Sylvain Gugger

Are you sure you are using PyTorch 1.3? The type-promotion should get rid of those errors. On Linux, I can do `x==y` with a tensor of type Int and a...

The first error should be fixed now btw.

We don't want to automatically convert to int64 tensors for users because it takes twice the space in GPU memory and sometimes they don't need the int64. I can't reproduce...

No, it's not a hard-dependency because it's only needed for `doc-builder preview`. We could add an extras for preview however with this dependency.

It doesn't look like we can easily do everything `literalinclude` does in just one command. What was the most useful feature you wanted @adrinjalali ? Highlighting lines? Showing line numbers?

We already have the ability to auto-generate notebooks from all the doc files, it sounds like what you want to do is similar, we just do it from a .mdx...

A few other comments: > The same holds true for cross-library references: [`~datasets.Dataset`] will do a link to the Dataset object documented in the datasets library. @LysandreJik note that this...

Good point! We can definitely add support for this and it shouldn't be difficult to parse such a py file to generate a MDX and the corresponding notebooks. The main...

I'm not entirely sure there is a point since the package doesn't use anything fancy, but okay to add 3.10.

> Can a collection of Tools be uploaded to a single space? Our thinking was that each space would be a demo of the Tool, which doesn't work with multiple...