Amol Kapoor

Hey Jiahui, Thanks for being so responsive on this thread about deepfillv2. Quick question: you mention in the paper that your model has 4.1M parameters. Is that just the generator...

A quick follow up. In the paper you describe using a coarse network and a refinement network from your previous work. Is the coarse network trained separately/has its own loss...

So in DeepfillV2, both the final output of the overall network _and_ the output of the coarse network get a pixelwise loss?

This may be a different bug, but figured I'd revive this thread. In my case, it seems like pip-check-reqs is returning _every_ package (used or not) as an extra requirement....

Any update on this? It's a pretty low key change, would love to get this merged.

Adding my voice to the chorus: my team has been stuck for a few days now because `prettier --check` is failing on Github Actions for files that we explicitly run...

By 'prettier throws errors' I mean that prettier --check indicates that there are problems. A diff would provide a starting point to find patterns for changes. For example, I used...

Maybe the output of the prettier check should include a note that says 'if you're still seeing errors, make sure your version is correct'? But either way I think the...

Hey @lmcinnes, was curious what your latest thoughts are on doing semi-supervised clustering with HDBSCAN. Is the method you laid out above still the best path? Also, I notice that...

Actually it looks like the version being distributed by pypi is not the version currently in this git repo. It would be great if that could be updated!