vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-11 01:24:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'master' into concedo_experimental

Browse source
This commit is contained in:
Concedo 2023-11-10 22:27:11 +08:00
commit a6e6b8b96b
14 changed files with 3105 additions and 3624 deletions
Download patch file Download diff file Expand all files Collapse all files

2
llama.cpp
View file

@ -4239,7 +4239,7 @@ struct llm_build_context {
struct ggml_tensor * Kcur = ggml_concat(ctx0, krotated, kpass);
cb(Kcur, "Kcur", il);
struct ggml_tensor * Q = ggml_cont(ctx0, ggml_permute(ctx0, Qcur, 1, 2, 0, 3));
struct ggml_tensor * Q = ggml_cont(ctx0, ggml_permute(ctx0, Qcur, 2, 1, 0, 3));
cb(Q, "Q", il);
Kcur = ggml_cont(ctx0, ggml_permute(ctx0, Kcur, 2, 1, 0, 3));