How To Participate

To get started and submit your first model, you will need to pass through the following steps.

Create an account on the Crunch Foundation platform

Creating an account on the CrunchDAO platform will allow you to be identified and get access to the competition dataset. Follow the link below to join the competition.

Submit your code through the crunch-cli

  1. Clone your workspace directory in a notebook or in your terminal.

All the following steps can be found in the Notebook Quick starters.

$pip install crunch-cli --upgrade
# Your token can be found in the submit page of the competition
$crunch setup <username> --token <token> 

The website generates new tokens every minute, and each token can only be used once within a 3-minute timeframe. This prevents any issues if your token is accidentally shared. This token allows the CLI to download data and submit it on your behalf.

# Example of a folder structure.
# The data files may change depending on the competition.
$ tree
.
β”œβ”€β”€ data
β”‚Β Β  β”œβ”€β”€ X_test.parquet
β”‚Β Β  β”œβ”€β”€ X_train.parquet
β”‚Β Β  └── y_train.parquet
β”œβ”€β”€ main.py
β”œβ”€β”€ requirements.txt
└── resources

3 directories, 5 files

If you need to save some files to run your code on the Crunch Foundation's servers, like the weights of your model, the tree structure... etc., you have to save them under the resources folder.

  1. Code 😎

  2. Test your solution locally

# Run the following command locally to test your code
$crunch test
  1. Push your work

# Run the following command to push your work onto the Crunch Foundation server
$crunch push

All the work that you push remains your exclusive property. The Crunch Foundation ensures the privacy of both the client data and of the competitors’ code.

Or upload your notebook directly on the platform

You can upload directly a notebook in the submit page.

The notebooks still needs the imports, train and infer functions.

In addition, it is highly recommended to freeze the versions on your imports adding a

# == [version_number] after each import. The specified version will be installed when run on the server side and avoid libraries versions bug.

Last updated