### Set up Conda Environment and Install Repositories

Source: https://github.com/robocasa/robocasa-gr1-tabletop-tasks/blob/main/README.md

Follow these steps to set up a conda environment and install the necessary repositories and dependencies for RoboCasa GR-1 Tabletop Tasks.

```bash
conda create -c conda-forge -n robocasa python=3.10
conda activate robocasa

git clone https://github.com/NVIDIA/Isaac-GR00T.git
cd Isaac-GR00T
pip install --upgrade setuptools
pip install -e .[base]
pip install --no-build-isolation flash-attn==2.7.1.post4 
cd ..

git clone https://github.com/ARISE-Initiative/robosuite.git
pip install -e robosuite

git clone https://github.com/robocasa/robocasa-gr1-tabletop-tasks.git
pip install -e robocasa-gr1-tabletop-tasks

cd robocasa-gr1-tabletop-tasks
python robocasa/scripts/download_tabletop_assets.py -y
```

--------------------------------

### Run Inference Server for Isaac-GR00T

Source: https://github.com/robocasa/robocasa-gr1-tabletop-tasks/blob/main/README.md

Starts the inference server for Isaac-GR00T. Ensure you replace `<MODEL_PATH>` with the actual path to your model.

```bash
cd Isaac-GR00T
python3 scripts/inference_service.py --server \
    --model_path <MODEL_PATH> \
    --data_config fourier_gr1_arms_waist
```

--------------------------------

### Demo All Tasks

Source: https://github.com/robocasa/robocasa-gr1-tabletop-tasks/blob/main/README.md

Run this command to demo any available task by sending random actions to the robot and rendering an egocentric video.

```bash
cd robocasa-gr1-tabletop-tasks
python3 robocasa/scripts/demo_task.py <TASK_NAME>
```

--------------------------------

### Playback Demonstration Trajectories

Source: https://github.com/robocasa/robocasa-gr1-tabletop-tasks/blob/main/README.md

Use this script to replay demonstration trajectories from the GR00T Teleop Simulation Dataset in simulation, rendering robot actions and camera observations.

```bash
cd robocasa-gr1-tabletop-tasks
python3 robocasa/scripts/playback_dataset.py --dataset <HDF5_FILE> --n 1
```

--------------------------------

### Finetune GR00T Model with Custom Datasets

Source: https://github.com/robocasa/robocasa-gr1-tabletop-tasks/blob/main/README.md

This script finetunes the GR00T model using specified dataset paths and training configurations. Ensure `YOUR_DATASET_PATH_1` and `YOUR_DATASET_PATH_2` are replaced with actual paths. The `--tune-visual` flag can be omitted if visual encoder tuning is not desired.

```bash
#!/bin/bash
ALL_DATASET_PATHS=(
  "YOUR_DATASET_PATH_1"
  "YOUR_DATASET_PATH_2"
)

python scripts/gr00t_finetune.py \
  --dataset-path "${ALL_DATASET_PATHS[@]}" \
  --num-gpus 8 --batch-size 60 --learning_rate 3e-5 \
  --output-dir OUTPUT_DIR \
  --data-config fourier_gr1_arms_waist --embodiment_tag gr1 \
  --tune-visual \
  --max-steps 30000 --save-steps 5000
```

--------------------------------

### List of GR1 Tabletop Task Names

Source: https://github.com/robocasa/robocasa-gr1-tabletop-tasks/blob/main/README.md

A comprehensive list of available task names for GR1 tabletop manipulation environments.

```bash
gr1_unified/PnPCupToDrawerClose_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PnPPotatoToMicrowaveClose_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PnPMilkToMicrowaveClose_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PnPBottleToCabinetClose_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PnPWineToCabinetClose_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PnPCanToDrawerClose_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromCuttingboardToBasketSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromCuttingboardToCardboardboxSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromCuttingboardToPanSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromCuttingboardToPotSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromCuttingboardToTieredbasketSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromPlacematToBasketSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromPlacematToBowlSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromPlacematToPlateSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromPlacematToTieredshelfSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromPlateToBowlSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromPlateToCardboardboxSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromPlateToPanSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromPlateToPlateSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromTrayToCardboardboxSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromTrayToPlateSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromTrayToPotSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromTrayToTieredbasketSplitA_GR1ArmsAndWaistFourierHands_Env
gr1_unified/PosttrainPnPNovelFromTrayToTieredshelfSplitA_GR1ArmsAndWaistFourierHands_Env
```

--------------------------------

### Run Simulation Evaluation for Isaac-GR00T

Source: https://github.com/robocasa/robocasa-gr1-tabletop-tasks/blob/main/README.md

Executes the simulation evaluation script for a single task with specified parameters. Replace `<TASK_NAME>` with the desired task.

```bash
cd Isaac-GR00T
python3 scripts/simulation_service.py --client \
    --env_name <TASK_NAME> \
    --video_dir ./videos \
    --max_episode_steps 720 \
    --n_envs 5 \
    --n_episodes 10
```

=== COMPLETE CONTENT === This response contains all available snippets from this library. No additional content exists. Do not make further requests.