otherhalf-dev
/

llama-3.2-1B-trt

Model card Files Files and versions Community

otherhalf-dev commited on Feb 5

Commit

0dd86cd

·

verified ·

1 Parent(s): 7567662

Update tensorrt_llm/config.pbtxt

update model path and kv cache memory utilization

Files changed (1) hide show

tensorrt_llm/config.pbtxt +2 -2

tensorrt_llm/config.pbtxt CHANGED Viewed

@@ -382,7 +382,7 @@ parameters: {
 parameters: {
   key: "gpt_model_path"
   value: {
-    string_value: "/all_models/inflight_batcher_llm/tensorrt_llm/1"
   }
 }
 parameters: {
@@ -418,7 +418,7 @@ parameters: {
 parameters: {
   key: "kv_cache_free_gpu_mem_fraction"
   value: {
-    string_value: "0.1"
   }
 }
 parameters: {

 parameters: {
   key: "gpt_model_path"
   value: {
+    string_value: "/data/llama/model_repository/tensorrt_llm/1"
   }
 }
 parameters: {
 parameters: {
   key: "kv_cache_free_gpu_mem_fraction"
   value: {
+    string_value: "0.9"
   }
 }
 parameters: {