CIFAR10_UCI_CS273A_Final_Project

Winter 2022 UCI CS273A Machine Learning Final Project

Prerequisite

• An NVIDIA GPU
• Enable GPU with docker
• Familiar with Pytorch

Our Enviroment

• Hardware: NVIDIA GEFORCE RTX 3060

Run the program

Step 1. Build the image

cd cifar10-uci-cs273a-final-project
docker build . -t final

Step 2. Run the container with gpu and mount all project to the container

docker run --gpus all -v $(pwd)/.:/source/. -it final /bin/bash

Step 3. Inside the container, run train.py

python3 train.py

We implement 4 different deep learning neaural network: basic CNN, RestNet, GoogleNet, VGGNet to do the image classfication for CIFAR10. Please check the source code for further imformation.

Step 4. Evaluate each models with test images

python3 test.py

Useful Links

• Get Docker
• How to Use the GPU within a Docker Container
• NVIDIA Container Toolkit
• Docker Images with Pytorch
• Pytorch Tutorial

Reference

• pytorch-cifar10
• pytorch-cifar
• CIFAR10-img-classification-tensorflow