Fix build requirements and enable trust_remote_code

app.py

@@ -75,11 +75,12 @@ def get_pipeline(model_id):
                 "text-generation",
                 model=model_id,
                 device_map="auto",
+                trust_remote_code=True,
                 torch_dtype=torch.float16 if torch.cuda.is_available() else torch.float32,
             )
             models_cache[model_id] = pipe
         except Exception as e:
-            raise gr.Error(f"Failed to load model {model_id}: {str(e)}")
+            raise gr.Error(f"Failed to load model {model_id} locally: {str(e)}")
     return models_cache[model_id]
 
 def respond(

requirements.txt

@@ -1,6 +1,6 @@
 gradio>=5.9.1
 huggingface_hub
 python-dotenv
-audioop-lts
 transformers
 torch
+accelerate