I'm receiving error with JPG (uppercase of jpg) image extension. This extension should be supported right?

ERROR Error resizing: Error: Expected one of: heic, heif, jpeg, jpg, png, raw, tiff, webp for format but received JPG of type string
at Object.invalidParameterError (/var/task/node_modules/sharp/lib/is.js:101:10)
at Sharp.toFormat (/var/task/node_modules/sharp/lib/output.js:168:

There is a set of Pixel Level transforms that is used in the work Benchmarking Neural Network Robustness to Common Corruptions and Perturbations

The authors also share the code => we can absorb some transforms that they have into the library.

https://github.com/hendrycks/robustness/blob/master/ImageNet-C/create_c/make_imagenet_c.py

• ShotNoise
• Defocus
• GlassBlur

A follow up on SixLabors/ImageSharp#1378 (comment).

Currently 32 bit test execution is only done for .NET Framework, with dotnet xunit which is an obsolete tool today, we need to adapt dotnet test, and add 32 bit CI targets for both net5.0 and netcoreapp3.1. Opening an issue to remember and track this debt.

See:
https://twitter.com/tannergooding/s

Description

There was a badass paper last year that unified and generalized Otsu and other thresholding methods ([arXiv]).

As usual, since it's a new algorithm the argument can be made that we should wait to see it in practice, but my gut feeling is that (a) it's really good, and (b) it's a generalization of an existing algorithm that we have in the library, so it does not add much code,

In kornia.color.rgb_to_hsv, the computation of s is leading to nan value due to division by 0 when variable v contains 0.
s: torch.Tensor = deltac / v # saturation
I don t know what is the best solution to solve this. On my local machine, I add a small eps=1e-7 as a workaround:
s: torch.Tensor = deltac / (v + 1e-7) # saturation