ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

llama : fix incorrect Qwen2Moe ffn_moe_out graph callback (#12631)

Browse source
This commit is contained in:
Sigbjørn Skjæret 2025-03-28 22:13:02 +01:00 • committed by GitHub
parent b4ae50810e
commit 3714c3ee1a
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 1 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

2
src/llama-model.cpp
View file

@ -6323,7 +6323,7 @@ struct llm_build_qwen2moe : public llm_graph_context {
false, 0.0,
LLAMA_EXPERT_GATING_FUNC_TYPE_SOFTMAX,
il);
cb(cur, "ffn_moe_out", il);
cb(moe_out, "ffn_moe_out", il);
// FFN shared expert
{