This repository is a template for AI research projects.

All environment configurations are provided in requirements.txt. To install all packages, you can create a conda environment and install the packages as follows:

conda create -n NAME python=3.8
conda activate NAME
pip3 install torch torchvision torchaudio
pip install -r requirements.txt

The config job_type is used to specify the type of job, which can be 'train', 'debug', or 'test'. There are train.sh and test.sh in the scripts folder. You can run the following command to train or test the model:

bash scripts/train.sh
bash scripts/test.sh

Remember to modify the ckpt_path in the test.sh before evaluation.

You can use the config resume to specify the path of checkpoint to resume training.

If you find this project helpful, please consider star or fork this repository.

Star, fork and follow me for more projects!