forked from EngineX-Hygon/enginex-hygon-vllm
37 lines
1.2 KiB
Python
37 lines
1.2 KiB
Python
# SPDX-License-Identifier: Apache-2.0
|
|
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
|
from typing import Optional
|
|
|
|
import torch
|
|
|
|
from vllm.v1.request import Request, RequestStatus
|
|
|
|
|
|
def check_stop(request: Request,
|
|
max_model_len: int,
|
|
pooler_output: Optional[torch.Tensor] = None) -> bool:
|
|
if (request.num_tokens >= max_model_len
|
|
or request.num_output_tokens >= request.max_tokens):
|
|
request.status = RequestStatus.FINISHED_LENGTH_CAPPED
|
|
return True
|
|
|
|
if request.pooling_params:
|
|
if pooler_output is not None:
|
|
request.status = RequestStatus.FINISHED_STOPPED
|
|
return True
|
|
return False
|
|
|
|
sampling_params = request.sampling_params
|
|
assert sampling_params is not None
|
|
last_token_id = request.output_token_ids[-1]
|
|
if (not sampling_params.ignore_eos
|
|
and last_token_id == request.eos_token_id):
|
|
request.status = RequestStatus.FINISHED_STOPPED
|
|
return True
|
|
|
|
if last_token_id in (sampling_params.stop_token_ids or ()):
|
|
request.status = RequestStatus.FINISHED_STOPPED
|
|
request.stop_reason = last_token_id
|
|
return True
|
|
return False
|