### Install Dependencies

Source: https://github.com/bytedance/infiniteyou/blob/main/README.md

Run this command to install all necessary Python packages. It is recommended to do this within a Python virtual environment.

```bash
pip install -r requirements.txt
```

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

### Launch Local Gradio Demo

Source: https://github.com/bytedance/infiniteyou/blob/main/README.md

Start the local Gradio demo application by running the app.py script. This provides a user-friendly interface for interacting with the model.

```bash
python app.py
```

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

### Combine CPU Offloading and 8-bit Quantization

Source: https://github.com/bytedance/infiniteyou/blob/main/README.md

Specify both `--cpu_offload` and `--quantize_8bit` arguments in test.py to further reduce peak VRAM usage to approximately 16GB with similar performance.

```python
--cpu_offload
--quantize_8bit
```

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

### Run Local Inference Script

Source: https://github.com/bytedance/infiniteyou/blob/main/README.md

Execute the local inference script with specified image, prompt, and output directory. Ensure all arguments are correctly provided for desired results.

```bash
python test.py --id_image ./assets/examples/man.jpg --prompt "A man, portrait, cinematic" --out_results_dir ./results
```

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

### Enable CPU Offloading for Reduced VRAM

Source: https://github.com/bytedance/infiniteyou/blob/main/README.md

Specify the `--cpu_offload` argument in test.py to reduce peak VRAM usage to approximately 30GB without performance loss.

```python
--cpu_offload
```

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

### Enable 8-bit Quantization for Reduced VRAM

Source: https://github.com/bytedance/infiniteyou/blob/main/README.md

Specify the `--quantize_8bit` argument in test.py to reduce peak VRAM usage to approximately 24GB with similar performance.

```python
--quantize_8bit
```

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