Nicolas Hug

Hi @Chen-Bo-Yang , thanks for the report. I am not sure why `pin_memory` would affect the result, but we can explain why setting `num_workers > 0` would produce different resuts:...

Thanks for the feature request @yggdrasil75 I don't think we'll support JXL in torchvision, as we'll soon be migrating the image decoders into TorchCodec. We'll keep JXL in mind when...

Hi @lfos , we have deprecated the torchvision C++ decoder part, and from 0.24 which we'll release in a few weeks, it will stop being built by default.

Hi @MHRDYN7 , thank you for the feature request. I think this may make sense, but adding a new interpolation mode is *a ton* of work, and it would have...

Hi @taozhiwei , can you please share more on your use case for this?

@taozhiwei can you please share more about the use-cases you have in mind? We currently support CPU and CUDA decoding, but CUDA decoding is explicitly enabled by passing the `device`...

Thanks for opening this. > 2. multiplying some sample weight by N is equivalent to repeating the corresponding samples N times To me, this is the core definition of what...

If we consider that 2. is the definition of what SW are, I'm not sure that 3. is something we can deduce for all estimators. Some models are actually sensitive...

That would indeed "fix" the issue I pointed out, but using a percentage wouldn't be a practical parameter. The reason `min_sample_leaf` exists is for users to say "I want the...

We should also discuss how subsampling the training set behaves w.r.t sample weights. This is relevant for any estimator that subsamples the training set, and particularly for random forests with...