FasterLivePortrait
Real time make the portrait move! Support onnx/sensorrt
- Real time running speed of 30+FPS on RTX 3090 GPU using TensorRT
- Convert LivePortrait model to Onnx model, supporting cross platform deployment
- Support native Gradio app, improve inference speed and support simultaneous inference of multiple faces
- Refactoring code structure, no longer relying on PyTorch, using onnx or tensorrt for inference
- Support Docker environment and provide runnable images
- Supports Windows and MacOS integration packages, supports one click operation
- Supports onnxruntime and TensorRT inference, providing detailed installation and usage guidelines
Product Details
FasterLivePortrait is a real-time portrait animation project based on deep learning. It achieves a real-time running speed of 30+FPS on RTX 3090 GPU using TensorRT, including pre-processing and post-processing, rather than just model inference speed. The project also implemented the conversion of LivePortrait model to Onnx model and achieved an inference speed of approximately 70ms/frame on RTX 3090 using onnxruntime gpu, supporting cross platform deployment. In addition, the project also supports native Gradio app, which increases speed several times and supports simultaneous inference of multiple faces. The code structure has been restructured and no longer relies on PyTorch. All models use onnx or tensorrt for inference.
