Mathis Chenuet

From my understanding, it's the number of texts to be processed in parallel (for each example) by Transformer, so 1 for text classification, 2 for Story Cloze Test (rocstories), n...

@jtpio Yes I have no problem accessing the notebook and lab interface (we have setup a password and I'm already authenticated).

No I don't know sorry, do you have ideas on how to investigate this?

Will try to reproduce it, but it may have been issues with reporting (matplotlib) figures created with seaborn https://seaborn.pydata.org/.

This could be an issue/particularities of python multiprocessing on windows, not pyfunctional. Can you try wrapping your code in a `def main` function and calling from the `if __main__` ?

Any news on this? @ainoam, + I just pushed a related PR https://github.com/allegroai/clearml-web/pull/58

A PR was created here https://github.com/erikbern/ann-benchmarks/pull/451 but I don't know its current status.