feat: add return hidden_states at async generation (#7507)
This commit is contained in:
@@ -422,7 +422,7 @@ class MooncakeKVManager(BaseKVManager):
|
||||
src_addr_list.append(src_slice_addr)
|
||||
dst_addr_list.append(dst_slice_addr)
|
||||
length_list.append(slice_lens_per_page)
|
||||
|
||||
|
||||
logger.debug(
|
||||
f"SYNC: sid={mooncake_session_id}, "
|
||||
f"src={src_slice_addr}, dst={dst_slice_addr}, len={slice_lens_per_page}"
|
||||
|
||||
@@ -242,6 +242,7 @@ class Engine(EngineBase):
|
||||
token_ids_logprob: Optional[Union[List[List[int]], List[int]]] = None,
|
||||
lora_path: Optional[List[Optional[str]]] = None,
|
||||
custom_logit_processor: Optional[Union[List[str], str]] = None,
|
||||
return_hidden_states: bool = False,
|
||||
stream: bool = False,
|
||||
bootstrap_host: Optional[Union[List[str], str]] = None,
|
||||
bootstrap_port: Optional[Union[List[int], int]] = None,
|
||||
@@ -274,6 +275,7 @@ class Engine(EngineBase):
|
||||
top_logprobs_num=top_logprobs_num,
|
||||
token_ids_logprob=token_ids_logprob,
|
||||
lora_path=lora_path,
|
||||
return_hidden_states=return_hidden_states,
|
||||
stream=stream,
|
||||
custom_logit_processor=custom_logit_processor,
|
||||
bootstrap_host=bootstrap_host,
|
||||
|
||||
Reference in New Issue
Block a user