Notebooks covering PyTorch fundamentals: tensors, neural networks, training loops, and image classification. Each part has an Exercises notebook and a Solution notebook (Part 6 has a single notebook).
Create the conda environment from environment.yml (PyTorch, torchvision and the packages the notebooks import):
conda env create -f environment.yml
conda activate pytorch| File | Description |
|---|---|
dl_pytorch_Part 1 - Tensors in PyTorch (Exercises / Solution).ipynb |
Tensor operations and autograd basics |
dl_pytorch_Part 2 - Neural Networks in PyTorch (Exercises / Solution).ipynb |
Building a feedforward network |
dl_pytorch_Part 3 - Training Neural Networks (Exercises / Solution).ipynb |
Training loop with loss and optimizer |
dl_pytorch_Part 4 - Fashion-MNIST (Exercises / Solution).ipynb |
Fashion-MNIST classification |
dl_pytorch_Part 5 - Inference and Validation (Exercises / Solution).ipynb |
Inference and validation |
dl_pytorch_Part 6 - Saving and Loading Models.ipynb |
Model persistence |
dl_pytorch_Part 7 - Loading Image Data (Exercises / Solution).ipynb |
Loading image data |
dl_pytorch_Part 8 - Transfer Learning (Exercises / Solution).ipynb |
Transfer learning |
dl_pytorch_fc_model.py |
Reusable fully-connected model class |
dl_pytorch_helper.py |
Helper functions |