This project aims at performing image classifications (Monkey dataset and Cars dataset) using transfer learning in deep neural networks along with TensorRT inference acceleration.
- Pytorch (GPU Version)
- TensorRT 5.x.x
Using Google Colab to run jupyter notebook files is highly recommended.
- training:
Open ResNet_152_Testing_on_Cars_Dataset.ipynb and run all. - testing:
Open ResNet_152_Train_on_Cars_Dataset.ipynb and run all.
- Save model to xx.onnx file
- python cars_152_build_engine.py to create tensorRT engine.
- python cars_152_run.py to run the inferencce acceleration.