tweak

app.py

@@ -8,7 +8,7 @@ from huggingface_hub import hf_hub_download
 # huggingface-cli download LoneStriker/OpenBioLLM-Llama3-8B-GGUF --local-dir ./llama3-gguf
 llm = Llama(
     # model_path="./Phi-3-mini-4k-instruct-q4.gguf",
-    model_path="LoneStriker/OpenBioLLM-Llama3-8B-GGUF",
+    model_path="LoneStriker/OpenBioLLM-Llama3-8B-Q5_K_M.gguf",
     n_ctx=2048,
     n_gpu_layers=50, # change n_gpu_layers if you have more or less VRAM 
 )