27 days to go!
You are encouraged to use Layer.ai to track your experiments.
In order to win you need to have used Layer and code submissions need to be supported by a Layer repo and report.
You can learn more about using Layer for machine learning, and sign up for Layer.ai here.
How do you create a layer repo ?
To what extent does layer need to be used?
Is simply doing some logging in the code enough and not using layer it for final model training?
Hi. I work at Layer.
It's ok to use layer only for logging. However it will still require you to init layer project (with layer.init) and then decorate your functions (with @dataset or @model). You can only log things to layer from layer decorated functions.