Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-18 16:55:14 +02:00
Code Issues Packages Projects Releases Wiki Activity

disable smartcache if slots is zero

Browse Source
This commit is contained in:
Concedo
2026-03-10 08:57:31 +08:00
parent 2bd6b87d5b
commit 500a1ab466
1 changed files with 6 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
gpttype_adapter.cpp
+6 -3
View File
@@ -2597,9 +2597,12 @@ ModelLoadResult gpttype_load_model(const load_model_inputs inputs, FileFormat in
//if RNN model AND shifting and fastforward is on, enable smartcache
if((llama_model_is_recurrent(llamamodel) || llama_model_is_hybrid(llamamodel)) && kcpp_data->use_fastforward && kcpp_data->use_contextshift)
{
printf("RNN or Hyrbid model with FF and shifting flags enabled - SmartCache will be enabled with extra slots. Disable CtxShift if you do not want this.\n",savestate_limit);
kcpp_data->smartcache = true;
savestate_limit += 3;
if(savestate_limit>0)
{
printf("RNN or Hyrbid model with FF and shifting flags enabled - SmartCache will be enabled with extra slots. Disable CtxShift if you do not want this.\n",savestate_limit);
kcpp_data->smartcache = true;
savestate_limit += 3;
}
}
savestates.resize(savestate_limit);
if(kcpp_data->smartcache)