jm_pytorch_tutorials

basic pytorch tutorials

Purpose

A crash course introduction for the Tjian/Darzacq Lab into computer vision and representation learning in PyTorch. A basic (MNIST training/eval and feature map intro), intermediate (CIFAR10 training/eval and feature maps), and advanced introduction (Various pretrained Imagenet1000 models inference on new images, feature maps, attention maps) are provided.

Clone repo from source

git clone https://github.com/jkmckenna/jm_pytorch_tutorials

Enter repo

cd jm_pytorch_tutorials

Make an environment from the environment.yaml file

conda env create -f environment.yml

Activate the environment

conda activate pytorch-tutorials

Register the kernel after creating the first time

python -m ipykernel install --user --name pytorch-tutorials --display-name "Python (PyTorch Tutorials)"

Install the jm_pytorch_tutorials python package in editable mode

pip install -e .

If updating the environment.yaml file later, run:

conda env update -f environment.yml --prune