mirror of
https://github.com/PaddlePaddle/FastDeploy.git
synced 2026-04-23 00:17:25 +08:00
merge logprob into batch_output (#3266)
This commit is contained in:
@@ -30,6 +30,7 @@ from fastdeploy.config import (
|
||||
TaskOption,
|
||||
)
|
||||
from fastdeploy.engine.config import Config
|
||||
from fastdeploy.platforms import current_platform
|
||||
from fastdeploy.scheduler.config import SchedulerConfig
|
||||
from fastdeploy.utils import DeprecatedOptionWarning, FlexibleArgumentParser
|
||||
|
||||
@@ -344,6 +345,13 @@ class EngineArgs:
|
||||
"""
|
||||
if not self.tokenizer:
|
||||
self.tokenizer = self.model
|
||||
if self.enable_logprob:
|
||||
if self.speculative_config is not None:
|
||||
raise NotImplementedError("Logprob does not support speculation_config.")
|
||||
if self.enable_expert_parallel:
|
||||
raise NotImplementedError("Logprob does not support enable_expert_parallel.")
|
||||
if not current_platform.is_cuda():
|
||||
raise NotImplementedError("Only CUDA platform supports logprob.")
|
||||
|
||||
@staticmethod
|
||||
def add_cli_args(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
|
||||
|
||||
Reference in New Issue
Block a user