happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

Update src/llama-model.cpp 

Co-authored-by: Sigbjørn Skjæret <sigbjorn.skjaeret@scala.com>
This commit is contained in:
Ryan Mangeno 2025-10-10 12:16:38 -04:00 committed by GitHub
parent edbe4d2c2f
commit 1f54cf4b73
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 3 additions and 7 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

10
src/llama-model.cpp
View File

@ -8071,13 +8071,9 @@ struct llm_build_modern_bert : public llm_graph_context {
const size_t type_size = ggml_type_size(cur->type);
Qcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd, n_tokens, cur->nb[1], 0*type_size*(n_embd)));
Kcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd, n_tokens, cur->nb[1], 1*type_size*(n_embd)));
Vcur = ggml_cont(ctx0, ggml_view_2d(ctx0, cur, n_embd, n_tokens, cur->nb[1], 2*type_size*(n_embd)));
Qcur = ggml_reshape_3d(ctx0, Qcur, n_embd_head, n_head, n_tokens);
Kcur = ggml_reshape_3d(ctx0, Kcur, n_embd_head, n_head_kv, n_tokens);
Vcur = ggml_reshape_3d(ctx0, Vcur, n_embd_head, n_head_kv, n_tokens);
Qcur = ggml_view_3d(ctx0, cur, n_embd_head, n_head, n_tokens, n_embd_head*type_size, cur->nb[1], 0*type_size*(n_embd));
Kcur = ggml_view_3d(ctx0, cur, n_embd_head, n_head_kv, n_tokens, n_embd_head*type_size, cur->nb[1], 1*type_size*(n_embd));
Vcur = ggml_view_3d(ctx0, cur, n_embd_head, n_head_kv, n_tokens, n_embd_head*type_size, cur->nb[1], 1*type_size*(n_embd + n_embd_gqa));
// RoPE
Qcur = ggml_rope_ext(