- /data: folder containg datasets used for benchmark tasks
- /notebooks: notebooks to run and test code written in /src
- /src: source code that will be autograded in GitHub
Unit tests will be run on code contained in /src folder. Nothing in /notebooks will be graded
- In
src/main.py
, fill out thecreate_model
function. This function should initialize and return the model that you want to use in the benchmarks. - Ensure that your model runs before submitting. Run
python benchmark_job.py
to double check that your model works. - Push your code to submit. The benchmarks will run in a GitHub action. You can submit as many times as you want. Your latest submission will appear on the leaderboard.
- After a few minutes, check the leaderboard site to see your results.
From the command line in the root directory run:
python3.9 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt