Nicolas Hug comments

Results 570 comments of


                                            Nicolas Hug

[WIP] Package torchvision with libjpeg-turbo instead of libjpeg

https://github.com/pytorch/vision/pull/7820

Evaluate models accuracy when using inference transforms on Tensors (instead of PIL images)

Closed by https://github.com/pytorch/vision/pull/7949 (background: https://github.com/pytorch/vision/issues/7093)

Evaluate models accuracy when using inference transforms on Tensors (instead of PIL images)

Hi @CA4GitHub , It's not about dtype,, it's really about antialiasing. In fact even if you passed a `uint8` tensor image it would be converted to `float` internally, resized, and...

decode_jpeg() creates tensors with different stride from PIL.Image.open()

I'll close this, a lot has happened since this issue was opened. We should still be concerned about the input / output format of the transforms / decoders, but this...

NOCATS: Categorical splits for tree-based learners (ctnd.)

@adrinjalali do you need an in-depth review at this point? If yes I'd be happy to try. Also, for core devs: let's say this PR is reviewed and accepted with...

NOCATS: Categorical splits for tree-based learners (ctnd.)

Thanks both for the feedback. @adrinjalali sure, LMK if there's anything I can help with!

NOCATS: Categorical splits for tree-based learners (ctnd.)

It seems that the `BitSet` class is just a wrapper on a uint64 right? And you dont' need to use it inside Python? In this case I would just directly...

NOCATS: Categorical splits for tree-based learners (ctnd.)

@adrinjalali considering that we decided not to implement categorical support in the `tree` module, I think we can close this one and #4899 ?

NOCATS: Categorical splits for tree-based learners (ctnd.)

> Even if we also implement the NOCATS approach, we could provide an estimator that does the "once and for all" ordering for regression and binary classification. What we plan...

NOCATS: Categorical splits for tree-based learners (ctnd.)

Also I'm super late to the party, but what is the benefit of NOCATs over One-Hot-Encoding the categories? As far as I understand the strategy proposed here is equivalent to...