### Run Web UI Server

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Starts the web server for the video2robot UI. Access it via a web browser.

```bash
uvicorn web.app:app --host 0.0.0.0 --port 8000
```

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

### Configure Google API Key

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Copies the example environment file and sets the Google API key for services like Veo.

```bash
cp .env.example .env
echo "GOOGLE_API_KEY=your-api-key" >> .env
```

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

### Run Full Pipeline from Existing Video

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Starts the pipeline from an existing video file to generate robot motion.

```bash
python scripts/run_pipeline.py --video /path/to/video.mp4
```

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

### Create and Activate GMR Conda Environment

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Sets up the conda environment for GMR (robot retargeting) and installs the project.

```bash
conda create -n gmr python=3.10 -y
conda activate gmr
pip install -e .
```

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

### Run Individual Pipeline Step: Generate Video

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Generates a video based on an action description.

```bash
python scripts/generate_video.py --action "Action sequence: The subject walks forward."
```

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

### Clone Repository and Initialize Submodules

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Clone the video2robot repository and ensure all submodules are initialized.

```bash
git clone --recursive https://github.com/AIM-Intelligence/video2robot.git
cd video2robot

git submodule update --init --recursive
```

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

### Run Full Pipeline from Action Prompt

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Executes the complete video2robot pipeline, converting an action description into robot motion. A base prompt is automatically applied.

```bash
python scripts/run_pipeline.py --action "Action sequence:
The subject walks forward with four steps."
```

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

### Run Full Pipeline Using Sora Provider

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Executes the full pipeline, specifying Sora as the video generation provider.

```bash
python scripts/run_pipeline.py --action "..." --provider sora
```

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

### Run Individual Pipeline Step: Extract Pose

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Extracts human pose information from a video project.

```bash
python scripts/extract_pose.py --project data/video_001
```

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

### Visualize Pipeline Results with Robot Motion Overlay

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Visualizes the pipeline results, including an overlay of the generated robot motion.

```bash
python scripts/visualize.py --project data/video_001 --robot
```

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

### Resume Pipeline from Existing Project

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Resumes a previously interrupted pipeline process using an existing project directory.

```bash
python scripts/run_pipeline.py --project data/video_001
```

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

### Run Individual Pipeline Step: Convert to Robot Motion

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Converts extracted pose information into robot motion data.

```bash
python scripts/convert_to_robot.py --project data/video_001
```

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

### Create and Activate PromptHMR Conda Environment for Blackwell GPUs

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Sets up the conda environment for PromptHMR (pose extraction) on Blackwell GPUs.

```bash
conda create -n phmr python=3.11 -y
conda activate phmr
cd third_party/PromptHMR
bash scripts/install_blackwell.sh
```

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

### Create and Activate PromptHMR Conda Environment for Other GPUs

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Sets up the conda environment for PromptHMR (pose extraction) on other GPU types.

```bash
conda create -n phmr python=3.10 -y
conda activate phmr
cd third_party/PromptHMR
pip install -e .
```

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

### Visualize Pipeline Results with Pose Overlay

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Visualizes the pipeline results, including an overlay of the extracted pose.

```bash
python scripts/visualize.py --project data/video_001 --pose
```

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

### Robot Motion Output Format

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Defines the structure of the output robot motion data, saved as a pickle file.

```python
# robot_motion.pkl
{
    "fps": 30.0,
    "robot_type": "unitree_g1",
    "num_frames": 240,
    "root_pos": np.ndarray,    # (N, 3)
    "root_rot": np.ndarray,    # (N, 4) quaternion xyzw
    "dof_pos": np.ndarray,     # (N, DOF)
}
```

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

### Visualize Pipeline Results

Source: https://github.com/aim-intelligence/video2robot/blob/main/README.md

Visualizes the results of the pipeline. The script automatically switches to the appropriate conda environment.

```bash
python scripts/visualize.py --project data/video_001
```

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