
+[2026-07-28 16:25:55,586][starVLA.training.train_starvla][INFO] - Step 3586, Loss: {'train/loss': 0.023444753140211105, 'train/grad_norm_pre_clip': 0.6039674878120422, 'timing/data': 0.00031030410900712013, 'timing/model': 0.42008902598172426, 'learning_rate/qwen_vl_interface': 9.640620967046186e-06, 'learning_rate/action_model': 9.640620967046186e-05, 'epoch': 0.66})
+[2026-07-28 16:26:01,384][starVLA.training.train_starvla][INFO] - Step 3587, Loss: {'train/loss': 0.008233508095145226, 'train/grad_norm_pre_clip': 0.39818060398101807, 'timing/data': 0.0004958887584507465, 'timing/model': 0.47656899876892567, 'learning_rate/qwen_vl_interface': 9.63915955402686e-06, 'learning_rate/action_model': 9.63915955402686e-05, 'epoch': 0.66})
+[2026-07-28 16:26:07,561][starVLA.training.train_starvla][INFO] - Step 3588, Loss: {'train/loss': 0.042987026274204254, 'train/grad_norm_pre_clip': 2.631842851638794, 'timing/data': 0.000451505184173584, 'timing/model': 0.6548346448689699, 'learning_rate/qwen_vl_interface': 9.637695292929552e-06, 'learning_rate/action_model': 9.637695292929552e-05, 'epoch': 0.66})
+[2026-07-28 16:26:13,836][starVLA.training.train_starvla][INFO] - Step 3589, Loss: {'train/loss': 0.033491820096969604, 'train/grad_norm_pre_clip': 1.2469984292984009, 'timing/data': 0.00030609965324401855, 'timing/model': 0.6746934582479298, 'learning_rate/qwen_vl_interface': 9.636228184704405e-06, 'learning_rate/action_model': 9.636228184704405e-05, 'epoch': 0.66})
+[2026-07-28 16:26:19,893][starVLA.training.train_starvla][INFO] - Step 3590, Loss: {'train/loss': 0.04576340690255165, 'train/grad_norm_pre_clip': 2.071505069732666, 'timing/data': 0.00025866739451885223, 'timing/model': 0.4210641640238464, 'learning_rate/qwen_vl_interface': 9.634758230303408e-06, 'learning_rate/action_model': 9.634758230303408e-05, 'epoch': 0.66})
+[2026-07-28 16:26:25,917][starVLA.training.train_starvla][INFO] - Step 3591, Loss: {'train/loss': 0.03380288928747177, 'train/grad_norm_pre_clip': 1.1769410371780396, 'timing/data': 0.0003348500467836857, 'timing/model': 0.43754856334999204, 'learning_rate/qwen_vl_interface': 9.633285430680401e-06, 'learning_rate/action_model': 9.633285430680401e-05, 'epoch': 0.66})
+[2026-07-28 16:26:31,952][starVLA.training.train_starvla][INFO] - Step 3592, Loss: {'train/loss': 0.03716754540801048, 'train/grad_norm_pre_clip': 0.9086735844612122, 'timing/data': 0.0002928832545876503, 'timing/model': 0.4338472760282457, 'learning_rate/qwen_vl_interface': 9.631809786791066e-06, 'learning_rate/action_model': 9.631809786791066e-05, 'epoch': 0.66})
+[2026-07-28 16:26:37,949][starVLA.training.train_starvla][INFO] - Step 3593, Loss: {'train/loss': 0.028964774683117867, 'train/grad_norm_pre_clip': 1.315617561340332, 'timing/data': 0.00030130287632346153, 'timing/model': 0.4206993211992085, 'learning_rate/qwen_vl_interface': 9.630331299592932e-06, 'learning_rate/action_model': 9.630331299592933e-05, 'epoch': 0.66})
+[2026-07-28 16:26:44,200][starVLA.training.train_starvla][INFO] - Step 3594, Loss: {'train/loss': 0.027220547199249268, 'train/grad_norm_pre_clip': 0.7037436962127686, 'timing/data': 0.000244850292801857, 'timing/model': 0.42049298621714115, 'learning_rate/qwen_vl_interface': 9.628849970045378e-06, 'learning_rate/action_model': 9.628849970045378e-05, 'epoch': 0.66})
+[2026-07-28 16:26:50,210][starVLA.training.train_starvla][INFO] - Step 3595, Loss: {'train/loss': 0.03943074867129326, 'train/grad_norm_pre_clip': 1.4979608058929443, 'timing/data': 0.0003800210542976856, 'timing/model': 0.43500413186848164, 'learning_rate/qwen_vl_interface': 9.627365799109617e-06, 'learning_rate/action_model': 9.627365799109617e-05, 'epoch': 0.66})
+[2026-07-28 16:26:56,240][starVLA.training.train_starvla][INFO] - Step 3596, Loss: {'train/loss': 0.060243889689445496, 'train/grad_norm_pre_clip': 1.7512813806533813, 'timing/data': 0.0003084912896156311, 'timing/model': 0.4217804353684187, 'learning_rate/qwen_vl_interface': 9.625878787748712e-06, 'learning_rate/action_model': 9.625878787748713e-05, 'epoch': 0.66})
+[2026-07-28 16:27:02,493][starVLA.training.train_starvla][INFO] - Step 3597, Loss: {'train/loss': 0.05753008648753166, 'train/grad_norm_pre_clip': 1.682041049003601, 'timing/data': 0.00025522196665406227, 'timing/model': 0.42140290373936296, 'learning_rate/qwen_vl_interface': 9.624388936927572e-06, 'learning_rate/action_model': 9.624388936927571e-05, 'epoch': 0.66})
+[2026-07-28 16:27:08,483][starVLA.training.train_starvla][INFO] - Step 3598, Loss: {'train/loss': 0.034942626953125, 'train/grad_norm_pre_clip': 1.9323097467422485, 'timing/data': 0.0003155623562633991, 'timing/model': 0.41959097189828753, 'learning_rate/qwen_vl_interface': 9.622896247612943e-06, 'learning_rate/action_model': 9.622896247612942e-05, 'epoch': 0.66})
+[2026-07-28 16:27:14,535][starVLA.training.train_starvla][INFO] - Step 3599, Loss: {'train/loss': 0.07410641014575958, 'train/grad_norm_pre_clip': 2.07413649559021, 'timing/data': 0.0002930592745542526, 'timing/model': 0.4226559232920408, 'learning_rate/qwen_vl_interface': 9.621400720773414e-06, 'learning_rate/action_model': 9.621400720773414e-05, 'epoch': 0.66})
+[2026-07-28 16:27:20,845][starVLA.training.train_starvla][INFO] - Step 3600, Loss: {'train/loss': 0.02872595749795437, 'train/grad_norm_pre_clip': 0.9574458599090576, 'timing/data': 0.0002640746533870697, 'timing/model': 0.4222723781131208, 'learning_rate/qwen_vl_interface': 9.619902357379416e-06, 'learning_rate/action_model': 9.619902357379415e-05, 'epoch': 0.66})
+[2026-07-28 16:27:27,100][starVLA.training.train_starvla][INFO] - Step 3601, Loss: {'train/loss': 0.043836530297994614, 'train/grad_norm_pre_clip': 1.6082996129989624, 'timing/data': 0.0003402251750230789, 'timing/model': 0.6609552958980203, 'learning_rate/qwen_vl_interface': 9.618401158403222e-06, 'learning_rate/action_model': 9.618401158403222e-05, 'epoch': 0.66})
+[2026-07-28 16:27:33,152][starVLA.training.train_starvla][INFO] - Step 3602, Loss: {'train/loss': 0.046450741589069366, 'train/grad_norm_pre_clip': 1.074546217918396, 'timing/data': 0.0002472386695444584, 'timing/model': 0.42028186470270157, 'learning_rate/qwen_vl_interface': 9.616897124818944e-06, 'learning_rate/action_model': 9.616897124818945e-05, 'epoch': 0.66})
+[2026-07-28 16:27:39,155][starVLA.training.train_starvla][INFO] - Step 3603, Loss: {'train/loss': 0.06114073470234871, 'train/grad_norm_pre_clip': 1.94314706325531, 'timing/data': 0.0002719550393521786, 'timing/model': 0.4195047481916845, 'learning_rate/qwen_vl_interface': 9.615390257602532e-06, 'learning_rate/action_model': 9.615390257602531e-05, 'epoch': 0.66})
+[2026-07-28 16:27:45,145][starVLA.training.train_starvla][INFO] - Step 3604, Loss: {'train/loss': 0.016712870448827744, 'train/grad_norm_pre_clip': 0.6447304487228394, 'timing/data': 0.000387389212846756, 'timing/model': 0.4317166511900723, 'learning_rate/qwen_vl_interface': 9.613880557731776e-06, 'learning_rate/action_model': 9.613880557731776e-05, 'epoch': 0.66})
+[2026-07-28 16:27:51,365][starVLA.training.train_starvla][INFO] - Step 3605, Loss: {'train/loss': 0.05035046115517616, 'train/grad_norm_pre_clip': 2.6289968490600586, 'timing/data': 0.0002766633406281471, 'timing/model': 0.4201097930781543, 'learning_rate/qwen_vl_interface': 9.612368026186304e-06, 'learning_rate/action_model': 9.612368026186304e-05, 'epoch': 0.66})
+[2026-07-28 16:27:57,325][starVLA.training.train_starvla][INFO] - Step 3606, Loss: {'train/loss': 0.04075450450181961, 'train/grad_norm_pre_clip': 0.8642945289611816, 'timing/data': 0.0003069671802222729, 'timing/model': 0.41964408615604043, 'learning_rate/qwen_vl_interface': 9.610852663947582e-06, 'learning_rate/action_model': 9.610852663947583e-05, 'epoch': 0.66})
+[2026-07-28 16:28:03,536][starVLA.training.train_starvla][INFO] - Step 3607, Loss: {'train/loss': 0.05543285980820656, 'train/grad_norm_pre_clip': 2.370155096054077, 'timing/data': 0.0002638939768075943, 'timing/model': 0.41951890103518963, 'learning_rate/qwen_vl_interface': 9.609334471998911e-06, 'learning_rate/action_model': 9.609334471998912e-05, 'epoch': 0.66})
+[2026-07-28 16:28:09,524][starVLA.training.train_starvla][INFO] - Step 3608, Loss: {'train/loss': 0.060026172548532486, 'train/grad_norm_pre_clip': 1.2260973453521729, 'timing/data': 0.0004039253108203411, 'timing/model': 0.4224369330331683, 'learning_rate/qwen_vl_interface': 9.607813451325431e-06, 'learning_rate/action_model': 9.60781345132543e-05, 'epoch': 0.66})
+[2026-07-28 16:28:15,894][starVLA.training.train_starvla][INFO] - Step 3609, Loss: {'train/loss': 0.05562116205692291, 'train/grad_norm_pre_clip': 1.2582882642745972, 'timing/data': 0.0002749459818005562, 'timing/model': 0.42034029122442007, 'learning_rate/qwen_vl_interface': 9.606289602914113e-06, 'learning_rate/action_model': 9.606289602914112e-05, 'epoch': 0.66})
+[2026-07-28 16:28:21,878][starVLA.training.train_starvla][INFO] - Step 3610, Loss: {'train/loss': 0.05544821172952652, 'train/grad_norm_pre_clip': 1.141707181930542, 'timing/data': 0.00032031117007136345, 'timing/model': 0.4245801172219217, 'learning_rate/qwen_vl_interface': 9.604762927753768e-06, 'learning_rate/action_model': 9.604762927753768e-05, 'epoch': 0.66})
+[2026-07-28 16:28:28,181][starVLA.training.train_starvla][INFO] - Step 3611, Loss: {'train/loss': 0.051838502287864685, 'train/grad_norm_pre_clip': 1.3870208263397217, 'timing/data': 0.0002930709160864353, 'timing/model': 0.42157639283686876, 'learning_rate/qwen_vl_interface': 9.603233426835037e-06, 'learning_rate/action_model': 9.603233426835036e-05, 'epoch': 0.66})
+[2026-07-28 16:28:34,457][starVLA.training.train_starvla][INFO] - Step 3612, Loss: {'train/loss': 0.040131308138370514, 'train/grad_norm_pre_clip': 1.120762825012207, 'timing/data': 0.0002556443214416504, 'timing/model': 0.42275393614545465, 'learning_rate/qwen_vl_interface': 9.601701101150399e-06, 'learning_rate/action_model': 9.601701101150398e-05, 'epoch': 0.66})
+[2026-07-28 16:28:40,494][starVLA.training.train_starvla][INFO] - Step 3613, Loss: {'train/loss': 0.046131424605846405, 'train/grad_norm_pre_clip': 2.0978219509124756, 'timing/data': 0.00032779108732938766, 'timing/model': 0.42058329936116934, 'learning_rate/qwen_vl_interface': 9.600165951694162e-06, 'learning_rate/action_model': 9.60016595169416e-05, 'epoch': 0.66})
+[2026-07-28 16:28:46,731][starVLA.training.train_starvla][INFO] - Step 3614, Loss: {'train/loss': 0.0794098824262619, 'train/grad_norm_pre_clip': 1.4613010883331299, 'timing/data': 0.00028744712471961975, 'timing/model': 0.4210265101864934, 'learning_rate/qwen_vl_interface': 9.598627979462466e-06, 'learning_rate/action_model': 9.598627979462467e-05, 'epoch': 0.66})
+[2026-07-28 16:28:52,800][starVLA.training.train_starvla][INFO] - Step 3615, Loss: {'train/loss': 0.02956199273467064, 'train/grad_norm_pre_clip': 1.0124881267547607, 'timing/data': 0.00043178070336580276, 'timing/model': 0.42477636504918337, 'learning_rate/qwen_vl_interface': 9.59708718545329e-06, 'learning_rate/action_model': 9.597087185453288e-05, 'epoch': 0.66})
+[2026-07-28 16:28:59,070][starVLA.training.train_starvla][INFO] - Step 3616, Loss: {'train/loss': 0.05087863653898239, 'train/grad_norm_pre_clip': 1.3597232103347778, 'timing/data': 0.0003039180301129818, 'timing/model': 0.6554896011948586, 'learning_rate/qwen_vl_interface': 9.595543570666433e-06, 'learning_rate/action_model': 9.595543570666432e-05, 'epoch': 0.66})
+[2026-07-28 16:29:05,292][starVLA.training.train_starvla][INFO] - Step 3617, Loss: {'train/loss': 0.04251764714717865, 'train/grad_norm_pre_clip': 0.983805239200592, 'timing/data': 0.00032868003472685814, 'timing/model': 0.5986054372042418, 'learning_rate/qwen_vl_interface': 9.593997136103533e-06, 'learning_rate/action_model': 9.593997136103533e-05, 'epoch': 0.66})
+[2026-07-28 16:29:11,286][starVLA.training.train_starvla][INFO] - Step 3618, Loss: {'train/loss': 0.04964964836835861, 'train/grad_norm_pre_clip': 0.9972090125083923, 'timing/data': 0.0003001280128955841, 'timing/model': 0.42208105279132724, 'learning_rate/qwen_vl_interface': 9.592447882768053e-06, 'learning_rate/action_model': 9.592447882768052e-05, 'epoch': 0.66})
+[2026-07-28 16:29:17,321][starVLA.training.train_starvla][INFO] - Step 3619, Loss: {'train/loss': 0.031202983111143112, 'train/grad_norm_pre_clip': 0.5086246728897095, 'timing/data': 0.00028309179469943047, 'timing/model': 0.4240002976730466, 'learning_rate/qwen_vl_interface': 9.590895811665284e-06, 'learning_rate/action_model': 9.590895811665285e-05, 'epoch': 0.66})
+[2026-07-28 16:29:23,630][starVLA.training.train_starvla][INFO] - Step 3620, Loss: {'train/loss': 0.03755136579275131, 'train/grad_norm_pre_clip': 1.0557023286819458, 'timing/data': 0.0003558071330189705, 'timing/model': 0.4220299022272229, 'learning_rate/qwen_vl_interface': 9.589340923802355e-06, 'learning_rate/action_model': 9.589340923802354e-05, 'epoch': 0.66})
+[2026-07-28 16:29:29,891][starVLA.training.train_starvla][INFO] - Step 3621, Loss: {'train/loss': 0.038185346871614456, 'train/grad_norm_pre_clip': 1.3750362396240234, 'timing/data': 0.000314484816044569, 'timing/model': 0.4340012362226844, 'learning_rate/qwen_vl_interface': 9.587783220188212e-06, 'learning_rate/action_model': 9.587783220188212e-05, 'epoch': 0.67})
+[2026-07-28 16:29:36,068][starVLA.training.train_starvla][INFO] - Step 3622, Loss: {'train/loss': 0.028728991746902466, 'train/grad_norm_pre_clip': 1.0723036527633667, 'timing/data': 0.000368588138371706, 'timing/model': 0.46708156960085034, 'learning_rate/qwen_vl_interface': 9.586222701833631e-06, 'learning_rate/action_model': 9.58622270183363e-05, 'epoch': 0.67})
+[2026-07-28 16:29:42,172][starVLA.training.train_starvla][INFO] - Step 3623, Loss: {'train/loss': 0.03542854264378548, 'train/grad_norm_pre_clip': 1.2415519952774048, 'timing/data': 0.000340957660228014, 'timing/model': 0.421655819285661, 'learning_rate/qwen_vl_interface': 9.584659369751216e-06, 'learning_rate/action_model': 9.584659369751215e-05, 'epoch': 0.67})
+[2026-07-28 16:29:48,715][starVLA.training.train_starvla][INFO] - Step 3624, Loss: {'train/loss': 0.023253239691257477, 'train/grad_norm_pre_clip': 0.7273549437522888, 'timing/data': 0.0003107530064880848, 'timing/model': 0.6636044150218368, 'learning_rate/qwen_vl_interface': 9.583093224955397e-06, 'learning_rate/action_model': 9.583093224955397e-05, 'epoch': 0.67})
+[2026-07-28 16:29:54,718][starVLA.training.train_starvla][INFO] - Step 3625, Loss: {'train/loss': 0.08842200040817261, 'train/grad_norm_pre_clip': 2.2253310680389404, 'timing/data': 0.0005849860608577728, 'timing/model': 0.42336967401206493, 'learning_rate/qwen_vl_interface': 9.58152426846243e-06, 'learning_rate/action_model': 9.581524268462429e-05, 'epoch': 0.67})
+[2026-07-28 16:30:00,706][starVLA.training.train_starvla][INFO] - Step 3626, Loss: {'train/loss': 0.06514547020196915, 'train/grad_norm_pre_clip': 1.7485755681991577, 'timing/data': 0.00036788685247302055, 'timing/model': 0.4208102962002158, 'learning_rate/qwen_vl_interface': 9.579952501290388e-06, 'learning_rate/action_model': 9.579952501290387e-05, 'epoch': 0.67})
+[2026-07-28 16:30:06,952][starVLA.training.train_starvla][INFO] - Step 3627, Loss: {'train/loss': 0.04957476630806923, 'train/grad_norm_pre_clip': 1.2272168397903442, 'timing/data': 0.00027698976919054985, 'timing/model': 0.41968079190701246, 'learning_rate/qwen_vl_interface': 9.57837792445918e-06, 'learning_rate/action_model': 9.578377924459181e-05, 'epoch': 0.67})
+[2026-07-28 16:30:12,979][starVLA.training.train_starvla][INFO] - Step 3628, Loss: {'train/loss': 0.04290393739938736, 'train/grad_norm_pre_clip': 0.9799311757087708, 'timing/data': 0.00039219576865434647, 'timing/model': 0.42320017609745264, 'learning_rate/qwen_vl_interface': 9.57680053899053e-06, 'learning_rate/action_model': 9.576800538990529e-05, 'epoch': 0.67})
+[2026-07-28 16:30:19,231][starVLA.training.train_starvla][INFO] - Step 3629, Loss: {'train/loss': 0.030475109815597534, 'train/grad_norm_pre_clip': 0.8591840267181396, 'timing/data': 0.0002665789797902107, 'timing/model': 0.4192649107426405, 'learning_rate/qwen_vl_interface': 9.575220345907984e-06, 'learning_rate/action_model': 9.575220345907984e-05, 'epoch': 0.67})
+[2026-07-28 16:30:25,228][starVLA.training.train_starvla][INFO] - Step 3630, Loss: {'train/loss': 0.07672674208879471, 'train/grad_norm_pre_clip': 1.3109792470932007, 'timing/data': 0.0002997778356075287, 'timing/model': 0.422881418839097, 'learning_rate/qwen_vl_interface': 9.573637346236915e-06, 'learning_rate/action_model': 9.573637346236915e-05, 'epoch': 0.67})
+[2026-07-28 16:30:31,502][starVLA.training.train_starvla][INFO] - Step 3631, Loss: {'train/loss': 0.02733459696173668, 'train/grad_norm_pre_clip': 0.5306158661842346, 'timing/data': 0.00040946993976831436, 'timing/model': 0.6522368360310793, 'learning_rate/qwen_vl_interface': 9.572051541004513e-06, 'learning_rate/action_model': 9.572051541004513e-05, 'epoch': 0.67})
+[2026-07-28 16:30:37,500][starVLA.training.train_starvla][INFO] - Step 3632, Loss: {'train/loss': 0.020341215655207634, 'train/grad_norm_pre_clip': 0.5845867395401001, 'timing/data': 0.0003019389696419239, 'timing/model': 0.4217169042676687, 'learning_rate/qwen_vl_interface': 9.57046293123979e-06, 'learning_rate/action_model': 9.570462931239789e-05, 'epoch': 0.67})
+[2026-07-28 16:30:43,698][starVLA.training.train_starvla][INFO] - Step 3633, Loss: {'train/loss': 0.02052238956093788, 'train/grad_norm_pre_clip': 1.0075327157974243, 'timing/data': 0.00032605789601802826, 'timing/model': 0.42113494127988815, 'learning_rate/qwen_vl_interface': 9.568871517973576e-06, 'learning_rate/action_model': 9.568871517973577e-05, 'epoch': 0.67})
+[2026-07-28 16:30:49,960][starVLA.training.train_starvla][INFO] - Step 3634, Loss: {'train/loss': 0.026139143854379654, 'train/grad_norm_pre_clip': 0.7604873776435852, 'timing/data': 0.00031503336504101753, 'timing/model': 0.660409573931247, 'learning_rate/qwen_vl_interface': 9.567277302238524e-06, 'learning_rate/action_model': 9.567277302238524e-05, 'epoch': 0.67})
+[2026-07-28 16:30:56,277][starVLA.training.train_starvla][INFO] - Step 3635, Loss: {'train/loss': 0.03301739692687988, 'train/grad_norm_pre_clip': 0.888529896736145, 'timing/data': 0.0002594590187072754, 'timing/model': 0.49086835281923413, 'learning_rate/qwen_vl_interface': 9.565680285069104e-06, 'learning_rate/action_model': 9.565680285069103e-05, 'epoch': 0.67})
+[2026-07-28 16:31:02,495][starVLA.training.train_starvla][INFO] - Step 3636, Loss: {'train/loss': 0.03424251079559326, 'train/grad_norm_pre_clip': 0.9594547748565674, 'timing/data': 0.0003940281458199024, 'timing/model': 0.4277962897904217, 'learning_rate/qwen_vl_interface': 9.564080467501601e-06, 'learning_rate/action_model': 9.5640804675016e-05, 'epoch': 0.67})
+[2026-07-28 16:31:08,249][starVLA.training.train_starvla][INFO] - Step 3637, Loss: {'train/loss': 0.024019064381718636, 'train/grad_norm_pre_clip': 0.7580776810646057, 'timing/data': 0.00029279710724949837, 'timing/model': 0.4205918861553073, 'learning_rate/qwen_vl_interface': 9.562477850574118e-06, 'learning_rate/action_model': 9.562477850574119e-05, 'epoch': 0.67})
+[2026-07-28 16:31:14,582][starVLA.training.train_starvla][INFO] - Step 3638, Loss: {'train/loss': 0.03505639359354973, 'train/grad_norm_pre_clip': 0.9370166659355164, 'timing/data': 0.00037488527595996857, 'timing/model': 0.6696348590776324, 'learning_rate/qwen_vl_interface': 9.56087243532658e-06, 'learning_rate/action_model': 9.56087243532658e-05, 'epoch': 0.67})
+[2026-07-28 16:31:21,111][starVLA.training.train_starvla][INFO] - Step 3639, Loss: {'train/loss': 0.05288734287023544, 'train/grad_norm_pre_clip': 4.017032146453857, 'timing/data': 0.0003632279112935066, 'timing/model': 0.48315050173550844, 'learning_rate/qwen_vl_interface': 9.55926422280072e-06, 'learning_rate/action_model': 9.55926422280072e-05, 'epoch': 0.67})
+[2026-07-28 16:31:27,193][starVLA.training.train_starvla][INFO] - Step 3640, Loss: {'train/loss': 0.12767350673675537, 'train/grad_norm_pre_clip': 3.785937786102295, 'timing/data': 0.00028228480368852615, 'timing/model': 0.4640346341766417, 'learning_rate/qwen_vl_interface': 9.55765321404009e-06, 'learning_rate/action_model': 9.557653214040091e-05, 'epoch': 0.67})
+[2026-07-28 16:31:33,036][starVLA.training.train_starvla][INFO] - Step 3641, Loss: {'train/loss': 0.009848486632108688, 'train/grad_norm_pre_clip': 0.48779937624931335, 'timing/data': 0.000297407153993845, 'timing/model': 0.42801621509715915, 'learning_rate/qwen_vl_interface': 9.556039410090061e-06, 'learning_rate/action_model': 9.556039410090061e-05, 'epoch': 0.67})
+[2026-07-28 16:31:39,272][starVLA.training.train_starvla][INFO] - Step 3642, Loss: {'train/loss': 0.031402282416820526, 'train/grad_norm_pre_clip': 0.8449011445045471, 'timing/data': 0.00025947578251361847, 'timing/model': 0.4233000809326768, 'learning_rate/qwen_vl_interface': 9.554422811997804e-06, 'learning_rate/action_model': 9.554422811997805e-05, 'epoch': 0.67})
+[2026-07-28 16:31:45,559][starVLA.training.train_starvla][INFO] - Step 3643, Loss: {'train/loss': 0.030402498319745064, 'train/grad_norm_pre_clip': 1.3061981201171875, 'timing/data': 0.00029971590265631676, 'timing/model': 0.6821088371798396, 'learning_rate/qwen_vl_interface': 9.55280342081232e-06, 'learning_rate/action_model': 9.55280342081232e-05, 'epoch': 0.67})
+[2026-07-28 16:31:51,621][starVLA.training.train_starvla][INFO] - Step 3644, Loss: {'train/loss': 0.01844109408557415, 'train/grad_norm_pre_clip': 0.5687083601951599, 'timing/data': 0.0002654450945556164, 'timing/model': 0.420473066624254, 'learning_rate/qwen_vl_interface': 9.55118123758441e-06, 'learning_rate/action_model': 9.551181237584409e-05, 'epoch': 0.67})
+[2026-07-28 16:31:57,841][starVLA.training.train_starvla][INFO] - Step 3645, Loss: {'train/loss': 0.057390034198760986, 'train/grad_norm_pre_clip': 3.025020122528076, 'timing/data': 0.00031168991699814796, 'timing/model': 0.5952848987653852, 'learning_rate/qwen_vl_interface': 9.549556263366691e-06, 'learning_rate/action_model': 9.549556263366692e-05, 'epoch': 0.67})
+[2026-07-28 16:32:03,878][starVLA.training.train_starvla][INFO] - Step 3646, Loss: {'train/loss': 0.04435006529092789, 'train/grad_norm_pre_clip': 1.3453947305679321, 'timing/data': 0.0002859700471162796, 'timing/model': 0.4198117679916322, 'learning_rate/qwen_vl_interface': 9.547928499213593e-06, 'learning_rate/action_model': 9.547928499213593e-05, 'epoch': 0.67})
+[2026-07-28 16:32:09,872][starVLA.training.train_starvla][INFO] - Step 3647, Loss: {'train/loss': 0.03355837240815163, 'train/grad_norm_pre_clip': 1.6089743375778198, 'timing/data': 0.00029893405735492706, 'timing/model': 0.4282279359176755, 'learning_rate/qwen_vl_interface': 9.546297946181354e-06, 'learning_rate/action_model': 9.546297946181354e-05, 'epoch': 0.67})
+[2026-07-28 16:32:16,061][starVLA.training.train_starvla][INFO] - Step 3648, Loss: {'train/loss': 0.045284681022167206, 'train/grad_norm_pre_clip': 0.8921685218811035, 'timing/data': 0.0002920101396739483, 'timing/model': 0.4188929758965969, 'learning_rate/qwen_vl_interface': 9.544664605328022e-06, 'learning_rate/action_model': 9.544664605328022e-05, 'epoch': 0.67})
+[2026-07-28 16:32:22,270][starVLA.training.train_starvla][INFO] - Step 3649, Loss: {'train/loss': 0.03211800381541252, 'train/grad_norm_pre_clip': 1.1891839504241943, 'timing/data': 0.0002664569765329361, 'timing/model': 0.4214236307889223, 'learning_rate/qwen_vl_interface': 9.543028477713455e-06, 'learning_rate/action_model': 9.543028477713453e-05, 'epoch': 0.67})
+[2026-07-28 16:32:28,508][starVLA.training.train_starvla][INFO] - Step 3650, Loss: {'train/loss': 0.026124659925699234, 'train/grad_norm_pre_clip': 1.380805253982544, 'timing/data': 0.00027083372697234154, 'timing/model': 0.42129699140787125, 'learning_rate/qwen_vl_interface': 9.541389564399316e-06, 'learning_rate/action_model': 9.541389564399315e-05, 'epoch': 0.67})
+[2026-07-28 16:32:34,542][starVLA.training.train_starvla][INFO] - Step 3651, Loss: {'train/loss': 0.0216220635920763, 'train/grad_norm_pre_clip': 1.3467663526535034, 'timing/data': 0.0002924739383161068, 'timing/model': 0.4365906147286296, 'learning_rate/qwen_vl_interface': 9.53974786644908e-06, 'learning_rate/action_model': 9.53974786644908e-05, 'epoch': 0.67})
+[2026-07-28 16:32:40,573][starVLA.training.train_starvla][INFO] - Step 3652, Loss: {'train/loss': 0.03367616608738899, 'train/grad_norm_pre_clip': 0.8233115673065186, 'timing/data': 0.0007319469004869461, 'timing/model': 0.4244495239108801, 'learning_rate/qwen_vl_interface': 9.538103384928029e-06, 'learning_rate/action_model': 9.538103384928027e-05, 'epoch': 0.67})
+[2026-07-28 16:32:46,790][starVLA.training.train_starvla][INFO] - Step 3653, Loss: {'train/loss': 0.04360201582312584, 'train/grad_norm_pre_clip': 1.0941516160964966, 'timing/data': 0.000274130143225193, 'timing/model': 0.6589431297034025, 'learning_rate/qwen_vl_interface': 9.536456120903244e-06, 'learning_rate/action_model': 9.536456120903244e-05, 'epoch': 0.67})
+[2026-07-28 16:32:52,996][starVLA.training.train_starvla][INFO] - Step 3654, Loss: {'train/loss': 0.04986557364463806, 'train/grad_norm_pre_clip': 2.35788893699646, 'timing/data': 0.0002735676243901253, 'timing/model': 0.6530560650862753, 'learning_rate/qwen_vl_interface': 9.534806075443619e-06, 'learning_rate/action_model': 9.534806075443619e-05, 'epoch': 0.67})
+[2026-07-28 16:32:58,944][starVLA.training.train_starvla][INFO] - Step 3655, Loss: {'train/loss': 0.02329934574663639, 'train/grad_norm_pre_clip': 1.1296899318695068, 'timing/data': 0.00029231607913970947, 'timing/model': 0.4205063208937645, 'learning_rate/qwen_vl_interface': 9.533153249619853e-06, 'learning_rate/action_model': 9.533153249619853e-05, 'epoch': 0.67})
+[2026-07-28 16:33:04,692][starVLA.training.train_starvla][INFO] - Step 3656, Loss: {'train/loss': 0.024481408298015594, 'train/grad_norm_pre_clip': 1.588990330696106, 'timing/data': 0.0002810060977935791, 'timing/model': 0.4211190938949585, 'learning_rate/qwen_vl_interface': 9.531497644504444e-06, 'learning_rate/action_model': 9.531497644504445e-05, 'epoch': 0.67})
+[2026-07-28 16:33:10,784][starVLA.training.train_starvla][INFO] - Step 3657, Loss: {'train/loss': 0.01517345942556858, 'train/grad_norm_pre_clip': 0.4899754524230957, 'timing/data': 0.00026773707941174507, 'timing/model': 0.4297103574499488, 'learning_rate/qwen_vl_interface': 9.529839261171698e-06, 'learning_rate/action_model': 9.529839261171696e-05, 'epoch': 0.67})
+[2026-07-28 16:33:17,169][starVLA.training.train_starvla][INFO] - Step 3658, Loss: {'train/loss': 0.04653048515319824, 'train/grad_norm_pre_clip': 1.4754815101623535, 'timing/data': 0.00033585075289011, 'timing/model': 0.4316984764300287, 'learning_rate/qwen_vl_interface': 9.52817810069772e-06, 'learning_rate/action_model': 9.52817810069772e-05, 'epoch': 0.67})
+[2026-07-28 16:33:23,589][starVLA.training.train_starvla][INFO] - Step 3659, Loss: {'train/loss': 0.07715418934822083, 'train/grad_norm_pre_clip': 2.7595887184143066, 'timing/data': 0.0008325902745127678, 'timing/model': 0.4244722039438784, 'learning_rate/qwen_vl_interface': 9.52651416416042e-06, 'learning_rate/action_model': 9.526514164160421e-05, 'epoch': 0.67})
+[2026-07-28 16:33:29,700][starVLA.training.train_starvla][INFO] - Step 3660, Loss: {'train/loss': 0.017349371686577797, 'train/grad_norm_pre_clip': 0.6382425427436829, 'timing/data': 0.0003081290051341057, 'timing/model': 0.42888057325035334, 'learning_rate/qwen_vl_interface': 9.524847452639509e-06, 'learning_rate/action_model': 9.524847452639509e-05, 'epoch': 0.67})
+[2026-07-28 16:33:35,881][starVLA.training.train_starvla][INFO] - Step 3661, Loss: {'train/loss': 0.04361873120069504, 'train/grad_norm_pre_clip': 1.9733728170394897, 'timing/data': 0.00036232778802514076, 'timing/model': 0.4296050239354372, 'learning_rate/qwen_vl_interface': 9.523177967216496e-06, 'learning_rate/action_model': 9.523177967216496e-05, 'epoch': 0.67})
+[2026-07-28 16:33:42,406][starVLA.training.train_starvla][INFO] - Step 3662, Loss: {'train/loss': 0.06467103213071823, 'train/grad_norm_pre_clip': 1.5228815078735352, 'timing/data': 0.0003334488719701767, 'timing/model': 0.4429031237959862, 'learning_rate/qwen_vl_interface': 9.521505708974694e-06, 'learning_rate/action_model': 9.521505708974694e-05, 'epoch': 0.67})
+[2026-07-28 16:33:48,715][starVLA.training.train_starvla][INFO] - Step 3663, Loss: {'train/loss': 0.024418596178293228, 'train/grad_norm_pre_clip': 1.2834683656692505, 'timing/data': 0.0003381301648914814, 'timing/model': 0.4249885780736804, 'learning_rate/qwen_vl_interface': 9.519830678999214e-06, 'learning_rate/action_model': 9.519830678999213e-05, 'epoch': 0.67})
+[2026-07-28 16:33:54,751][starVLA.training.train_starvla][INFO] - Step 3664, Loss: {'train/loss': 0.034227609634399414, 'train/grad_norm_pre_clip': 1.2495653629302979, 'timing/data': 0.00030290568247437477, 'timing/model': 0.4446107493713498, 'learning_rate/qwen_vl_interface': 9.518152878376961e-06, 'learning_rate/action_model': 9.51815287837696e-05, 'epoch': 0.67})
+[2026-07-28 16:34:01,054][starVLA.training.train_starvla][INFO] - Step 3665, Loss: {'train/loss': 0.055354200303554535, 'train/grad_norm_pre_clip': 4.525459289550781, 'timing/data': 0.0003784610889852047, 'timing/model': 0.6672395137138665, 'learning_rate/qwen_vl_interface': 9.516472308196646e-06, 'learning_rate/action_model': 9.516472308196645e-05, 'epoch': 0.67})
+[2026-07-28 16:34:07,426][starVLA.training.train_starvla][INFO] - Step 3666, Loss: {'train/loss': 0.059513211250305176, 'train/grad_norm_pre_clip': 3.4067423343658447, 'timing/data': 0.00031810905784368515, 'timing/model': 0.6590709611773491, 'learning_rate/qwen_vl_interface': 9.51478896954877e-06, 'learning_rate/action_model': 9.51478896954877e-05, 'epoch': 0.67})
+[2026-07-28 16:34:13,502][starVLA.training.train_starvla][INFO] - Step 3667, Loss: {'train/loss': 0.017819637432694435, 'train/grad_norm_pre_clip': 0.7350365519523621, 'timing/data': 0.0003157341852784157, 'timing/model': 0.42021856969222426, 'learning_rate/qwen_vl_interface': 9.51310286352563e-06, 'learning_rate/action_model': 9.51310286352563e-05, 'epoch': 0.67})
+[2026-07-28 16:34:19,521][starVLA.training.train_starvla][INFO] - Step 3668, Loss: {'train/loss': 0.042739275842905045, 'train/grad_norm_pre_clip': 1.006776213645935, 'timing/data': 0.00028662336990237236, 'timing/model': 0.4237957396544516, 'learning_rate/qwen_vl_interface': 9.51141399122133e-06, 'learning_rate/action_model': 9.511413991221331e-05, 'epoch': 0.67})
+[2026-07-28 16:34:25,598][starVLA.training.train_starvla][INFO] - Step 3669, Loss: {'train/loss': 0.04404683783650398, 'train/grad_norm_pre_clip': 0.9833567142486572, 'timing/data': 0.00029197242110967636, 'timing/model': 0.4201643969863653, 'learning_rate/qwen_vl_interface': 9.509722353731757e-06, 'learning_rate/action_model': 9.509722353731756e-05, 'epoch': 0.67})
+[2026-07-28 16:34:31,598][starVLA.training.train_starvla][INFO] - Step 3670, Loss: {'train/loss': 0.010737650096416473, 'train/grad_norm_pre_clip': 0.47710707783699036, 'timing/data': 0.0003131916746497154, 'timing/model': 0.6742244479246438, 'learning_rate/qwen_vl_interface': 9.508027952154595e-06, 'learning_rate/action_model': 9.508027952154593e-05, 'epoch': 0.67})
+[2026-07-28 16:34:37,405][starVLA.training.train_starvla][INFO] - Step 3671, Loss: {'train/loss': 0.017796071246266365, 'train/grad_norm_pre_clip': 1.0121960639953613, 'timing/data': 0.00034859031438827515, 'timing/model': 0.4208322362974286, 'learning_rate/qwen_vl_interface': 9.506330787589321e-06, 'learning_rate/action_model': 9.506330787589322e-05, 'epoch': 0.67})
+[2026-07-28 16:34:43,152][starVLA.training.train_starvla][INFO] - Step 3672, Loss: {'train/loss': 0.053592026233673096, 'train/grad_norm_pre_clip': 1.463625192642212, 'timing/data': 0.0003211372531950474, 'timing/model': 0.4211704642511904, 'learning_rate/qwen_vl_interface': 9.504630861137212e-06, 'learning_rate/action_model': 9.504630861137212e-05, 'epoch': 0.67})
+[2026-07-28 16:34:49,177][starVLA.training.train_starvla][INFO] - Step 3673, Loss: {'train/loss': 0.02068229392170906, 'train/grad_norm_pre_clip': 0.6981096863746643, 'timing/data': 0.0003928598016500473, 'timing/model': 0.42506704293191433, 'learning_rate/qwen_vl_interface': 9.50292817390133e-06, 'learning_rate/action_model': 9.502928173901329e-05, 'epoch': 0.67})
+[2026-07-28 16:34:55,394][starVLA.training.train_starvla][INFO] - Step 3674, Loss: {'train/loss': 0.023989055305719376, 'train/grad_norm_pre_clip': 0.780337929725647, 'timing/data': 0.0003712298348546028, 'timing/model': 0.45395555393770337, 'learning_rate/qwen_vl_interface': 9.501222726986525e-06, 'learning_rate/action_model': 9.501222726986524e-05, 'epoch': 0.67})
+[2026-07-28 16:35:01,476][starVLA.training.train_starvla][INFO] - Step 3675, Loss: {'train/loss': 0.01775258593261242, 'train/grad_norm_pre_clip': 0.6654789447784424, 'timing/data': 0.0003613326698541641, 'timing/model': 0.4231257326900959, 'learning_rate/qwen_vl_interface': 9.49951452149945e-06, 'learning_rate/action_model': 9.499514521499451e-05, 'epoch': 0.68})
+[2026-07-28 16:35:08,060][starVLA.training.train_starvla][INFO] - Step 3676, Loss: {'train/loss': 0.051209092140197754, 'train/grad_norm_pre_clip': 1.001194953918457, 'timing/data': 0.00027635088190436363, 'timing/model': 0.6674242038279772, 'learning_rate/qwen_vl_interface': 9.497803558548539e-06, 'learning_rate/action_model': 9.497803558548539e-05, 'epoch': 0.68})
+[2026-07-28 16:35:14,213][starVLA.training.train_starvla][INFO] - Step 3677, Loss: {'train/loss': 0.05626814067363739, 'train/grad_norm_pre_clip': 1.1523797512054443, 'timing/data': 0.0003474736586213112, 'timing/model': 0.422176671680063, 'learning_rate/qwen_vl_interface': 9.496089839244015e-06, 'learning_rate/action_model': 9.496089839244015e-05, 'epoch': 0.68})
+[2026-07-28 16:35:20,306][starVLA.training.train_starvla][INFO] - Step 3678, Loss: {'train/loss': 0.05997418612241745, 'train/grad_norm_pre_clip': 1.2913134098052979, 'timing/data': 0.00026794103905558586, 'timing/model': 0.4205356687307358, 'learning_rate/qwen_vl_interface': 9.494373364697895e-06, 'learning_rate/action_model': 9.494373364697894e-05, 'epoch': 0.68})
+[2026-07-28 16:35:26,595][starVLA.training.train_starvla][INFO] - Step 3679, Loss: {'train/loss': 0.01999884657561779, 'train/grad_norm_pre_clip': 0.7040262818336487, 'timing/data': 0.0008643288165330887, 'timing/model': 0.4271414107643068, 'learning_rate/qwen_vl_interface': 9.492654136023978e-06, 'learning_rate/action_model': 9.492654136023979e-05, 'epoch': 0.68})
+[2026-07-28 16:35:32,956][starVLA.training.train_starvla][INFO] - Step 3680, Loss: {'train/loss': 0.03828335925936699, 'train/grad_norm_pre_clip': 1.9552446603775024, 'timing/data': 0.00032980088144540787, 'timing/model': 0.42248686123639345, 'learning_rate/qwen_vl_interface': 9.490932154337855e-06, 'learning_rate/action_model': 9.490932154337855e-05, 'epoch': 0.68})
+[2026-07-28 16:35:39,463][starVLA.training.train_starvla][INFO] - Step 3681, Loss: {'train/loss': 0.022226650267839432, 'train/grad_norm_pre_clip': 0.612461268901825, 'timing/data': 0.00028680311515927315, 'timing/model': 0.4193375501781702, 'learning_rate/qwen_vl_interface': 9.4892074207569e-06, 'learning_rate/action_model': 9.489207420756899e-05, 'epoch': 0.68})
+[2026-07-28 16:35:45,485][starVLA.training.train_starvla][INFO] - Step 3682, Loss: {'train/loss': 0.027777455747127533, 'train/grad_norm_pre_clip': 1.1192505359649658, 'timing/data': 0.0002861623652279377, 'timing/model': 0.42531179217621684, 'learning_rate/qwen_vl_interface': 9.487479936400274e-06, 'learning_rate/action_model': 9.487479936400272e-05, 'epoch': 0.68})
+[2026-07-28 16:35:51,849][starVLA.training.train_starvla][INFO] - Step 3683, Loss: {'train/loss': 0.025974448770284653, 'train/grad_norm_pre_clip': 0.538274884223938, 'timing/data': 0.00026131095364689827, 'timing/model': 0.41918201791122556, 'learning_rate/qwen_vl_interface': 9.485749702388923e-06, 'learning_rate/action_model': 9.485749702388923e-05, 'epoch': 0.68})
+[2026-07-28 16:35:57,993][starVLA.training.train_starvla][INFO] - Step 3684, Loss: {'train/loss': 0.02698291279375553, 'train/grad_norm_pre_clip': 1.0942707061767578, 'timing/data': 0.00029482273384928703, 'timing/model': 0.42113648913800716, 'learning_rate/qwen_vl_interface': 9.484016719845576e-06, 'learning_rate/action_model': 9.484016719845576e-05, 'epoch': 0.68})
+[2026-07-28 16:36:04,308][starVLA.training.train_starvla][INFO] - Step 3685, Loss: {'train/loss': 0.047440774738788605, 'train/grad_norm_pre_clip': 0.998695969581604, 'timing/data': 0.0002782619558274746, 'timing/model': 0.41940460354089737, 'learning_rate/qwen_vl_interface': 9.482280989894748e-06, 'learning_rate/action_model': 9.482280989894748e-05, 'epoch': 0.68})
+[2026-07-28 16:36:10,305][starVLA.training.train_starvla][INFO] - Step 3686, Loss: {'train/loss': 0.02347276732325554, 'train/grad_norm_pre_clip': 0.6310237646102905, 'timing/data': 0.0003074207343161106, 'timing/model': 0.4219574579037726, 'learning_rate/qwen_vl_interface': 9.480542513662734e-06, 'learning_rate/action_model': 9.480542513662734e-05, 'epoch': 0.68})
+[2026-07-28 16:36:16,499][starVLA.training.train_starvla][INFO] - Step 3687, Loss: {'train/loss': 0.01861489564180374, 'train/grad_norm_pre_clip': 0.8634653687477112, 'timing/data': 0.0002940930426120758, 'timing/model': 0.4211268029175699, 'learning_rate/qwen_vl_interface': 9.478801292277617e-06, 'learning_rate/action_model': 9.478801292277616e-05, 'epoch': 0.68})
+[2026-07-28 16:36:22,510][starVLA.training.train_starvla][INFO] - Step 3688, Loss: {'train/loss': 0.04246421903371811, 'train/grad_norm_pre_clip': 0.8644474744796753, 'timing/data': 0.0003148876130580902, 'timing/model': 0.42405062168836594, 'learning_rate/qwen_vl_interface': 9.477057326869249e-06, 'learning_rate/action_model': 9.477057326869249e-05, 'epoch': 0.68})
+[2026-07-28 16:36:28,602][starVLA.training.train_starvla][INFO] - Step 3689, Loss: {'train/loss': 0.04414791613817215, 'train/grad_norm_pre_clip': 1.2917754650115967, 'timing/data': 0.000287884846329689, 'timing/model': 0.43594080209732056, 'learning_rate/qwen_vl_interface': 9.475310618569274e-06, 'learning_rate/action_model': 9.475310618569274e-05, 'epoch': 0.68})
+[2026-07-28 16:36:34,880][starVLA.training.train_starvla][INFO] - Step 3690, Loss: {'train/loss': 0.03825998306274414, 'train/grad_norm_pre_clip': 1.0355541706085205, 'timing/data': 0.0003964751958847046, 'timing/model': 0.6590251959860325, 'learning_rate/qwen_vl_interface': 9.473561168511115e-06, 'learning_rate/action_model': 9.473561168511115e-05, 'epoch': 0.68})
+[2026-07-28 16:36:41,149][starVLA.training.train_starvla][INFO] - Step 3691, Loss: {'train/loss': 0.04842822253704071, 'train/grad_norm_pre_clip': 1.2474874258041382, 'timing/data': 0.0002750968560576439, 'timing/model': 0.42102203192189336, 'learning_rate/qwen_vl_interface': 9.471808977829967e-06, 'learning_rate/action_model': 9.471808977829967e-05, 'epoch': 0.68})
+[2026-07-28 16:36:47,266][starVLA.training.train_starvla][INFO] - Step 3692, Loss: {'train/loss': 0.04936247318983078, 'train/grad_norm_pre_clip': 1.0681084394454956, 'timing/data': 0.0003107958473265171, 'timing/model': 0.42100644018501043, 'learning_rate/qwen_vl_interface': 9.470054047662808e-06, 'learning_rate/action_model': 9.470054047662808e-05, 'epoch': 0.68})
+[2026-07-28 16:36:53,286][starVLA.training.train_starvla][INFO] - Step 3693, Loss: {'train/loss': 0.059751883149147034, 'train/grad_norm_pre_clip': 1.074091911315918, 'timing/data': 0.00033030612394213676, 'timing/model': 0.44086073199287057, 'learning_rate/qwen_vl_interface': 9.468296379148394e-06, 'learning_rate/action_model': 9.468296379148393e-05, 'epoch': 0.68})
+[2026-07-28 16:36:59,101][starVLA.training.train_starvla][INFO] - Step 3694, Loss: {'train/loss': 0.04748830199241638, 'train/grad_norm_pre_clip': 1.2097336053848267, 'timing/data': 0.0003154478035867214, 'timing/model': 0.4200510582886636, 'learning_rate/qwen_vl_interface': 9.466535973427259e-06, 'learning_rate/action_model': 9.466535973427258e-05, 'epoch': 0.68})
+[2026-07-28 16:37:04,895][starVLA.training.train_starvla][INFO] - Step 3695, Loss: {'train/loss': 0.03571508452296257, 'train/grad_norm_pre_clip': 1.4407955408096313, 'timing/data': 0.0002675233408808708, 'timing/model': 0.42024926422163844, 'learning_rate/qwen_vl_interface': 9.464772831641706e-06, 'learning_rate/action_model': 9.464772831641707e-05, 'epoch': 0.68})
+[2026-07-28 16:37:10,950][starVLA.training.train_starvla][INFO] - Step 3696, Loss: {'train/loss': 0.06953712552785873, 'train/grad_norm_pre_clip': 3.308563232421875, 'timing/data': 0.00039243418723344803, 'timing/model': 0.6779886381700635, 'learning_rate/qwen_vl_interface': 9.463006954935823e-06, 'learning_rate/action_model': 9.463006954935822e-05, 'epoch': 0.68})
+[2026-07-28 16:37:17,046][starVLA.training.train_starvla][INFO] - Step 3697, Loss: {'train/loss': 0.0465267151594162, 'train/grad_norm_pre_clip': 1.7971751689910889, 'timing/data': 0.0002679810859262943, 'timing/model': 0.4364434336312115, 'learning_rate/qwen_vl_interface': 9.461238344455468e-06, 'learning_rate/action_model': 9.461238344455468e-05, 'epoch': 0.68})
+[2026-07-28 16:37:22,866][starVLA.training.train_starvla][INFO] - Step 3698, Loss: {'train/loss': 0.0652635395526886, 'train/grad_norm_pre_clip': 2.978166341781616, 'timing/data': 0.0003850129432976246, 'timing/model': 0.4297550772316754, 'learning_rate/qwen_vl_interface': 9.459467001348271e-06, 'learning_rate/action_model': 9.45946700134827e-05, 'epoch': 0.68})
+[2026-07-28 16:37:28,928][starVLA.training.train_starvla][INFO] - Step 3699, Loss: {'train/loss': 0.047713689506053925, 'train/grad_norm_pre_clip': 1.2258007526397705, 'timing/data': 0.00028621964156627655, 'timing/model': 0.42326384922489524, 'learning_rate/qwen_vl_interface': 9.457692926763638e-06, 'learning_rate/action_model': 9.457692926763638e-05, 'epoch': 0.68})
+[2026-07-28 16:37:35,141][starVLA.training.train_starvla][INFO] - Step 3700, Loss: {'train/loss': 0.027830054983496666, 'train/grad_norm_pre_clip': 0.8755773305892944, 'timing/data': 0.00040740473195910454, 'timing/model': 0.4382463260553777, 'learning_rate/qwen_vl_interface': 9.455916121852748e-06, 'learning_rate/action_model': 9.455916121852747e-05, 'epoch': 0.68})
+[2026-07-28 16:37:41,766][starVLA.training.train_starvla][INFO] - Step 3701, Loss: {'train/loss': 0.06420595943927765, 'train/grad_norm_pre_clip': 1.9464532136917114, 'timing/data': 0.00026665208861231804, 'timing/model': 0.6765852249227464, 'learning_rate/qwen_vl_interface': 9.454136587768549e-06, 'learning_rate/action_model': 9.454136587768548e-05, 'epoch': 0.68})
+[2026-07-28 16:37:47,833][starVLA.training.train_starvla][INFO] - Step 3702, Loss: {'train/loss': 0.037802234292030334, 'train/grad_norm_pre_clip': 1.6271377801895142, 'timing/data': 0.0002961438149213791, 'timing/model': 0.4403862259350717, 'learning_rate/qwen_vl_interface': 9.452354325665763e-06, 'learning_rate/action_model': 9.452354325665763e-05, 'epoch': 0.68})
+[2026-07-28 16:37:53,866][starVLA.training.train_starvla][INFO] - Step 3703, Loss: {'train/loss': 0.02695649489760399, 'train/grad_norm_pre_clip': 0.8632903695106506, 'timing/data': 0.00028616469353437424, 'timing/model': 0.43772578006610274, 'learning_rate/qwen_vl_interface': 9.450569336700878e-06, 'learning_rate/action_model': 9.450569336700878e-05, 'epoch': 0.68})
+[2026-07-28 16:38:00,151][starVLA.training.train_starvla][INFO] - Step 3704, Loss: {'train/loss': 0.0730937048792839, 'train/grad_norm_pre_clip': 1.679161787033081, 'timing/data': 0.0002632550895214081, 'timing/model': 0.42316626384854317, 'learning_rate/qwen_vl_interface': 9.448781622032155e-06, 'learning_rate/action_model': 9.448781622032155e-05, 'epoch': 0.68})
+[2026-07-28 16:38:06,463][starVLA.training.train_starvla][INFO] - Step 3705, Loss: {'train/loss': 0.037659842520952225, 'train/grad_norm_pre_clip': 0.7414984107017517, 'timing/data': 0.00027811480686068535, 'timing/model': 0.4238995499908924, 'learning_rate/qwen_vl_interface': 9.446991182819624e-06, 'learning_rate/action_model': 9.446991182819623e-05, 'epoch': 0.68})
+[2026-07-28 16:38:12,538][starVLA.training.train_starvla][INFO] - Step 3706, Loss: {'train/loss': 0.03376791626214981, 'train/grad_norm_pre_clip': 1.196045994758606, 'timing/data': 0.0008663530461490154, 'timing/model': 0.42234249599277973, 'learning_rate/qwen_vl_interface': 9.44519802022508e-06, 'learning_rate/action_model': 9.44519802022508e-05, 'epoch': 0.68})
+[2026-07-28 16:38:18,598][starVLA.training.train_starvla][INFO] - Step 3707, Loss: {'train/loss': 0.020106162875890732, 'train/grad_norm_pre_clip': 0.6129071712493896, 'timing/data': 0.0003051050007343292, 'timing/model': 0.42521468503400683, 'learning_rate/qwen_vl_interface': 9.443402135412088e-06, 'learning_rate/action_model': 9.443402135412088e-05, 'epoch': 0.68})
+[2026-07-28 16:38:24,952][starVLA.training.train_starvla][INFO] - Step 3708, Loss: {'train/loss': 0.03898703306913376, 'train/grad_norm_pre_clip': 1.7224534749984741, 'timing/data': 0.00028357328847050667, 'timing/model': 0.6800363599322736, 'learning_rate/qwen_vl_interface': 9.441603529545977e-06, 'learning_rate/action_model': 9.441603529545976e-05, 'epoch': 0.68})
+[2026-07-28 16:38:30,995][starVLA.training.train_starvla][INFO] - Step 3709, Loss: {'train/loss': 0.057886406779289246, 'train/grad_norm_pre_clip': 1.7972203493118286, 'timing/data': 0.00026223016902804375, 'timing/model': 0.42131385300308466, 'learning_rate/qwen_vl_interface': 9.439802203793843e-06, 'learning_rate/action_model': 9.439802203793842e-05, 'epoch': 0.68})
+[2026-07-28 16:38:37,039][starVLA.training.train_starvla][INFO] - Step 3710, Loss: {'train/loss': 0.026100439950823784, 'train/grad_norm_pre_clip': 0.6773970127105713, 'timing/data': 0.0003018002025783062, 'timing/model': 0.4367892178706825, 'learning_rate/qwen_vl_interface': 9.437998159324547e-06, 'learning_rate/action_model': 9.437998159324547e-05, 'epoch': 0.68})
+[2026-07-28 16:38:43,064][starVLA.training.train_starvla][INFO] - Step 3711, Loss: {'train/loss': 0.07651036977767944, 'train/grad_norm_pre_clip': 1.2614496946334839, 'timing/data': 0.00028696097433567047, 'timing/model': 0.44802428781986237, 'learning_rate/qwen_vl_interface': 9.436191397308713e-06, 'learning_rate/action_model': 9.436191397308713e-05, 'epoch': 0.68})
+[2026-07-28 16:38:49,303][starVLA.training.train_starvla][INFO] - Step 3712, Loss: {'train/loss': 0.04160730540752411, 'train/grad_norm_pre_clip': 1.1122212409973145, 'timing/data': 0.0002676490694284439, 'timing/model': 0.6615948700346053, 'learning_rate/qwen_vl_interface': 9.434381918918731e-06, 'learning_rate/action_model': 9.434381918918731e-05, 'epoch': 0.68})
+[2026-07-28 16:38:55,293][starVLA.training.train_starvla][INFO] - Step 3713, Loss: {'train/loss': 0.0416233167052269, 'train/grad_norm_pre_clip': 1.0387746095657349, 'timing/data': 0.0003741821274161339, 'timing/model': 0.42654173308983445, 'learning_rate/qwen_vl_interface': 9.432569725328753e-06, 'learning_rate/action_model': 9.432569725328753e-05, 'epoch': 0.68})
+[2026-07-28 16:39:01,647][starVLA.training.train_starvla][INFO] - Step 3714, Loss: {'train/loss': 0.02066943421959877, 'train/grad_norm_pre_clip': 0.8320801258087158, 'timing/data': 0.00026399921625852585, 'timing/model': 0.4321564696729183, 'learning_rate/qwen_vl_interface': 9.430754817714686e-06, 'learning_rate/action_model': 9.430754817714687e-05, 'epoch': 0.68})
+[2026-07-28 16:39:07,756][starVLA.training.train_starvla][INFO] - Step 3715, Loss: {'train/loss': 0.06823770701885223, 'train/grad_norm_pre_clip': 2.1518537998199463, 'timing/data': 0.0003742547705769539, 'timing/model': 0.4971769731491804, 'learning_rate/qwen_vl_interface': 9.42893719725421e-06, 'learning_rate/action_model': 9.42893719725421e-05, 'epoch': 0.68})
+[2026-07-28 16:39:13,801][starVLA.training.train_starvla][INFO] - Step 3716, Loss: {'train/loss': 0.08034954220056534, 'train/grad_norm_pre_clip': 2.3280417919158936, 'timing/data': 0.00030581606552004814, 'timing/model': 0.42074755812063813, 'learning_rate/qwen_vl_interface': 9.427116865126757e-06, 'learning_rate/action_model': 9.427116865126756e-05, 'epoch': 0.68})
+[2026-07-28 16:39:19,808][starVLA.training.train_starvla][INFO] - Step 3717, Loss: {'train/loss': 0.02333468571305275, 'train/grad_norm_pre_clip': 1.03266441822052, 'timing/data': 0.0003664949908852577, 'timing/model': 0.41998308431357145, 'learning_rate/qwen_vl_interface': 9.42529382251352e-06, 'learning_rate/action_model': 9.42529382251352e-05, 'epoch': 0.68})
+[2026-07-28 16:39:26,043][starVLA.training.train_starvla][INFO] - Step 3718, Loss: {'train/loss': 0.038821086287498474, 'train/grad_norm_pre_clip': 0.964514434337616, 'timing/data': 0.0002627340145409107, 'timing/model': 0.4213545061647892, 'learning_rate/qwen_vl_interface': 9.423468070597453e-06, 'learning_rate/action_model': 9.423468070597452e-05, 'epoch': 0.68})
+[2026-07-28 16:39:32,114][starVLA.training.train_starvla][INFO] - Step 3719, Loss: {'train/loss': 0.03218851983547211, 'train/grad_norm_pre_clip': 1.816027283668518, 'timing/data': 0.0003770110197365284, 'timing/model': 0.4214679100550711, 'learning_rate/qwen_vl_interface': 9.421639610563264e-06, 'learning_rate/action_model': 9.421639610563264e-05, 'epoch': 0.68})
+[2026-07-28 16:39:38,346][starVLA.training.train_starvla][INFO] - Step 3720, Loss: {'train/loss': 0.05548573657870293, 'train/grad_norm_pre_clip': 2.2811050415039062, 'timing/data': 0.0002911994233727455, 'timing/model': 0.4870053078047931, 'learning_rate/qwen_vl_interface': 9.419808443597422e-06, 'learning_rate/action_model': 9.419808443597423e-05, 'epoch': 0.68})
+[2026-07-28 16:39:44,581][starVLA.training.train_starvla][INFO] - Step 3721, Loss: {'train/loss': 0.06028937175869942, 'train/grad_norm_pre_clip': 2.3612871170043945, 'timing/data': 0.00027005188167095184, 'timing/model': 0.42151490272954106, 'learning_rate/qwen_vl_interface': 9.417974570888152e-06, 'learning_rate/action_model': 9.417974570888153e-05, 'epoch': 0.68})
+[2026-07-28 16:39:50,599][starVLA.training.train_starvla][INFO] - Step 3722, Loss: {'train/loss': 0.04744185134768486, 'train/grad_norm_pre_clip': 1.7595347166061401, 'timing/data': 0.00032114703208208084, 'timing/model': 0.4210986909456551, 'learning_rate/qwen_vl_interface': 9.416137993625437e-06, 'learning_rate/action_model': 9.416137993625437e-05, 'epoch': 0.68})
+[2026-07-28 16:39:56,718][starVLA.training.train_starvla][INFO] - Step 3723, Loss: {'train/loss': 0.04067414999008179, 'train/grad_norm_pre_clip': 1.7725213766098022, 'timing/data': 0.0004808199591934681, 'timing/model': 0.42320869537070394, 'learning_rate/qwen_vl_interface': 9.414298713001007e-06, 'learning_rate/action_model': 9.414298713001006e-05, 'epoch': 0.68})
+[2026-07-28 16:40:03,199][starVLA.training.train_starvla][INFO] - Step 3724, Loss: {'train/loss': 0.023257378488779068, 'train/grad_norm_pre_clip': 0.9036561846733093, 'timing/data': 0.0002576801925897598, 'timing/model': 0.5951011078432202, 'learning_rate/qwen_vl_interface': 9.412456730208352e-06, 'learning_rate/action_model': 9.412456730208352e-05, 'epoch': 0.68})
+[2026-07-28 16:40:09,566][starVLA.training.train_starvla][INFO] - Step 3725, Loss: {'train/loss': 0.046042513102293015, 'train/grad_norm_pre_clip': 1.7661583423614502, 'timing/data': 0.002048634923994541, 'timing/model': 0.5473294872790575, 'learning_rate/qwen_vl_interface': 9.410612046442719e-06, 'learning_rate/action_model': 9.410612046442718e-05, 'epoch': 0.68})
+[2026-07-28 16:40:16,028][starVLA.training.train_starvla][INFO] - Step 3726, Loss: {'train/loss': 0.04157605022192001, 'train/grad_norm_pre_clip': 1.0488206148147583, 'timing/data': 0.0002936068922281265, 'timing/model': 0.42357328487560153, 'learning_rate/qwen_vl_interface': 9.408764662901097e-06, 'learning_rate/action_model': 9.408764662901098e-05, 'epoch': 0.68})
+[2026-07-28 16:40:22,072][starVLA.training.train_starvla][INFO] - Step 3727, Loss: {'train/loss': 0.02593517303466797, 'train/grad_norm_pre_clip': 0.5449258089065552, 'timing/data': 0.00041410792618989944, 'timing/model': 0.42445860523730516, 'learning_rate/qwen_vl_interface': 9.406914580782238e-06, 'learning_rate/action_model': 9.406914580782238e-05, 'epoch': 0.68})
+[2026-07-28 16:40:28,386][starVLA.training.train_starvla][INFO] - Step 3728, Loss: {'train/loss': 0.054141778498888016, 'train/grad_norm_pre_clip': 1.8908238410949707, 'timing/data': 0.0003519570454955101, 'timing/model': 0.422529854811728, 'learning_rate/qwen_vl_interface': 9.405061801286638e-06, 'learning_rate/action_model': 9.405061801286638e-05, 'epoch': 0.68})
+[2026-07-28 16:40:34,675][starVLA.training.train_starvla][INFO] - Step 3729, Loss: {'train/loss': 0.03706015273928642, 'train/grad_norm_pre_clip': 0.7434017658233643, 'timing/data': 0.00035622669383883476, 'timing/model': 0.4236391973681748, 'learning_rate/qwen_vl_interface': 9.403206325616546e-06, 'learning_rate/action_model': 9.403206325616545e-05, 'epoch': 0.68})
+[2026-07-28 16:40:40,737][starVLA.training.train_starvla][INFO] - Step 3730, Loss: {'train/loss': 0.014247494749724865, 'train/grad_norm_pre_clip': 0.5199837684631348, 'timing/data': 0.00035357382148504257, 'timing/model': 0.4336326429620385, 'learning_rate/qwen_vl_interface': 9.40134815497596e-06, 'learning_rate/action_model': 9.40134815497596e-05, 'epoch': 0.69})
+[2026-07-28 16:40:47,066][starVLA.training.train_starvla][INFO] - Step 3731, Loss: {'train/loss': 0.0330640971660614, 'train/grad_norm_pre_clip': 1.2168930768966675, 'timing/data': 0.00031446199864149094, 'timing/model': 0.6584770050831139, 'learning_rate/qwen_vl_interface': 9.399487290570629e-06, 'learning_rate/action_model': 9.399487290570629e-05, 'epoch': 0.69})
+[2026-07-28 16:40:53,143][starVLA.training.train_starvla][INFO] - Step 3732, Loss: {'train/loss': 0.05470827966928482, 'train/grad_norm_pre_clip': 2.48880672454834, 'timing/data': 0.0003365837037563324, 'timing/model': 0.42432333808392286, 'learning_rate/qwen_vl_interface': 9.397623733608042e-06, 'learning_rate/action_model': 9.397623733608042e-05, 'epoch': 0.69})
+[2026-07-28 16:40:59,569][starVLA.training.train_starvla][INFO] - Step 3733, Loss: {'train/loss': 0.08427461981773376, 'train/grad_norm_pre_clip': 2.0098001956939697, 'timing/data': 0.00044410908594727516, 'timing/model': 0.42395371524617076, 'learning_rate/qwen_vl_interface': 9.395757485297447e-06, 'learning_rate/action_model': 9.395757485297446e-05, 'epoch': 0.69})
+[2026-07-28 16:41:05,556][starVLA.training.train_starvla][INFO] - Step 3734, Loss: {'train/loss': 0.06416129320859909, 'train/grad_norm_pre_clip': 1.6427441835403442, 'timing/data': 0.00031651416793465614, 'timing/model': 0.4206500928848982, 'learning_rate/qwen_vl_interface': 9.393888546849828e-06, 'learning_rate/action_model': 9.393888546849828e-05, 'epoch': 0.69})
+[2026-07-28 16:41:11,827][starVLA.training.train_starvla][INFO] - Step 3735, Loss: {'train/loss': 0.05917380005121231, 'train/grad_norm_pre_clip': 1.8697842359542847, 'timing/data': 0.0004982063546776772, 'timing/model': 0.4510873397812247, 'learning_rate/qwen_vl_interface': 9.392016919477922e-06, 'learning_rate/action_model': 9.392016919477921e-05, 'epoch': 0.69})
+[2026-07-28 16:41:18,075][starVLA.training.train_starvla][INFO] - Step 3736, Loss: {'train/loss': 0.011581158265471458, 'train/grad_norm_pre_clip': 0.617728590965271, 'timing/data': 0.0002628718502819538, 'timing/model': 0.42584112705662847, 'learning_rate/qwen_vl_interface': 9.390142604396206e-06, 'learning_rate/action_model': 9.390142604396205e-05, 'epoch': 0.69})
+[2026-07-28 16:41:24,119][starVLA.training.train_starvla][INFO] - Step 3737, Loss: {'train/loss': 0.02966087870299816, 'train/grad_norm_pre_clip': 1.3166226148605347, 'timing/data': 0.0003847437910735607, 'timing/model': 0.42028952576220036, 'learning_rate/qwen_vl_interface': 9.388265602820901e-06, 'learning_rate/action_model': 9.3882656028209e-05, 'epoch': 0.69})
+[2026-07-28 16:41:30,353][starVLA.training.train_starvla][INFO] - Step 3738, Loss: {'train/loss': 0.04026508703827858, 'train/grad_norm_pre_clip': 1.0461472272872925, 'timing/data': 0.00027999794110655785, 'timing/model': 0.4201668002642691, 'learning_rate/qwen_vl_interface': 9.386385915969975e-06, 'learning_rate/action_model': 9.386385915969976e-05, 'epoch': 0.69})
+[2026-07-28 16:41:36,392][starVLA.training.train_starvla][INFO] - Step 3739, Loss: {'train/loss': 0.08416114747524261, 'train/grad_norm_pre_clip': 1.7864567041397095, 'timing/data': 0.00029677199199795723, 'timing/model': 0.4197873258963227, 'learning_rate/qwen_vl_interface': 9.384503545063138e-06, 'learning_rate/action_model': 9.384503545063137e-05, 'epoch': 0.69})
+[2026-07-28 16:41:42,700][starVLA.training.train_starvla][INFO] - Step 3740, Loss: {'train/loss': 0.017936334013938904, 'train/grad_norm_pre_clip': 1.21625816822052, 'timing/data': 0.0004123058170080185, 'timing/model': 0.7220023637637496, 'learning_rate/qwen_vl_interface': 9.382618491321837e-06, 'learning_rate/action_model': 9.382618491321837e-05, 'epoch': 0.69})
+[2026-07-28 16:41:48,949][starVLA.training.train_starvla][INFO] - Step 3741, Loss: {'train/loss': 0.017827481031417847, 'train/grad_norm_pre_clip': 0.6959522366523743, 'timing/data': 0.0002971263602375984, 'timing/model': 0.6687516230158508, 'learning_rate/qwen_vl_interface': 9.380730755969265e-06, 'learning_rate/action_model': 9.380730755969264e-05, 'epoch': 0.69})
+[2026-07-28 16:41:54,932][starVLA.training.train_starvla][INFO] - Step 3742, Loss: {'train/loss': 0.05621667951345444, 'train/grad_norm_pre_clip': 2.359286308288574, 'timing/data': 0.00030814995989203453, 'timing/model': 0.4214507667347789, 'learning_rate/qwen_vl_interface': 9.378840340230353e-06, 'learning_rate/action_model': 9.378840340230352e-05, 'epoch': 0.69})
+[2026-07-28 16:42:00,995][starVLA.training.train_starvla][INFO] - Step 3743, Loss: {'train/loss': 0.05157802253961563, 'train/grad_norm_pre_clip': 1.7847903966903687, 'timing/data': 0.00029664672911167145, 'timing/model': 0.42109693959355354, 'learning_rate/qwen_vl_interface': 9.376947245331769e-06, 'learning_rate/action_model': 9.376947245331767e-05, 'epoch': 0.69})
+[2026-07-28 16:42:07,263][starVLA.training.train_starvla][INFO] - Step 3744, Loss: {'train/loss': 0.05254318565130234, 'train/grad_norm_pre_clip': 2.115328788757324, 'timing/data': 0.0002629249356687069, 'timing/model': 0.41998021909967065, 'learning_rate/qwen_vl_interface': 9.375051472501922e-06, 'learning_rate/action_model': 9.375051472501923e-05, 'epoch': 0.69})
+[2026-07-28 16:42:13,463][starVLA.training.train_starvla][INFO] - Step 3745, Loss: {'train/loss': 0.07745490968227386, 'train/grad_norm_pre_clip': 7.127865791320801, 'timing/data': 0.0002585751935839653, 'timing/model': 0.4201668887399137, 'learning_rate/qwen_vl_interface': 9.373153022970961e-06, 'learning_rate/action_model': 9.373153022970961e-05, 'epoch': 0.69})
+[2026-07-28 16:42:19,590][starVLA.training.train_starvla][INFO] - Step 3746, Loss: {'train/loss': 0.06047233194112778, 'train/grad_norm_pre_clip': 1.5365453958511353, 'timing/data': 0.0004069199785590172, 'timing/model': 0.4232345097698271, 'learning_rate/qwen_vl_interface': 9.371251897970768e-06, 'learning_rate/action_model': 9.371251897970768e-05, 'epoch': 0.69})
+[2026-07-28 16:42:25,621][starVLA.training.train_starvla][INFO] - Step 3747, Loss: {'train/loss': 0.04064857214689255, 'train/grad_norm_pre_clip': 1.7756916284561157, 'timing/data': 0.00029971636831760406, 'timing/model': 0.42076505813747644, 'learning_rate/qwen_vl_interface': 9.369348098734961e-06, 'learning_rate/action_model': 9.369348098734961e-05, 'epoch': 0.69})
+[2026-07-28 16:42:32,350][starVLA.training.train_starvla][INFO] - Step 3748, Loss: {'train/loss': 0.010585838928818703, 'train/grad_norm_pre_clip': 0.4941149055957794, 'timing/data': 0.0003003990277647972, 'timing/model': 0.42006492195650935, 'learning_rate/qwen_vl_interface': 9.367441626498897e-06, 'learning_rate/action_model': 9.367441626498896e-05, 'epoch': 0.69})
+[2026-07-28 16:42:38,376][starVLA.training.train_starvla][INFO] - Step 3749, Loss: {'train/loss': 0.03533368557691574, 'train/grad_norm_pre_clip': 1.3931193351745605, 'timing/data': 0.0003178580664098263, 'timing/model': 0.43422794714570045, 'learning_rate/qwen_vl_interface': 9.365532482499663e-06, 'learning_rate/action_model': 9.365532482499662e-05, 'epoch': 0.69})
+[2026-07-28 16:42:44,440][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 3750
+[2026-07-28 16:42:49,584][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 3750
+[2026-07-28 16:49:09,918][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 3750
+[2026-07-28 16:49:09,919][starVLA.training.train_starvla][INFO] - Step 3750, Loss: {'train/loss': 0.04100137576460838, 'train/grad_norm_pre_clip': 0.9292654395103455, 'eval/action_loss/seconds': 5.139298448804766, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.04523986950516701, 'eval/latency_3/loss': 0.04523986950516701, 'eval/flappy/loss': 0.04523986950516701, 'eval/flappy/latency_3/loss': 0.04523986950516701, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 277.18000326827166, 'rl_games_eval/mid_train/mean_length': 2252.45, 'rl_games_eval/mid_train/std_reward': 168.96234570997657, 'rl_games_eval/mid_train/std_length': 1361.875525699761, 'rl_games_eval/mid_train/macro_mean_reward': 277.18000326827166, 'rl_games_eval/mid_train/macro_mean_length': 2252.45, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 277.18000326827166, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2252.45, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 168.96234570997657, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1361.875525699761, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 277.18000326827166, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.0004000863991677761, 'timing/model': 0.44179888581857085, 'learning_rate/qwen_vl_interface': 9.363620667976085e-06, 'learning_rate/action_model': 9.363620667976085e-05, 'epoch': 0.69})
+[2026-07-28 16:49:16,278][starVLA.training.train_starvla][INFO] - Step 3751, Loss: {'train/loss': 0.04733259230852127, 'train/grad_norm_pre_clip': 1.4317272901535034, 'timing/data': 0.00034919288009405136, 'timing/model': 0.419749241322279, 'learning_rate/qwen_vl_interface': 9.361706184168714e-06, 'learning_rate/action_model': 9.361706184168713e-05, 'epoch': 0.69})
+[2026-07-28 16:49:22,557][starVLA.training.train_starvla][INFO] - Step 3752, Loss: {'train/loss': 0.047590017318725586, 'train/grad_norm_pre_clip': 2.240264415740967, 'timing/data': 0.0004494301974773407, 'timing/model': 0.42181107914075255, 'learning_rate/qwen_vl_interface': 9.359789032319843e-06, 'learning_rate/action_model': 9.359789032319842e-05, 'epoch': 0.69})
+[2026-07-28 16:49:28,569][starVLA.training.train_starvla][INFO] - Step 3753, Loss: {'train/loss': 0.012770279310643673, 'train/grad_norm_pre_clip': 1.0377206802368164, 'timing/data': 0.00039727194234728813, 'timing/model': 0.42229281598702073, 'learning_rate/qwen_vl_interface': 9.357869213673484e-06, 'learning_rate/action_model': 9.357869213673485e-05, 'epoch': 0.69})
+[2026-07-28 16:49:34,329][starVLA.training.train_starvla][INFO] - Step 3754, Loss: {'train/loss': 0.04340265691280365, 'train/grad_norm_pre_clip': 0.7301391959190369, 'timing/data': 0.00031830277293920517, 'timing/model': 0.4222511472180486, 'learning_rate/qwen_vl_interface': 9.355946729475393e-06, 'learning_rate/action_model': 9.355946729475393e-05, 'epoch': 0.69})
+[2026-07-28 16:49:40,613][starVLA.training.train_starvla][INFO] - Step 3755, Loss: {'train/loss': 0.013164732605218887, 'train/grad_norm_pre_clip': 0.5471816658973694, 'timing/data': 0.0002595246769487858, 'timing/model': 0.42214125115424395, 'learning_rate/qwen_vl_interface': 9.354021580973049e-06, 'learning_rate/action_model': 9.354021580973048e-05, 'epoch': 0.69})
+[2026-07-28 16:49:46,904][starVLA.training.train_starvla][INFO] - Step 3756, Loss: {'train/loss': 0.01623065397143364, 'train/grad_norm_pre_clip': 0.7096923589706421, 'timing/data': 0.00031103333458304405, 'timing/model': 0.6639955043792725, 'learning_rate/qwen_vl_interface': 9.352093769415655e-06, 'learning_rate/action_model': 9.352093769415655e-05, 'epoch': 0.69})
+[2026-07-28 16:49:52,961][starVLA.training.train_starvla][INFO] - Step 3757, Loss: {'train/loss': 0.04032110795378685, 'train/grad_norm_pre_clip': 1.5695724487304688, 'timing/data': 0.00029321014881134033, 'timing/model': 0.4226340698078275, 'learning_rate/qwen_vl_interface': 9.35016329605415e-06, 'learning_rate/action_model': 9.350163296054151e-05, 'epoch': 0.69})
+[2026-07-28 16:49:59,165][starVLA.training.train_starvla][INFO] - Step 3758, Loss: {'train/loss': 0.005124771036207676, 'train/grad_norm_pre_clip': 0.2713462710380554, 'timing/data': 0.00031221192330121994, 'timing/model': 0.5957414787262678, 'learning_rate/qwen_vl_interface': 9.348230162141198e-06, 'learning_rate/action_model': 9.348230162141198e-05, 'epoch': 0.69})
+[2026-07-28 16:50:05,182][starVLA.training.train_starvla][INFO] - Step 3759, Loss: {'train/loss': 0.0150682982057333, 'train/grad_norm_pre_clip': 0.6290521025657654, 'timing/data': 0.0003805807791650295, 'timing/model': 0.4202187838964164, 'learning_rate/qwen_vl_interface': 9.346294368931188e-06, 'learning_rate/action_model': 9.346294368931188e-05, 'epoch': 0.69})
+[2026-07-28 16:50:11,216][starVLA.training.train_starvla][INFO] - Step 3760, Loss: {'train/loss': 0.04279468581080437, 'train/grad_norm_pre_clip': 1.1745718717575073, 'timing/data': 0.0003637303598225117, 'timing/model': 0.42091158498078585, 'learning_rate/qwen_vl_interface': 9.344355917680235e-06, 'learning_rate/action_model': 9.344355917680235e-05, 'epoch': 0.69})
+[2026-07-28 16:50:17,213][starVLA.training.train_starvla][INFO] - Step 3761, Loss: {'train/loss': 0.06013479828834534, 'train/grad_norm_pre_clip': 2.682260274887085, 'timing/data': 0.000348441768437624, 'timing/model': 0.42088690865784883, 'learning_rate/qwen_vl_interface': 9.342414809646179e-06, 'learning_rate/action_model': 9.342414809646179e-05, 'epoch': 0.69})
+[2026-07-28 16:50:23,434][starVLA.training.train_starvla][INFO] - Step 3762, Loss: {'train/loss': 0.0569009967148304, 'train/grad_norm_pre_clip': 2.11014461517334, 'timing/data': 0.0002933819778263569, 'timing/model': 0.42059379164129496, 'learning_rate/qwen_vl_interface': 9.340471046088581e-06, 'learning_rate/action_model': 9.34047104608858e-05, 'epoch': 0.69})
+[2026-07-28 16:50:29,666][starVLA.training.train_starvla][INFO] - Step 3763, Loss: {'train/loss': 0.050736311823129654, 'train/grad_norm_pre_clip': 1.2545983791351318, 'timing/data': 0.00025709299370646477, 'timing/model': 0.4203487290069461, 'learning_rate/qwen_vl_interface': 9.338524628268734e-06, 'learning_rate/action_model': 9.338524628268733e-05, 'epoch': 0.69})
+[2026-07-28 16:50:35,659][starVLA.training.train_starvla][INFO] - Step 3764, Loss: {'train/loss': 0.06951461732387543, 'train/grad_norm_pre_clip': 1.948830485343933, 'timing/data': 0.0002924259752035141, 'timing/model': 0.4275361211039126, 'learning_rate/qwen_vl_interface': 9.336575557449641e-06, 'learning_rate/action_model': 9.336575557449641e-05, 'epoch': 0.69})
+[2026-07-28 16:50:42,007][starVLA.training.train_starvla][INFO] - Step 3765, Loss: {'train/loss': 0.04722195863723755, 'train/grad_norm_pre_clip': 1.7496814727783203, 'timing/data': 0.0003031543456017971, 'timing/model': 0.7034810041077435, 'learning_rate/qwen_vl_interface': 9.334623834896039e-06, 'learning_rate/action_model': 9.334623834896038e-05, 'epoch': 0.69})
+[2026-07-28 16:50:48,034][starVLA.training.train_starvla][INFO] - Step 3766, Loss: {'train/loss': 0.07257629930973053, 'train/grad_norm_pre_clip': 2.4539706707000732, 'timing/data': 0.0003054719418287277, 'timing/model': 0.42397857923060656, 'learning_rate/qwen_vl_interface': 9.332669461874376e-06, 'learning_rate/action_model': 9.332669461874374e-05, 'epoch': 0.69})
+[2026-07-28 16:50:54,280][starVLA.training.train_starvla][INFO] - Step 3767, Loss: {'train/loss': 0.06975588947534561, 'train/grad_norm_pre_clip': 2.1225368976593018, 'timing/data': 0.00028293998911976814, 'timing/model': 0.6659796196036041, 'learning_rate/qwen_vl_interface': 9.330712439652824e-06, 'learning_rate/action_model': 9.330712439652823e-05, 'epoch': 0.69})
+[2026-07-28 16:51:00,295][starVLA.training.train_starvla][INFO] - Step 3768, Loss: {'train/loss': 0.06153194606304169, 'train/grad_norm_pre_clip': 1.440902590751648, 'timing/data': 0.0003179451450705528, 'timing/model': 0.42124939477071166, 'learning_rate/qwen_vl_interface': 9.328752769501272e-06, 'learning_rate/action_model': 9.328752769501272e-05, 'epoch': 0.69})
+[2026-07-28 16:51:06,377][starVLA.training.train_starvla][INFO] - Step 3769, Loss: {'train/loss': 0.06433910131454468, 'train/grad_norm_pre_clip': 1.4746030569076538, 'timing/data': 0.00033179111778736115, 'timing/model': 0.477010696195066, 'learning_rate/qwen_vl_interface': 9.326790452691332e-06, 'learning_rate/action_model': 9.326790452691332e-05, 'epoch': 0.69})
+[2026-07-28 16:51:12,681][starVLA.training.train_starvla][INFO] - Step 3770, Loss: {'train/loss': 0.06416893005371094, 'train/grad_norm_pre_clip': 1.941598892211914, 'timing/data': 0.00024887919425964355, 'timing/model': 0.42562447721138597, 'learning_rate/qwen_vl_interface': 9.324825490496327e-06, 'learning_rate/action_model': 9.324825490496327e-05, 'epoch': 0.69})
+[2026-07-28 16:51:18,974][starVLA.training.train_starvla][INFO] - Step 3771, Loss: {'train/loss': 0.03765708580613136, 'train/grad_norm_pre_clip': 0.9580520391464233, 'timing/data': 0.0002565477043390274, 'timing/model': 0.4701390997506678, 'learning_rate/qwen_vl_interface': 9.322857884191303e-06, 'learning_rate/action_model': 9.322857884191302e-05, 'epoch': 0.69})
+[2026-07-28 16:51:25,035][starVLA.training.train_starvla][INFO] - Step 3772, Loss: {'train/loss': 0.05788802728056908, 'train/grad_norm_pre_clip': 2.308316707611084, 'timing/data': 0.00031637540087103844, 'timing/model': 0.42110152496024966, 'learning_rate/qwen_vl_interface': 9.320887635053013e-06, 'learning_rate/action_model': 9.320887635053014e-05, 'epoch': 0.69})
+[2026-07-28 16:51:31,115][starVLA.training.train_starvla][INFO] - Step 3773, Loss: {'train/loss': 0.07217811048030853, 'train/grad_norm_pre_clip': 1.2414332628250122, 'timing/data': 0.00028936099261045456, 'timing/model': 0.45273466501384974, 'learning_rate/qwen_vl_interface': 9.318914744359936e-06, 'learning_rate/action_model': 9.318914744359937e-05, 'epoch': 0.69})
+[2026-07-28 16:51:37,043][starVLA.training.train_starvla][INFO] - Step 3774, Loss: {'train/loss': 0.050609882920980453, 'train/grad_norm_pre_clip': 1.5707634687423706, 'timing/data': 0.00036249076947569847, 'timing/model': 0.43081704527139664, 'learning_rate/qwen_vl_interface': 9.316939213392257e-06, 'learning_rate/action_model': 9.316939213392257e-05, 'epoch': 0.69})
+[2026-07-28 16:51:43,510][starVLA.training.train_starvla][INFO] - Step 3775, Loss: {'train/loss': 0.04279293864965439, 'train/grad_norm_pre_clip': 1.3521586656570435, 'timing/data': 0.00025429995730519295, 'timing/model': 0.6609861669130623, 'learning_rate/qwen_vl_interface': 9.314961043431877e-06, 'learning_rate/action_model': 9.314961043431876e-05, 'epoch': 0.69})
+[2026-07-28 16:51:49,461][starVLA.training.train_starvla][INFO] - Step 3776, Loss: {'train/loss': 0.02397257462143898, 'train/grad_norm_pre_clip': 1.054160475730896, 'timing/data': 0.00031046103686094284, 'timing/model': 0.4185889381915331, 'learning_rate/qwen_vl_interface': 9.312980235762408e-06, 'learning_rate/action_model': 9.312980235762408e-05, 'epoch': 0.69})
+[2026-07-28 16:51:55,449][starVLA.training.train_starvla][INFO] - Step 3777, Loss: {'train/loss': 0.03819984942674637, 'train/grad_norm_pre_clip': 1.382837176322937, 'timing/data': 0.0003056460991501808, 'timing/model': 0.4204501761123538, 'learning_rate/qwen_vl_interface': 9.310996791669175e-06, 'learning_rate/action_model': 9.310996791669175e-05, 'epoch': 0.69})
+[2026-07-28 16:52:01,939][starVLA.training.train_starvla][INFO] - Step 3778, Loss: {'train/loss': 0.03156464546918869, 'train/grad_norm_pre_clip': 1.322493314743042, 'timing/data': 0.00031420309096574783, 'timing/model': 0.7202734556049109, 'learning_rate/qwen_vl_interface': 9.309010712439213e-06, 'learning_rate/action_model': 9.309010712439213e-05, 'epoch': 0.69})
+[2026-07-28 16:52:07,973][starVLA.training.train_starvla][INFO] - Step 3779, Loss: {'train/loss': 0.05366174504160881, 'train/grad_norm_pre_clip': 1.616525650024414, 'timing/data': 0.0002982430160045624, 'timing/model': 0.42030300898477435, 'learning_rate/qwen_vl_interface': 9.307021999361267e-06, 'learning_rate/action_model': 9.307021999361267e-05, 'epoch': 0.69})
+[2026-07-28 16:52:14,272][starVLA.training.train_starvla][INFO] - Step 3780, Loss: {'train/loss': 0.0662994384765625, 'train/grad_norm_pre_clip': 1.2071688175201416, 'timing/data': 0.0003600679337978363, 'timing/model': 0.43330685002729297, 'learning_rate/qwen_vl_interface': 9.305030653725794e-06, 'learning_rate/action_model': 9.305030653725794e-05, 'epoch': 0.69})
+[2026-07-28 16:52:20,330][starVLA.training.train_starvla][INFO] - Step 3781, Loss: {'train/loss': 0.04092475771903992, 'train/grad_norm_pre_clip': 0.7348496317863464, 'timing/data': 0.0004299762658774853, 'timing/model': 0.42041204776614904, 'learning_rate/qwen_vl_interface': 9.303036676824954e-06, 'learning_rate/action_model': 9.303036676824953e-05, 'epoch': 0.69})
+[2026-07-28 16:52:26,373][starVLA.training.train_starvla][INFO] - Step 3782, Loss: {'train/loss': 0.027089424431324005, 'train/grad_norm_pre_clip': 0.5857565402984619, 'timing/data': 0.0003240327350795269, 'timing/model': 0.421591155230999, 'learning_rate/qwen_vl_interface': 9.301040069952616e-06, 'learning_rate/action_model': 9.301040069952616e-05, 'epoch': 0.69})
+[2026-07-28 16:52:32,902][starVLA.training.train_starvla][INFO] - Step 3783, Loss: {'train/loss': 0.01780470833182335, 'train/grad_norm_pre_clip': 0.8886632323265076, 'timing/data': 0.00029836175963282585, 'timing/model': 0.6541393361985683, 'learning_rate/qwen_vl_interface': 9.299040834404362e-06, 'learning_rate/action_model': 9.299040834404361e-05, 'epoch': 0.69})
+[2026-07-28 16:52:39,009][starVLA.training.train_starvla][INFO] - Step 3784, Loss: {'train/loss': 0.04136146605014801, 'train/grad_norm_pre_clip': 1.4273675680160522, 'timing/data': 0.00026174401864409447, 'timing/model': 0.420236777048558, 'learning_rate/qwen_vl_interface': 9.297038971477468e-06, 'learning_rate/action_model': 9.297038971477466e-05, 'epoch': 0.7})
+[2026-07-28 16:52:45,059][starVLA.training.train_starvla][INFO] - Step 3785, Loss: {'train/loss': 0.03516649454832077, 'train/grad_norm_pre_clip': 1.922409176826477, 'timing/data': 0.0004169088788330555, 'timing/model': 0.42089530685916543, 'learning_rate/qwen_vl_interface': 9.295034482470922e-06, 'learning_rate/action_model': 9.295034482470923e-05, 'epoch': 0.7})
+[2026-07-28 16:52:51,252][starVLA.training.train_starvla][INFO] - Step 3786, Loss: {'train/loss': 0.01690790429711342, 'train/grad_norm_pre_clip': 0.7479543089866638, 'timing/data': 0.0004252600483596325, 'timing/model': 0.6085821851156652, 'learning_rate/qwen_vl_interface': 9.293027368685418e-06, 'learning_rate/action_model': 9.293027368685419e-05, 'epoch': 0.7})
+[2026-07-28 16:52:57,365][starVLA.training.train_starvla][INFO] - Step 3787, Loss: {'train/loss': 0.051787227392196655, 'train/grad_norm_pre_clip': 1.9874075651168823, 'timing/data': 0.0006329012103378773, 'timing/model': 0.4201935171149671, 'learning_rate/qwen_vl_interface': 9.29101763142335e-06, 'learning_rate/action_model': 9.29101763142335e-05, 'epoch': 0.7})
+[2026-07-28 16:53:03,398][starVLA.training.train_starvla][INFO] - Step 3788, Loss: {'train/loss': 0.027783259749412537, 'train/grad_norm_pre_clip': 0.7348886132240295, 'timing/data': 0.0004399670287966728, 'timing/model': 0.42222614819183946, 'learning_rate/qwen_vl_interface': 9.28900527198881e-06, 'learning_rate/action_model': 9.289005271988811e-05, 'epoch': 0.7})
+[2026-07-28 16:53:09,675][starVLA.training.train_starvla][INFO] - Step 3789, Loss: {'train/loss': 0.06597750633955002, 'train/grad_norm_pre_clip': 3.4585845470428467, 'timing/data': 0.0003161327913403511, 'timing/model': 0.7170817302539945, 'learning_rate/qwen_vl_interface': 9.286990291687605e-06, 'learning_rate/action_model': 9.286990291687604e-05, 'epoch': 0.7})
+[2026-07-28 16:53:15,928][starVLA.training.train_starvla][INFO] - Step 3790, Loss: {'train/loss': 0.04657992348074913, 'train/grad_norm_pre_clip': 0.8542358875274658, 'timing/data': 0.00026366906240582466, 'timing/model': 0.42045554611831903, 'learning_rate/qwen_vl_interface': 9.284972691827226e-06, 'learning_rate/action_model': 9.284972691827224e-05, 'epoch': 0.7})
+[2026-07-28 16:53:21,949][starVLA.training.train_starvla][INFO] - Step 3791, Loss: {'train/loss': 0.05324186384677887, 'train/grad_norm_pre_clip': 1.9812941551208496, 'timing/data': 0.00038426509127020836, 'timing/model': 0.42518180701881647, 'learning_rate/qwen_vl_interface': 9.282952473716873e-06, 'learning_rate/action_model': 9.282952473716874e-05, 'epoch': 0.7})
+[2026-07-28 16:53:27,922][starVLA.training.train_starvla][INFO] - Step 3792, Loss: {'train/loss': 0.04782232642173767, 'train/grad_norm_pre_clip': 1.6376248598098755, 'timing/data': 0.00043341517448425293, 'timing/model': 0.4202309479005635, 'learning_rate/qwen_vl_interface': 9.280929638667447e-06, 'learning_rate/action_model': 9.280929638667447e-05, 'epoch': 0.7})
+[2026-07-28 16:53:34,016][starVLA.training.train_starvla][INFO] - Step 3793, Loss: {'train/loss': 0.042969051748514175, 'train/grad_norm_pre_clip': 2.005990743637085, 'timing/data': 0.0003134319558739662, 'timing/model': 0.42194680497050285, 'learning_rate/qwen_vl_interface': 9.278904187991542e-06, 'learning_rate/action_model': 9.278904187991541e-05, 'epoch': 0.7})
+[2026-07-28 16:53:40,286][starVLA.training.train_starvla][INFO] - Step 3794, Loss: {'train/loss': 0.030880235135555267, 'train/grad_norm_pre_clip': 1.1359448432922363, 'timing/data': 0.0003010649234056473, 'timing/model': 0.6610015691258013, 'learning_rate/qwen_vl_interface': 9.27687612300345e-06, 'learning_rate/action_model': 9.27687612300345e-05, 'epoch': 0.7})
+[2026-07-28 16:53:46,300][starVLA.training.train_starvla][INFO] - Step 3795, Loss: {'train/loss': 0.05397751182317734, 'train/grad_norm_pre_clip': 1.571067452430725, 'timing/data': 0.0003452589735388756, 'timing/model': 0.42009106604382396, 'learning_rate/qwen_vl_interface': 9.27484544501916e-06, 'learning_rate/action_model': 9.274845445019159e-05, 'epoch': 0.7})
+[2026-07-28 16:53:52,286][starVLA.training.train_starvla][INFO] - Step 3796, Loss: {'train/loss': 0.04133090749382973, 'train/grad_norm_pre_clip': 1.759638786315918, 'timing/data': 0.0002989610657095909, 'timing/model': 0.42018906166777015, 'learning_rate/qwen_vl_interface': 9.27281215535636e-06, 'learning_rate/action_model': 9.272812155356358e-05, 'epoch': 0.7})
+[2026-07-28 16:53:58,327][starVLA.training.train_starvla][INFO] - Step 3797, Loss: {'train/loss': 0.01993265561759472, 'train/grad_norm_pre_clip': 1.0403333902359009, 'timing/data': 0.00032483693212270737, 'timing/model': 0.4237222261726856, 'learning_rate/qwen_vl_interface': 9.270776255334425e-06, 'learning_rate/action_model': 9.270776255334425e-05, 'epoch': 0.7})
+[2026-07-28 16:54:04,620][starVLA.training.train_starvla][INFO] - Step 3798, Loss: {'train/loss': 0.04541340470314026, 'train/grad_norm_pre_clip': 2.198519229888916, 'timing/data': 0.0002775699831545353, 'timing/model': 0.42272707587108016, 'learning_rate/qwen_vl_interface': 9.268737746274432e-06, 'learning_rate/action_model': 9.268737746274432e-05, 'epoch': 0.7})
+[2026-07-28 16:54:10,667][starVLA.training.train_starvla][INFO] - Step 3799, Loss: {'train/loss': 0.02386349067091942, 'train/grad_norm_pre_clip': 0.9566600918769836, 'timing/data': 0.00043638283386826515, 'timing/model': 0.4195426367223263, 'learning_rate/qwen_vl_interface': 9.266696629499146e-06, 'learning_rate/action_model': 9.266696629499145e-05, 'epoch': 0.7})
+[2026-07-28 16:54:16,946][starVLA.training.train_starvla][INFO] - Step 3800, Loss: {'train/loss': 0.02157137356698513, 'train/grad_norm_pre_clip': 0.7903593182563782, 'timing/data': 0.00024937698617577553, 'timing/model': 0.41903649363666773, 'learning_rate/qwen_vl_interface': 9.264652906333026e-06, 'learning_rate/action_model': 9.264652906333026e-05, 'epoch': 0.7})
+[2026-07-28 16:54:22,942][starVLA.training.train_starvla][INFO] - Step 3801, Loss: {'train/loss': 0.029549457132816315, 'train/grad_norm_pre_clip': 0.9132830500602722, 'timing/data': 0.0004451097920536995, 'timing/model': 0.4233276299200952, 'learning_rate/qwen_vl_interface': 9.26260657810222e-06, 'learning_rate/action_model': 9.262606578102221e-05, 'epoch': 0.7})
+[2026-07-28 16:54:28,930][starVLA.training.train_starvla][INFO] - Step 3802, Loss: {'train/loss': 0.01735331490635872, 'train/grad_norm_pre_clip': 1.0315088033676147, 'timing/data': 0.0003056386485695839, 'timing/model': 0.42063052114099264, 'learning_rate/qwen_vl_interface': 9.260557646134572e-06, 'learning_rate/action_model': 9.260557646134572e-05, 'epoch': 0.7})
+[2026-07-28 16:54:35,169][starVLA.training.train_starvla][INFO] - Step 3803, Loss: {'train/loss': 0.0307135172188282, 'train/grad_norm_pre_clip': 1.0566248893737793, 'timing/data': 0.00041255727410316467, 'timing/model': 0.6764297699555755, 'learning_rate/qwen_vl_interface': 9.258506111759608e-06, 'learning_rate/action_model': 9.258506111759609e-05, 'epoch': 0.7})
+[2026-07-28 16:54:41,193][starVLA.training.train_starvla][INFO] - Step 3804, Loss: {'train/loss': 0.01652372069656849, 'train/grad_norm_pre_clip': 0.8030068278312683, 'timing/data': 0.0002650446258485317, 'timing/model': 0.4198791729286313, 'learning_rate/qwen_vl_interface': 9.256451976308548e-06, 'learning_rate/action_model': 9.25645197630855e-05, 'epoch': 0.7})
+[2026-07-28 16:54:47,389][starVLA.training.train_starvla][INFO] - Step 3805, Loss: {'train/loss': 0.03965669125318527, 'train/grad_norm_pre_clip': 2.093414783477783, 'timing/data': 0.00035293400287628174, 'timing/model': 0.42001121677458286, 'learning_rate/qwen_vl_interface': 9.2543952411143e-06, 'learning_rate/action_model': 9.254395241114299e-05, 'epoch': 0.7})
+[2026-07-28 16:54:53,361][starVLA.training.train_starvla][INFO] - Step 3806, Loss: {'train/loss': 0.06078134849667549, 'train/grad_norm_pre_clip': 2.0504279136657715, 'timing/data': 0.00031124474480748177, 'timing/model': 0.4274175679311156, 'learning_rate/qwen_vl_interface': 9.252335907511452e-06, 'learning_rate/action_model': 9.252335907511453e-05, 'epoch': 0.7})
+[2026-07-28 16:54:59,625][starVLA.training.train_starvla][INFO] - Step 3807, Loss: {'train/loss': 0.021284999325871468, 'train/grad_norm_pre_clip': 0.6009011268615723, 'timing/data': 0.0003026006743311882, 'timing/model': 0.6594137880019844, 'learning_rate/qwen_vl_interface': 9.250273976836289e-06, 'learning_rate/action_model': 9.250273976836289e-05, 'epoch': 0.7})
+[2026-07-28 16:55:05,794][starVLA.training.train_starvla][INFO] - Step 3808, Loss: {'train/loss': 0.025152016431093216, 'train/grad_norm_pre_clip': 0.7055969834327698, 'timing/data': 0.0003921627067029476, 'timing/model': 0.4195547220297158, 'learning_rate/qwen_vl_interface': 9.248209450426772e-06, 'learning_rate/action_model': 9.248209450426772e-05, 'epoch': 0.7})
+[2026-07-28 16:55:11,812][starVLA.training.train_starvla][INFO] - Step 3809, Loss: {'train/loss': 0.026918655261397362, 'train/grad_norm_pre_clip': 0.9759955406188965, 'timing/data': 0.0003707949072122574, 'timing/model': 0.4211527141742408, 'learning_rate/qwen_vl_interface': 9.24614232962255e-06, 'learning_rate/action_model': 9.24614232962255e-05, 'epoch': 0.7})
+[2026-07-28 16:55:18,060][starVLA.training.train_starvla][INFO] - Step 3810, Loss: {'train/loss': 0.05828192085027695, 'train/grad_norm_pre_clip': 1.7307766675949097, 'timing/data': 0.0003808862529695034, 'timing/model': 0.42002200381830335, 'learning_rate/qwen_vl_interface': 9.244072615764957e-06, 'learning_rate/action_model': 9.244072615764956e-05, 'epoch': 0.7})
+[2026-07-28 16:55:24,353][starVLA.training.train_starvla][INFO] - Step 3811, Loss: {'train/loss': 0.03089209832251072, 'train/grad_norm_pre_clip': 1.0653783082962036, 'timing/data': 0.0003511039540171623, 'timing/model': 0.4202908310107887, 'learning_rate/qwen_vl_interface': 9.242000310197005e-06, 'learning_rate/action_model': 9.242000310197004e-05, 'epoch': 0.7})
+[2026-07-28 16:55:30,420][starVLA.training.train_starvla][INFO] - Step 3812, Loss: {'train/loss': 0.028705641627311707, 'train/grad_norm_pre_clip': 1.5816094875335693, 'timing/data': 0.0003853021189570427, 'timing/model': 0.4205937488004565, 'learning_rate/qwen_vl_interface': 9.239925414263392e-06, 'learning_rate/action_model': 9.239925414263392e-05, 'epoch': 0.7})
+[2026-07-28 16:55:36,198][starVLA.training.train_starvla][INFO] - Step 3813, Loss: {'train/loss': 0.06587396562099457, 'train/grad_norm_pre_clip': 1.9028359651565552, 'timing/data': 0.00028459494933485985, 'timing/model': 0.42046189587563276, 'learning_rate/qwen_vl_interface': 9.237847929310497e-06, 'learning_rate/action_model': 9.237847929310496e-05, 'epoch': 0.7})
+[2026-07-28 16:55:42,455][starVLA.training.train_starvla][INFO] - Step 3814, Loss: {'train/loss': 0.013539023697376251, 'train/grad_norm_pre_clip': 0.5296452641487122, 'timing/data': 0.00043109292164444923, 'timing/model': 0.652247202116996, 'learning_rate/qwen_vl_interface': 9.235767856686375e-06, 'learning_rate/action_model': 9.235767856686375e-05, 'epoch': 0.7})
+[2026-07-28 16:55:48,476][starVLA.training.train_starvla][INFO] - Step 3815, Loss: {'train/loss': 0.024591512978076935, 'train/grad_norm_pre_clip': 0.8022700548171997, 'timing/data': 0.0003665359690785408, 'timing/model': 0.4206853178329766, 'learning_rate/qwen_vl_interface': 9.233685197740765e-06, 'learning_rate/action_model': 9.233685197740764e-05, 'epoch': 0.7})
+[2026-07-28 16:55:54,280][starVLA.training.train_starvla][INFO] - Step 3816, Loss: {'train/loss': 0.02640308439731598, 'train/grad_norm_pre_clip': 1.1255719661712646, 'timing/data': 0.0003072097897529602, 'timing/model': 0.4201660049147904, 'learning_rate/qwen_vl_interface': 9.23159995382508e-06, 'learning_rate/action_model': 9.231599953825079e-05, 'epoch': 0.7})
+[2026-07-28 16:56:00,500][starVLA.training.train_starvla][INFO] - Step 3817, Loss: {'train/loss': 0.02001115307211876, 'train/grad_norm_pre_clip': 0.3759182095527649, 'timing/data': 0.0003458620049059391, 'timing/model': 0.6053386912681162, 'learning_rate/qwen_vl_interface': 9.22951212629241e-06, 'learning_rate/action_model': 9.229512126292411e-05, 'epoch': 0.7})
+[2026-07-28 16:56:06,781][starVLA.training.train_starvla][INFO] - Step 3818, Loss: {'train/loss': 0.05324837565422058, 'train/grad_norm_pre_clip': 1.3805930614471436, 'timing/data': 0.0002491059713065624, 'timing/model': 0.4236101699061692, 'learning_rate/qwen_vl_interface': 9.227421716497531e-06, 'learning_rate/action_model': 9.227421716497531e-05, 'epoch': 0.7})
+[2026-07-28 16:56:12,777][starVLA.training.train_starvla][INFO] - Step 3819, Loss: {'train/loss': 0.05064881592988968, 'train/grad_norm_pre_clip': 1.2316522598266602, 'timing/data': 0.0003177565522491932, 'timing/model': 0.4201252330094576, 'learning_rate/qwen_vl_interface': 9.225328725796879e-06, 'learning_rate/action_model': 9.22532872579688e-05, 'epoch': 0.7})
+[2026-07-28 16:56:18,946][starVLA.training.train_starvla][INFO] - Step 3820, Loss: {'train/loss': 0.04786486551165581, 'train/grad_norm_pre_clip': 1.832436442375183, 'timing/data': 0.00028728973120450974, 'timing/model': 0.4199542310088873, 'learning_rate/qwen_vl_interface': 9.223233155548579e-06, 'learning_rate/action_model': 9.223233155548578e-05, 'epoch': 0.7})
+[2026-07-28 16:56:25,400][starVLA.training.train_starvla][INFO] - Step 3821, Loss: {'train/loss': 0.018518049269914627, 'train/grad_norm_pre_clip': 1.0585628747940063, 'timing/data': 0.0002719312906265259, 'timing/model': 0.4231806378811598, 'learning_rate/qwen_vl_interface': 9.22113500711242e-06, 'learning_rate/action_model': 9.221135007112419e-05, 'epoch': 0.7})
+[2026-07-28 16:56:31,775][starVLA.training.train_starvla][INFO] - Step 3822, Loss: {'train/loss': 0.04448104277253151, 'train/grad_norm_pre_clip': 1.5507395267486572, 'timing/data': 0.0003327629528939724, 'timing/model': 0.42061959533020854, 'learning_rate/qwen_vl_interface': 9.219034281849867e-06, 'learning_rate/action_model': 9.219034281849867e-05, 'epoch': 0.7})
+[2026-07-28 16:56:37,764][starVLA.training.train_starvla][INFO] - Step 3823, Loss: {'train/loss': 0.0326341837644577, 'train/grad_norm_pre_clip': 0.748500406742096, 'timing/data': 0.00029494985938072205, 'timing/model': 0.42007420025765896, 'learning_rate/qwen_vl_interface': 9.216930981124062e-06, 'learning_rate/action_model': 9.216930981124062e-05, 'epoch': 0.7})
+[2026-07-28 16:56:44,106][starVLA.training.train_starvla][INFO] - Step 3824, Loss: {'train/loss': 0.039553381502628326, 'train/grad_norm_pre_clip': 0.7745221853256226, 'timing/data': 0.00029537733644247055, 'timing/model': 0.6599518121220171, 'learning_rate/qwen_vl_interface': 9.214825106299808e-06, 'learning_rate/action_model': 9.214825106299808e-05, 'epoch': 0.7})
+[2026-07-28 16:56:50,429][starVLA.training.train_starvla][INFO] - Step 3825, Loss: {'train/loss': 0.08703745901584625, 'train/grad_norm_pre_clip': 1.5969507694244385, 'timing/data': 0.00041353795677423477, 'timing/model': 0.6787171112373471, 'learning_rate/qwen_vl_interface': 9.212716658743592e-06, 'learning_rate/action_model': 9.212716658743592e-05, 'epoch': 0.7})
+[2026-07-28 16:56:56,434][starVLA.training.train_starvla][INFO] - Step 3826, Loss: {'train/loss': 0.03446761146187782, 'train/grad_norm_pre_clip': 1.1693391799926758, 'timing/data': 0.00031352415680885315, 'timing/model': 0.42347160493955016, 'learning_rate/qwen_vl_interface': 9.210605639823558e-06, 'learning_rate/action_model': 9.210605639823557e-05, 'epoch': 0.7})
+[2026-07-28 16:57:02,459][starVLA.training.train_starvla][INFO] - Step 3827, Loss: {'train/loss': 0.03277251124382019, 'train/grad_norm_pre_clip': 1.0086055994033813, 'timing/data': 0.0002967389300465584, 'timing/model': 0.42221246287226677, 'learning_rate/qwen_vl_interface': 9.208492050909523e-06, 'learning_rate/action_model': 9.208492050909523e-05, 'epoch': 0.7})
+[2026-07-28 16:57:08,715][starVLA.training.train_starvla][INFO] - Step 3828, Loss: {'train/loss': 0.03124106302857399, 'train/grad_norm_pre_clip': 1.3165605068206787, 'timing/data': 0.0004358869045972824, 'timing/model': 0.42899344908073545, 'learning_rate/qwen_vl_interface': 9.206375893372973e-06, 'learning_rate/action_model': 9.206375893372973e-05, 'epoch': 0.7})
+[2026-07-28 16:57:14,721][starVLA.training.train_starvla][INFO] - Step 3829, Loss: {'train/loss': 0.023682331666350365, 'train/grad_norm_pre_clip': 0.4939335882663727, 'timing/data': 0.0003185737878084183, 'timing/model': 0.42346446216106415, 'learning_rate/qwen_vl_interface': 9.204257168587061e-06, 'learning_rate/action_model': 9.204257168587061e-05, 'epoch': 0.7})
+[2026-07-28 16:57:21,021][starVLA.training.train_starvla][INFO] - Step 3830, Loss: {'train/loss': 0.05939171090722084, 'train/grad_norm_pre_clip': 1.5114911794662476, 'timing/data': 0.0003452412784099579, 'timing/model': 0.6661216802895069, 'learning_rate/qwen_vl_interface': 9.2021358779266e-06, 'learning_rate/action_model': 9.202135877926601e-05, 'epoch': 0.7})
+[2026-07-28 16:57:27,245][starVLA.training.train_starvla][INFO] - Step 3831, Loss: {'train/loss': 0.07336480170488358, 'train/grad_norm_pre_clip': 1.3336626291275024, 'timing/data': 0.0003619459457695484, 'timing/model': 0.4199523380957544, 'learning_rate/qwen_vl_interface': 9.200012022768077e-06, 'learning_rate/action_model': 9.200012022768077e-05, 'epoch': 0.7})
+[2026-07-28 16:57:33,238][starVLA.training.train_starvla][INFO] - Step 3832, Loss: {'train/loss': 0.04319965839385986, 'train/grad_norm_pre_clip': 1.4413166046142578, 'timing/data': 0.00038365693762898445, 'timing/model': 0.4211844499222934, 'learning_rate/qwen_vl_interface': 9.19788560448964e-06, 'learning_rate/action_model': 9.197885604489639e-05, 'epoch': 0.7})
+[2026-07-28 16:57:39,230][starVLA.training.train_starvla][INFO] - Step 3833, Loss: {'train/loss': 0.034477610141038895, 'train/grad_norm_pre_clip': 1.0487828254699707, 'timing/data': 0.0002819220535457134, 'timing/model': 0.42460959730669856, 'learning_rate/qwen_vl_interface': 9.195756624471094e-06, 'learning_rate/action_model': 9.195756624471094e-05, 'epoch': 0.7})
+[2026-07-28 16:57:45,449][starVLA.training.train_starvla][INFO] - Step 3834, Loss: {'train/loss': 0.060428015887737274, 'train/grad_norm_pre_clip': 2.6759209632873535, 'timing/data': 0.0002583172172307968, 'timing/model': 0.4244964919053018, 'learning_rate/qwen_vl_interface': 9.193625084093912e-06, 'learning_rate/action_model': 9.193625084093911e-05, 'epoch': 0.7})
+[2026-07-28 16:57:51,498][starVLA.training.train_starvla][INFO] - Step 3835, Loss: {'train/loss': 0.027081724256277084, 'train/grad_norm_pre_clip': 1.4592372179031372, 'timing/data': 0.00046183494850993156, 'timing/model': 0.4201964111998677, 'learning_rate/qwen_vl_interface': 9.191490984741227e-06, 'learning_rate/action_model': 9.191490984741227e-05, 'epoch': 0.7})
+[2026-07-28 16:57:57,769][starVLA.training.train_starvla][INFO] - Step 3836, Loss: {'train/loss': 0.015035774558782578, 'train/grad_norm_pre_clip': 0.8935834765434265, 'timing/data': 0.0003079809248447418, 'timing/model': 0.6895939488895237, 'learning_rate/qwen_vl_interface': 9.189354327797835e-06, 'learning_rate/action_model': 9.189354327797836e-05, 'epoch': 0.7})
+[2026-07-28 16:58:03,772][starVLA.training.train_starvla][INFO] - Step 3837, Loss: {'train/loss': 0.02417089231312275, 'train/grad_norm_pre_clip': 1.0838617086410522, 'timing/data': 0.0002993829548358917, 'timing/model': 0.4218584499321878, 'learning_rate/qwen_vl_interface': 9.18721511465019e-06, 'learning_rate/action_model': 9.187215114650191e-05, 'epoch': 0.7})
+[2026-07-28 16:58:10,214][starVLA.training.train_starvla][INFO] - Step 3838, Loss: {'train/loss': 0.028241019695997238, 'train/grad_norm_pre_clip': 1.365193247795105, 'timing/data': 0.0003506438806653023, 'timing/model': 0.7687640697695315, 'learning_rate/qwen_vl_interface': 9.185073346686402e-06, 'learning_rate/action_model': 9.185073346686401e-05, 'epoch': 0.7})
+[2026-07-28 16:58:16,357][starVLA.training.train_starvla][INFO] - Step 3839, Loss: {'train/loss': 0.0680726170539856, 'train/grad_norm_pre_clip': 2.5814809799194336, 'timing/data': 0.00024721911177039146, 'timing/model': 0.4837187733501196, 'learning_rate/qwen_vl_interface': 9.182929025296242e-06, 'learning_rate/action_model': 9.182929025296242e-05, 'epoch': 0.71})
+[2026-07-28 16:58:22,408][starVLA.training.train_starvla][INFO] - Step 3840, Loss: {'train/loss': 0.09591034054756165, 'train/grad_norm_pre_clip': 2.40152645111084, 'timing/data': 0.00036671385169029236, 'timing/model': 0.4300399082712829, 'learning_rate/qwen_vl_interface': 9.180782151871138e-06, 'learning_rate/action_model': 9.180782151871137e-05, 'epoch': 0.71})
+[2026-07-28 16:58:28,678][starVLA.training.train_starvla][INFO] - Step 3841, Loss: {'train/loss': 0.0491025410592556, 'train/grad_norm_pre_clip': 0.877710223197937, 'timing/data': 0.00036755017936229706, 'timing/model': 0.4272626410238445, 'learning_rate/qwen_vl_interface': 9.17863272780417e-06, 'learning_rate/action_model': 9.178632727804168e-05, 'epoch': 0.71})
+[2026-07-28 16:58:34,722][starVLA.training.train_starvla][INFO] - Step 3842, Loss: {'train/loss': 0.04218994826078415, 'train/grad_norm_pre_clip': 1.8410125970840454, 'timing/data': 0.0003172950819134712, 'timing/model': 0.421146668959409, 'learning_rate/qwen_vl_interface': 9.176480754490078e-06, 'learning_rate/action_model': 9.176480754490078e-05, 'epoch': 0.71})
+[2026-07-28 16:58:40,946][starVLA.training.train_starvla][INFO] - Step 3843, Loss: {'train/loss': 0.020859135314822197, 'train/grad_norm_pre_clip': 1.3549922704696655, 'timing/data': 0.00030341092497110367, 'timing/model': 0.43152284529060125, 'learning_rate/qwen_vl_interface': 9.174326233325253e-06, 'learning_rate/action_model': 9.174326233325253e-05, 'epoch': 0.71})
+[2026-07-28 16:58:47,276][starVLA.training.train_starvla][INFO] - Step 3844, Loss: {'train/loss': 0.0372132882475853, 'train/grad_norm_pre_clip': 1.0978935956954956, 'timing/data': 0.000256482046097517, 'timing/model': 0.42023602174595, 'learning_rate/qwen_vl_interface': 9.17216916570774e-06, 'learning_rate/action_model': 9.172169165707741e-05, 'epoch': 0.71})
+[2026-07-28 16:58:53,493][starVLA.training.train_starvla][INFO] - Step 3845, Loss: {'train/loss': 0.06809772551059723, 'train/grad_norm_pre_clip': 1.54783296585083, 'timing/data': 0.000265663955360651, 'timing/model': 0.42206059768795967, 'learning_rate/qwen_vl_interface': 9.170009553037239e-06, 'learning_rate/action_model': 9.170009553037238e-05, 'epoch': 0.71})
+[2026-07-28 16:58:59,672][starVLA.training.train_starvla][INFO] - Step 3846, Loss: {'train/loss': 0.06426940858364105, 'train/grad_norm_pre_clip': 1.2133368253707886, 'timing/data': 0.00025732675567269325, 'timing/model': 0.45798156689852476, 'learning_rate/qwen_vl_interface': 9.167847396715098e-06, 'learning_rate/action_model': 9.167847396715098e-05, 'epoch': 0.71})
+[2026-07-28 16:59:05,783][starVLA.training.train_starvla][INFO] - Step 3847, Loss: {'train/loss': 0.031548626720905304, 'train/grad_norm_pre_clip': 1.1533191204071045, 'timing/data': 0.0003139739856123924, 'timing/model': 0.42473089322447777, 'learning_rate/qwen_vl_interface': 9.165682698144317e-06, 'learning_rate/action_model': 9.165682698144317e-05, 'epoch': 0.71})
+[2026-07-28 16:59:12,212][starVLA.training.train_starvla][INFO] - Step 3848, Loss: {'train/loss': 0.04422681778669357, 'train/grad_norm_pre_clip': 2.0517704486846924, 'timing/data': 0.00035215914249420166, 'timing/model': 0.4957043780013919, 'learning_rate/qwen_vl_interface': 9.163515458729545e-06, 'learning_rate/action_model': 9.163515458729545e-05, 'epoch': 0.71})
+[2026-07-28 16:59:18,284][starVLA.training.train_starvla][INFO] - Step 3849, Loss: {'train/loss': 0.032635994255542755, 'train/grad_norm_pre_clip': 1.4241045713424683, 'timing/data': 0.0002907607704401016, 'timing/model': 0.42098509473726153, 'learning_rate/qwen_vl_interface': 9.16134567987708e-06, 'learning_rate/action_model': 9.16134567987708e-05, 'epoch': 0.71})
+[2026-07-28 16:59:24,609][starVLA.training.train_starvla][INFO] - Step 3850, Loss: {'train/loss': 0.038465362042188644, 'train/grad_norm_pre_clip': 0.7514036297798157, 'timing/data': 0.0002995510585606098, 'timing/model': 0.6968618761748075, 'learning_rate/qwen_vl_interface': 9.159173362994868e-06, 'learning_rate/action_model': 9.159173362994869e-05, 'epoch': 0.71})
+[2026-07-28 16:59:30,850][starVLA.training.train_starvla][INFO] - Step 3851, Loss: {'train/loss': 0.029243793338537216, 'train/grad_norm_pre_clip': 1.0420494079589844, 'timing/data': 0.000250769779086113, 'timing/model': 0.42075832933187485, 'learning_rate/qwen_vl_interface': 9.156998509492503e-06, 'learning_rate/action_model': 9.156998509492503e-05, 'epoch': 0.71})
+[2026-07-28 16:59:36,945][starVLA.training.train_starvla][INFO] - Step 3852, Loss: {'train/loss': 0.03413255140185356, 'train/grad_norm_pre_clip': 0.7814542651176453, 'timing/data': 0.00031514326110482216, 'timing/model': 0.42071457114070654, 'learning_rate/qwen_vl_interface': 9.154821120781221e-06, 'learning_rate/action_model': 9.154821120781222e-05, 'epoch': 0.71})
+[2026-07-28 16:59:42,953][starVLA.training.train_starvla][INFO] - Step 3853, Loss: {'train/loss': 0.04242267459630966, 'train/grad_norm_pre_clip': 2.2978248596191406, 'timing/data': 0.0003942381590604782, 'timing/model': 0.4220984657295048, 'learning_rate/qwen_vl_interface': 9.152641198273911e-06, 'learning_rate/action_model': 9.15264119827391e-05, 'epoch': 0.71})
+[2026-07-28 16:59:49,147][starVLA.training.train_starvla][INFO] - Step 3854, Loss: {'train/loss': 0.05351390689611435, 'train/grad_norm_pre_clip': 0.7061545252799988, 'timing/data': 0.00040318118408322334, 'timing/model': 0.6032156320288777, 'learning_rate/qwen_vl_interface': 9.150458743385095e-06, 'learning_rate/action_model': 9.150458743385094e-05, 'epoch': 0.71})
+[2026-07-28 16:59:55,212][starVLA.training.train_starvla][INFO] - Step 3855, Loss: {'train/loss': 0.04195306450128555, 'train/grad_norm_pre_clip': 1.4090254306793213, 'timing/data': 0.00039386097341775894, 'timing/model': 0.4216332221403718, 'learning_rate/qwen_vl_interface': 9.148273757530948e-06, 'learning_rate/action_model': 9.148273757530948e-05, 'epoch': 0.71})
+[2026-07-28 17:00:01,257][starVLA.training.train_starvla][INFO] - Step 3856, Loss: {'train/loss': 0.05813799425959587, 'train/grad_norm_pre_clip': 0.913838803768158, 'timing/data': 0.00040588341653347015, 'timing/model': 0.42086961306631565, 'learning_rate/qwen_vl_interface': 9.146086242129283e-06, 'learning_rate/action_model': 9.146086242129282e-05, 'epoch': 0.71})
+[2026-07-28 17:00:07,535][starVLA.training.train_starvla][INFO] - Step 3857, Loss: {'train/loss': 0.02342999540269375, 'train/grad_norm_pre_clip': 0.6445357203483582, 'timing/data': 0.0003401204012334347, 'timing/model': 0.4197975932620466, 'learning_rate/qwen_vl_interface': 9.143896198599553e-06, 'learning_rate/action_model': 9.143896198599553e-05, 'epoch': 0.71})
+[2026-07-28 17:00:13,804][starVLA.training.train_starvla][INFO] - Step 3858, Loss: {'train/loss': 0.025513222441077232, 'train/grad_norm_pre_clip': 0.797077476978302, 'timing/data': 0.0002969861961901188, 'timing/model': 0.6610199161805212, 'learning_rate/qwen_vl_interface': 9.141703628362858e-06, 'learning_rate/action_model': 9.141703628362858e-05, 'epoch': 0.71})
+[2026-07-28 17:00:19,800][starVLA.training.train_starvla][INFO] - Step 3859, Loss: {'train/loss': 0.025180907920002937, 'train/grad_norm_pre_clip': 1.4442569017410278, 'timing/data': 0.00031397677958011627, 'timing/model': 0.42276676883921027, 'learning_rate/qwen_vl_interface': 9.139508532841928e-06, 'learning_rate/action_model': 9.139508532841927e-05, 'epoch': 0.71})
+[2026-07-28 17:00:25,900][starVLA.training.train_starvla][INFO] - Step 3860, Loss: {'train/loss': 0.06617486476898193, 'train/grad_norm_pre_clip': 1.672804832458496, 'timing/data': 0.0003320667892694473, 'timing/model': 0.4217109871096909, 'learning_rate/qwen_vl_interface': 9.137310913461139e-06, 'learning_rate/action_model': 9.137310913461139e-05, 'epoch': 0.71})
+[2026-07-28 17:00:32,218][starVLA.training.train_starvla][INFO] - Step 3861, Loss: {'train/loss': 0.02381734549999237, 'train/grad_norm_pre_clip': 1.1103124618530273, 'timing/data': 0.00035638967528939247, 'timing/model': 0.6649113246239722, 'learning_rate/qwen_vl_interface': 9.135110771646502e-06, 'learning_rate/action_model': 9.135110771646502e-05, 'epoch': 0.71})
+[2026-07-28 17:00:38,305][starVLA.training.train_starvla][INFO] - Step 3862, Loss: {'train/loss': 0.018369633704423904, 'train/grad_norm_pre_clip': 0.9436424970626831, 'timing/data': 0.0002950923517346382, 'timing/model': 0.42402155324816704, 'learning_rate/qwen_vl_interface': 9.132908108825667e-06, 'learning_rate/action_model': 9.132908108825667e-05, 'epoch': 0.71})
+[2026-07-28 17:00:44,158][starVLA.training.train_starvla][INFO] - Step 3863, Loss: {'train/loss': 0.02428889274597168, 'train/grad_norm_pre_clip': 0.5834981799125671, 'timing/data': 0.00032020360231399536, 'timing/model': 0.42033185716718435, 'learning_rate/qwen_vl_interface': 9.130702926427915e-06, 'learning_rate/action_model': 9.130702926427914e-05, 'epoch': 0.71})
+[2026-07-28 17:00:50,245][starVLA.training.train_starvla][INFO] - Step 3864, Loss: {'train/loss': 0.024025345221161842, 'train/grad_norm_pre_clip': 0.6239306330680847, 'timing/data': 0.0004633590579032898, 'timing/model': 0.41945367865264416, 'learning_rate/qwen_vl_interface': 9.128495225884167e-06, 'learning_rate/action_model': 9.128495225884166e-05, 'epoch': 0.71})
+[2026-07-28 17:00:56,524][starVLA.training.train_starvla][INFO] - Step 3865, Loss: {'train/loss': 0.04865533858537674, 'train/grad_norm_pre_clip': 1.3394116163253784, 'timing/data': 0.00024919304996728897, 'timing/model': 0.42152760503813624, 'learning_rate/qwen_vl_interface': 9.126285008626975e-06, 'learning_rate/action_model': 9.126285008626974e-05, 'epoch': 0.71})
+[2026-07-28 17:01:02,839][starVLA.training.train_starvla][INFO] - Step 3866, Loss: {'train/loss': 0.07524053007364273, 'train/grad_norm_pre_clip': 1.9322518110275269, 'timing/data': 0.00025822408497333527, 'timing/model': 0.42367233987897635, 'learning_rate/qwen_vl_interface': 9.124072276090524e-06, 'learning_rate/action_model': 9.124072276090524e-05, 'epoch': 0.71})
+[2026-07-28 17:01:08,912][starVLA.training.train_starvla][INFO] - Step 3867, Loss: {'train/loss': 0.06989401578903198, 'train/grad_norm_pre_clip': 2.6852529048919678, 'timing/data': 0.00032020313665270805, 'timing/model': 0.4213217282667756, 'learning_rate/qwen_vl_interface': 9.121857029710632e-06, 'learning_rate/action_model': 9.121857029710633e-05, 'epoch': 0.71})
+[2026-07-28 17:01:15,170][starVLA.training.train_starvla][INFO] - Step 3868, Loss: {'train/loss': 0.04519839584827423, 'train/grad_norm_pre_clip': 1.1652532815933228, 'timing/data': 0.0004980908706784248, 'timing/model': 0.4618552098982036, 'learning_rate/qwen_vl_interface': 9.119639270924754e-06, 'learning_rate/action_model': 9.119639270924754e-05, 'epoch': 0.71})
+[2026-07-28 17:01:21,526][starVLA.training.train_starvla][INFO] - Step 3869, Loss: {'train/loss': 0.08234763145446777, 'train/grad_norm_pre_clip': 2.0750622749328613, 'timing/data': 0.00032950984314084053, 'timing/model': 0.4199102269485593, 'learning_rate/qwen_vl_interface': 9.11741900117196e-06, 'learning_rate/action_model': 9.11741900117196e-05, 'epoch': 0.71})
+[2026-07-28 17:01:27,577][starVLA.training.train_starvla][INFO] - Step 3870, Loss: {'train/loss': 0.05410011485219002, 'train/grad_norm_pre_clip': 1.7924456596374512, 'timing/data': 0.00036818813532590866, 'timing/model': 0.42100423388183117, 'learning_rate/qwen_vl_interface': 9.115196221892969e-06, 'learning_rate/action_model': 9.115196221892969e-05, 'epoch': 0.71})
+[2026-07-28 17:01:33,890][starVLA.training.train_starvla][INFO] - Step 3871, Loss: {'train/loss': 0.01147688739001751, 'train/grad_norm_pre_clip': 0.9095443487167358, 'timing/data': 0.0003389548510313034, 'timing/model': 0.6664245692081749, 'learning_rate/qwen_vl_interface': 9.112970934530111e-06, 'learning_rate/action_model': 9.112970934530111e-05, 'epoch': 0.71})
+[2026-07-28 17:01:40,135][starVLA.training.train_starvla][INFO] - Step 3872, Loss: {'train/loss': 0.041885100305080414, 'train/grad_norm_pre_clip': 2.674142599105835, 'timing/data': 0.0003714826889336109, 'timing/model': 0.6556236511096358, 'learning_rate/qwen_vl_interface': 9.110743140527353e-06, 'learning_rate/action_model': 9.110743140527352e-05, 'epoch': 0.71})
+[2026-07-28 17:01:46,203][starVLA.training.train_starvla][INFO] - Step 3873, Loss: {'train/loss': 0.04276624694466591, 'train/grad_norm_pre_clip': 1.1645878553390503, 'timing/data': 0.0003633960150182247, 'timing/model': 0.42236730782315135, 'learning_rate/qwen_vl_interface': 9.108512841330288e-06, 'learning_rate/action_model': 9.108512841330287e-05, 'epoch': 0.71})
+[2026-07-28 17:01:52,287][starVLA.training.train_starvla][INFO] - Step 3874, Loss: {'train/loss': 0.03770706430077553, 'train/grad_norm_pre_clip': 1.853800892829895, 'timing/data': 0.0002903030253946781, 'timing/model': 0.4190148808993399, 'learning_rate/qwen_vl_interface': 9.106280038386129e-06, 'learning_rate/action_model': 9.10628003838613e-05, 'epoch': 0.71})
+[2026-07-28 17:01:58,303][starVLA.training.train_starvla][INFO] - Step 3875, Loss: {'train/loss': 0.06643767654895782, 'train/grad_norm_pre_clip': 2.1679093837738037, 'timing/data': 0.0003376118838787079, 'timing/model': 0.4277683119289577, 'learning_rate/qwen_vl_interface': 9.104044733143723e-06, 'learning_rate/action_model': 9.104044733143723e-05, 'epoch': 0.71})
+[2026-07-28 17:02:04,615][starVLA.training.train_starvla][INFO] - Step 3876, Loss: {'train/loss': 0.013540536165237427, 'train/grad_norm_pre_clip': 0.7698279023170471, 'timing/data': 0.0003433600068092346, 'timing/model': 0.4209198812022805, 'learning_rate/qwen_vl_interface': 9.10180692705353e-06, 'learning_rate/action_model': 9.10180692705353e-05, 'epoch': 0.71})
+[2026-07-28 17:02:10,866][starVLA.training.train_starvla][INFO] - Step 3877, Loss: {'train/loss': 0.02794152870774269, 'train/grad_norm_pre_clip': 1.8681113719940186, 'timing/data': 0.00057225301861763, 'timing/model': 0.671653431840241, 'learning_rate/qwen_vl_interface': 9.099566621567643e-06, 'learning_rate/action_model': 9.099566621567642e-05, 'epoch': 0.71})
+[2026-07-28 17:02:16,880][starVLA.training.train_starvla][INFO] - Step 3878, Loss: {'train/loss': 0.07186596095561981, 'train/grad_norm_pre_clip': 1.707045316696167, 'timing/data': 0.00026098359376192093, 'timing/model': 0.4271155153401196, 'learning_rate/qwen_vl_interface': 9.097323818139767e-06, 'learning_rate/action_model': 9.097323818139767e-05, 'epoch': 0.71})
+[2026-07-28 17:02:23,163][starVLA.training.train_starvla][INFO] - Step 3879, Loss: {'train/loss': 0.03903593868017197, 'train/grad_norm_pre_clip': 1.7970691919326782, 'timing/data': 0.00025585200637578964, 'timing/model': 0.42165092332288623, 'learning_rate/qwen_vl_interface': 9.095078518225237e-06, 'learning_rate/action_model': 9.095078518225236e-05, 'epoch': 0.71})
+[2026-07-28 17:02:29,203][starVLA.training.train_starvla][INFO] - Step 3880, Loss: {'train/loss': 0.05518657714128494, 'train/grad_norm_pre_clip': 1.8731167316436768, 'timing/data': 0.0003047441132366657, 'timing/model': 0.4264079499989748, 'learning_rate/qwen_vl_interface': 9.092830723281e-06, 'learning_rate/action_model': 9.092830723281e-05, 'epoch': 0.71})
+[2026-07-28 17:02:35,196][starVLA.training.train_starvla][INFO] - Step 3881, Loss: {'train/loss': 0.018832992762327194, 'train/grad_norm_pre_clip': 1.2883307933807373, 'timing/data': 0.00035143597051501274, 'timing/model': 0.4225799790583551, 'learning_rate/qwen_vl_interface': 9.09058043476563e-06, 'learning_rate/action_model': 9.09058043476563e-05, 'epoch': 0.71})
+[2026-07-28 17:02:41,347][starVLA.training.train_starvla][INFO] - Step 3882, Loss: {'train/loss': 0.03564893454313278, 'train/grad_norm_pre_clip': 1.1483896970748901, 'timing/data': 0.00029561808332800865, 'timing/model': 0.5993734910152853, 'learning_rate/qwen_vl_interface': 9.088327654139311e-06, 'learning_rate/action_model': 9.08832765413931e-05, 'epoch': 0.71})
+[2026-07-28 17:02:47,564][starVLA.training.train_starvla][INFO] - Step 3883, Loss: {'train/loss': 0.026317927986383438, 'train/grad_norm_pre_clip': 0.7358435988426208, 'timing/data': 0.00030112219974398613, 'timing/model': 0.6564964358694851, 'learning_rate/qwen_vl_interface': 9.08607238286385e-06, 'learning_rate/action_model': 9.08607238286385e-05, 'epoch': 0.71})
+[2026-07-28 17:02:53,571][starVLA.training.train_starvla][INFO] - Step 3884, Loss: {'train/loss': 0.059137649834156036, 'train/grad_norm_pre_clip': 1.8226052522659302, 'timing/data': 0.0002974933013319969, 'timing/model': 0.41989061003550887, 'learning_rate/qwen_vl_interface': 9.083814622402667e-06, 'learning_rate/action_model': 9.083814622402666e-05, 'epoch': 0.71})
+[2026-07-28 17:02:59,575][starVLA.training.train_starvla][INFO] - Step 3885, Loss: {'train/loss': 0.03646143898367882, 'train/grad_norm_pre_clip': 1.3158071041107178, 'timing/data': 0.0003201579675078392, 'timing/model': 0.4206769159063697, 'learning_rate/qwen_vl_interface': 9.0815543742208e-06, 'learning_rate/action_model': 9.081554374220799e-05, 'epoch': 0.71})
+[2026-07-28 17:03:05,847][starVLA.training.train_starvla][INFO] - Step 3886, Loss: {'train/loss': 0.075601726770401, 'train/grad_norm_pre_clip': 2.3160579204559326, 'timing/data': 0.0003113248385488987, 'timing/model': 0.4204480308108032, 'learning_rate/qwen_vl_interface': 9.079291639784897e-06, 'learning_rate/action_model': 9.079291639784896e-05, 'epoch': 0.71})
+[2026-07-28 17:03:11,854][starVLA.training.train_starvla][INFO] - Step 3887, Loss: {'train/loss': 0.017720364034175873, 'train/grad_norm_pre_clip': 0.6362945437431335, 'timing/data': 0.00029882509261369705, 'timing/model': 0.41944354167208076, 'learning_rate/qwen_vl_interface': 9.077026420563223e-06, 'learning_rate/action_model': 9.077026420563223e-05, 'epoch': 0.71})
+[2026-07-28 17:03:18,201][starVLA.training.train_starvla][INFO] - Step 3888, Loss: {'train/loss': 0.06480318307876587, 'train/grad_norm_pre_clip': 2.3734359741210938, 'timing/data': 0.000259897205978632, 'timing/model': 0.42209825199097395, 'learning_rate/qwen_vl_interface': 9.074758718025652e-06, 'learning_rate/action_model': 9.074758718025652e-05, 'epoch': 0.71})
+[2026-07-28 17:03:24,491][starVLA.training.train_starvla][INFO] - Step 3889, Loss: {'train/loss': 0.03576991707086563, 'train/grad_norm_pre_clip': 0.764013946056366, 'timing/data': 0.0003015729598701, 'timing/model': 0.6892315931618214, 'learning_rate/qwen_vl_interface': 9.072488533643678e-06, 'learning_rate/action_model': 9.072488533643677e-05, 'epoch': 0.71})
+[2026-07-28 17:03:30,521][starVLA.training.train_starvla][INFO] - Step 3890, Loss: {'train/loss': 0.02980407327413559, 'train/grad_norm_pre_clip': 0.7731711864471436, 'timing/data': 0.0003030211664736271, 'timing/model': 0.4274668088182807, 'learning_rate/qwen_vl_interface': 9.070215868890391e-06, 'learning_rate/action_model': 9.070215868890391e-05, 'epoch': 0.71})
+[2026-07-28 17:03:36,851][starVLA.training.train_starvla][INFO] - Step 3891, Loss: {'train/loss': 0.040321797132492065, 'train/grad_norm_pre_clip': 0.7551571726799011, 'timing/data': 0.00027479929849505424, 'timing/model': 0.43223227886483073, 'learning_rate/qwen_vl_interface': 9.067940725240503e-06, 'learning_rate/action_model': 9.067940725240503e-05, 'epoch': 0.71})
+[2026-07-28 17:03:43,180][starVLA.training.train_starvla][INFO] - Step 3892, Loss: {'train/loss': 0.03731381148099899, 'train/grad_norm_pre_clip': 0.9351431727409363, 'timing/data': 0.0002713180147111416, 'timing/model': 0.4190680612809956, 'learning_rate/qwen_vl_interface': 9.06566310417033e-06, 'learning_rate/action_model': 9.06566310417033e-05, 'epoch': 0.71})
+[2026-07-28 17:03:49,244][starVLA.training.train_starvla][INFO] - Step 3893, Loss: {'train/loss': 0.038656122982501984, 'train/grad_norm_pre_clip': 1.3013805150985718, 'timing/data': 0.0005813529714941978, 'timing/model': 0.4295531981624663, 'learning_rate/qwen_vl_interface': 9.063383007157795e-06, 'learning_rate/action_model': 9.063383007157796e-05, 'epoch': 0.72})
+[2026-07-28 17:03:55,491][starVLA.training.train_starvla][INFO] - Step 3894, Loss: {'train/loss': 0.042591631412506104, 'train/grad_norm_pre_clip': 1.2580512762069702, 'timing/data': 0.0002594650723040104, 'timing/model': 0.4204864539206028, 'learning_rate/qwen_vl_interface': 9.061100435682429e-06, 'learning_rate/action_model': 9.061100435682429e-05, 'epoch': 0.72})
+[2026-07-28 17:04:01,491][starVLA.training.train_starvla][INFO] - Step 3895, Loss: {'train/loss': 0.050832100212574005, 'train/grad_norm_pre_clip': 1.3481491804122925, 'timing/data': 0.0004600388929247856, 'timing/model': 0.4214691831730306, 'learning_rate/qwen_vl_interface': 9.058815391225363e-06, 'learning_rate/action_model': 9.058815391225362e-05, 'epoch': 0.72})
+[2026-07-28 17:04:07,586][starVLA.training.train_starvla][INFO] - Step 3896, Loss: {'train/loss': 0.03010905161499977, 'train/grad_norm_pre_clip': 0.7104496955871582, 'timing/data': 0.00029887259006500244, 'timing/model': 0.43223102390766144, 'learning_rate/qwen_vl_interface': 9.056527875269344e-06, 'learning_rate/action_model': 9.056527875269344e-05, 'epoch': 0.72})
+[2026-07-28 17:04:13,830][starVLA.training.train_starvla][INFO] - Step 3897, Loss: {'train/loss': 0.040342241525650024, 'train/grad_norm_pre_clip': 1.0284323692321777, 'timing/data': 0.00030515994876623154, 'timing/model': 0.6598193538375199, 'learning_rate/qwen_vl_interface': 9.054237889298714e-06, 'learning_rate/action_model': 9.054237889298714e-05, 'epoch': 0.72})
+[2026-07-28 17:04:19,918][starVLA.training.train_starvla][INFO] - Step 3898, Loss: {'train/loss': 0.04412326216697693, 'train/grad_norm_pre_clip': 2.083483934402466, 'timing/data': 0.00024938397109508514, 'timing/model': 0.42890859628096223, 'learning_rate/qwen_vl_interface': 9.05194543479942e-06, 'learning_rate/action_model': 9.05194543479942e-05, 'epoch': 0.72})
+[2026-07-28 17:04:26,225][starVLA.training.train_starvla][INFO] - Step 3899, Loss: {'train/loss': 0.017838623374700546, 'train/grad_norm_pre_clip': 0.8505579233169556, 'timing/data': 0.00031530437991023064, 'timing/model': 0.6908175898715854, 'learning_rate/qwen_vl_interface': 9.049650513259007e-06, 'learning_rate/action_model': 9.049650513259008e-05, 'epoch': 0.72})
+[2026-07-28 17:04:32,416][starVLA.training.train_starvla][INFO] - Step 3900, Loss: {'train/loss': 0.05389873683452606, 'train/grad_norm_pre_clip': 1.1956868171691895, 'timing/data': 0.0003358903340995312, 'timing/model': 0.42291163094341755, 'learning_rate/qwen_vl_interface': 9.047353126166632e-06, 'learning_rate/action_model': 9.047353126166632e-05, 'epoch': 0.72})
+[2026-07-28 17:04:38,484][starVLA.training.train_starvla][INFO] - Step 3901, Loss: {'train/loss': 0.021523617208003998, 'train/grad_norm_pre_clip': 0.6985459923744202, 'timing/data': 0.0002941410057246685, 'timing/model': 0.42417104495689273, 'learning_rate/qwen_vl_interface': 9.045053275013038e-06, 'learning_rate/action_model': 9.045053275013038e-05, 'epoch': 0.72})
+[2026-07-28 17:04:44,678][starVLA.training.train_starvla][INFO] - Step 3902, Loss: {'train/loss': 0.0505431666970253, 'train/grad_norm_pre_clip': 0.8816497921943665, 'timing/data': 0.0002621249295771122, 'timing/model': 0.4209332326427102, 'learning_rate/qwen_vl_interface': 9.042750961290578e-06, 'learning_rate/action_model': 9.042750961290578e-05, 'epoch': 0.72})
+[2026-07-28 17:04:50,671][starVLA.training.train_starvla][INFO] - Step 3903, Loss: {'train/loss': 0.023968970403075218, 'train/grad_norm_pre_clip': 1.052856206893921, 'timing/data': 0.000303611159324646, 'timing/model': 0.4214542298577726, 'learning_rate/qwen_vl_interface': 9.040446186493195e-06, 'learning_rate/action_model': 9.040446186493194e-05, 'epoch': 0.72})
+[2026-07-28 17:04:56,968][starVLA.training.train_starvla][INFO] - Step 3904, Loss: {'train/loss': 0.044145725667476654, 'train/grad_norm_pre_clip': 1.0175789594650269, 'timing/data': 0.00032064924016594887, 'timing/model': 0.43759520910680294, 'learning_rate/qwen_vl_interface': 9.038138952116435e-06, 'learning_rate/action_model': 9.038138952116435e-05, 'epoch': 0.72})
+[2026-07-28 17:05:02,999][starVLA.training.train_starvla][INFO] - Step 3905, Loss: {'train/loss': 0.03398418799042702, 'train/grad_norm_pre_clip': 0.8946465849876404, 'timing/data': 0.00029099779203534126, 'timing/model': 0.42570748925209045, 'learning_rate/qwen_vl_interface': 9.035829259657437e-06, 'learning_rate/action_model': 9.035829259657436e-05, 'epoch': 0.72})
+[2026-07-28 17:05:09,021][starVLA.training.train_starvla][INFO] - Step 3906, Loss: {'train/loss': 0.031569577753543854, 'train/grad_norm_pre_clip': 1.9218997955322266, 'timing/data': 0.00033929385244846344, 'timing/model': 0.4263147758319974, 'learning_rate/qwen_vl_interface': 9.033517110614933e-06, 'learning_rate/action_model': 9.033517110614933e-05, 'epoch': 0.72})
+[2026-07-28 17:05:15,283][starVLA.training.train_starvla][INFO] - Step 3907, Loss: {'train/loss': 0.048580676317214966, 'train/grad_norm_pre_clip': 5.318053722381592, 'timing/data': 0.00025280285626649857, 'timing/model': 0.4209336661733687, 'learning_rate/qwen_vl_interface': 9.031202506489254e-06, 'learning_rate/action_model': 9.031202506489253e-05, 'epoch': 0.72})
+[2026-07-28 17:05:21,573][starVLA.training.train_starvla][INFO] - Step 3908, Loss: {'train/loss': 0.05604717880487442, 'train/grad_norm_pre_clip': 4.6145477294921875, 'timing/data': 0.0002704691141843796, 'timing/model': 0.6644825097173452, 'learning_rate/qwen_vl_interface': 9.02888544878232e-06, 'learning_rate/action_model': 9.028885448782319e-05, 'epoch': 0.72})
+[2026-07-28 17:05:27,881][starVLA.training.train_starvla][INFO] - Step 3909, Loss: {'train/loss': 0.03863042593002319, 'train/grad_norm_pre_clip': 1.2224940061569214, 'timing/data': 0.0003737206570804119, 'timing/model': 0.4199390127323568, 'learning_rate/qwen_vl_interface': 9.026565938997644e-06, 'learning_rate/action_model': 9.026565938997645e-05, 'epoch': 0.72})
+[2026-07-28 17:05:34,155][starVLA.training.train_starvla][INFO] - Step 3910, Loss: {'train/loss': 0.022367099300026894, 'train/grad_norm_pre_clip': 0.7723922729492188, 'timing/data': 0.00031745992600917816, 'timing/model': 0.4559338609687984, 'learning_rate/qwen_vl_interface': 9.024243978640334e-06, 'learning_rate/action_model': 9.024243978640333e-05, 'epoch': 0.72})
+[2026-07-28 17:05:40,397][starVLA.training.train_starvla][INFO] - Step 3911, Loss: {'train/loss': 0.029356257990002632, 'train/grad_norm_pre_clip': 1.0692490339279175, 'timing/data': 0.00037055695429444313, 'timing/model': 0.4194856369867921, 'learning_rate/qwen_vl_interface': 9.02191956921708e-06, 'learning_rate/action_model': 9.02191956921708e-05, 'epoch': 0.72})
+[2026-07-28 17:05:46,711][starVLA.training.train_starvla][INFO] - Step 3912, Loss: {'train/loss': 0.0335640087723732, 'train/grad_norm_pre_clip': 0.8110020756721497, 'timing/data': 0.0002641156315803528, 'timing/model': 0.4197619743645191, 'learning_rate/qwen_vl_interface': 9.019592712236172e-06, 'learning_rate/action_model': 9.019592712236171e-05, 'epoch': 0.72})
+[2026-07-28 17:05:52,761][starVLA.training.train_starvla][INFO] - Step 3913, Loss: {'train/loss': 0.05214840546250343, 'train/grad_norm_pre_clip': 1.9203100204467773, 'timing/data': 0.00034651393070816994, 'timing/model': 0.42148197535425425, 'learning_rate/qwen_vl_interface': 9.017263409207478e-06, 'learning_rate/action_model': 9.017263409207478e-05, 'epoch': 0.72})
+[2026-07-28 17:05:58,876][starVLA.training.train_starvla][INFO] - Step 3914, Loss: {'train/loss': 0.02681731805205345, 'train/grad_norm_pre_clip': 0.6749088168144226, 'timing/data': 0.00028902385383844376, 'timing/model': 0.421297300606966, 'learning_rate/qwen_vl_interface': 9.014931661642455e-06, 'learning_rate/action_model': 9.014931661642455e-05, 'epoch': 0.72})
+[2026-07-28 17:06:05,180][starVLA.training.train_starvla][INFO] - Step 3915, Loss: {'train/loss': 0.02854527346789837, 'train/grad_norm_pre_clip': 0.8720462322235107, 'timing/data': 0.00026082899421453476, 'timing/model': 0.4356626239605248, 'learning_rate/qwen_vl_interface': 9.012597471054154e-06, 'learning_rate/action_model': 9.012597471054154e-05, 'epoch': 0.72})
+[2026-07-28 17:06:11,428][starVLA.training.train_starvla][INFO] - Step 3916, Loss: {'train/loss': 0.022508081048727036, 'train/grad_norm_pre_clip': 0.8192386627197266, 'timing/data': 0.0002688150852918625, 'timing/model': 0.42066678311675787, 'learning_rate/qwen_vl_interface': 9.010260838957206e-06, 'learning_rate/action_model': 9.010260838957205e-05, 'epoch': 0.72})
+[2026-07-28 17:06:17,411][starVLA.training.train_starvla][INFO] - Step 3917, Loss: {'train/loss': 0.04133941978216171, 'train/grad_norm_pre_clip': 0.7668497562408447, 'timing/data': 0.0002940199337899685, 'timing/model': 0.41970649315044284, 'learning_rate/qwen_vl_interface': 9.007921766867823e-06, 'learning_rate/action_model': 9.007921766867822e-05, 'epoch': 0.72})
+[2026-07-28 17:06:23,155][starVLA.training.train_starvla][INFO] - Step 3918, Loss: {'train/loss': 0.02290041744709015, 'train/grad_norm_pre_clip': 0.5800212621688843, 'timing/data': 0.000306163914501667, 'timing/model': 0.4207391981035471, 'learning_rate/qwen_vl_interface': 9.005580256303804e-06, 'learning_rate/action_model': 9.005580256303803e-05, 'epoch': 0.72})
+[2026-07-28 17:06:29,379][starVLA.training.train_starvla][INFO] - Step 3919, Loss: {'train/loss': 0.018524233251810074, 'train/grad_norm_pre_clip': 0.505055844783783, 'timing/data': 0.0003842739388346672, 'timing/model': 0.6613249517977238, 'learning_rate/qwen_vl_interface': 9.00323630878453e-06, 'learning_rate/action_model': 9.00323630878453e-05, 'epoch': 0.72})
+[2026-07-28 17:06:35,615][starVLA.training.train_starvla][INFO] - Step 3920, Loss: {'train/loss': 0.03541172295808792, 'train/grad_norm_pre_clip': 1.1515707969665527, 'timing/data': 0.00043720565736293793, 'timing/model': 0.6632485860027373, 'learning_rate/qwen_vl_interface': 9.000889925830964e-06, 'learning_rate/action_model': 9.000889925830965e-05, 'epoch': 0.72})
+[2026-07-28 17:06:41,589][starVLA.training.train_starvla][INFO] - Step 3921, Loss: {'train/loss': 0.0662769004702568, 'train/grad_norm_pre_clip': 3.307640790939331, 'timing/data': 0.00025390321388840675, 'timing/model': 0.42097967816516757, 'learning_rate/qwen_vl_interface': 8.998541108965648e-06, 'learning_rate/action_model': 8.998541108965648e-05, 'epoch': 0.72})
+[2026-07-28 17:06:47,572][starVLA.training.train_starvla][INFO] - Step 3922, Loss: {'train/loss': 0.02158265933394432, 'train/grad_norm_pre_clip': 1.6202526092529297, 'timing/data': 0.0007478031329810619, 'timing/model': 0.42197923781350255, 'learning_rate/qwen_vl_interface': 8.996189859712704e-06, 'learning_rate/action_model': 8.996189859712703e-05, 'epoch': 0.72})
+[2026-07-28 17:06:53,543][starVLA.training.train_starvla][INFO] - Step 3923, Loss: {'train/loss': 0.03490537032485008, 'train/grad_norm_pre_clip': 2.182985305786133, 'timing/data': 0.000294999685138464, 'timing/model': 0.42067847587168217, 'learning_rate/qwen_vl_interface': 8.993836179597833e-06, 'learning_rate/action_model': 8.993836179597833e-05, 'epoch': 0.72})
+[2026-07-28 17:06:59,818][starVLA.training.train_starvla][INFO] - Step 3924, Loss: {'train/loss': 0.049127183854579926, 'train/grad_norm_pre_clip': 2.0970237255096436, 'timing/data': 0.0004059639759361744, 'timing/model': 0.7112911683507264, 'learning_rate/qwen_vl_interface': 8.99148007014831e-06, 'learning_rate/action_model': 8.99148007014831e-05, 'epoch': 0.72})
+[2026-07-28 17:07:05,806][starVLA.training.train_starvla][INFO] - Step 3925, Loss: {'train/loss': 0.027745068073272705, 'train/grad_norm_pre_clip': 1.3743585348129272, 'timing/data': 0.0002593020908534527, 'timing/model': 0.41993721295148134, 'learning_rate/qwen_vl_interface': 8.989121532892989e-06, 'learning_rate/action_model': 8.989121532892988e-05, 'epoch': 0.72})
+[2026-07-28 17:07:11,832][starVLA.training.train_starvla][INFO] - Step 3926, Loss: {'train/loss': 0.009687880054116249, 'train/grad_norm_pre_clip': 0.4701337516307831, 'timing/data': 0.0002953018993139267, 'timing/model': 0.42139603570103645, 'learning_rate/qwen_vl_interface': 8.986760569362302e-06, 'learning_rate/action_model': 8.986760569362301e-05, 'epoch': 0.72})
+[2026-07-28 17:07:17,813][starVLA.training.train_starvla][INFO] - Step 3927, Loss: {'train/loss': 0.03556507080793381, 'train/grad_norm_pre_clip': 2.875155210494995, 'timing/data': 0.000295471865683794, 'timing/model': 0.4204422878101468, 'learning_rate/qwen_vl_interface': 8.984397181088251e-06, 'learning_rate/action_model': 8.984397181088251e-05, 'epoch': 0.72})
+[2026-07-28 17:07:24,139][starVLA.training.train_starvla][INFO] - Step 3928, Loss: {'train/loss': 0.03790879249572754, 'train/grad_norm_pre_clip': 4.677823066711426, 'timing/data': 0.0002589072100818157, 'timing/model': 0.44004935398697853, 'learning_rate/qwen_vl_interface': 8.982031369604412e-06, 'learning_rate/action_model': 8.982031369604411e-05, 'epoch': 0.72})
+[2026-07-28 17:07:30,179][starVLA.training.train_starvla][INFO] - Step 3929, Loss: {'train/loss': 0.023799799382686615, 'train/grad_norm_pre_clip': 1.1767642498016357, 'timing/data': 0.00034309783950448036, 'timing/model': 0.42109307600185275, 'learning_rate/qwen_vl_interface': 8.979663136445936e-06, 'learning_rate/action_model': 8.979663136445935e-05, 'epoch': 0.72})
+[2026-07-28 17:07:36,386][starVLA.training.train_starvla][INFO] - Step 3930, Loss: {'train/loss': 0.024566613137722015, 'train/grad_norm_pre_clip': 0.896710216999054, 'timing/data': 0.0002926541492342949, 'timing/model': 0.6554974452592432, 'learning_rate/qwen_vl_interface': 8.977292483149542e-06, 'learning_rate/action_model': 8.977292483149542e-05, 'epoch': 0.72})
+[2026-07-28 17:07:42,379][starVLA.training.train_starvla][INFO] - Step 3931, Loss: {'train/loss': 0.031746841967105865, 'train/grad_norm_pre_clip': 0.9398274421691895, 'timing/data': 0.0003318311646580696, 'timing/model': 0.42057296819984913, 'learning_rate/qwen_vl_interface': 8.974919411253523e-06, 'learning_rate/action_model': 8.974919411253523e-05, 'epoch': 0.72})
+[2026-07-28 17:07:48,622][starVLA.training.train_starvla][INFO] - Step 3932, Loss: {'train/loss': 0.017178673297166824, 'train/grad_norm_pre_clip': 1.1450276374816895, 'timing/data': 0.0003028218634426594, 'timing/model': 0.6623709369450808, 'learning_rate/qwen_vl_interface': 8.972543922297739e-06, 'learning_rate/action_model': 8.972543922297739e-05, 'epoch': 0.72})
+[2026-07-28 17:07:54,615][starVLA.training.train_starvla][INFO] - Step 3933, Loss: {'train/loss': 0.03686048835515976, 'train/grad_norm_pre_clip': 1.2400153875350952, 'timing/data': 0.0002642981708049774, 'timing/model': 0.42059040209278464, 'learning_rate/qwen_vl_interface': 8.970166017823621e-06, 'learning_rate/action_model': 8.97016601782362e-05, 'epoch': 0.72})
+[2026-07-28 17:08:00,670][starVLA.training.train_starvla][INFO] - Step 3934, Loss: {'train/loss': 0.033330436795949936, 'train/grad_norm_pre_clip': 1.0965538024902344, 'timing/data': 0.0002757762558758259, 'timing/model': 0.4216881599277258, 'learning_rate/qwen_vl_interface': 8.96778569937416e-06, 'learning_rate/action_model': 8.96778569937416e-05, 'epoch': 0.72})
+[2026-07-28 17:08:07,019][starVLA.training.train_starvla][INFO] - Step 3935, Loss: {'train/loss': 0.04785195365548134, 'train/grad_norm_pre_clip': 1.6259101629257202, 'timing/data': 0.00032445183023810387, 'timing/model': 0.41868396988138556, 'learning_rate/qwen_vl_interface': 8.965402968493922e-06, 'learning_rate/action_model': 8.965402968493922e-05, 'epoch': 0.72})
+[2026-07-28 17:08:13,233][starVLA.training.train_starvla][INFO] - Step 3936, Loss: {'train/loss': 0.0415976382791996, 'train/grad_norm_pre_clip': 1.4708847999572754, 'timing/data': 0.0004513259045779705, 'timing/model': 0.4496767120435834, 'learning_rate/qwen_vl_interface': 8.96301782672904e-06, 'learning_rate/action_model': 8.963017826729038e-05, 'epoch': 0.72})
+[2026-07-28 17:08:19,387][starVLA.training.train_starvla][INFO] - Step 3937, Loss: {'train/loss': 0.024624492973089218, 'train/grad_norm_pre_clip': 0.6522114872932434, 'timing/data': 0.00033190008252859116, 'timing/model': 0.43773410096764565, 'learning_rate/qwen_vl_interface': 8.960630275627197e-06, 'learning_rate/action_model': 8.960630275627197e-05, 'epoch': 0.72})
+[2026-07-28 17:08:25,628][starVLA.training.train_starvla][INFO] - Step 3938, Loss: {'train/loss': 0.04581082612276077, 'train/grad_norm_pre_clip': 1.776613473892212, 'timing/data': 0.0002684551291167736, 'timing/model': 0.42997128795832396, 'learning_rate/qwen_vl_interface': 8.958240316737655e-06, 'learning_rate/action_model': 8.958240316737656e-05, 'epoch': 0.72})
+[2026-07-28 17:08:31,892][starVLA.training.train_starvla][INFO] - Step 3939, Loss: {'train/loss': 0.056888435035943985, 'train/grad_norm_pre_clip': 1.8685413599014282, 'timing/data': 0.0002508261241018772, 'timing/model': 0.4200494489632547, 'learning_rate/qwen_vl_interface': 8.955847951611233e-06, 'learning_rate/action_model': 8.955847951611232e-05, 'epoch': 0.72})
+[2026-07-28 17:08:37,942][starVLA.training.train_starvla][INFO] - Step 3940, Loss: {'train/loss': 0.06467816978693008, 'train/grad_norm_pre_clip': 1.7828478813171387, 'timing/data': 0.0003064167685806751, 'timing/model': 0.4368619229644537, 'learning_rate/qwen_vl_interface': 8.953453181800307e-06, 'learning_rate/action_model': 8.953453181800306e-05, 'epoch': 0.72})
+[2026-07-28 17:08:43,984][starVLA.training.train_starvla][INFO] - Step 3941, Loss: {'train/loss': 0.03527968376874924, 'train/grad_norm_pre_clip': 1.401137351989746, 'timing/data': 0.0003778468817472458, 'timing/model': 0.4334614919498563, 'learning_rate/qwen_vl_interface': 8.951056008858819e-06, 'learning_rate/action_model': 8.95105600885882e-05, 'epoch': 0.72})
+[2026-07-28 17:08:50,207][starVLA.training.train_starvla][INFO] - Step 3942, Loss: {'train/loss': 0.034540675580501556, 'train/grad_norm_pre_clip': 0.9880772829055786, 'timing/data': 0.0002815118059515953, 'timing/model': 0.4286427260376513, 'learning_rate/qwen_vl_interface': 8.94865643434227e-06, 'learning_rate/action_model': 8.948656434342269e-05, 'epoch': 0.72})
+[2026-07-28 17:08:56,191][starVLA.training.train_starvla][INFO] - Step 3943, Loss: {'train/loss': 0.03913937509059906, 'train/grad_norm_pre_clip': 1.2143994569778442, 'timing/data': 0.0003143046051263809, 'timing/model': 0.41844332637265325, 'learning_rate/qwen_vl_interface': 8.946254459807717e-06, 'learning_rate/action_model': 8.946254459807717e-05, 'epoch': 0.72})
+[2026-07-28 17:09:02,434][starVLA.training.train_starvla][INFO] - Step 3944, Loss: {'train/loss': 0.041856564581394196, 'train/grad_norm_pre_clip': 1.595380187034607, 'timing/data': 0.000289207324385643, 'timing/model': 0.6522307130508125, 'learning_rate/qwen_vl_interface': 8.943850086813773e-06, 'learning_rate/action_model': 8.943850086813773e-05, 'epoch': 0.72})
+[2026-07-28 17:09:08,726][starVLA.training.train_starvla][INFO] - Step 3945, Loss: {'train/loss': 0.05923088639974594, 'train/grad_norm_pre_clip': 1.5095328092575073, 'timing/data': 0.0003559477627277374, 'timing/model': 0.42085855500772595, 'learning_rate/qwen_vl_interface': 8.941443316920613e-06, 'learning_rate/action_model': 8.941443316920612e-05, 'epoch': 0.72})
+[2026-07-28 17:09:14,798][starVLA.training.train_starvla][INFO] - Step 3946, Loss: {'train/loss': 0.02582605928182602, 'train/grad_norm_pre_clip': 1.2213658094406128, 'timing/data': 0.0003710431046783924, 'timing/model': 0.4195024841465056, 'learning_rate/qwen_vl_interface': 8.93903415168996e-06, 'learning_rate/action_model': 8.93903415168996e-05, 'epoch': 0.72})
+[2026-07-28 17:09:21,071][starVLA.training.train_starvla][INFO] - Step 3947, Loss: {'train/loss': 0.03408582881093025, 'train/grad_norm_pre_clip': 0.641594409942627, 'timing/data': 0.0002811579033732414, 'timing/model': 0.4188007772900164, 'learning_rate/qwen_vl_interface': 8.9366225926851e-06, 'learning_rate/action_model': 8.936622592685098e-05, 'epoch': 0.72})
+[2026-07-28 17:09:27,340][starVLA.training.train_starvla][INFO] - Step 3948, Loss: {'train/loss': 0.0643511712551117, 'train/grad_norm_pre_clip': 0.9771072864532471, 'timing/data': 0.00029755523428320885, 'timing/model': 0.6010138657875359, 'learning_rate/qwen_vl_interface': 8.934208641470863e-06, 'learning_rate/action_model': 8.934208641470863e-05, 'epoch': 0.73})
+[2026-07-28 17:09:33,353][starVLA.training.train_starvla][INFO] - Step 3949, Loss: {'train/loss': 0.04580148681998253, 'train/grad_norm_pre_clip': 0.7189145684242249, 'timing/data': 0.0005363309755921364, 'timing/model': 0.4211134589277208, 'learning_rate/qwen_vl_interface': 8.931792299613637e-06, 'learning_rate/action_model': 8.931792299613636e-05, 'epoch': 0.73})
+[2026-07-28 17:09:39,421][starVLA.training.train_starvla][INFO] - Step 3950, Loss: {'train/loss': 0.03139315918087959, 'train/grad_norm_pre_clip': 1.5598363876342773, 'timing/data': 0.0002839770168066025, 'timing/model': 0.4248879458755255, 'learning_rate/qwen_vl_interface': 8.92937356868136e-06, 'learning_rate/action_model': 8.92937356868136e-05, 'epoch': 0.73})
+[2026-07-28 17:09:45,667][starVLA.training.train_starvla][INFO] - Step 3951, Loss: {'train/loss': 0.03032299503684044, 'train/grad_norm_pre_clip': 0.7857069969177246, 'timing/data': 0.0002572750672698021, 'timing/model': 0.42678758688271046, 'learning_rate/qwen_vl_interface': 8.926952450243522e-06, 'learning_rate/action_model': 8.926952450243522e-05, 'epoch': 0.73})
+[2026-07-28 17:09:51,939][starVLA.training.train_starvla][INFO] - Step 3952, Loss: {'train/loss': 0.02338501438498497, 'train/grad_norm_pre_clip': 0.8102083802223206, 'timing/data': 0.00034697819501161575, 'timing/model': 0.43084075720980763, 'learning_rate/qwen_vl_interface': 8.924528945871157e-06, 'learning_rate/action_model': 8.924528945871156e-05, 'epoch': 0.73})
+[2026-07-28 17:09:58,042][starVLA.training.train_starvla][INFO] - Step 3953, Loss: {'train/loss': 0.04215165972709656, 'train/grad_norm_pre_clip': 1.117065191268921, 'timing/data': 0.0003033778630197048, 'timing/model': 0.42700980231165886, 'learning_rate/qwen_vl_interface': 8.922103057136857e-06, 'learning_rate/action_model': 8.922103057136856e-05, 'epoch': 0.73})
+[2026-07-28 17:10:04,051][starVLA.training.train_starvla][INFO] - Step 3954, Loss: {'train/loss': 0.021800735965371132, 'train/grad_norm_pre_clip': 0.6581146121025085, 'timing/data': 0.00032764021307229996, 'timing/model': 0.4204237381927669, 'learning_rate/qwen_vl_interface': 8.919674785614747e-06, 'learning_rate/action_model': 8.919674785614747e-05, 'epoch': 0.73})
+[2026-07-28 17:10:10,334][starVLA.training.train_starvla][INFO] - Step 3955, Loss: {'train/loss': 0.05203362554311752, 'train/grad_norm_pre_clip': 1.8395756483078003, 'timing/data': 0.00030186912044882774, 'timing/model': 0.6681582271121442, 'learning_rate/qwen_vl_interface': 8.91724413288051e-06, 'learning_rate/action_model': 8.91724413288051e-05, 'epoch': 0.73})
+[2026-07-28 17:10:16,540][starVLA.training.train_starvla][INFO] - Step 3956, Loss: {'train/loss': 0.04060344770550728, 'train/grad_norm_pre_clip': 2.504770517349243, 'timing/data': 0.0002549756318330765, 'timing/model': 0.42051053885370493, 'learning_rate/qwen_vl_interface': 8.91481110051137e-06, 'learning_rate/action_model': 8.91481110051137e-05, 'epoch': 0.73})
+[2026-07-28 17:10:22,539][starVLA.training.train_starvla][INFO] - Step 3957, Loss: {'train/loss': 0.018195906654000282, 'train/grad_norm_pre_clip': 1.5949662923812866, 'timing/data': 0.00035079289227724075, 'timing/model': 0.41946163587272167, 'learning_rate/qwen_vl_interface': 8.912375690086097e-06, 'learning_rate/action_model': 8.912375690086096e-05, 'epoch': 0.73})
+[2026-07-28 17:10:28,538][starVLA.training.train_starvla][INFO] - Step 3958, Loss: {'train/loss': 0.03466120362281799, 'train/grad_norm_pre_clip': 1.184141993522644, 'timing/data': 0.0002676309086382389, 'timing/model': 0.4205387579277158, 'learning_rate/qwen_vl_interface': 8.909937903185e-06, 'learning_rate/action_model': 8.909937903185e-05, 'epoch': 0.73})
+[2026-07-28 17:10:34,786][starVLA.training.train_starvla][INFO] - Step 3959, Loss: {'train/loss': 0.020295750349760056, 'train/grad_norm_pre_clip': 0.7327247858047485, 'timing/data': 0.0003564748913049698, 'timing/model': 0.41957403998821974, 'learning_rate/qwen_vl_interface': 8.907497741389932e-06, 'learning_rate/action_model': 8.907497741389932e-05, 'epoch': 0.73})
+[2026-07-28 17:10:40,805][starVLA.training.train_starvla][INFO] - Step 3960, Loss: {'train/loss': 0.028819140046834946, 'train/grad_norm_pre_clip': 0.8568570613861084, 'timing/data': 0.00032085785642266273, 'timing/model': 0.4378986689262092, 'learning_rate/qwen_vl_interface': 8.905055206284289e-06, 'learning_rate/action_model': 8.905055206284288e-05, 'epoch': 0.73})
+[2026-07-28 17:10:47,017][starVLA.training.train_starvla][INFO] - Step 3961, Loss: {'train/loss': 0.03886912763118744, 'train/grad_norm_pre_clip': 1.3703621625900269, 'timing/data': 0.0002820957452058792, 'timing/model': 0.6627175426110625, 'learning_rate/qwen_vl_interface': 8.902610299453005e-06, 'learning_rate/action_model': 8.902610299453004e-05, 'epoch': 0.73})
+[2026-07-28 17:10:52,982][starVLA.training.train_starvla][INFO] - Step 3962, Loss: {'train/loss': 0.04188421368598938, 'train/grad_norm_pre_clip': 0.9499300718307495, 'timing/data': 0.00038370396941900253, 'timing/model': 0.4192388248629868, 'learning_rate/qwen_vl_interface': 8.900163022482554e-06, 'learning_rate/action_model': 8.900163022482553e-05, 'epoch': 0.73})
+[2026-07-28 17:10:59,269][starVLA.training.train_starvla][INFO] - Step 3963, Loss: {'train/loss': 0.01774299517273903, 'train/grad_norm_pre_clip': 0.8924928903579712, 'timing/data': 0.00026434101164340973, 'timing/model': 0.4199054082855582, 'learning_rate/qwen_vl_interface': 8.897713376960946e-06, 'learning_rate/action_model': 8.897713376960946e-05, 'epoch': 0.73})
+[2026-07-28 17:11:05,266][starVLA.training.train_starvla][INFO] - Step 3964, Loss: {'train/loss': 0.03460141643881798, 'train/grad_norm_pre_clip': 1.0185472965240479, 'timing/data': 0.0003051520325243473, 'timing/model': 0.42083931900560856, 'learning_rate/qwen_vl_interface': 8.895261364477731e-06, 'learning_rate/action_model': 8.895261364477731e-05, 'epoch': 0.73})
+[2026-07-28 17:11:11,233][starVLA.training.train_starvla][INFO] - Step 3965, Loss: {'train/loss': 0.029931768774986267, 'train/grad_norm_pre_clip': 1.5432652235031128, 'timing/data': 0.0002874149940907955, 'timing/model': 0.4195455079898238, 'learning_rate/qwen_vl_interface': 8.892806986623993e-06, 'learning_rate/action_model': 8.892806986623992e-05, 'epoch': 0.73})
+[2026-07-28 17:11:17,704][starVLA.training.train_starvla][INFO] - Step 3966, Loss: {'train/loss': 0.05125953629612923, 'train/grad_norm_pre_clip': 1.3603553771972656, 'timing/data': 0.00024802563712000847, 'timing/model': 0.6530294045805931, 'learning_rate/qwen_vl_interface': 8.890350244992349e-06, 'learning_rate/action_model': 8.890350244992348e-05, 'epoch': 0.73})
+[2026-07-28 17:11:23,665][starVLA.training.train_starvla][INFO] - Step 3967, Loss: {'train/loss': 0.029431696981191635, 'train/grad_norm_pre_clip': 1.0584337711334229, 'timing/data': 0.00027144188061356544, 'timing/model': 0.41994104627519846, 'learning_rate/qwen_vl_interface': 8.887891141176954e-06, 'learning_rate/action_model': 8.887891141176954e-05, 'epoch': 0.73})
+[2026-07-28 17:11:29,789][starVLA.training.train_starvla][INFO] - Step 3968, Loss: {'train/loss': 0.06367293000221252, 'train/grad_norm_pre_clip': 2.2441723346710205, 'timing/data': 0.00032054074108600616, 'timing/model': 0.4191752769984305, 'learning_rate/qwen_vl_interface': 8.885429676773495e-06, 'learning_rate/action_model': 8.885429676773495e-05, 'epoch': 0.73})
+[2026-07-28 17:11:36,078][starVLA.training.train_starvla][INFO] - Step 3969, Loss: {'train/loss': 0.04474179446697235, 'train/grad_norm_pre_clip': 1.6601834297180176, 'timing/data': 0.0002759220078587532, 'timing/model': 0.41967070708051324, 'learning_rate/qwen_vl_interface': 8.882965853379188e-06, 'learning_rate/action_model': 8.882965853379187e-05, 'epoch': 0.73})
+[2026-07-28 17:11:42,123][starVLA.training.train_starvla][INFO] - Step 3970, Loss: {'train/loss': 0.05491228401660919, 'train/grad_norm_pre_clip': 1.2827394008636475, 'timing/data': 0.0003060358576476574, 'timing/model': 0.42216907488182187, 'learning_rate/qwen_vl_interface': 8.88049967259278e-06, 'learning_rate/action_model': 8.880499672592778e-05, 'epoch': 0.73})
+[2026-07-28 17:11:48,393][starVLA.training.train_starvla][INFO] - Step 3971, Loss: {'train/loss': 0.017528772354125977, 'train/grad_norm_pre_clip': 0.6933481693267822, 'timing/data': 0.00030383095145225525, 'timing/model': 0.6688529662787914, 'learning_rate/qwen_vl_interface': 8.878031136014552e-06, 'learning_rate/action_model': 8.878031136014551e-05, 'epoch': 0.73})
+[2026-07-28 17:11:54,438][starVLA.training.train_starvla][INFO] - Step 3972, Loss: {'train/loss': 0.030595581978559494, 'train/grad_norm_pre_clip': 0.8803893327713013, 'timing/data': 0.0002464042045176029, 'timing/model': 0.4399680900387466, 'learning_rate/qwen_vl_interface': 8.875560245246307e-06, 'learning_rate/action_model': 8.875560245246307e-05, 'epoch': 0.73})
+[2026-07-28 17:12:00,686][starVLA.training.train_starvla][INFO] - Step 3973, Loss: {'train/loss': 0.03755209594964981, 'train/grad_norm_pre_clip': 1.3418574333190918, 'timing/data': 0.0003122701309621334, 'timing/model': 0.6598701938055456, 'learning_rate/qwen_vl_interface': 8.873087001891381e-06, 'learning_rate/action_model': 8.87308700189138e-05, 'epoch': 0.73})
+[2026-07-28 17:12:06,898][starVLA.training.train_starvla][INFO] - Step 3974, Loss: {'train/loss': 0.036211416125297546, 'train/grad_norm_pre_clip': 1.565881609916687, 'timing/data': 0.00025916099548339844, 'timing/model': 0.41995399398729205, 'learning_rate/qwen_vl_interface': 8.870611407554635e-06, 'learning_rate/action_model': 8.870611407554636e-05, 'epoch': 0.73})
+[2026-07-28 17:12:13,078][starVLA.training.train_starvla][INFO] - Step 3975, Loss: {'train/loss': 0.06881938874721527, 'train/grad_norm_pre_clip': 2.2079710960388184, 'timing/data': 0.00028156209737062454, 'timing/model': 0.41999709978699684, 'learning_rate/qwen_vl_interface': 8.868133463842456e-06, 'learning_rate/action_model': 8.868133463842457e-05, 'epoch': 0.73})
+[2026-07-28 17:12:19,244][starVLA.training.train_starvla][INFO] - Step 3976, Loss: {'train/loss': 0.022965647280216217, 'train/grad_norm_pre_clip': 0.822935938835144, 'timing/data': 0.00041608698666095734, 'timing/model': 0.42230026610195637, 'learning_rate/qwen_vl_interface': 8.865653172362755e-06, 'learning_rate/action_model': 8.865653172362754e-05, 'epoch': 0.73})
+[2026-07-28 17:12:25,524][starVLA.training.train_starvla][INFO] - Step 3977, Loss: {'train/loss': 0.03884833678603172, 'train/grad_norm_pre_clip': 1.2440515756607056, 'timing/data': 0.00039357272908091545, 'timing/model': 0.6604791060090065, 'learning_rate/qwen_vl_interface': 8.863170534724962e-06, 'learning_rate/action_model': 8.863170534724961e-05, 'epoch': 0.73})
+[2026-07-28 17:12:31,496][starVLA.training.train_starvla][INFO] - Step 3978, Loss: {'train/loss': 0.03830115497112274, 'train/grad_norm_pre_clip': 0.9434704780578613, 'timing/data': 0.0002633999101817608, 'timing/model': 0.4192925379611552, 'learning_rate/qwen_vl_interface': 8.860685552540038e-06, 'learning_rate/action_model': 8.860685552540037e-05, 'epoch': 0.73})
+[2026-07-28 17:12:37,743][starVLA.training.train_starvla][INFO] - Step 3979, Loss: {'train/loss': 0.01380912121385336, 'train/grad_norm_pre_clip': 0.5056700110435486, 'timing/data': 0.00026348233222961426, 'timing/model': 0.4191527720540762, 'learning_rate/qwen_vl_interface': 8.858198227420461e-06, 'learning_rate/action_model': 8.85819822742046e-05, 'epoch': 0.73})
+[2026-07-28 17:12:43,786][starVLA.training.train_starvla][INFO] - Step 3980, Loss: {'train/loss': 0.038763388991355896, 'train/grad_norm_pre_clip': 0.9516602158546448, 'timing/data': 0.00030715111643075943, 'timing/model': 0.4377331603318453, 'learning_rate/qwen_vl_interface': 8.855708560980225e-06, 'learning_rate/action_model': 8.855708560980225e-05, 'epoch': 0.73})
+[2026-07-28 17:12:49,776][starVLA.training.train_starvla][INFO] - Step 3981, Loss: {'train/loss': 0.021562092006206512, 'train/grad_norm_pre_clip': 1.019474983215332, 'timing/data': 0.0003079203888773918, 'timing/model': 0.42004957515746355, 'learning_rate/qwen_vl_interface': 8.853216554834852e-06, 'learning_rate/action_model': 8.85321655483485e-05, 'epoch': 0.73})
+[2026-07-28 17:12:56,131][starVLA.training.train_starvla][INFO] - Step 3982, Loss: {'train/loss': 0.04001146927475929, 'train/grad_norm_pre_clip': 1.450153112411499, 'timing/data': 0.00025292905047535896, 'timing/model': 0.4194985078647733, 'learning_rate/qwen_vl_interface': 8.850722210601376e-06, 'learning_rate/action_model': 8.850722210601377e-05, 'epoch': 0.73})
+[2026-07-28 17:13:02,446][starVLA.training.train_starvla][INFO] - Step 3983, Loss: {'train/loss': 0.04287870600819588, 'train/grad_norm_pre_clip': 1.335869550704956, 'timing/data': 0.0003056759014725685, 'timing/model': 0.4218286108225584, 'learning_rate/qwen_vl_interface': 8.848225529898351e-06, 'learning_rate/action_model': 8.84822552989835e-05, 'epoch': 0.73})
+[2026-07-28 17:13:08,431][starVLA.training.train_starvla][INFO] - Step 3984, Loss: {'train/loss': 0.028895849362015724, 'train/grad_norm_pre_clip': 0.6551904678344727, 'timing/data': 0.00035560503602027893, 'timing/model': 0.41874172212556005, 'learning_rate/qwen_vl_interface': 8.845726514345845e-06, 'learning_rate/action_model': 8.845726514345843e-05, 'epoch': 0.73})
+[2026-07-28 17:13:14,245][starVLA.training.train_starvla][INFO] - Step 3985, Loss: {'train/loss': 0.03384681046009064, 'train/grad_norm_pre_clip': 1.9428013563156128, 'timing/data': 0.0003185882233083248, 'timing/model': 0.420023525133729, 'learning_rate/qwen_vl_interface': 8.84322516556544e-06, 'learning_rate/action_model': 8.84322516556544e-05, 'epoch': 0.73})
+[2026-07-28 17:13:20,444][starVLA.training.train_starvla][INFO] - Step 3986, Loss: {'train/loss': 0.05736537277698517, 'train/grad_norm_pre_clip': 2.5017967224121094, 'timing/data': 0.0002455776557326317, 'timing/model': 0.42070405511185527, 'learning_rate/qwen_vl_interface': 8.840721485180237e-06, 'learning_rate/action_model': 8.840721485180236e-05, 'epoch': 0.73})
+[2026-07-28 17:13:26,201][starVLA.training.train_starvla][INFO] - Step 3987, Loss: {'train/loss': 0.020471587777137756, 'train/grad_norm_pre_clip': 0.5584606528282166, 'timing/data': 0.0002755140885710716, 'timing/model': 0.42126811435446143, 'learning_rate/qwen_vl_interface': 8.838215474814848e-06, 'learning_rate/action_model': 8.838215474814847e-05, 'epoch': 0.73})
+[2026-07-28 17:13:31,957][starVLA.training.train_starvla][INFO] - Step 3988, Loss: {'train/loss': 0.03715964034199715, 'train/grad_norm_pre_clip': 0.9960744976997375, 'timing/data': 0.0003553573042154312, 'timing/model': 0.4197356211952865, 'learning_rate/qwen_vl_interface': 8.835707136095392e-06, 'learning_rate/action_model': 8.835707136095391e-05, 'epoch': 0.73})
+[2026-07-28 17:13:38,153][starVLA.training.train_starvla][INFO] - Step 3989, Loss: {'train/loss': 0.039383526891469955, 'train/grad_norm_pre_clip': 1.6483100652694702, 'timing/data': 0.00027759792283177376, 'timing/model': 0.44884435273706913, 'learning_rate/qwen_vl_interface': 8.833196470649505e-06, 'learning_rate/action_model': 8.833196470649505e-05, 'epoch': 0.73})
+[2026-07-28 17:13:44,366][starVLA.training.train_starvla][INFO] - Step 3990, Loss: {'train/loss': 0.059521373361349106, 'train/grad_norm_pre_clip': 1.7468854188919067, 'timing/data': 0.0003862217999994755, 'timing/model': 0.4201759430579841, 'learning_rate/qwen_vl_interface': 8.83068348010633e-06, 'learning_rate/action_model': 8.83068348010633e-05, 'epoch': 0.73})
+[2026-07-28 17:13:50,732][starVLA.training.train_starvla][INFO] - Step 3991, Loss: {'train/loss': 0.05891487002372742, 'train/grad_norm_pre_clip': 1.631344199180603, 'timing/data': 0.00030771782621741295, 'timing/model': 0.6592009011656046, 'learning_rate/qwen_vl_interface': 8.82816816609652e-06, 'learning_rate/action_model': 8.828168166096521e-05, 'epoch': 0.73})
+[2026-07-28 17:13:56,748][starVLA.training.train_starvla][INFO] - Step 3992, Loss: {'train/loss': 0.02391006052494049, 'train/grad_norm_pre_clip': 0.7627032995223999, 'timing/data': 0.0003545517101883888, 'timing/model': 0.4205976598896086, 'learning_rate/qwen_vl_interface': 8.825650530252238e-06, 'learning_rate/action_model': 8.825650530252238e-05, 'epoch': 0.73})
+[2026-07-28 17:14:02,763][starVLA.training.train_starvla][INFO] - Step 3993, Loss: {'train/loss': 0.03586945682764053, 'train/grad_norm_pre_clip': 1.2461087703704834, 'timing/data': 0.00029750680550932884, 'timing/model': 0.4224686645902693, 'learning_rate/qwen_vl_interface': 8.823130574207144e-06, 'learning_rate/action_model': 8.823130574207143e-05, 'epoch': 0.73})
+[2026-07-28 17:14:09,031][starVLA.training.train_starvla][INFO] - Step 3994, Loss: {'train/loss': 0.08185252547264099, 'train/grad_norm_pre_clip': 2.9023377895355225, 'timing/data': 0.00025347108021378517, 'timing/model': 0.43173054978251457, 'learning_rate/qwen_vl_interface': 8.82060829959641e-06, 'learning_rate/action_model': 8.820608299596411e-05, 'epoch': 0.73})
+[2026-07-28 17:14:15,092][starVLA.training.train_starvla][INFO] - Step 3995, Loss: {'train/loss': 0.04088079184293747, 'train/grad_norm_pre_clip': 1.379929780960083, 'timing/data': 0.000304357148706913, 'timing/model': 0.4214992201887071, 'learning_rate/qwen_vl_interface': 8.81808370805672e-06, 'learning_rate/action_model': 8.818083708056721e-05, 'epoch': 0.73})
+[2026-07-28 17:14:21,167][starVLA.training.train_starvla][INFO] - Step 3996, Loss: {'train/loss': 0.04376034066081047, 'train/grad_norm_pre_clip': 1.4251253604888916, 'timing/data': 0.00029551004990935326, 'timing/model': 0.42283411184325814, 'learning_rate/qwen_vl_interface': 8.815556801226251e-06, 'learning_rate/action_model': 8.815556801226251e-05, 'epoch': 0.73})
+[2026-07-28 17:14:27,173][starVLA.training.train_starvla][INFO] - Step 3997, Loss: {'train/loss': 0.014045443385839462, 'train/grad_norm_pre_clip': 1.0165026187896729, 'timing/data': 0.0002721180208027363, 'timing/model': 0.4224203210324049, 'learning_rate/qwen_vl_interface': 8.813027580744685e-06, 'learning_rate/action_model': 8.813027580744684e-05, 'epoch': 0.73})
+[2026-07-28 17:14:33,439][starVLA.training.train_starvla][INFO] - Step 3998, Loss: {'train/loss': 0.04088365286588669, 'train/grad_norm_pre_clip': 1.6392385959625244, 'timing/data': 0.00026987725868821144, 'timing/model': 0.4224556451663375, 'learning_rate/qwen_vl_interface': 8.810496048253203e-06, 'learning_rate/action_model': 8.810496048253202e-05, 'epoch': 0.73})
+[2026-07-28 17:14:39,714][starVLA.training.train_starvla][INFO] - Step 3999, Loss: {'train/loss': 0.03552690148353577, 'train/grad_norm_pre_clip': 1.1227449178695679, 'timing/data': 0.00028708716854453087, 'timing/model': 0.6775439088232815, 'learning_rate/qwen_vl_interface': 8.80796220539449e-06, 'learning_rate/action_model': 8.80796220539449e-05, 'epoch': 0.73})
+[2026-07-28 17:14:45,723][starVLA.training.train_starvla][INFO] - Starting held-out action loss eval at step 4000
+[2026-07-28 17:14:50,835][starVLA.training.train_starvla][INFO] - Finished held-out action loss eval at step 4000
+[2026-07-28 17:21:57,560][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000
+[2026-07-28 17:21:57,561][starVLA.training.train_starvla][INFO] - Step 4000, Loss: {'train/loss': 0.021803952753543854, 'train/grad_norm_pre_clip': 0.6649140119552612, 'eval/action_loss/seconds': 5.108460139017552, 'eval/action_loss/samples': 800.0, 'eval/loss': 0.030299033969640732, 'eval/latency_3/loss': 0.030299033969640732, 'eval/flappy/loss': 0.030299033969640732, 'eval/flappy/latency_3/loss': 0.030299033969640732, 'rl_games_eval/mid_train/total_episodes': 20.0, 'rl_games_eval/mid_train/mean_reward': 323.67500381283463, 'rl_games_eval/mid_train/mean_length': 2627.55, 'rl_games_eval/mid_train/std_reward': 145.51053697177636, 'rl_games_eval/mid_train/std_length': 1171.6111759026542, 'rl_games_eval/mid_train/macro_mean_reward': 323.67500381283463, 'rl_games_eval/mid_train/macro_mean_length': 2627.55, 'rl_games_eval/mid_train/task_count': 1.0, 'rl_games_eval/mid_train/flappy__latency_3/mean_reward': 323.67500381283463, 'rl_games_eval/mid_train/flappy__latency_3/mean_length': 2627.55, 'rl_games_eval/mid_train/flappy__latency_3/std_reward': 145.51053697177636, 'rl_games_eval/mid_train/flappy__latency_3/std_length': 1171.6111759026542, 'rl_games_eval/mid_train/flappy__latency_3/num_episodes': 20.0, 'checkpoint/current_eval_score': 323.67500381283463, 'checkpoint/best_score': -inf, 'checkpoint/best_step': 0.0, 'checkpoint/is_new_best': 0.0, 'timing/data': 0.00025958102196455, 'timing/model': 0.42208191473037004, 'learning_rate/qwen_vl_interface': 8.805426053812731e-06, 'learning_rate/action_model': 8.805426053812731e-05, 'epoch': 0.73})
+[2026-07-28 17:21:57,564][accelerate.accelerator][INFO] - Saving current state to /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_4000_state
+[2026-07-28 17:21:57,564][accelerate.accelerator][INFO] - Saving DeepSpeed Model and Optimizer
+[2026-07-28 17:22:35,265][accelerate.accelerator][INFO] - DeepSpeed Model and Optimizer saved to output dir /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_4000_state/pytorch_model
+[2026-07-28 17:22:35,266][accelerate.checkpointing][INFO] - Sampler state for dataloader 0 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_4000_state/sampler.bin
+[2026-07-28 17:22:35,266][accelerate.checkpointing][INFO] - Sampler state for dataloader 1 saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_4000_state/sampler_1.bin
+[2026-07-28 17:22:35,270][accelerate.checkpointing][INFO] - Random states saved in /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints/steps_4000_state/random_states_0.pkl
+[2026-07-28 17:22:43,354][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping upload for step 4000
+[2026-07-28 17:22:43,354][starVLA.training.train_starvla][INFO] - 📊 Saving accessed configuration...
+[2026-07-28 17:22:43,376][starVLA.training.train_starvla][INFO] - ✅ Configuration files saved
+[2026-07-28 17:22:43,377][starVLA.training.train_starvla][INFO] - Training complete. Final training state is saved under /lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/checkpoints
+[2026-07-28 17:39:31,738][starVLA.training.rl_games.checkpoint_sync][INFO] - HF checkpoint sync disabled; skipping eval result upload for step 4000
+[2026-07-28 17:39:33,370][starVLA.training.train_starvla][INFO] - ... and that's all, folks!
+[2026-07-28 17:39:33,370][starVLA.training.train_starvla][INFO] - ... and that's all, folks!
+[2026-07-28 17:39:38,967][starVLA.training.train_starvla][INFO] - Destroyed distributed process group
+[2026-07-28 17:39:39,017][starVLA.training.train_starvla][INFO] - Destroyed distributed process group
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..7162feba5aef32cc64f851565fcc454676fc579d
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 79.00000095367432, "episode_return_env": 79.00000095367432, "game_score": 16.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 657, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 657}
+{"episode_id": 2, "episode_return": 431.30000507086515, "episode_return_env": 431.30000507086515, "game_score": 92.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3496, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3496}
+{"episode_id": 4, "episode_return": 41.900000520050526, "episode_return_env": 41.900000520050526, "game_score": 8.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 358, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 358}
+{"episode_id": 6, "episode_return": 6.800000101327896, "episode_return_env": 6.800000101327896, "game_score": 1.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 70, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 70}
+{"episode_id": 8, "episode_return": 345.5000040754676, "episode_return_env": 345.5000040754676, "game_score": 73.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2809, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2809}
+{"episode_id": 10, "episode_return": 257.0000030398369, "episode_return_env": 257.0000030398369, "game_score": 54.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2095, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2095}
+{"episode_id": 12, "episode_return": 210.00000248849392, "episode_return_env": 210.00000248849392, "game_score": 44.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1715, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1715}
+{"episode_id": 14, "episode_return": 224.0000026524067, "episode_return_env": 224.0000026524067, "game_score": 47.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1828, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1828}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 80.70000096410513, "episode_return_env": 80.70000096410513, "game_score": 17.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 665, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 665}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..0459fd76411c85a2746ff03ffe1c4732863e69c1
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 335.6000039577484, "episode_return_env": 335.6000039577484, "game_score": 71.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2728, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2728}
+{"episode_id": 7, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 9, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 11, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 17, "episode_return": 74.8000009059906, "episode_return_env": 74.8000009059906, "game_score": 15.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 624, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 624}
+{"episode_id": 19, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..1e206494faa327387723a7e2473db297d2a10398
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 230.80000272393227, "episode_return_env": 230.80000272393227, "game_score": 49.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1878, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1878}
+{"episode_id": 3, "episode_return": 440.70000518113375, "episode_return_env": 440.70000518113375, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3572, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3572}
+{"episode_id": 5, "episode_return": 200.00000236928463, "episode_return_env": 200.00000236928463, "game_score": 42.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1633, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1633}
+{"episode_id": 7, "episode_return": 419.2000049352646, "episode_return_env": 419.2000049352646, "game_score": 89.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3402, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3402}
+{"episode_id": 9, "episode_return": 11.30000015348196, "episode_return_env": 11.30000015348196, "game_score": 2.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 106, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 106}
+{"episode_id": 11, "episode_return": 69.40000084042549, "episode_return_env": 69.40000084042549, "game_score": 14.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 579, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 579}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 132.80000157654285, "episode_return_env": 132.80000157654285, "game_score": 28.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1087, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1087}
+{"episode_id": 17, "episode_return": 90.80000108480453, "episode_return_env": 90.80000108480453, "game_score": 19.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 748, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 748}
+{"episode_id": 19, "episode_return": 256.10000302642584, "episode_return_env": 256.10000302642584, "game_score": 54.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1500/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2086, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2086}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..1d0fffa4c6a475739a4e09c8ae902e2f6d1dc1fe
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 139.0000016540289, "episode_return_env": 139.0000016540289, "game_score": 29.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1140, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1140}
+{"episode_id": 3, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 5, "episode_return": 3.90000007301569, "episode_return_env": 3.90000007301569, "game_score": 0.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 50, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 50}
+{"episode_id": 7, "episode_return": 3.90000007301569, "episode_return_env": 3.90000007301569, "game_score": 0.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 50, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 50}
+{"episode_id": 9, "episode_return": 181.40000215172768, "episode_return_env": 181.40000215172768, "game_score": 38.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1483, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1483}
+{"episode_id": 11, "episode_return": 207.30000244826078, "episode_return_env": 207.30000244826078, "game_score": 44.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1688, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1688}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 3.90000007301569, "episode_return_env": 3.90000007301569, "game_score": 0.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 50, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 50}
+{"episode_id": 17, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 19, "episode_return": 53.30000064522028, "episode_return_env": 53.30000064522028, "game_score": 11.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_1750/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 445, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 445}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..12edf04d479a2866a50d3c6a96f5579e459ff229
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 0, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 42, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 2, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 44, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 4, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 46, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 6, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 48, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 8, "episode_return": 140.10000167042017, "episode_return_env": 140.10000167042017, "game_score": 30.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 50, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1151, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1151}
+{"episode_id": 10, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 52, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 12, "episode_return": 424.90000500530005, "episode_return_env": 424.90000500530005, "game_score": 90.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 54, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3450, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3450}
+{"episode_id": 14, "episode_return": 263.3000031039119, "episode_return_env": 263.3000031039119, "game_score": 56.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 56, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2140, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2140}
+{"episode_id": 16, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 58, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 18, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 60, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2000/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/_progress/rank_0.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/_progress/rank_0.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/_progress/rank_0.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/_progress/rank_1.json b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/_progress/rank_1.json
new file mode 100644
index 0000000000000000000000000000000000000000..52c9bdd9607475e8de5d6891cd1fcaedc191d85a
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/_progress/rank_1.json
@@ -0,0 +1 @@
+{"completed": 10}
\ No newline at end of file
diff --git a/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3/episode_metrics.jsonl b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3/episode_metrics.jsonl
new file mode 100644
index 0000000000000000000000000000000000000000..2a20c34fb0fe3aa2f6f43b3f998c630a09c35a69
--- /dev/null
+++ b/evaluations/flappy/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3/episode_metrics.jsonl
@@ -0,0 +1,10 @@
+{"episode_id": 1, "episode_return": 92.90000111609697, "episode_return_env": 92.90000111609697, "game_score": 19.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 43, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 769, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 769}
+{"episode_id": 3, "episode_return": 367.9000043347478, "episode_return_env": 367.9000043347478, "game_score": 78.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 45, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 2988, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 2988}
+{"episode_id": 5, "episode_return": 31.900000400841236, "episode_return_env": 31.900000400841236, "game_score": 6.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 47, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 276, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 276}
+{"episode_id": 7, "episode_return": 181.40000215172768, "episode_return_env": 181.40000215172768, "game_score": 38.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 49, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1483, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1483}
+{"episode_id": 9, "episode_return": 214.90000254660845, "episode_return_env": 214.90000254660845, "game_score": 45.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 51, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1755, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1755}
+{"episode_id": 11, "episode_return": 134.9000016078353, "episode_return_env": 134.9000016078353, "game_score": 28.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 53, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1108, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1108}
+{"episode_id": 13, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 55, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 15, "episode_return": 139.100001655519, "episode_return_env": 139.100001655519, "game_score": 29.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 57, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1141, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1141}
+{"episode_id": 17, "episode_return": 444.6000052243471, "episode_return_env": 444.6000052243471, "game_score": 94.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 59, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 3600, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 3600}
+{"episode_id": 19, "episode_return": 125.20000149309635, "episode_return_env": 125.20000149309635, "game_score": 26.0, "mean_latency_ms": 100.0, "metadata": {"checkpoint_path": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "config_name": "starvla_flappy_train_eval", "env_fps": 30.0, "env_id": "flappy", "episode_seed": 61, "frame_ms": 33.333333333333336, "instance_id": null, "instance_origin": null, "latency_type": "fixed", "mode": "simulated", "obs_fps": 30.0, "output_dir": "/lustre/fsw/portfolios/nvr/projects/nvr_lacr_llm/users/zihwang/latency-sensitive-bench/third_party/starVLA/results/Checkpoints/flappy_fix_latency_3_200ep_7k2steps_single_baseline/latency_bench_eval/mid_train/step_2250/flappy/latency_3", "policy_id": "starvla", "profile_id": null, "profile_path": null, "replay_host": null, "run_name": "starvla_flappy_train_eval", "target_device": null}, "num_actions": 1029, "num_dropped_actions": 2, "num_invalid_actions": 0, "p90_latency_ms": 100.0, "p99_latency_ms": 100.0, "return_raw": null, "survival_steps": 1029}