]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/commit
overview / pkg / ggml / sources / llama.cpp / commit
summary | shortlog | log | commit | commitdiff | tree
(parent: 5c4d767)
llama : add gguf_remove_key + remove split meta during quantize (#6591)
authorjiez <redacted>
Fri, 12 Apr 2024 10:45:06 +0000 (18:45 +0800)
committerGitHub <redacted>
Fri, 12 Apr 2024 10:45:06 +0000 (13:45 +0300)
commit91c736015b66ba1d0b82cbae6313b6d5eaa61b68
tree098b60b95e78a1062daf0fe2b362de506eb23df7tree
parent5c4d767ac028c0f9c31cba3fceaf765c6097abfccommit | diff
llama : add gguf_remove_key + remove split meta during quantize (#6591)

* Remove split metadata when quantize model shards

* Find metadata key by enum

* Correct loop range for gguf_remove_key and code format

* Free kv memory

---------

Co-authored-by: z5269887 <redacted>
ggml.c diff | blob | history
ggml.h diff | blob | history
llama.cpp diff | blob | history
Packaging of ggml-org/llama.cpp