]> git.djapps.eu Git - pkg/ggml/sources/whisper.cpp/commit
vulkan: support mixed/deepseekR1 FA head sizes (llama/14509)
authorJeff Bolz <redacted>
Thu, 3 Jul 2025 18:21:14 +0000 (13:21 -0500)
committerGeorgi Gerganov <redacted>
Sat, 12 Jul 2025 16:23:56 +0000 (19:23 +0300)
commita432929d585993de8ee8ca97a5d2e7f43e05756e
tree314ecbf5d0f78d8f45d524233460ebcfb79d0f68
parent4aaf8114e7953715ef9880463d0de46dd294e13d
vulkan: support mixed/deepseekR1 FA head sizes (llama/14509)

* vulkan: better parameterize FA by head sizes

* vulkan: support mixed/deepseekR1 FA head sizes
ggml/src/ggml-vulkan/ggml-vulkan.cpp
ggml/src/ggml-vulkan/vulkan-shaders/flash_attn.comp
ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_base.comp
ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm1.comp
ggml/src/ggml-vulkan/vulkan-shaders/flash_attn_cm2.comp