Expose priority parameter in Engine.generate() and Engine.async_generate() (#18944)
Signed-off-by: PeaBrane <peabrane@gmail.com> Co-authored-by: Cursor <cursoragent@cursor.com>
This commit is contained in:
@@ -30,6 +30,7 @@ class EngineBase(ABC):
|
||||
bootstrap_room: Optional[Union[List[int], int]] = None,
|
||||
data_parallel_rank: Optional[int] = None,
|
||||
rid: Optional[Union[List[str], str]] = None,
|
||||
priority: Optional[int] = None,
|
||||
) -> Union[Dict, Iterator[Dict]]:
|
||||
"""Generate outputs based on given inputs."""
|
||||
pass
|
||||
|
||||
@@ -236,6 +236,7 @@ class Engine(EngineBase):
|
||||
external_trace_header: Optional[Dict] = None,
|
||||
rid: Optional[Union[List[str], str]] = None,
|
||||
session_params: Optional[Dict] = None,
|
||||
priority: Optional[int] = None,
|
||||
) -> Union[Dict, Iterator[Dict]]:
|
||||
"""
|
||||
The arguments of this function is the same as `sglang/srt/managers/io_struct.py::GenerateReqInput`.
|
||||
@@ -274,6 +275,7 @@ class Engine(EngineBase):
|
||||
external_trace_header=external_trace_header,
|
||||
rid=rid,
|
||||
session_params=session_params,
|
||||
priority=priority,
|
||||
)
|
||||
generator = self.tokenizer_manager.generate_request(obj, None)
|
||||
|
||||
@@ -325,6 +327,7 @@ class Engine(EngineBase):
|
||||
external_trace_header: Optional[Dict] = None,
|
||||
rid: Optional[Union[List[str], str]] = None,
|
||||
session_params: Optional[Dict] = None,
|
||||
priority: Optional[int] = None,
|
||||
) -> Union[Dict, AsyncIterator[Dict]]:
|
||||
"""
|
||||
The arguments of this function is the same as `sglang/srt/managers/io_struct.py::GenerateReqInput`.
|
||||
@@ -364,6 +367,7 @@ class Engine(EngineBase):
|
||||
external_trace_header=external_trace_header,
|
||||
rid=rid,
|
||||
session_params=session_params,
|
||||
priority=priority,
|
||||
)
|
||||
generator = self.tokenizer_manager.generate_request(obj, None)
|
||||
|
||||
|
||||
@@ -108,6 +108,7 @@ class HttpServerEngineAdapter(EngineBase):
|
||||
token_ids_logprob=None,
|
||||
lora_path=None,
|
||||
custom_logit_processor=None,
|
||||
priority=None,
|
||||
):
|
||||
payload = {
|
||||
"text": prompt,
|
||||
@@ -120,6 +121,7 @@ class HttpServerEngineAdapter(EngineBase):
|
||||
"token_ids_logprob": token_ids_logprob,
|
||||
"lora_path": lora_path,
|
||||
"custom_logit_processor": custom_logit_processor,
|
||||
"priority": priority,
|
||||
}
|
||||
# Filter out None values
|
||||
payload = {k: v for k, v in payload.items() if v is not None}
|
||||
|
||||
Reference in New Issue
Block a user