Update modeling_quiet.py

modeling_quiet.py

@@ -1371,7 +1371,7 @@ class QuietForCausalLM(QuietPreTrainedModel, GenerationMixin):
             next_token_id = torch.argmax(next_token_logits, dim=-1)
 
             # Append the generated token to the input sequence
-            input_ids = torch.cat([input_ids, next_token_id.unsqueeze(-1).to(input_ids.device)], dim=-1)
+            # input_ids = torch.cat([input_ids, next_token_id.unsqueeze(-1).to(input_ids.device)], dim=-1)
             seq_len += 1
 
             # Update the attention mask