/cifar-classificiation

PyTorch codes for CIFAR-10 classification and ResNet

Primary LanguageJupyter Notebook

CNN for CIFAR-10 dataset

This repository is to practice my PyTorch for image classficiation problem.

The dataset is from https://www.cs.toronto.edu/~kriz/cifar.html

Downloaded data needs to be stored in ./data folder

Files to execute:

  1. EDA.ipynb: explore the raw data
  2. preprocess.ipynb: preprocess the raw data
  3. sanity_checks.py: sanity checks for the implementation
  4. hyper_tune.py: tune hyperparameters
  5. train.py: train the model based on best hyperparameters searched in step 4

The implemeted model here is based on the ResNet paper: Deep Residual Learning for Image Recognition [He et al. 2015]