James Lamb

> `TabularPredictor.fit(..., num_cpus=..., num_gpus=...)` (global control) Thank you for this! I suspected that what I was doing with those per-model-type overrides was overcomplicated 😅 I think I did not find...

> *Is this also part of this task?* I think yes, because that section only updates the `[test]` extra for `dask-cuda` wheels, which wouldn't be used until we're running wheel...

I love this! I have wanted tab-per-language tutorials like this in LightGBM for a while, never took the time to investigate it. I'd love to adopt something like what you...

Tonight, I see a similar issue to https://github.com/scipy/scipy/issues/21623#issuecomment-2372240678 trying to *download* some (but not all) `scipy` nightly wheels. (ref: https://github.com/microsoft/LightGBM/pull/6663#issuecomment-2394945653) steps to reproduce: 1. Navigate to https://pypi.anaconda.org/scientific-python-nightly-wheels/simple/scipy 2. click on...

I just tried https://github.com/scipy/scipy/issues/21623#issuecomment-2394945406 again today and did not have any issues 😁

@hcho3 let me start by saying I am SO SORRY it took over a month to get back to you! Thanks for redirecting me here from #10803. I've read this...

I think that's a good path forward! Thanks for writing that out.

/ok to test 37e3ff2393c89760348e601c247e954bd450e2a3

/ok to test b5ff465d85af3a5722b73a3829a2e9d0b91ba2fa

/ok to test https://github.com/rapidsai/integration/pull/764/commits/aa3c1f507af2dbaf149573afdae9bab650effbe3