For this repository, I implemented the ResNet paper in PyTorch from scratch without referencing any prior implementations. I trained the model on the full CIFAR-10 dataset.
Before cloning the repository, ensure that you have PyTorch installed, preferably in a conda environment (see here for setup details).
To set up this repository, first clone it on your local machine. Then, navigate to CIFAR-10 webpage and downlaod "CIFAR-10 python version" (163 MB). Put the files data_batch_* and test_batch into the same folder as resnet.ipynb, or alternatively modify the file names in the notebook to properly load these files.
After properly loading the data, you should be able to run all of the blocks in resnet.ipynb.