chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 14 Packages Projects Releases Wiki Activity

[misc] add model arch and type to server info and use it for harmony (#14456)

Browse Source
This commit is contained in:
Simo Lin
2025-12-05 06:51:00 -08:00
committed by GitHub
parent 8fce9e7b2a
commit 7235a7fbe9
11 changed files with 205 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files
python/sglang/srt/grpc/sglang_scheduler.proto
+1
View File
@@ -427,6 +427,7 @@ message GetModelInfoResponse {
int32 pad_token_id = 12;
int32 bos_token_id = 13;
int32 max_req_input_len = 14;
repeated string architectures = 15;
}
// Get server information
python/sglang/srt/grpc/sglang_scheduler_pb2.py
+8 -8
View File
File diff suppressed because one or more lines are too long
python/sglang/srt/grpc/sglang_scheduler_pb2.pyi
+4 -2
View File
@@ -432,7 +432,7 @@ class GetModelInfoRequest(_message.Message):
def __init__(self) -> None: ...
class GetModelInfoResponse(_message.Message):
__slots__ = ("model_path", "tokenizer_path", "is_generation", "preferred_sampling_params", "weight_version", "served_model_name", "max_context_length", "vocab_size", "supports_vision", "model_type", "eos_token_ids", "pad_token_id", "bos_token_id", "max_req_input_len")
__slots__ = ("model_path", "tokenizer_path", "is_generation", "preferred_sampling_params", "weight_version", "served_model_name", "max_context_length", "vocab_size", "supports_vision", "model_type", "eos_token_ids", "pad_token_id", "bos_token_id", "max_req_input_len", "architectures")
MODEL_PATH_FIELD_NUMBER: _ClassVar[int]
TOKENIZER_PATH_FIELD_NUMBER: _ClassVar[int]
IS_GENERATION_FIELD_NUMBER: _ClassVar[int]
@@ -447,6 +447,7 @@ class GetModelInfoResponse(_message.Message):
PAD_TOKEN_ID_FIELD_NUMBER: _ClassVar[int]
BOS_TOKEN_ID_FIELD_NUMBER: _ClassVar[int]
MAX_REQ_INPUT_LEN_FIELD_NUMBER: _ClassVar[int]
ARCHITECTURES_FIELD_NUMBER: _ClassVar[int]
model_path: str
tokenizer_path: str
is_generation: bool
@@ -461,7 +462,8 @@ class GetModelInfoResponse(_message.Message):
pad_token_id: int
bos_token_id: int
max_req_input_len: int
def __init__(self, model_path: _Optional[str] = ..., tokenizer_path: _Optional[str] = ..., is_generation: bool = ..., preferred_sampling_params: _Optional[str] = ..., weight_version: _Optional[str] = ..., served_model_name: _Optional[str] = ..., max_context_length: _Optional[int] = ..., vocab_size: _Optional[int] = ..., supports_vision: bool = ..., model_type: _Optional[str] = ..., eos_token_ids: _Optional[_Iterable[int]] = ..., pad_token_id: _Optional[int] = ..., bos_token_id: _Optional[int] = ..., max_req_input_len: _Optional[int] = ...) -> None: ...
architectures: _containers.RepeatedScalarFieldContainer[str]
def __init__(self, model_path: _Optional[str] = ..., tokenizer_path: _Optional[str] = ..., is_generation: bool = ..., preferred_sampling_params: _Optional[str] = ..., weight_version: _Optional[str] = ..., served_model_name: _Optional[str] = ..., max_context_length: _Optional[int] = ..., vocab_size: _Optional[int] = ..., supports_vision: bool = ..., model_type: _Optional[str] = ..., eos_token_ids: _Optional[_Iterable[int]] = ..., pad_token_id: _Optional[int] = ..., bos_token_id: _Optional[int] = ..., max_req_input_len: _Optional[int] = ..., architectures: _Optional[_Iterable[str]] = ...) -> None: ...
class GetServerInfoRequest(_message.Message):
__slots__ = ()