]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 14b1d7e)
ggml : support CUDA's half type for aarch64(#1455) (#2670)
authorKylin <redacted>
Tue, 22 Aug 2023 07:14:23 +0000 (15:14 +0800)
committerGitHub <redacted>
Tue, 22 Aug 2023 07:14:23 +0000 (10:14 +0300)
commit1e3bc523d8053a77df3ac7126a84d0297ee97ef6
tree4fcee8fbe2d776a99d44a4c6beee727e858c5b6atree
parent14b1d7e6f720dee41ce5a826376df738096d9033commit | diff
ggml : support CUDA's half type for aarch64(#1455) (#2670)

* ggml: support CUDA's half type for aarch64(#1455)
support CUDA's half type for aarch64 in ggml_fp16_t definition

* ggml: use __CUDACC__ to recognise nvcc compiler
ggml.h diff | blob | history
Packaging of ggml-org/llama.cpp