]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 6b5c2ef)
sycl: fix UE4M3 parsing (#25608)
authorChris Lee <redacted>
Fri, 7 Aug 2026 05:28:53 +0000 (23:28 -0600)
committerGitHub <redacted>
Fri, 7 Aug 2026 05:28:53 +0000 (08:28 +0300)
commitfc3f10b3895ebb0ddfe1fcb7fd5950f2c1719339
tree2e55d77a31afecfdd07f9dd464a763f51bd5932dtree
parent6b5c2efb4e2f43e6bdd07eaacb25789eea294983commit | diff
sycl: fix UE4M3 parsing (#25608)

The NVFP4 quantization format stores a scaling factor for every group of
16 weights, packed into a single UE4M3 byte.

The SYCL GPU code was converting these scale values using the E4M3 path,
but that's *signed*, and these are unsigned values.
ggml/src/ggml-sycl/common.hpp diff | blob | history
Packaging of ggml-org/llama.cpp