This project performs checks on the ResNeXt model: Aggregated residual transformations for deep neural networks

The resnet.pytorch-directory has been checked out from another gitlab repository (with slight custom modifications to adapt to custom use): https://github.com/prlz77/ResNeXt.pytorch

============================================

Running the jupyter notebook ResNeXt_tests_ABurger.ipynb needs the TestFigures/-directory and the resnext.pytorch- directory.

GoogleColabNotebook/ResNeXt_TrainOnFlower-102_Data_ABurger.ipynb has been tested on google colab and runs without any local files.

============================================

The jupyter notebook ResNeXt_tests_ABurger.ipynb loads a pre-trained pytorch model link and performs some basic tests:

• load the model
• have the pre-trained model classify some random images
• Test the content of the flower-102 dataset
• Training of the model on the flowers-102 dataset is performed using the script GoogleColabNotebook/ResNeXt_TrainOnFlower-102_Data_ABurger.ipynb on the google colab environment (not in this notebook)
• The logs of the trained models have been saved in resnext.pytorch/logs/. This notebook runs plotting scripts taking the files in resnext.pytorch/logs/ as input.

The jupyter notebook GoogleColabNotebook/ResNeXt_TrainOnFlower-102_Data_ABurger.ipynb trains and tests the ResNeXt model using the flowers-102 dataset.