root/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 19:34:35 +00:00
Code Issues Actions 4 Packages Projects Releases Wiki Activity

fix auto merge

Browse Source
This commit is contained in:
ngxson 2024-07-07 16:36:50 +02:00
parent a1666aaaca
commit 30faf1f3de
1 changed files with 3 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

6
src/llama.cpp
View File

@ -10863,7 +10863,7 @@ struct llm_build_context {
// special-case: the up and gate tensors are merged into a single tensor // special-case: the up and gate tensors are merged into a single tensor
// TOOD: support into llm_build_ffn // TOOD: support into llm_build_ffn
{ {
cur = llm_build_ffn(ctx0, cur, cur = llm_build_ffn(lctx, ctx0, cur,
model.layers[il].ffn_up, NULL, NULL, model.layers[il].ffn_up, NULL, NULL,
NULL, NULL, NULL, NULL, NULL, NULL,
model.layers[il].ffn_down, NULL, NULL, model.layers[il].ffn_down, NULL, NULL,
@ -13622,7 +13622,7 @@ struct llm_build_context {
); );
cb(Kcur, "Kcur_rope", il); cb(Kcur, "Kcur_rope", il);
cur = llm_build_kv(ctx0, model, hparams, cparams, kv_self, gf, cur = llm_build_kv(lctx, ctx0, model, hparams, cparams, kv_self, gf,
model.layers[il].wo, NULL, model.layers[il].wo, NULL,
Kcur, Vcur, Qcur, KQ_mask, n_tokens, kv_head, n_kv, 1.0f/sqrtf(float(n_embd_head)), cb, il); Kcur, Vcur, Qcur, KQ_mask, n_tokens, kv_head, n_kv, 1.0f/sqrtf(float(n_embd_head)), cb, il);
@ -13647,7 +13647,7 @@ struct llm_build_context {
LLM_NORM_RMS, cb, il); LLM_NORM_RMS, cb, il);
cb(cur, "ffn_norm", il); cb(cur, "ffn_norm", il);
cur = llm_build_ffn(ctx0, cur, cur = llm_build_ffn(lctx, ctx0, cur,
model.layers[il].ffn_up, NULL, NULL, model.layers[il].ffn_up, NULL, NULL,
NULL, NULL, NULL, NULL, NULL, NULL,
model.layers[il].ffn_down, NULL, NULL, model.layers[il].ffn_down, NULL, NULL,