more

app.py

@@ -37,10 +37,10 @@ def respond(message, history, system_message, max_tokens, temperature, top_p):
     ).to("cuda" if torch.cuda.is_available() else "cpu")
 
     # Generate the response
-    attention_mask = inputs.ne(tokenizer.pad_token_id).long()
+    #attention_mask = inputs.ne(tokenizer.pad_token_id).long()
     generated_tokens = model.generate(
         input_ids=inputs,
-        attention_mask=attention_mask,
+        #attention_mask=attention_mask,
         max_new_tokens=max_tokens,
         use_cache=True,
         temperature=temperature,