Andrew Yates

Any update on this? Given that PyPi packages can't declare dependencies on the git repo directly, it would be very helpful to make a new release.

Besides the test issue (hopefully just due to black), is this ready to merge?

Thanks for pointing this out. I don't remember why we avoid using the dict from hgf's `tokenizer` class directly, but this is something we should look into in the future...

PR #102 avoids this issue by default, but we should still document to avoid unnecessary slowdowns

I think adding type annotations is a good idea. This is something we considered but haven't prioritized, so any contributions adding them would be welcome. Just let us know if...

Hi Martin, Thank you for pointing this out. While Sean is looking into the third question (#21), I'll try to provide some information about the others. > ## Experiment 1:...

> Hi Andrew, thanks for your quick and detailed reply! > > > Some background: this repository is a simplified version of a (in-house) toolkit called srank, which is what...

> One more question regarding > > > In this setting, I get P@20 = 0.4167 and nDCG@20 = 0.4826 with CEDR-KNRM, which are higher than the metrics you obtained....

I've spent some time looking into the reproduction issues from a different angle by implementing CEDR-KNRM in a toolkit [1] known to work well with Transformer-based models like PARADE [2]....

Could you commit the HTML scraping code somewhere for reference? From what I understand, there's no way around the fact that private tweets and deleted tweets aren't available.