| { |
| "type": "fineart_vla", |
| "n_obs_steps": 1, |
| "input_features": { |
| "observation.state": { |
| "type": "STATE", |
| "shape": [ |
| 14 |
| ] |
| }, |
| "observation.images.cam_high": { |
| "type": "VISUAL", |
| "shape": [ |
| 3, |
| 360, |
| 640 |
| ] |
| }, |
| "observation.images.cam_left_wrist": { |
| "type": "VISUAL", |
| "shape": [ |
| 3, |
| 360, |
| 640 |
| ] |
| }, |
| "observation.images.cam_right_wrist": { |
| "type": "VISUAL", |
| "shape": [ |
| 3, |
| 360, |
| 640 |
| ] |
| } |
| }, |
| "output_features": { |
| "action": { |
| "type": "ACTION", |
| "shape": [ |
| 14 |
| ] |
| } |
| }, |
| "device": "cuda", |
| "use_amp": false, |
| "use_peft": false, |
| "push_to_hub": false, |
| "repo_id": null, |
| "private": null, |
| "tags": null, |
| "license": null, |
| "pretrained_path": "lerobot/FineART-VLA-subtask-midtrained", |
| "pretrained_revision": null, |
| "paligemma_variant": "gemma_2b", |
| "action_expert_variant": "gemma_300m", |
| "dtype": "bfloat16", |
| "chunk_size": 50, |
| "n_action_steps": 50, |
| "max_state_dim": 32, |
| "max_action_dim": 32, |
| "num_inference_steps": 10, |
| "time_sampling_beta_alpha": 1.5, |
| "time_sampling_beta_beta": 1.0, |
| "time_sampling_scale": 0.999, |
| "time_sampling_offset": 0.001, |
| "min_period": 0.004, |
| "max_period": 4.0, |
| "use_relative_actions": false, |
| "relative_exclude_joints": [ |
| "gripper" |
| ], |
| "action_feature_names": [ |
| "left_shoulder_pan", |
| "left_shoulder_lift", |
| "left_elbow_flex", |
| "left_wrist_angle", |
| "left_wrist_rotate", |
| "left_gripper", |
| "left_gripper_finger", |
| "right_shoulder_pan", |
| "right_shoulder_lift", |
| "right_elbow_flex", |
| "right_wrist_angle", |
| "right_wrist_rotate", |
| "right_gripper", |
| "right_gripper_finger" |
| ], |
| "rtc_config": null, |
| "image_resolution": [ |
| 224, |
| 224 |
| ], |
| "empty_cameras": 0, |
| "tokenizer_max_length": 128, |
| "normalization_mapping": { |
| "VISUAL": "IDENTITY", |
| "STATE": "QUANTILES", |
| "ACTION": "QUANTILES" |
| }, |
| "gradient_checkpointing": true, |
| "compile_model": true, |
| "compile_mode": "default", |
| "freeze_vision_encoder": false, |
| "train_expert_only": false, |
| "optimizer_lr": 2.5e-05, |
| "optimizer_betas": [ |
| 0.9, |
| 0.95 |
| ], |
| "optimizer_eps": 1e-08, |
| "optimizer_weight_decay": 1e-10, |
| "optimizer_grad_clip_norm": 1.0, |
| "scheduler_warmup_steps": 2000, |
| "scheduler_decay_steps": 300000, |
| "scheduler_decay_lr": 5e-06, |
| "recipe_path": "recipes/subtask.yaml", |
| "text_loss_weight": 1.0, |
| "flow_loss_weight": 10.0, |
| "unfreeze_lm_head": true, |
| "enable_fast_action_loss": true, |
| "action_tokenizer_name": "lerobot/fast-action-tokenizer", |
| "max_action_tokens": 384, |
| "fast_skip_tokens": 128, |
| "fast_action_loss_weight": 1.0, |
| "auto_fit_fast_tokenizer": true, |
| "fast_tokenizer_cache_dir": "~/.cache/lerobot/fast_tokenizers", |
| "fast_tokenizer_fit_samples": 16384, |
| "fast_tokenizer_validation_samples": 256, |
| "fast_tokenizer_max_reconstruction_rmse": 0.1, |
| "fast_tokenizer_max_dim_rmse": 0.2, |
| "knowledge_insulation": true, |
| "use_flashrt_adarms": false, |
| "use_compiled_text_ce": true, |
| "use_compiled_vision": false, |
| "use_flex_attention": false, |
| "use_manual_attention": false, |
| "manual_attention_scope": "all", |
| "lm_head_lr_scale": 1.0, |
| "backbone_lr_scale": 1.0, |
| "action_expert_lr_scale": 1.0, |
| "flow_num_repeats": 1, |
| "text_ce_z_loss_weight": 0.0001, |
| "optimizer_foreach": false, |
| "optimizer_fused": true |
| } |
|
|
