Valter Hudovernik

Will the weights for the bottom up backbone be released?

1

Hello, great work on both the ViTPose and + papers. I have already used it in a project of mine and cited your work. I would like to know if...

[Re] Learning with Noisy Labels Revisited: A Study Using Real-World Human Annotations

13

**Original article:** [Learning with Noisy Labels Revisited: A Study Using Real-World Human Annotations](https://openreview.net/pdf?id=TBWA6PLJZQm) **PDF URL:** https://github.com/KlemenVovk/noisy-labels/blob/master/rescience/rescience.pdf **Metadata URL:** https://github.com/KlemenVovk/noisy-labels/blob/master/rescience/metadata.yaml **Code URL:** https://github.com/KlemenVovk/noisy-labels **Openreview:** https://openreview.net/forum?id=GKZ2leags0 **Scientific domain:** Machine Learning **Programming language:**...

Adds a keyword argument `metric` to the `oob_dropcol_importances` method to support passing a custom metric to the method. Closes #58

The default implementation of `oob_dropcol_importances` sets the `oob_score` attribute to true. This results the method always returning the default ($R^2$) score. Adding a keyword argument for metric of interest would...

Limiting pyarrow version to avoid integration issues with other packages due to conflicting versions of libarrow. Based on discussion on this issue (https://github.com/apache/arrow/issues/44696).

make `coalesce` compile friendly

2

Update coalesce to be torch compiled without graph breaks.