This repository is the machine learning codebase as described in the following guide, which contains good practices for a machine learning researcher to structure day-to-day work.
- Clone the repo and create a conda environment by running:
conda env create. - Run training with
python run_training.py --config experiments/cifar.yml. This will download CIFAR10 data in a new folder./cifar10/datasetand save the experiment outputs in./cifar10/experiments/.
For a new project, create a new trainer class in the trainers folder and implement the abstract methods of the general
Trainer class. See trainers/trainer_cifar.py for a detailed example.