A clean and modular deep learning project built using PyTorch. This repo is designed to be beginner-friendly and scalable for more complex tasks in the future.
- Modular PyTorch pipeline
- Dataset loaders and transformations
- Custom model architectures
- Training and validation loops
- Checkpoint saving & loading
- Logging with tqdm
- GPU support
deep-learning-pytorch/ │ ├── data/ # Datasets or data loading scripts ├── models/ # Custom model architectures ├── utils/ # Helper functions (metrics, plotting, etc.) ├── checkpoints/ # Saved model weights ├── train.py # Training loop ├── evaluate.py # Evaluation script ├── config.py # Configuration (hyperparameters, paths) └── requirements.txt # All dependencies
| Model | Description |
|---|---|
| CNN | Simple convolutional network |
| MLP | Multi-layer perceptron |
| ResNet (custom) | Residual block architecture |
git clone https://github.com/Abisheck007/PyTorch_Deeplearning.git
cd deep-learning-pytorch
Hyper parameteres tuning
python train.py --lr 0.001 --epochs 20 --model mlp