Sid Ravinutala

Hi all - We came across the same issue using gensim 3.8.3 and numpy 1.22.2 I removed numpy from `requirements.txt` (it gets installed as a dependency by other packages anyway)...

Hi - Is this PR still being worked on? Happy to take over if needed. I just swapped out pyemd for POT and it improved performance by around a factor...

Thanks @TLouf! One other thing we may wish to do - there is a function provided by POT to calculate [distances](https://pythonot.github.io/all.html?highlight=dist#ot.dist). It falls back on [scipy's cdist](https://pythonot.github.io/_modules/ot/utils.html#dist) if metric is...

> Switch to it in which places [This line](https://github.com/RaRe-Technologies/gensim/blob/a435f24fe25e17f473e71af3468660512c2606cb/gensim/models/keyedvectors.py#L990) to `ot.dist`. > in pursuit of what advantage? First, since we are already using POT for `emd`, we might as well...

Hi all - Can this be merged? I'd love to use this from the dev branch. Thanks for your work on this.

> Is this for metrics tracking so that the Danswer admin of a team can know how often an answer is good/bad? Yes - this would be the immediate use...

This [PR](https://github.com/danswer-ai/danswer/pull/238) allows for the 1st one.

Thanks! I'll let you know if I find time to put together a PR. It probably won't be till mid-Sept.