[fix] qwen output inconsistency when top_p=0 (#3634) (#3662)

* [fix] qwen output inconsistency when top_p=0

* [fix] remove decode pre_id code

fastdeploy/worker/gpu_model_runner.py

@@ -283,6 +283,7 @@ class GPUModelRunner(ModelRunnerBase):
                 self.share_inputs["step_idx"][idx : idx + 1] = (
                     len(request.output_token_ids) if prefill_end_index >= len(input_ids) else 0
                 )
+                self.share_inputs["pre_ids"][idx : idx + 1] = -1
                 has_prefill_task = True
             elif request.task_type.value == RequestType.DECODE.value:  # decode task
                 logger.debug(f"Handle decode request {request} at idx {idx}")