switched to hf_hub_download

backend.py

@@ -23,7 +23,7 @@ hf_hub_download(
 )
 
 llm = Llama(
-            model_path=f"models/{"2b_it_v2.gguf"}",
+            model_path=f"models/2b_it_v2.gguf",
             #flash_attn=True,
             #_gpu_layers=81,
             n_batch=1024,