root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2025-01-13 12:10:18 +00:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

no more transpose A

Browse Source
This commit is contained in:
ngxson 2024-07-08 16:48:55 +02:00
parent 79e2982788
commit 712fecba61
1 changed files with 1 additions and 2 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

3
src/llama.cpp
View File

@ -7867,10 +7867,9 @@ static struct ggml_tensor * llm_build_lora_mm(
continue;
}
// TODO: check if lora_a need transpose
struct ggml_tensor * a = ggml_cont(ctx0, ggml_transpose(ctx0, lora->a));
struct ggml_tensor * ab_cur = ggml_mul_mat(
ctx0, lora->b,
ggml_mul_mat(ctx0, a, cur)
ggml_mul_mat(ctx0, lora->a, cur)
);
ab_cur = ggml_scale_inplace(ctx0, ab_cur, scale);
res = ggml_add(ctx0, res, ab_cur);