Optimize Gallery UI and enable Ctrl+V image pasting

Update the PyTorch version to CUDA 12.8 in the requirements.

feat: Add option to force PNG output for best quality

Fixed the handling behavior when face_model is set to None.

Upgraded Gradio version to 5.25.1.

Added support for using gallery as an image input.

Improve RRDB and ESRGAN model loading methods.

Add two option parameters: Face eye dist threshold and Face detection only center.

Overall improvements with added support for various face restoration models and upscale models.

1. Change model download from the wget command to the requests module functionality.

Update README.md

initial commit