conda create -n ONNX python=3.8
conda activate ONNX
conda install pytorch torchvision torchaudio pytorch-cuda=11.6 -c pytorch -c nvidia
pip install onnxruntime-gpu==1.14.0
pip install opencv-python==4.5.5.64
- Real-time face detection using ONNX model.
- Emotion recognition from detected faces using another ONNX model.
- Visualization of emotion scores on the video stream.
- Place the downloaded models in the weights directory
Download | weight |
---|
$ python demo.py
- This repo supports only inference, see reference for more details