Skip to content

vllm.v1.core.sched.utils

check_stop 

check_stop(
    request: Request,
    max_model_len: int,
    pooler_output: Optional[Tensor] = None,
) -> bool
Source code in vllm/v1/core/sched/utils.py 
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
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