Matthew Rocklin

Serialization is also maybe something that we could ask for help from the RAPIDS folks like @quasiben @jakirkham @pentschev . It's not RAPIDS obviously, but these are often GPU related...

Current pytorch serialization is here: https://github.com/dask/distributed/blob/master/distributed/protocol/torch.py It looks like it forces things to numpy though, and so may not be GPU-optimized. Rather than scikeras I'm still curious if we can...

It looks like this has gone a bit stale. @stsievert @TomAugspurger is this still worth pursuing?

test_basic is genuinely hanging locally for me

I don't have a clear memory of this, but my guess is that these were causing some error that I didn't have time to track down at the time. If...

It looks like scikit-image uses an `io` module?

Is there a good reason to avoid retries in `HTTPFileSystem`?

A link from @ian-r-rose about how Jupyter handles translation: https://jupyter.readthedocs.io/en/latest/contributing/docs-contributions/doc-new-translation.html

Yes, to me this notebook seems perhaps overly-specific to a single use case. I'm having trouble finding ways to generalize this notebook to other situations. I think that a general...

Although in general in many of these cases I think that we can improve them just by expanding Blockwise and HighLevelGraph operator fusion out to data access operations On Fri,...