A standard template for any PyTorch project 🔥
Includes boilerplate code for: training, evaluation, transformation, model checkpointing etc.
Use config.yaml for constants or global variables in general.
To setup dependencies and train:
virtualenv -p python3 venv
source venv/bin/activate
pip install -r requirements.txt
python -m train- The features are ..
- Setup training and validation pipeline ..
- Perform image augmentation ..
Add evaluations and details of experiments
| Experiment | Description | Hyperparameters | Accuracy | F1 score |
|---|