tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-08-18 05:56:00 -04:00
Code Issues Packages Projects Releases Wiki Activity

Add --no-op-offload to improve -ot pp perf in MoE models like llama4 400B (#13386)

Browse Source
This commit is contained in:
David Huang
2025-05-11 20:18:39 +08:00
committed by GitHub
parent 3eac209319
commit 7f323a589f
11 changed files with 57 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
common/common.cpp
View File

@@ -1113,6 +1113,7 @@ struct llama_context_params common_context_params_to_llama(const common_params &
cparams.offload_kqv = !params.no_kv_offload;
cparams.flash_attn = params.flash_attn;
cparams.no_perf = params.no_perf;
cparams.op_offload = !params.no_op_offload;
if (params.reranking) {
cparams.embeddings = true;