Download _STRICT_COMPLETE.json from ldunvidia/MemTTT-full16-PatternLock-stage2-10000: direct link, hf CLI and curl.
- Browser
- Download file 19.2 kB
-
https://huggingface.co/ldunvidia/MemTTT-full16-PatternLock-stage2-10000/resolve/main/_STRICT_COMPLETE.json
- Command line
-
hf download hf://ldunvidia/MemTTT-full16-PatternLock-stage2-10000/_STRICT_COMPLETE.json
-
curl -L -o _STRICT_COMPLETE.json https://huggingface.co/ldunvidia/MemTTT-full16-PatternLock-stage2-10000/resolve/main/_STRICT_COMPLETE.json
19.2 kB
| { | |
| "code_fingerprint": "759603bdc0b9bfe134c97d559c6746a4f6f5fb87ff3d4077093e7aec2403b322", | |
| "completed_steps": 10000, | |
| "config": { | |
| "allow_fresh_lora_init": false, | |
| "b1": 0.9, | |
| "b2": 0.95, | |
| "backbone_lr": 3e-05, | |
| "buckets": { | |
| "lengths": [ | |
| 224, | |
| 304, | |
| 368, | |
| 496, | |
| 672, | |
| 1104 | |
| ], | |
| "micro_batch": [ | |
| 4, | |
| 4, | |
| 4, | |
| 4, | |
| 4, | |
| 4 | |
| ] | |
| }, | |
| "cache_marker_contract": [ | |
| [ | |
| "tasks", | |
| [ | |
| "PatternLock" | |
| ] | |
| ], | |
| [ | |
| "index_train_total", | |
| 600 | |
| ], | |
| [ | |
| "index_test_total", | |
| 0 | |
| ], | |
| [ | |
| "memory_stride", | |
| 16 | |
| ], | |
| [ | |
| "execution_stride", | |
| 1 | |
| ] | |
| ], | |
| "cache_marker_sha256": "e4c6c510e83b58f31f66413e85e37ab9bb446ac73e84e3e7d653ed9dec639e4a", | |
| "clip_norm": 1.0, | |
| "curriculum_bucket": 32, | |
| "curriculum_micro_batch": 4, | |
| "curriculum_steps": 0, | |
| "data_artifact_hashes": [ | |
| [ | |
| "annotations.jsonl", | |
| "8ada0d376ea0a51de4cbcaac4ffff627c5a540e8771888735adfa53f9874b6ae" | |
| ], | |
| [ | |
| "index_train.json", | |
| "c305382df52afaf2c6a397988087038e2a817e506e97053ab3d121217df82138" | |
| ], | |
| [ | |
| "index_test.json", | |
| "ca3d163bab055381827226140568f3bef7eaac187cebd76878e0b63e9e442356" | |
| ], | |
| [ | |
| "source_manifest_mapping.jsonl", | |
| "b640267fb888329fa6dab3ce87c2a711955183698a7953eccd05ca66078b3937" | |
| ], | |
| [ | |
| "runtime_stage_report.json", | |
| "23de9e81bb797b5b8b94ff0dd2f214777ed73374d3be95f3af2e34127581ad68" | |
| ] | |
| ], | |
| "data_fingerprint": "a931c397233e45b936dd20a3a1ab13b86d5bc2c4fa7040bc4c801bd441f95e5b", | |
| "data_manifest_contract": [ | |
| [ | |
| "tasks", | |
| [ | |
| "PatternLock" | |
| ] | |
| ], | |
| [ | |
| "index_train_total", | |
| 600 | |
| ], | |
| [ | |
| "index_test_total", | |
| 0 | |
| ], | |
| [ | |
| "memory_stride", | |
| 16 | |
| ], | |
| [ | |
| "execution_stride", | |
| 1 | |
| ] | |
| ], | |
| "data_manifest_sha256": "23de9e81bb797b5b8b94ff0dd2f214777ed73374d3be95f3af2e34127581ad68", | |
| "decision_annotations_difficulty_only": true, | |
| "decision_annotations_sha256": "8ada0d376ea0a51de4cbcaac4ffff627c5a540e8771888735adfa53f9874b6ae", | |
| "ema_decay": null, | |
| "exp_name": "lane1", | |
| "gradient_accumulation_steps": 1, | |
| "h5_handle_cache_size": 32, | |
| "hard_two_decision_fraction": 0.0, | |
| "init_params_path": "/workspace/scratch/memttt-single-task-full16/selected7_single_task_5dfd876_20260917T164712Z/lane1/stage1_20000/params", | |
| "init_provenance_path": "", | |
| "init_provenance_sha256": "", | |
| "jitter": false, | |
| "lr": 6e-05, | |
| "max_episodes_per_task": 0, | |
| "model": { | |
| "action_dim": 32, | |
| "action_expert_variant": "gemma_300m", | |
| "action_horizon": 64, | |
| "action_loss_dims": 8, | |
| "decision_loss_weight": 1.0, | |
| "decision_only_supervision": false, | |
| "dtype": "bfloat16", | |
| "exec_horizon": 32, | |
| "execution_stride": 1, | |
| "flow_tau_min": 0.001, | |
| "max_token_len": 64, | |
| "mem_enable": true, | |
| "mem_fast_delta_max_norm": 4.0, | |
| "mem_inner_lr": 0.01, | |
| "mem_num_heads": 16, | |
| "mem_registers": 16, | |
| "mem_ttt_layers": [ | |
| 2, | |
| 5, | |
| 8, | |
| 11, | |
| 14, | |
| 17 | |
| ], | |
| "memory_architecture": "robottt_vl_register_only_v2", | |
| "memory_stride": 16, | |
| "paligemma_variant": "gemma_2b_lora", | |
| "pi05": true, | |
| "remat_ticks": true, | |
| "robottt_action_side_layers": 18, | |
| "robottt_adaptive_inner_min_scale": 0.1, | |
| "robottt_adaptive_inner_residual_reference": 1.0, | |
| "robottt_adaptive_inner_step": true, | |
| "robottt_anchor_tau_min": 0.0, | |
| "robottt_counterfactual_delta_weight": 0.0, | |
| "robottt_dense_video_prefix_tasks": [ | |
| "VideoUnmask", | |
| "VideoUnmaskSwap", | |
| "VideoPlaceOrder", | |
| "VideoPlaceButton", | |
| "VideoRepick", | |
| "InsertPeg", | |
| "MoveCube", | |
| "PatternLock", | |
| "RouteStick" | |
| ], | |
| "robottt_dense_video_prefix_ticks": 80, | |
| "robottt_gate_init": 0.1, | |
| "robottt_hidden_multiplier": 4, | |
| "robottt_inner_curvature_target": 0.5, | |
| "robottt_inner_lr": 0.1, | |
| "robottt_inner_step_normalization": true, | |
| "robottt_input_rms_norm": true, | |
| "robottt_isolate_invalid_action_inputs": true, | |
| "robottt_max_inner_step": 1000.0, | |
| "robottt_num_heads": 16, | |
| "robottt_pair_anchor_tau": 0.0, | |
| "robottt_read_before_write": true, | |
| "robottt_register_lora_alpha": 16.0, | |
| "robottt_register_lora_rank": 16, | |
| "robottt_register_only_read": true, | |
| "robottt_register_only_write": true, | |
| "robottt_register_vl_only_attention": true, | |
| "robottt_rope_theta": 10000.0, | |
| "robottt_sequence_action_forcing": true, | |
| "robottt_silence_weight": 0.0, | |
| "robottt_tau_scale": 0.999, | |
| "robottt_train_dense_read_block": 16, | |
| "robottt_train_reuse_prefix_cache": false, | |
| "robottt_train_scan_unroll": 1, | |
| "robottt_train_skip_empty_commit": true, | |
| "robottt_train_tbptt_frames": 0, | |
| "robottt_train_true_read_only": true, | |
| "robottt_trainable_ttt_layers": 17, | |
| "robottt_unfreeze_action_expert": false, | |
| "sequence_backprop_mode": "whole_short_sequence", | |
| "siglip_time_block": 8, | |
| "siglip_variant": "So400m/14", | |
| "supervision_post_anchor_steps": 0, | |
| "timeline_mode": "robottt_dense_v1", | |
| "video_stride": 16 | |
| }, | |
| "name": "single_PatternLock_stage2", | |
| "norm_stats_sha256": "f332bbd34ace1b6837cdc415b44f680896070a41564f9ce39016f1ebf99d1be5", | |
| "num_steps": 10000, | |
| "optimization_phases": [ | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.25 | |
| ], | |
| [ | |
| "medium", | |
| 0.25 | |
| ], | |
| [ | |
| "hard", | |
| 0.5 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 6e-05, | |
| "name": "ttt_cycle_1", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "ttt", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.25 | |
| ], | |
| [ | |
| "medium", | |
| 0.25 | |
| ], | |
| [ | |
| "hard", | |
| 0.5 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 3e-05, | |
| "name": "backbone_cycle_1", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "backbone", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.25 | |
| ], | |
| [ | |
| "medium", | |
| 0.25 | |
| ], | |
| [ | |
| "hard", | |
| 0.5 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 6e-05, | |
| "name": "ttt_cycle_2", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "ttt", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.25 | |
| ], | |
| [ | |
| "medium", | |
| 0.25 | |
| ], | |
| [ | |
| "hard", | |
| 0.5 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 3e-05, | |
| "name": "backbone_cycle_2", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "backbone", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.25 | |
| ], | |
| [ | |
| "medium", | |
| 0.25 | |
| ], | |
| [ | |
| "hard", | |
| 0.5 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 6e-05, | |
| "name": "ttt_cycle_3", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "ttt", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.25 | |
| ], | |
| [ | |
| "medium", | |
| 0.25 | |
| ], | |
| [ | |
| "hard", | |
| 0.5 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 3e-05, | |
| "name": "backbone_cycle_3", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "backbone", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.25 | |
| ], | |
| [ | |
| "medium", | |
| 0.25 | |
| ], | |
| [ | |
| "hard", | |
| 0.5 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 6e-05, | |
| "name": "ttt_cycle_4", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "ttt", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.25 | |
| ], | |
| [ | |
| "medium", | |
| 0.25 | |
| ], | |
| [ | |
| "hard", | |
| 0.5 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 3e-05, | |
| "name": "backbone_cycle_4", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "backbone", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 6e-05, | |
| "name": "ttt_cycle_5", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "ttt", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 3e-05, | |
| "name": "backbone_cycle_5", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "backbone", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 6e-05, | |
| "name": "ttt_cycle_6", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "ttt", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 3e-05, | |
| "name": "backbone_cycle_6", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "backbone", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 6e-05, | |
| "name": "ttt_cycle_7", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "ttt", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 3e-05, | |
| "name": "backbone_cycle_7", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "backbone", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 6e-05, | |
| "name": "ttt_cycle_8", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "ttt", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 3e-05, | |
| "name": "backbone_cycle_8", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "backbone", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 6e-05, | |
| "name": "ttt_cycle_9", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "ttt", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 3e-05, | |
| "name": "backbone_cycle_9", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "backbone", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 6e-05, | |
| "name": "ttt_cycle_10", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "ttt", | |
| "warmup_steps": 25 | |
| }, | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.15 | |
| ], | |
| [ | |
| "medium", | |
| 0.15 | |
| ], | |
| [ | |
| "hard", | |
| 0.7 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 3e-05, | |
| "name": "backbone_cycle_10", | |
| "recovery_episode_fraction": 0.0, | |
| "steps": 500, | |
| "trainable_group": "backbone", | |
| "warmup_steps": 25 | |
| } | |
| ], | |
| "pair_fraction": 0.0, | |
| "pair_manifest_path": "", | |
| "pair_manifest_sha256": "", | |
| "pair_repeats": 1, | |
| "pair_second_decision_fraction": 0.0, | |
| "per_device_batch": 1, | |
| "require_decision_annotations": true, | |
| "require_fresh_memory_init": true, | |
| "require_native_pi05_init": false, | |
| "require_stage1_lora_init": true, | |
| "required_device_count": 4, | |
| "required_parent_code_fingerprint": "009fc34f99b2fa370b9839b3e54ec194589435f97fe03aab4fa378e4ca5be4e9", | |
| "required_parent_completed_steps": 20000, | |
| "required_parent_config_name": "robomme_pi05_full16_h64_native_framewise40k_4gpu_accum8_v4", | |
| "required_parent_data_fingerprint": "09e8aac209c0657379161d165b3a9645876c652f4c880683aa558a851ca33edd", | |
| "required_parent_marker_sha256": "bde356ac58babcf9e530d5a097de8c594f2b5d91a2db68ff4ceae9bd5e71be74", | |
| "required_parent_params_tree_sha256": "96a2c89d022428f9f89b37dc0371f36b1229501e802e1a613ee26cc88e6cb6f3", | |
| "robottt_alpha_lr_multiplier": 8.333333333333334, | |
| "sampler_episode_allowlist": [], | |
| "sampler_max_episodes_per_task": 0, | |
| "save_steps": [ | |
| 5000, | |
| 10000 | |
| ], | |
| "seed": 7, | |
| "split": "train", | |
| "stage": 2, | |
| "stage1_framewise": false, | |
| "stage1_task_balanced": false, | |
| "tasks": [ | |
| "PatternLock" | |
| ], | |
| "ttt_phase_atomic_counterfactual_pairs": false, | |
| "ttt_phase_decision_only_supervision": false, | |
| "ttt_phase_pair_fraction": 0.0, | |
| "ttt_phase_periodic_supervision_stride": 0, | |
| "ttt_phase_require_equal_pair_ticks": false, | |
| "ttt_phase_timeline_mode": "", | |
| "verify_gradient_families": false, | |
| "verify_history_causality": false, | |
| "verify_parameter_update": false, | |
| "w0_weight_decay": 0.0, | |
| "warmup_ratio": 0.0 | |
| }, | |
| "config_fingerprint": "142e5d9857fed395881a44992abe7e9d9c0743879ac7201f464139f3ceb5ff49", | |
| "data_state": { | |
| "completed_steps": 10000, | |
| "optimization_phase": { | |
| "index": 20, | |
| "local_step": 0, | |
| "name": "complete", | |
| "trainable_group": "none" | |
| }, | |
| "phase": "full", | |
| "sampler_seed": 7, | |
| "windows_in_phase": 10000 | |
| }, | |
| "has_ema": false, | |
| "params_tree_sha256": "87bfcca0d3f6cd084faaffd8c09a4a41a481ecfdb1694456214da9bf84ed1bb1", | |
| "schema": 1, | |
| "trainer_state_tree_sha256": "0989f44800c37489ebe4959c864e7b13235e02845e8aab0df32c1d5265caa10a" | |
| } |