14 lines
363 B
Python
14 lines
363 B
Python
|
|
|
|
|
|
from dataclasses import dataclass
|
|
from vllm.v1.outputs import ModelRunnerOutput
|
|
|
|
@dataclass
|
|
class ZeroV1ModelRunnerOutput(ModelRunnerOutput):
|
|
# [num_reqs]
|
|
fix_req_ids: list[str] = None
|
|
fix_sampled_token_ids:list[list[int]] = None
|
|
fix_draft_req_ids:list[str] = None
|
|
fix_draft_tokens_ids:list[list[int]] = None
|
|
is_output_valid:bool = True |