root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 11:24:35 +00:00
Code Issues Actions 2 Packages Projects Releases Wiki Activity

Fixed incorrectly applying RMS norm twice (#1925)

Browse Source
This commit is contained in:
Johannes Gäßler 2023-06-18 16:07:09 +02:00 committed by GitHub
parent 8596af4277
commit 0ede372a51
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 1 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
llama.cpp
View File

@ -1657,11 +1657,7 @@ static bool llama_eval_internal(
{ {
cur = ggml_rms_norm(ctx0, inpL); cur = ggml_rms_norm(ctx0, inpL);
offload_func_nr(cur); offload_func_nr(cur);
ggml_set_name(cur, "rms_norm_inpL"); ggml_set_name(cur, "rms_norm_2");
cur = ggml_rms_norm(ctx0, cur);
offload_func_nr(cur);
ggml_set_name(cur, "rms_norm_after");
// cur = cur*norm(broadcasted) // cur = cur*norm(broadcasted)
cur = ggml_mul(ctx0, cur, model.norm); cur = ggml_mul(ctx0, cur, model.norm);