]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 4f31eed)
cuda : prevent integer truncation and overflow errors when using KQ mask strides...
authorfairydreaming <redacted>
Tue, 30 Jun 2026 18:47:05 +0000 (20:47 +0200)
committerGitHub <redacted>
Tue, 30 Jun 2026 18:47:05 +0000 (20:47 +0200)
commit0eca4d490e591d4e93058d07540cf47278a72577
treeaadab70264fa84010f4245c90b664ea702bb403dtree
parent4f31eedb0ccf546b7e8d6bb243b170f12522f54dcommit | diff
cuda : prevent integer truncation and overflow errors when using KQ mask strides in flash_attn_mask_to_KV_max kernel (#24945)

Co-authored-by: Stanisław Szymczyk <redacted>
ggml/src/ggml-cuda/fattn-common.cuh diff | blob | history
Packaging of ggml-org/llama.cpp