https://github.com/lmamon/vitransformer
GHW machine learning workshop using CIFAR10
https://github.com/lmamon/vitransformer
computer-vision machine-learning
Last synced: 6 months ago
JSON representation
GHW machine learning workshop using CIFAR10
- Host: GitHub
- URL: https://github.com/lmamon/vitransformer
- Owner: LMamon
- Created: 2024-11-17T01:43:05.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-08-21T05:09:09.000Z (11 months ago)
- Last Synced: 2025-08-21T07:27:13.658Z (11 months ago)
- Topics: computer-vision, machine-learning
- Language: Jupyter Notebook
- Homepage: http://www.cs.toronto.edu/~kriz/cifar.html
- Size: 177 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Vision Transformer (ViT) on CIFAR-10
deprecated and moved to HuggingFace https://huggingface.co/roylvzn/vit-cifar10
This repository contains the implementation of a Vision Transformer (ViT) trained on the CIFAR-10 dataset. This project was developed during the MLH Global Hack Week as an introduction to leveraging transformer models for computer vision tasks.
## Features
### Vision Transformer (ViT):
Implementation of a ViT from scratch.
Training and evaluation on the CIFAR-10 dataset.
### Deep Learning Frameworks:
PyTorch used for model creation, training, and evaluation.
## CIFAR-10 Dataset:
Used to benchmark the ViT's performance on image classification tasks.
## Reproducible Results:
Code for loading, training, and testing the model.
Pre-trained model checkpoint included for quick testing.
## Usage
1. **Clone the repository**:
```bash
git clone
3. **Navigate to the project directory**:
```bash
cd ViT-CIFAR10
3. **Run the Jupyter Notebook**:
```bash
jupyter notebook ViTransformer.ipynb
4. **For standalone scripts**:
Train the model: Use vit_model.py.
Load and test a pre-trained model: Use load_vit_model.py.
## Learning Objectives
Understand the architecture and implementation of Vision Transformers.
Learn to preprocess and train on CIFAR-10 using PyTorch.
Experiment with transformer-based models for image classification.
## Results
Accuracy: Achieved 93% accuracy with the CIFAR-10 test set.
Visualization: Includes attention maps and classification results.
## Contributing
Contributions are welcome! Please open an issue or submit a pull request with your improvements.
## License
This project is licensed under the MIT License. See the LICENSE file for details.