Train ResNet18 on AFAD dataset for gender and age estimation with Pytorch
Environment
python==3.6.9
Pytorch==1.7.0
onnx==1.7.0
onnxruntime==1.5.2
opencv-python
Getting Started
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.