[robosuite WARNING] No private macro file found! (macros.py:57) [robosuite WARNING] It is recommended to use a private macro file (macros.py:58) [robosuite WARNING] To setup, run: python /home/seonho/clvla/benchmarks/robocasa_v2/robosuite/robosuite/scripts/setup_macros.py (macros.py:59) [robosuite WARNING] Could not import robosuite_models. Some robots may not be available. If you want to use these robots, please install robosuite_models from source (https://github.com/ARISE-Initiative/robosuite_models) or through pip install. (__init__.py:30) [robosuite WARNING] Could not load the mink-based whole-body IK. Make sure you install related import properly (e.g. pip install mink==0.0.5), otherwise you will not be able to use the default IK controller setting for GR1 robot. (__init__.py:40) /home/seonho/miniconda3/envs/robocasa/lib/python3.11/site-packages/albumentations/__init__.py:13: UserWarning: A new version of Albumentations is available: 2.0.8 (you have 1.4.18). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1. check_for_updates() `use_fast` is set to `True` but the image processor class does not have a fast version. Falling back to the slow version. ***************************************** Setting OMP_NUM_THREADS environment variable for each process to be 1 in default, to avoid your system being overloaded, please further tune the variable for optimal performance in your application as needed. ***************************************** [robosuite WARNING] No private macro file found! (macros.py:57) [robosuite WARNING] It is recommended to use a private macro file (macros.py:58) [robosuite WARNING] To setup, run: python /home/seonho/clvla/benchmarks/robocasa_v2/robosuite/robosuite/scripts/setup_macros.py (macros.py:59) [robosuite WARNING] No private macro file found! (macros.py:57) [robosuite WARNING] It is recommended to use a private macro file (macros.py:58) [robosuite WARNING] To setup, run: python /home/seonho/clvla/benchmarks/robocasa_v2/robosuite/robosuite/scripts/setup_macros.py (macros.py:59) [robosuite WARNING] Could not import robosuite_models. Some robots may not be available. If you want to use these robots, please install robosuite_models from source (https://github.com/ARISE-Initiative/robosuite_models) or through pip install. (__init__.py:30) [robosuite WARNING] Could not load the mink-based whole-body IK. Make sure you install related import properly (e.g. pip install mink==0.0.5), otherwise you will not be able to use the default IK controller setting for GR1 robot. (__init__.py:40) [robosuite WARNING] Could not import robosuite_models. Some robots may not be available. If you want to use these robots, please install robosuite_models from source (https://github.com/ARISE-Initiative/robosuite_models) or through pip install. (__init__.py:30) [robosuite WARNING] Could not load the mink-based whole-body IK. Make sure you install related import properly (e.g. pip install mink==0.0.5), otherwise you will not be able to use the default IK controller setting for GR1 robot. (__init__.py:40) WARNING: mimicgen environments not imported since mimicgen is not installed! WARNING: mimicgen environments not imported since mimicgen is not installed! /home/seonho/miniconda3/envs/robocasa/lib/python3.11/site-packages/albumentations/__init__.py:13: UserWarning: A new version of Albumentations is available: 2.0.8 (you have 1.4.18). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1. check_for_updates() /home/seonho/miniconda3/envs/robocasa/lib/python3.11/site-packages/albumentations/__init__.py:13: UserWarning: A new version of Albumentations is available: 2.0.8 (you have 1.4.18). Upgrade using: pip install -U albumentations. To disable automatic update checks, set the environment variable NO_ALBUMENTATIONS_UPDATE to 1. check_for_updates() `use_fast` is set to `True` but the image processor class does not have a fast version. Falling back to the slow version. `use_fast` is set to `True` but the image processor class does not have a fast version. Falling back to the slow version. ================================================== GR00T FINE-TUNING CONFIGURATION: ================================================== config: /home/seonho/clvla/benchmarks/robocasa_v2/experiment_cfg/processing_line_only/mgd_0.3_skip.yaml dataset_soup: None output_dir: /tmp/gr00t output_root: None data_config: panda_omron batch_size: 64 max_steps: 300000 num_gpus: 2 save_steps: 20000 run_name: None save_total_limit: 100 seed: 42 base_model_path: nvidia/GR00T-N1.5-3B tune_llm: False tune_visual: False tune_projector: True tune_diffusion_model: True resume: False learning_rate: 3e-05 weight_decay: 1e-05 warmup_ratio: 0.05 lora_rank: 0 lora_alpha: 16 lora_dropout: 0.1 lora_full_model: False dataloader_num_workers: 8 report_to: wandb embodiment_tag: new_embodiment video_backend: opencv balance_dataset_weights: True balance_trajectory_weights: True ds_weights_alpha: 0.4 ================================================== Using 2 GPUs ================================================================================ Starting sweep branch: mgd_token_mask_ratio_0.3 Sweep vars: {'mgd_token_mask_ratio': 0.3} ================================================================================ -------------------------------------------------------------------------------- Running phase 1: phase3_mgd_fm Policy type: groot_mgd Base model path: /home/seonho/groot_robocasa/robocasa_v2/MGD_Processing_line_Only/mgd_token_mask_ratio_0.3/phase2 Output dir: /home/seonho/groot_robocasa/robocasa_v2/MGD_Processing_line_Only/mgd_token_mask_ratio_0.3/phase3 Trainable preset: None Policy overrides: {'mgd_enabled': True, 'mgd_fm_loss_weight': 1.0, 'mgd_loss_weight': 1.0, 'mgd_token_mask_ratio': 0.3, 'mgd_use_cosine_loss': True, 'mgd_use_mse_loss': False} -------------------------------------------------------------------------------- [{'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/CloseBlenderLid/20250822/lerobot', 'horizon': 900, 'filter_key': '100_demos', 'task': 'CloseBlenderLid', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/CloseFridge/20250819/lerobot', 'horizon': 900, 'filter_key': '100_demos', 'task': 'CloseFridge', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/CloseToasterOvenDoor/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'CloseToasterOvenDoor', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/CoffeeSetupMug/20250819/lerobot', 'horizon': 600, 'filter_key': '100_demos', 'task': 'CoffeeSetupMug', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/NavigateKitchen/20250821/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'NavigateKitchen', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/OpenCabinet/20250819/lerobot', 'horizon': 1050, 'filter_key': '100_demos', 'task': 'OpenCabinet', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/OpenDrawer/20250819/lerobot', 'horizon': 750, 'filter_key': '100_demos', 'task': 'OpenDrawer', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/OpenStandMixerHead/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'OpenStandMixerHead', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceCounterToCabinet/20250819/lerobot', 'horizon': 750, 'filter_key': '100_demos', 'task': 'PickPlaceCounterToCabinet', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceCounterToStove/20250819/lerobot', 'horizon': 600, 'filter_key': '100_demos', 'task': 'PickPlaceCounterToStove', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceDrawerToCounter/20250820/lerobot', 'horizon': 750, 'filter_key': '100_demos', 'task': 'PickPlaceDrawerToCounter', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceSinkToCounter/20250819/lerobot', 'horizon': 900, 'filter_key': '100_demos', 'task': 'PickPlaceSinkToCounter', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceToasterToCounter/20250819/lerobot', 'horizon': 600, 'filter_key': '100_demos', 'task': 'PickPlaceToasterToCounter', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceCabinetToCounter/20250819/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'PickPlaceCabinetToCounter', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOnElectricKettle/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'TurnOnElectricKettle', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOnMicrowave/20250819/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'TurnOnMicrowave', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOffMicrowave/20250819/lerobot', 'horizon': 300, 'filter_key': '100_demos', 'task': 'TurnOffMicrowave', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/OpenElectricKettleLid/20250820/lerobot', 'horizon': 300, 'filter_key': '100_demos', 'task': 'OpenElectricKettleLid', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/SlideDishwasherRack/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'SlideDishwasherRack', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOnSinkFaucet/20250819/lerobot', 'horizon': 600, 'filter_key': '100_demos', 'task': 'TurnOnSinkFaucet', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnSinkSpout/20250820/lerobot', 'horizon': 300, 'filter_key': '100_demos', 'task': 'TurnSinkSpout', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOffSinkFaucet/20250819/lerobot', 'horizon': 300, 'filter_key': '100_demos', 'task': 'TurnOffSinkFaucet', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/AdjustWaterTemperature/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'AdjustWaterTemperature', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOffStove/20250819/lerobot', 'horizon': 750, 'filter_key': '100_demos', 'task': 'TurnOffStove', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOnStove/20250819/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'TurnOnStove', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/AdjustToasterOvenTemperature/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'AdjustToasterOvenTemperature', 'split': 'pretrain', 'source': 'human'}] Using 100 subset demos for filter_key: 100_demos /home/seonho/clvla/benchmarks/robocasa_v2/Isaac-GR00T/gr00t/data/transform/state_action.py:257: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.detach().clone() or sourceTensor.detach().clone().requires_grad_(True), rather than torch.tensor(sourceTensor). self.statistics[key] = torch.tensor(value) Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos ================================================== GR00T FINE-TUNING CONFIGURATION: ================================================== config: /home/seonho/clvla/benchmarks/robocasa_v2/experiment_cfg/processing_line_only/mgd_0.3_skip.yaml dataset_soup: None output_dir: /tmp/gr00t output_root: None data_config: panda_omron batch_size: 64 max_steps: 300000 num_gpus: 2 save_steps: 20000 run_name: None save_total_limit: 100 seed: 42 base_model_path: nvidia/GR00T-N1.5-3B tune_llm: False tune_visual: False tune_projector: True tune_diffusion_model: True resume: False learning_rate: 3e-05 weight_decay: 1e-05 warmup_ratio: 0.05 lora_rank: 0 lora_alpha: 16 lora_dropout: 0.1 lora_full_model: False dataloader_num_workers: 8 report_to: wandb embodiment_tag: new_embodiment video_backend: opencv balance_dataset_weights: True balance_trajectory_weights: True ds_weights_alpha: 0.4 ================================================== Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 2 GPUs Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT ================================================================================ Starting sweep branch: mgd_token_mask_ratio_0.3 Sweep vars: {'mgd_token_mask_ratio': 0.3} ================================================================================ -------------------------------------------------------------------------------- Running phase 1: phase3_mgd_fm Policy type: groot_mgd Base model path: /home/seonho/groot_robocasa/robocasa_v2/MGD_Processing_line_Only/mgd_token_mask_ratio_0.3/phase2 Output dir: /home/seonho/groot_robocasa/robocasa_v2/MGD_Processing_line_Only/mgd_token_mask_ratio_0.3/phase3 Trainable preset: None Policy overrides: {'mgd_enabled': True, 'mgd_fm_loss_weight': 1.0, 'mgd_loss_weight': 1.0, 'mgd_token_mask_ratio': 0.3, 'mgd_use_cosine_loss': True, 'mgd_use_mse_loss': False} -------------------------------------------------------------------------------- [{'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/CloseBlenderLid/20250822/lerobot', 'horizon': 900, 'filter_key': '100_demos', 'task': 'CloseBlenderLid', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/CloseFridge/20250819/lerobot', 'horizon': 900, 'filter_key': '100_demos', 'task': 'CloseFridge', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/CloseToasterOvenDoor/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'CloseToasterOvenDoor', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/CoffeeSetupMug/20250819/lerobot', 'horizon': 600, 'filter_key': '100_demos', 'task': 'CoffeeSetupMug', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/NavigateKitchen/20250821/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'NavigateKitchen', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/OpenCabinet/20250819/lerobot', 'horizon': 1050, 'filter_key': '100_demos', 'task': 'OpenCabinet', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/OpenDrawer/20250819/lerobot', 'horizon': 750, 'filter_key': '100_demos', 'task': 'OpenDrawer', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/OpenStandMixerHead/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'OpenStandMixerHead', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceCounterToCabinet/20250819/lerobot', 'horizon': 750, 'filter_key': '100_demos', 'task': 'PickPlaceCounterToCabinet', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceCounterToStove/20250819/lerobot', 'horizon': 600, 'filter_key': '100_demos', 'task': 'PickPlaceCounterToStove', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceDrawerToCounter/20250820/lerobot', 'horizon': 750, 'filter_key': '100_demos', 'task': 'PickPlaceDrawerToCounter', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceSinkToCounter/20250819/lerobot', 'horizon': 900, 'filter_key': '100_demos', 'task': 'PickPlaceSinkToCounter', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceToasterToCounter/20250819/lerobot', 'horizon': 600, 'filter_key': '100_demos', 'task': 'PickPlaceToasterToCounter', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/PickPlaceCabinetToCounter/20250819/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'PickPlaceCabinetToCounter', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOnElectricKettle/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'TurnOnElectricKettle', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOnMicrowave/20250819/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'TurnOnMicrowave', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOffMicrowave/20250819/lerobot', 'horizon': 300, 'filter_key': '100_demos', 'task': 'TurnOffMicrowave', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/OpenElectricKettleLid/20250820/lerobot', 'horizon': 300, 'filter_key': '100_demos', 'task': 'OpenElectricKettleLid', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/SlideDishwasherRack/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'SlideDishwasherRack', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOnSinkFaucet/20250819/lerobot', 'horizon': 600, 'filter_key': '100_demos', 'task': 'TurnOnSinkFaucet', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnSinkSpout/20250820/lerobot', 'horizon': 300, 'filter_key': '100_demos', 'task': 'TurnSinkSpout', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOffSinkFaucet/20250819/lerobot', 'horizon': 300, 'filter_key': '100_demos', 'task': 'TurnOffSinkFaucet', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/AdjustWaterTemperature/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'AdjustWaterTemperature', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOffStove/20250819/lerobot', 'horizon': 750, 'filter_key': '100_demos', 'task': 'TurnOffStove', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/TurnOnStove/20250819/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'TurnOnStove', 'split': 'pretrain', 'source': 'human'}, {'path': '/home/seonho/clvla/benchmarks/robocasa_v2/robocasa/datasets/v1.0/pretrain/atomic/AdjustToasterOvenTemperature/20250820/lerobot', 'horizon': 450, 'filter_key': '100_demos', 'task': 'AdjustToasterOvenTemperature', 'split': 'pretrain', 'source': 'human'}] Using 100 subset demos for filter_key: 100_demos Using 100 subset demos for filter_key: 100_demos /home/seonho/clvla/benchmarks/robocasa_v2/Isaac-GR00T/gr00t/data/transform/state_action.py:257: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.detach().clone() or sourceTensor.detach().clone().requires_grad_(True), rather than torch.tensor(sourceTensor). self.statistics[key] = torch.tensor(value) Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT dataset weights: [1. 0.88494669 0.76443286 0.83888607 0.73785464 1.00501644 0.80029956 0.66065525 0.83967491 0.83515002 0.95047759 0.86791692 0.88329477 0.78510653 0.64381843 0.67404766 0.69697218 0.60443684 0.78448103 0.83447486 0.61146215 0.64384058 0.79545025 0.93845856 0.75517122 0.7973985 ] Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Loaded 26 datasets Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Loading pretrained dual brain from /home/seonho/groot_robocasa/robocasa_v2/MGD_Processing_line_Only/mgd_token_mask_ratio_0.3/phase2 Tune backbone vision tower: False Tune backbone LLM: False Tune action head projector: True Tune action head DiT: True Model not found or avail in the huggingface hub. Loading from local path: /home/seonho/groot_robocasa/robocasa_v2/MGD_Processing_line_Only/mgd_token_mask_ratio_0.3/phase2 Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT Using 100 subset demos for filter_key: 100_demos Initialized dataset lerobot with EmbodimentTag.NEW_EMBODIMENT dataset weights: [1. 0.88494669 0.76443286 0.83888607 0.73785464 1.00501644 0.80029956 0.66065525 0.83967491 0.83515002 0.95047759 0.86791692 0.88329477 0.78510653 0.64381843 0.67404766 0.69697218 0.60443684 0.78448103 0.83447486 0.61146215 0.64384058 0.79545025 0.93845856 0.75517122 0.7973985 ] Loaded 26 datasets Loading pretrained dual brain from /home/seonho/groot_robocasa/robocasa_v2/MGD_Processing_line_Only/mgd_token_mask_ratio_0.3/phase2 Tune backbone vision tower: False Tune backbone LLM: False Tune action head projector: True Tune action head DiT: True Model not found or avail in the huggingface hub. Loading from local path: /home/seonho/groot_robocasa/robocasa_v2/MGD_Processing_line_Only/mgd_token_mask_ratio_0.3/phase2 Tune backbone llm: False Tune backbone visual: True Total number of DiT parameters: 550386688 Tune backbone llm: False Tune backbone visual: True Total number of DiT parameters: 550386688 Total number of SelfAttentionTransformer parameters: 201433088 Tune action head projector: True Tune action head diffusion model: True Loading checkpoint shards: 0%| | 0/2 [00:00