ikawrakow/ik_llama.cpp
1
0
Fork 0
mirror of https://github.com/ikawrakow/ik_llama.cpp.git synced 2026-04-29 19:01:47 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix rtr when mqkv is enabled (#971)

Browse Source 
Co-authored-by: Iwan Kawrakow <iwan.kawrakow@gmail.com>
This commit is contained in:
Kawrakow
2025-11-16 16:51:45 +02:00
committed by GitHub
parent eafa77c412
commit dffb45d44a
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/llama.cpp
View File

@@ -1907,9 +1907,9 @@ static bool llm_load_tensors(
for (auto& it : model.tensors_by_name) { for (auto& it : model.tensors_by_name) {
if (ggml_backend_buffer_is_host(it.second->buffer)) { if (ggml_backend_buffer_is_host(it.second->buffer)) {
auto orig_type = it.second->type; auto orig_type = it.second->type;
if (it.second->view_src) continue;
iqk_repack_tensor(it.second); iqk_repack_tensor(it.second);
if (it.second->type != orig_type) ++n_repacked; if (it.second->type != orig_type) ++n_repacked;
//printf("Repacking tensor %s\n", it.first.c_str());
} }
} }
if (n_repacked > 0) printf("============ Repacked %d tensors\n", n_repacked); if (n_repacked > 0) printf("============ Repacked %d tensors\n", n_repacked);