{ "dataset": { "repo_id": "Jiamo0912/so101-stack-cubes", "repo_type": "dataset", "root": "/scratch/e1583535/datasets/so101-stack-cubes", "episodes": null, "image_transforms": { "enable": true, "max_num_transforms": 3, "random_order": false, "tfs": { "crop": { "weight": 1.0, "type": "RandomResizedCrop", "kwargs": { "size": [ 480, 640 ], "scale": [ 0.9025, 0.9025 ], "ratio": [ 1.3333333, 1.3333333 ] } }, "rotation": { "weight": 1.0, "type": "RandomRotation", "kwargs": { "degrees": [ -5.0, 5.0 ], "interpolation": 2 } }, "color": { "weight": 1.0, "type": "ColorJitter", "kwargs": { "brightness": 0.2, "contrast": [ 0.8, 1.2 ], "saturation": [ 0.8, 1.2 ], "hue": 0.05 } } } }, "revision": null, "use_imagenet_stats": true, "video_backend": "pyav", "return_uint8": false, "depth_output_unit": "mm", "streaming": false, "eval_split": 0.0 }, "env": null, "policy": { "type": "molmoact2", "n_obs_steps": 1, "input_features": { "observation.state": { "type": "STATE", "shape": [ 6 ] }, "observation.images.front": { "type": "VISUAL", "shape": [ 3, 480, 640 ] }, "observation.images.wrist": { "type": "VISUAL", "shape": [ 3, 480, 640 ] } }, "output_features": { "action": { "type": "ACTION", "shape": [ 6 ] } }, "device": "cuda", "use_amp": false, "use_peft": false, "push_to_hub": true, "repo_id": "tsangb34/molmoact2-so101-stack-cubes-40k", "private": false, "tags": [ "molmoact2", "so101", "robocolosseum", "stack-cubes" ], "license": null, "pretrained_path": null, "pretrained_revision": null, "checkpoint_path": "allenai/MolmoAct2", "checkpoint_revision": null, "checkpoint_force_download": false, "chunk_size": 32, "n_action_steps": 32, "action_mode": "continuous", "inference_action_mode": "continuous", "discrete_action_tokenizer": "allenai/MolmoAct2-FAST-Tokenizer", "discrete_generation_max_steps": null, "norm_tag": null, "setup_type": "single so100/so101 robotic arm in molmoact2", "control_mode": "absolute joint pose", "image_keys": [ "observation.images.front", "observation.images.wrist" ], "normalize_language": true, "add_setup_tokens": true, "add_control_tokens": true, "normalize_gripper": true, "num_state_tokens": 256, "max_sequence_length": null, "expected_max_action_dim": 32, "num_flow_timesteps": 8, "flow_matching_cutoff": 1.0, "flow_matching_time_offset": 0.001, "flow_matching_time_scale": 0.999, "flow_matching_beta_alpha": 1.0, "flow_matching_beta_beta": 1.5, "num_inference_steps": null, "mask_action_dim_padding": true, "enable_inference_cuda_graph": true, "per_episode_seed": false, "eval_seed": null, "rtc_config": null, "joint_signs": null, "joint_offsets": null, "enable_lora_vlm": false, "lora_rank": 64, "lora_alpha": 16, "lora_dropout": 0.05, "lora_bias": "none", "enable_lora_action_expert": false, "enable_knowledge_insulation": false, "freeze_embedding": true, "train_action_expert_only": false, "gradient_checkpointing": true, "model_dtype": "bfloat16", "softmax_auxiliary_loss": true, "softmax_auxiliary_loss_scale": 0.0001, "discrete_loss_token_weighting": "root_subsegments_root_tokens", "optimizer_lr": 1e-05, "optimizer_vit_lr": 5e-06, "optimizer_connector_lr": 5e-06, "optimizer_action_expert_lr": 5e-05, "optimizer_betas": [ 0.9, 0.95 ], "optimizer_eps": 1e-06, "optimizer_weight_decay": 0.0, "optimizer_grad_clip_norm": 1.0, "scheduler_warmup_steps": 200, "scheduler_decay_steps": 40000, "scheduler_decay_lr": 1e-06, "normalization_mapping": { "ACTION": "QUANTILES", "STATE": "QUANTILES", "VISUAL": "IDENTITY" }, "dataset_feature_names": { "action": [ "shoulder_pan.pos", "shoulder_lift.pos", "elbow_flex.pos", "wrist_flex.pos", "wrist_roll.pos", "gripper.pos" ], "observation.state": [ "shoulder_pan.pos", "shoulder_lift.pos", "elbow_flex.pos", "wrist_flex.pos", "wrist_roll.pos", "gripper.pos" ] } }, "reward_model": null, "output_dir": "/scratch/Projects/CFP-05/CFP05-CF-002/robocolosseum-finetuned-models/molmoact2-so101-stack-cubes-40k", "job_name": "molmoact2-so101-stack-cubes-40k", "resume": false, "seed": 1000, "cudnn_deterministic": false, "num_workers": 4, "batch_size": 16, "prefetch_factor": 4, "persistent_workers": true, "dataloader_multiprocessing_context": "spawn", "steps": 40000, "env_eval_freq": 20000, "log_freq": 20, "eval_steps": 0, "max_eval_samples": 0, "tolerance_s": 0.0001, "save_checkpoint": true, "save_freq": 20000, "use_policy_training_preset": true, "optimizer": { "type": "adamw", "lr": 1e-05, "weight_decay": 0.0, "grad_clip_norm": 1.0, "betas": [ 0.9, 0.95 ], "eps": 1e-06 }, "scheduler": { "type": "cosine_decay_with_warmup", "num_warmup_steps": 200, "num_decay_steps": 40000, "peak_lr": 1e-05, "decay_lr": 1e-06 }, "eval": { "n_episodes": 50, "batch_size": 50, "use_async_envs": true, "recording": false, "recording_repo_id": null, "recording_private": false }, "wandb": { "enable": true, "disable_artifact": true, "project": "RoboColosseum", "entity": "tsangb34-national-university-of-singapore-students-union", "notes": null, "run_id": "molmoact2-so101-stack-cubes-40k", "mode": null, "add_tags": true }, "peft": null, "job": { "target": null, "image": "huggingface/lerobot-gpu:latest", "timeout": "2d", "detach": false, "tags": [] }, "save_checkpoint_to_hub": true, "sample_weighting": null, "rename_map": {}, "checkpoint_path": null }