# Custom CNN Model for Image Classification

This repository contains a custom ResNet-like CNN trained for image classification. The model is provided in both PyTorch (`pytorch_model.bin`) and Safetensors (`model.safetensors`) formats for easy inference and transfer.

## Usage

- Load model weights and architecture from Hugging Face Hub.
- Preprocess images using the provided mean/std and img_size in `preprocessor_config.json`.
- See `train.py` for full training and inference pipeline.

## Files
- `pytorch_model.bin`: PyTorch weights
- `model.safetensors`: Safetensors weights
- `config.json`: Model config
- `preprocessor_config.json`: Preprocessing config
- `train.py`: Source code
- `prediction.csv`: Example predictions

## Citation
If you use this model, please cite the repository: https://huggingface.co/DungND1107/IntroAIver2