gradio-llama-cpp-2

Sleeping

pngwn HF staff commited on May 11

Commit

d1a85cc

•

1 Parent(s): 48db1b5

Update app.py

Files changed (1) hide show

app.py CHANGED Viewed

@@ -11,7 +11,7 @@ llm = Llama(
         filename=os.environ.get("MODEL_FILE", "Phi-3-mini-4k-instruct-q4.gguf"),
     ),
     n_ctx=2048,
-    n_gpu_layers=50, # change n_gpu_layers if you have more or less VRAM
 )

         filename=os.environ.get("MODEL_FILE", "Phi-3-mini-4k-instruct-q4.gguf"),
     ),
     n_ctx=2048,
+    n_gpu_layers=-1, # change n_gpu_layers if you have more or less VRAM
 )