redo the test

tikikun · tikikun · commit a1c6b973419c · 2023-12-10T19:01:34.000+07:00
diff --git a/controllers/llamaCPP.cc b/controllers/llamaCPP.cc
@@ -399,7 +399,7 @@ bool llamaCPP::loadModelImpl(const Json::Value &jsonBody) {
   }
 #ifdef GGML_USE_CUBLAS
   LOG_INFO << "Setting up GGML CUBLAS PARAMS";
-  params.mul_mat_q = true;
+  params.mul_mat_q = false;
 #endif // GGML_USE_CUBLAS
   if (params.model_alias == "unknown") {
     params.model_alias = params.model;

Original file line number	Diff line number	Diff line change
`@@ -399,7 +399,7 @@ bool llamaCPP::loadModelImpl(const Json::Value &jsonBody) {`
`399`	`399`	`}`
`400`	`400`	`#ifdef GGML_USE_CUBLAS`
`401`	`401`	`LOG_INFO << "Setting up GGML CUBLAS PARAMS";`
`402`		`- params.mul_mat_q = true;`
	`402`	`+ params.mul_mat_q = false;`
`403`	`403`	`#endif // GGML_USE_CUBLAS`
`404`	`404`	`if (params.model_alias == "unknown") {`
`405`	`405`	`params.model_alias = params.model;`