leideng/QCFuse / srt /speculative /base_spec_worker.py
leideng's picture
download
raw
698 Bytes
from __future__ import annotations
from abc import ABC, abstractmethod
from typing import TYPE_CHECKING
if TYPE_CHECKING:
from sglang.srt.managers.tp_worker import TpModelWorker
class BaseDraftWorker(ABC):
@abstractmethod
def draft():
pass
@abstractmethod
def draft_extend():
pass
class BaseSpecWorker(ABC):
@property
@abstractmethod
def target_worker(self) -> TpModelWorker:
pass
@property
@abstractmethod
def draft_worker(self) -> BaseDraftWorker:
pass
@abstractmethod
def clear_cache_pool(self):
# TODO: move this abstract method to BaseTpWorker and call through self.model_runner
pass

Xet Storage Details

Size:
698 Bytes
·
Xet hash:
523d1b9fc064c7176a1efb905063e7dd9cd2c79bd7e0c8f3742640cc79916dde

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.