1

I have 115 notebooks that run the same logic on different data. One notebook takes on average 20 minutes to run. I'd like to use Jupyterbook to create a shareable/clean version of all the information together. Jupyterbook has currently been running for 3 days, and will run ~12 to finish if it is doing it sequentially (in theory it should take 20 minutes with no compute limit and parallelization).

Questions: Is there a way to run in parallel? Another way could be to run the notebooks prior and not have JupyterBook run them. Is there a third-option? Is there a better tool?

user123328
  • 63
  • 1
  • 10
  • Here is an idea instead of an answer: 1. Use `jupyter nbconvert --execute` to execute all your notebooks: https://nbconvert.readthedocs.io/en/latest/execute_api.html 2. Set `execute_notebooks: off` in _config.yml: https://jupyterbook.org/en/stable/customize/config.html 3. Build your book as normal – Richard Herron Sep 13 '22 at 13:37

0 Answers0