Vensen

Fix(models/siglip): Add compatibility for Gemma models quantized by llm-compressor

4

### Description This PR resolves a `KeyError` that occurs when attempting to serve a Gemma-3 model that has been quantized by the `vllm-project/llm-compressor` library. It makes the SigLIP vision model...

[Bugfix][cache_kernels]: Fix OOB in cache_kernels.cu

1

## Purpose This PR fixes a potential out-of-bounds (OOB) memory access in the `gather_and_maybe_dequant_cache` CUDA kernel, as originally reported in Issue #27909. The bug was identified by static analysis. The...

fix(nnUNet): Correct background mask in BraTS preprocessor

1

Hi, Maintainer. This PR fixes the issue described in #1438. Problem: The preprocessing code for the BraTS nnU-Net example was using `np.where(image[i]

Fix broken links from issue #3605

3

Fixes #3605 ## Description This PR addresses the issue of broken links detected by the `github-actions` bot. It fixes three incorrect URLs in the documentation: - In `CONTRIBUTING.md`, removed an...