Upload starVLA checkpoint piper_demos_classify_blocks_30hz_qwenpi_uniform_left_s2_h30_c18_2gpu_gb32_30k step 30000

StarVLA/piper_demos_classify_blocks_30hz_qwenpi_uniform_left_s2_h30_c18_2gpu_gb32_30k/checkpoints/steps_30000_pytorch_model.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:da944bcd79fa54bd62191045d5e9e8f3730a014f7e74e377c36dc574723597a1
+size 10920597149

StarVLA/piper_demos_classify_blocks_30hz_qwenpi_uniform_left_s2_h30_c18_2gpu_gb32_30k/config.full.yaml

@@ -0,0 +1,115 @@
+run_id: piper_demos_classify_blocks_30hz_qwenpi_uniform_left_s2_h30_c18_2gpu_gb32_30k
+run_root_dir: /scratch/wangpc/DiscreteRTCv2/results/Checkpoints
+seed: 42
+wandb_entity: 401910710-university-of-california-berkeley
+wandb_project: starVLA_piper
+is_debug: false
+version_id: '0.21'
+framework:
+  name: QwenPI_v3_UniformBSpline
+  qwenvl:
+    base_vlm: ./playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    attn_implementation: flash_attention_2
+    vl_hidden_dim: 2560
+    num_vl_layers: 36
+  action_model:
+    action_model_type: LayerwiseFM
+    action_dim: 7
+    state_dim: 7
+    action_horizon: 18
+    repeated_diffusion_steps: 2
+    num_inference_timesteps: 4
+    add_pos_embed: true
+    max_seq_len: 1024
+    num_target_vision_tokens: 32
+    noise_beta_alpha: 1.5
+    noise_beta_beta: 1.0
+    noise_s: 0.999
+    num_timestep_buckets: 1000
+    diffusion_model_cfg:
+      action_dit_hidden_dim: 1024
+      dropout: 0.2
+      final_dropout: true
+      interleave_self_attention: true
+      norm_type: ada_norm
+      positional_embeddings: null
+      attention_head_dim: 64
+      num_layers: 36
+      input_embedding_dim: 1024
+      cross_attention_dim: 1024
+      num_attention_heads: 16
+    future_action_window_size: 17
+    past_action_window_size: 0
+  action_representation:
+    type: continuous_uniform_bspline
+    mode: uniform_left
+    action_dim: 7
+    frequency_hz: 30.0
+    executable_horizon: 30
+    num_control_points: 18
+    degree: 3
+    span_length_steps: 2
+    regularization: 0.0001
+    end_padding: true
+    implementation_version: uniform_left_direct_fit_v1
+datasets:
+  vla_data:
+    dataset_py: lerobot_datasets
+    data_root_dir: ./data
+    data_mix: piper_demos_classify_blocks_30hz_uniform_left_bspline
+    controls_dataset_path: ./data/processed/piper_demos/ActionCommand/State_Joint/classify_blocks_30hz_cleanup_spline_uniform_left_s2_h30_c18
+    include_state: true
+    episode_split: train
+    action_type: abs_qpos
+    action_mode: abs
+    sequential_step_sampling: false
+    per_device_batch_size: 16
+    load_all_data_for_training: true
+    obs_image_size:
+    - 224
+    - 224
+    delete_pause_frame: false
+    video_backend: torchvision_av
+    num_workers: 8
+    persistent_workers: true
+    prefetch_factor: 2
+    pin_memory: true
+trainer:
+  pretrained_checkpoint: null
+  freeze_modules: qwen_vl_interface
+  max_train_steps: 30000
+  num_warmup_steps: 500
+  save_interval: 5000
+  eval_interval: 1000
+  repeated_diffusion_steps: 2
+  learning_rate:
+    base: 1.0e-05
+    qwen_vl_interface: 1.0e-06
+    action_model: 0.0001
+    project_layers: 0.0001
+  lr_scheduler_type: cosine_with_min_lr
+  scheduler_specific_kwargs:
+    min_lr: 1.0e-06
+  loss_scale:
+    vla: 1.0
+    vlm: 0.0
+  max_grad_norm: 1.0
+  weight_decay: 0.0
+  logging_frequency: 20
+  gradient_clipping: 1.0
+  gradient_accumulation_steps: 1
+  gradient_checkpointing: true
+  optimizer:
+    name: AdamW
+    betas:
+    - 0.9
+    - 0.95
+    eps: 1.0e-08
+    weight_decay: 1.0e-08
+  save_format: pt
+  is_resume: false
+  max_wall_time_hours: 10
+  wall_time_safety_minutes: 20
+  save_training_state: true
+config_yaml: /scratch/wangpc/DiscreteRTCv2/examples/realRobots/Piper/train_files/starvla_qwenpi_piper_demos_classify_blocks_30hz_uniform_left_s2_h30_c18_2gpu_gb32_30k.yaml
+output_dir: /scratch/wangpc/DiscreteRTCv2/results/Checkpoints/piper_demos_classify_blocks_30hz_qwenpi_uniform_left_s2_h30_c18_2gpu_gb32_30k

StarVLA/piper_demos_classify_blocks_30hz_qwenpi_uniform_left_s2_h30_c18_2gpu_gb32_30k/config.yaml

@@ -0,0 +1,94 @@
+datasets:
+  vla_data:
+    action_mode: abs
+    controls_dataset_path: ./data/processed/piper_demos/ActionCommand/State_Joint/classify_blocks_30hz_cleanup_spline_uniform_left_s2_h30_c18
+    data_mix: piper_demos_classify_blocks_30hz_uniform_left_bspline
+    data_root_dir: ./data
+    dataset_py: lerobot_datasets
+    delete_pause_frame: false
+    episode_split: train
+    num_workers: 8
+    obs_image_size:
+    - 224
+    - 224
+    per_device_batch_size: 16
+    persistent_workers: true
+    pin_memory: true
+    prefetch_factor: 2
+    video_backend: torchvision_av
+framework:
+  action_model:
+    action_dim: 7
+    action_horizon: 18
+    add_pos_embed: true
+    diffusion_model_cfg:
+      action_dit_hidden_dim: 1024
+      attention_head_dim: 64
+      cross_attention_dim: 1024
+      dropout: 0.2
+      final_dropout: true
+      input_embedding_dim: 1024
+      interleave_self_attention: true
+      norm_type: ada_norm
+      num_attention_heads: 16
+      num_layers: 36
+      positional_embeddings: null
+    max_seq_len: 1024
+    noise_beta_alpha: 1.5
+    noise_beta_beta: 1.0
+    noise_s: 0.999
+    num_inference_timesteps: 4
+    num_target_vision_tokens: 32
+    num_timestep_buckets: 1000
+    state_dim: 7
+  action_representation:
+    action_dim: 7
+    degree: 3
+    end_padding: true
+    executable_horizon: 30
+    frequency_hz: 30.0
+    implementation_version: uniform_left_direct_fit_v1
+    num_control_points: 18
+    regularization: 0.0001
+    span_length_steps: 2
+  name: QwenPI_v3_UniformBSpline
+  qwenvl:
+    attn_implementation: flash_attention_2
+    base_vlm: ./playground/Pretrained_models/Qwen3-VL-4B-Instruct
+    num_vl_layers: 36
+    vl_hidden_dim: 2560
+output_dir: /scratch/wangpc/DiscreteRTCv2/results/Checkpoints/piper_demos_classify_blocks_30hz_qwenpi_uniform_left_s2_h30_c18_2gpu_gb32_30k
+run_id: piper_demos_classify_blocks_30hz_qwenpi_uniform_left_s2_h30_c18_2gpu_gb32_30k
+run_root_dir: /scratch/wangpc/DiscreteRTCv2/results/Checkpoints
+seed: 42
+trainer:
+  eval_interval: 1000
+  freeze_modules: qwen_vl_interface
+  gradient_clipping: 1.0
+  is_resume: false
+  learning_rate:
+    action_model: 0.0001
+    base: 1.0e-05
+    project_layers: 0.0001
+    qwen_vl_interface: 1.0e-06
+  logging_frequency: 20
+  lr_scheduler_type: cosine_with_min_lr
+  max_train_steps: 30000
+  max_wall_time_hours: 10
+  num_warmup_steps: 500
+  optimizer:
+    betas:
+    - 0.9
+    - 0.95
+    eps: 1.0e-08
+    weight_decay: 1.0e-08
+  pretrained_checkpoint: null
+  repeated_diffusion_steps: 2
+  save_format: pt
+  save_interval: 5000
+  save_training_state: true
+  scheduler_specific_kwargs:
+    min_lr: 1.0e-06
+  wall_time_safety_minutes: 20
+wandb_entity: 401910710-university-of-california-berkeley
+wandb_project: starVLA_piper

StarVLA/piper_demos_classify_blocks_30hz_qwenpi_uniform_left_s2_h30_c18_2gpu_gb32_30k/dataset_statistics.json

@@ -0,0 +1,127 @@
+{
+  "new_embodiment": {
+    "action": {
+      "mean": [
+        -0.09834051877260208,
+        1.839445948600769,
+        -1.454539179801941,
+        0.07625001668930054,
+        0.6646823287010193,
+        0.004705857019871473,
+        0.3798874616622925
+      ],
+      "std": [
+        0.3213742673397064,
+        0.45298922061920166,
+        0.5084811449050903,
+        0.21163690090179443,
+        0.2981576919555664,
+        0.3729538321495056,
+        0.4721868634223938
+      ],
+      "max": [
+        0.6822819709777832,
+        2.5526516437530518,
+        0.007414266001433134,
+        0.9107677340507507,
+        1.6138538122177124,
+        1.3140977621078491,
+        1.0
+      ],
+      "min": [
+        -1.2183202505111694,
+        -0.02290080301463604,
+        -2.6815927028656006,
+        -0.9883002042770386,
+        -0.3849231004714966,
+        -1.440931797027588,
+        0.0
+      ],
+      "q01": [
+        -0.7213104367256165,
+        -0.010630491189658642,
+        -2.403943202495575,
+        -0.41912044703960416,
+        -0.09879948891699311,
+        -0.9439219832420349,
+        0.0
+      ],
+      "q99": [
+        0.439912990629673,
+        2.417661190032959,
+        -0.0017282594926655293,
+        0.645389199256897,
+        1.2549023628234863,
+        0.9198646545410156,
+        1.0
+      ],
+      "mask": [
+        true,
+        true,
+        true,
+        true,
+        true,
+        true,
+        true
+      ]
+    },
+    "state": {
+      "mean": [
+        -0.09691226482391357,
+        1.841578483581543,
+        -1.4529383182525635,
+        0.0789480060338974,
+        0.6674876809120178,
+        0.008207720704376698,
+        0.29442355036735535
+      ],
+      "std": [
+        0.3209986686706543,
+        0.45203259587287903,
+        0.5068883895874023,
+        0.20927873253822327,
+        0.2927248775959015,
+        0.37215012311935425,
+        0.35192400217056274
+      ],
+      "max": [
+        0.6779382228851318,
+        2.5556507110595703,
+        0.0019373154500499368,
+        0.907466471195221,
+        1.245327353477478,
+        1.3127843141555786,
+        1.0
+      ],
+      "min": [
+        -1.2123056650161743,
+        -0.00205948855727911,
+        -2.674123764038086,
+        -0.9800024032592773,
+        -0.3712140917778015,
+        -1.4272429943084717,
+        0.004000000189989805
+      ],
+      "q01": [
+        -0.718255341053009,
+        0.0036128314677625895,
+        -2.3986440181732176,
+        -0.41029199957847595,
+        -0.08815658092498779,
+        -0.9336115121841431,
+        0.004999999888241291
+      ],
+      "q99": [
+        0.44024184346199036,
+        2.4192892026901243,
+        -0.006806784309446812,
+        0.6413069176673888,
+        1.2160232067108154,
+        0.9206437468528748,
+        0.753000020980835
+      ]
+    },
+    "num_transitions": 174348,
+    "num_trajectories": 80
+  }
+}

StarVLA/piper_demos_classify_blocks_30hz_qwenpi_uniform_left_s2_h30_c18_2gpu_gb32_30k/summary.jsonl

@@ -0,0 +1,6 @@
+{"steps": 5000}
+{"steps": 10000}
+{"steps": 15000}
+{"steps": 20000}
+{"steps": 25000}
+{"steps": 30000}

StarVLA/piper_demos_classify_blocks_30hz_qwenpi_uniform_left_s2_h30_c18_2gpu_gb32_30k/wall_time_ledger.json

@@ -0,0 +1,15 @@
+{
+  "budget_hours": 10.0,
+  "safety_minutes": 20.0,
+  "sessions": [
+    {
+      "started_at": "2026-10-15T17:49:02+00:00",
+      "elapsed_seconds": 24001.2,
+      "start_step": 0,
+      "last_step": 30000,
+      "end_reason": "completed",
+      "signal": null
+    }
+  ],
+  "total_elapsed_seconds": 24001.2
+}