zoexiao0516 / capstone_2023_part_a

Overview

This git repo consists of code for experimenting with various CNN architectures to analyze representations from different learning rules.

Env Set-up

conda create -n capstone_2023 python=3.10
conda activate capstone_2023
pip install -r requirements.txt

Dataset

We are using the ILSVRC 2012 dataset, also known as the 'ImageNet 2012 dataset'. The data size is dreadfully large (138G!), but this amount of dataset is required for successful training of NN.

Training images

There is a tar file for each synset, named by its WNID. The image files are named as x_y.JPEG, where x is WNID of the synset and y is an integer (not fixed width and not necessarily consecutive). All images are in JPEG format.

There are a total of 1281167 images for training. The number of images for each synset ranges from 732 to 1300

Validation images

*update later

Testing images

*update later

The images vary in dimensions and resolution. Many applications resize / crop all of the images to 256x256 pixels.

ImageNet 2012's dataset structure is already arranged as /root/[class]/[img_id].jpeg, so using torchvision.datasets.ImageFolder is convenient.