Update app.py

app.py

@@ -69,15 +69,15 @@ def generate_response(user_message, messages_list, image_pil, max_new_tokens, to
             top_p=top_p,
             top_k=top_k,
             temperature=temperature,
-            stop_strings="<|endoftext|>"
+            eos_token_id=processor.tokenizer.eos_token_id,
+            pad_token_id=processor.tokenizer.pad_token_id
         )
 
         # Generate output using model's specific method
         generate_kwargs = dict(
-            **inputs,
+            batch=inputs,
             streamer=streamer,
-            generation_config=generation_config,
-            tokenizer=processor.tokenizer
+            generation_config=generation_config
         )
 
         # Start the generation in a separate thread to allow streaming