search

Georgetown-IR-Lab / cedr

Code for CEDR: Contextualized Embeddings for Document Ranking, accepted at SIGIR 2019.
MIT License
155 stars 28 forks source link

tqdm #38

Open cmacdonald opened 3 years ago

cmacdonald commented 3 years ago

Sean, there is some peculiar tqdm setups in CEDR. Could we remove the leave=false and the ncols; perhaps set the units and some descriptions?

If you agree, I can submit a PR.

seanmacavaney commented 3 years ago

Yes, adding units and writing better descriptions would be helpful.

The leave and ncols settings are important for the usability of the tool on the command line. Removing ncols makes the tqdms really ugly (and the tool borderline unusable if you dare resize your console), and setting leave=False makes it much easier to scan the train/validation performance over time without dead tqdms getting in the way (which themselves no longer provide much/any value).

I know that ncols in notebooks produces some pretty ugly bars (it's the number of pixels, not the number of characters). What effect does leave=False have? Is there an easy way to detect that it's running in a notebook? If so, these could be set conditionally.

cmacdonald commented 3 years ago

leave=False means the progress bar dissapears after use. If tqdm is obtained in imported in one single place, it could be overwritten?

seanmacavaney commented 3 years ago

That sounds fair. So tqdm could be set with CLI defaults in CEDR, and then PyTerrier could overwrite that reference.