PyTorch intro
Lab
Read over (or run if you need practice) notebooks 1-Basics. Run all of notebook 2-NN basics and workflows. Rerun notebook 2 and make the task a bit more difficult by:
adding Gaussian standard noise to the data. (w/ at least sigma=0.01)
adjust training hyperparameters as necessary to achieve good results.
For extra credit run notebook 3. If you do not have access to a GPU, use free tier Google collab GPU with instructions outlined in https://flexie.github.io/CSE-8803-Twin/schedule.html.
Depending on your execution environment (local machine vs. Google Colab), you must configure PyTorch to use the appropriate GPU backend (CUDA for NVIDIA GPUs, MPS for Apple Silicon) in the Torch_test_GPU_CPU.ipynb notebook.