Train ResNet18 on AFAD dataset for gender and age estimation with Pytorch
- python==3.6.9
- Pytorch==1.7.0
- onnx==1.7.0
- onnxruntime==1.5.2
- opencv-python
- Configure parameters and dataset dictionary in "config.yaml".
- Run "main.py" to train the network.
- Run "export.py" to convert the .pt model into onnx format. (a trained model with 93.14% validation gender estimation accuracy has been placed and converted already).
- Run "image_demo.py" and "video_demo.py" to test the onnx model with a single image or USB camera.