David Gilbertson

Thanks for your enthusiasm! My performance decisions are all data-based. So if the `App interactive (ms)` in the console doesn't change by a mile then everything is fine. The focus...

Ah, that's helpful, thanks @albanD, I didn't think to try a 0D tensor.

Ah, I've re-read through the parameter lists for everything and found `remove_unused_columns=False` in `TrainingArguments`. Setting this resolves the issue, so I guess this won't be considered a bug. I think...

Sorry, I've got a full plate at the moment.

I appreciate the reluctance to 'move fast and break things' - nice to see :) As a user finding his way around the Hugging Face packages, it did strike me...

I have two options when it comes to score inheritance (that I can think of). # Option 1 - be loose Let people propagate any of the four scores down...

That's not very good, thanks for logging (and thanks for the versions). Does it crash immediately or after some expanding/scoring? Is dev tools open by any chance?

Does this mean that a chart like one isn't possible with Altair?  https://vega.github.io/vega-lite-v4/docs/text.html#properties That's from V4 but uses `"params"` in the chart spec which as far as I can...

OK cool, thanks!

Yes I was surprised to find "persist to disk" as an option that defaulted to off. My setup was taking 10 minutes to index on every start (on a i7...