Download _STRICT_COMPLETE.json from ldunvidia/MemTTT-full16-PatternLock-stage3-1000: direct link, hf CLI and curl.
- Browser
- Download file 8.57 kB
-
https://huggingface.co/ldunvidia/MemTTT-full16-PatternLock-stage3-1000/resolve/main/_STRICT_COMPLETE.json
- Command line
-
hf download hf://ldunvidia/MemTTT-full16-PatternLock-stage3-1000/_STRICT_COMPLETE.json
-
curl -L -o _STRICT_COMPLETE.json https://huggingface.co/ldunvidia/MemTTT-full16-PatternLock-stage3-1000/resolve/main/_STRICT_COMPLETE.json
8.57 kB
| { | |
| "code_fingerprint": "759603bdc0b9bfe134c97d559c6746a4f6f5fb87ff3d4077093e7aec2403b322", | |
| "completed_steps": 1000, | |
| "config": { | |
| "allow_fresh_lora_init": false, | |
| "b1": 0.9, | |
| "b2": 0.95, | |
| "backbone_lr": 0.0, | |
| "buckets": { | |
| "lengths": [ | |
| 224, | |
| 304, | |
| 368, | |
| 496, | |
| 672, | |
| 1104 | |
| ], | |
| "micro_batch": [ | |
| 4, | |
| 4, | |
| 4, | |
| 4, | |
| 4, | |
| 4 | |
| ] | |
| }, | |
| "cache_marker_contract": [ | |
| [ | |
| "tasks", | |
| [ | |
| "PatternLock" | |
| ] | |
| ], | |
| [ | |
| "index_train_total", | |
| 700 | |
| ], | |
| [ | |
| "index_test_total", | |
| 0 | |
| ], | |
| [ | |
| "memory_stride", | |
| 16 | |
| ], | |
| [ | |
| "execution_stride", | |
| 1 | |
| ] | |
| ], | |
| "cache_marker_sha256": "fc479646b8dd7003100d39df4792e2b5108b1e3e547d4924a0d7f5633e1d8a53", | |
| "clip_norm": 1.0, | |
| "curriculum_bucket": 32, | |
| "curriculum_micro_batch": 4, | |
| "curriculum_steps": 0, | |
| "data_artifact_hashes": [ | |
| [ | |
| "annotations.jsonl", | |
| "8d95c7de7467d61cf27094735424aa39603da448c032765fafd25686098a6be2" | |
| ], | |
| [ | |
| "index_train.json", | |
| "cf58e60753ed0e8dfbc2bae48cf196acb9b3dc6342eb1c478ee9c46aaca47a3d" | |
| ], | |
| [ | |
| "index_test.json", | |
| "ca3d163bab055381827226140568f3bef7eaac187cebd76878e0b63e9e442356" | |
| ], | |
| [ | |
| "source_manifest_mapping.jsonl", | |
| "46094fea8e0c7a364bc6c9af9bac2f1f9d873382c6a5b2e2b5b10603fb6a5a9b" | |
| ], | |
| [ | |
| "runtime_stage_report.json", | |
| "184fbfa4b7b9131eab68470c62d08046994cdf003f10a1a96eed19281e7c1742" | |
| ] | |
| ], | |
| "data_fingerprint": "fb5693799b7050bd508c791c7b8171e4e4a15bc01210ffc9043ec72fe3f5969e", | |
| "data_manifest_contract": [ | |
| [ | |
| "tasks", | |
| [ | |
| "PatternLock" | |
| ] | |
| ], | |
| [ | |
| "index_train_total", | |
| 700 | |
| ], | |
| [ | |
| "index_test_total", | |
| 0 | |
| ], | |
| [ | |
| "memory_stride", | |
| 16 | |
| ], | |
| [ | |
| "execution_stride", | |
| 1 | |
| ] | |
| ], | |
| "data_manifest_sha256": "184fbfa4b7b9131eab68470c62d08046994cdf003f10a1a96eed19281e7c1742", | |
| "decision_annotations_difficulty_only": true, | |
| "decision_annotations_sha256": "8d95c7de7467d61cf27094735424aa39603da448c032765fafd25686098a6be2", | |
| "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/checkpoints/memttt-single-task-full16/selected7_single_task_5dfd876_20260917T164712Z/lane1/single_PatternLock_stage2/lane1/10000/params", | |
| "init_provenance_path": "", | |
| "init_provenance_sha256": "", | |
| "jitter": false, | |
| "lr": 7e-06, | |
| "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_stage3", | |
| "norm_stats_sha256": "f332bbd34ace1b6837cdc415b44f680896070a41564f9ce39016f1ebf99d1be5", | |
| "num_steps": 1000, | |
| "optimization_phases": [ | |
| { | |
| "backbone_learning_rate": null, | |
| "data_mode": "dense", | |
| "difficulty_weights": [ | |
| [ | |
| "easy", | |
| 0.1 | |
| ], | |
| [ | |
| "medium", | |
| 0.1 | |
| ], | |
| [ | |
| "hard", | |
| 0.8 | |
| ] | |
| ], | |
| "execution_write_perturbation": "none", | |
| "learning_rate": 7e-06, | |
| "name": "action_expert_only_1k", | |
| "recovery_episode_fraction": 0.2, | |
| "steps": 1000, | |
| "trainable_group": "action", | |
| "warmup_steps": 50 | |
| } | |
| ], | |
| "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": false, | |
| "require_native_pi05_init": false, | |
| "require_stage1_lora_init": true, | |
| "required_device_count": 4, | |
| "required_parent_code_fingerprint": "759603bdc0b9bfe134c97d559c6746a4f6f5fb87ff3d4077093e7aec2403b322", | |
| "required_parent_completed_steps": 10000, | |
| "required_parent_config_name": "single_PatternLock_stage2", | |
| "required_parent_data_fingerprint": "a931c397233e45b936dd20a3a1ab13b86d5bc2c4fa7040bc4c801bd441f95e5b", | |
| "required_parent_marker_sha256": "47f864879374d40b850093247e4bee4c05fa01a98ba01034b5ce6ec6b2b43e5a", | |
| "required_parent_params_tree_sha256": "87bfcca0d3f6cd084faaffd8c09a4a41a481ecfdb1694456214da9bf84ed1bb1", | |
| "robottt_alpha_lr_multiplier": 1.0, | |
| "sampler_episode_allowlist": [], | |
| "sampler_max_episodes_per_task": 0, | |
| "save_steps": [ | |
| 1000 | |
| ], | |
| "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": "d55a690c7b239cf857089d6cd9dde137381c08ac4ba4c4c861502e30410e8fd2", | |
| "data_state": { | |
| "completed_steps": 1000, | |
| "optimization_phase": { | |
| "index": 1, | |
| "local_step": 0, | |
| "name": "complete", | |
| "trainable_group": "none" | |
| }, | |
| "phase": "full", | |
| "sampler_seed": 7, | |
| "windows_in_phase": 1000 | |
| }, | |
| "has_ema": false, | |
| "params_tree_sha256": "742ffe8fc901cc019d03d779e81e0a93c461a16159e8584359ec49624f1aca5b", | |
| "schema": 1, | |
| "trainer_state_tree_sha256": "b5dc6870c7de8b2e75261a5758e53c45c4367e2da53d61f4da339fb0bc2861fc" | |
| } |