This commit is contained in:
@@ -6,7 +6,7 @@ from functools import partial
|
||||
from typing import (TYPE_CHECKING, Any, Callable, ClassVar, Deque, Dict,
|
||||
Iterable, List, Mapping, NamedTuple, Optional)
|
||||
from typing import Sequence as GenericSequence
|
||||
from typing import Set, Type, Union, overload
|
||||
from typing import Set, Type, Union
|
||||
|
||||
import torch
|
||||
from typing_extensions import TypeVar
|
||||
@@ -29,7 +29,7 @@ from vllm.executor.executor_base import ExecutorBase
|
||||
from vllm.executor.gpu_executor import GPUExecutor
|
||||
from vllm.executor.ray_utils import initialize_ray_cluster
|
||||
from vllm.inputs import (INPUT_REGISTRY, EncoderDecoderLLMInputs,
|
||||
InputRegistry, LLMInputs, PromptType)
|
||||
InputRegistry, LLMInputs, PromptInputs)
|
||||
from vllm.inputs.preprocess import InputPreprocessor
|
||||
from vllm.logger import init_logger
|
||||
from vllm.lora.request import LoRARequest
|
||||
@@ -51,7 +51,7 @@ from vllm.transformers_utils.tokenizer_group import (
|
||||
BaseTokenizerGroup, init_tokenizer_from_configs)
|
||||
from vllm.usage.usage_lib import (UsageContext, is_usage_stats_enabled,
|
||||
usage_message)
|
||||
from vllm.utils import Counter, Device, deprecate_kwargs, weak_bind
|
||||
from vllm.utils import Counter, Device, weak_bind
|
||||
from vllm.version import __version__ as VLLM_VERSION
|
||||
|
||||
logger = init_logger(__name__)
|
||||
@@ -689,51 +689,16 @@ class LLMEngine:
|
||||
def stop_remote_worker_execution_loop(self) -> None:
|
||||
self.model_executor.stop_remote_worker_execution_loop()
|
||||
|
||||
@overload # DEPRECATED
|
||||
def add_request(
|
||||
self,
|
||||
request_id: str,
|
||||
*,
|
||||
inputs: PromptType,
|
||||
inputs: PromptInputs,
|
||||
params: Union[SamplingParams, PoolingParams],
|
||||
arrival_time: Optional[float] = None,
|
||||
lora_request: Optional[LoRARequest] = None,
|
||||
trace_headers: Optional[Mapping[str, str]] = None,
|
||||
prompt_adapter_request: Optional[PromptAdapterRequest] = None,
|
||||
priority: int = 0,
|
||||
) -> None:
|
||||
...
|
||||
|
||||
@overload
|
||||
def add_request(
|
||||
self,
|
||||
request_id: str,
|
||||
prompt: PromptType,
|
||||
params: Union[SamplingParams, PoolingParams],
|
||||
arrival_time: Optional[float] = None,
|
||||
lora_request: Optional[LoRARequest] = None,
|
||||
trace_headers: Optional[Mapping[str, str]] = None,
|
||||
prompt_adapter_request: Optional[PromptAdapterRequest] = None,
|
||||
priority: int = 0,
|
||||
) -> None:
|
||||
...
|
||||
|
||||
@deprecate_kwargs(
|
||||
"inputs",
|
||||
additional_message="Please use the 'prompt' parameter instead.",
|
||||
)
|
||||
def add_request(
|
||||
self,
|
||||
request_id: str,
|
||||
prompt: Optional[PromptType] = None,
|
||||
params: Optional[Union[SamplingParams, PoolingParams]] = None,
|
||||
arrival_time: Optional[float] = None,
|
||||
lora_request: Optional[LoRARequest] = None,
|
||||
trace_headers: Optional[Mapping[str, str]] = None,
|
||||
prompt_adapter_request: Optional[PromptAdapterRequest] = None,
|
||||
priority: int = 0,
|
||||
*,
|
||||
inputs: Optional[PromptType] = None, # DEPRECATED
|
||||
) -> None:
|
||||
"""Add a request to the engine's request pool.
|
||||
|
||||
@@ -743,7 +708,8 @@ class LLMEngine:
|
||||
|
||||
Args:
|
||||
request_id: The unique ID of the request.
|
||||
prompt: The prompt to the LLM. See :class:`~vllm.inputs.PromptType`
|
||||
inputs: The inputs to the LLM. See
|
||||
:class:`~vllm.inputs.PromptInputs`
|
||||
for more details about the format of each input.
|
||||
params: Parameters for sampling or pooling.
|
||||
:class:`~vllm.SamplingParams` for text generation.
|
||||
@@ -778,10 +744,6 @@ class LLMEngine:
|
||||
>>> # continue the request processing
|
||||
>>> ...
|
||||
"""
|
||||
if inputs is not None:
|
||||
prompt = inputs
|
||||
assert prompt is not None and params is not None
|
||||
|
||||
if lora_request is not None and not self.lora_config:
|
||||
raise ValueError(f"Got lora_request {lora_request} but LoRA is "
|
||||
"not enabled!")
|
||||
@@ -794,7 +756,7 @@ class LLMEngine:
|
||||
arrival_time = time.time()
|
||||
|
||||
preprocessed_inputs = self.input_preprocessor.preprocess(
|
||||
prompt,
|
||||
inputs,
|
||||
request_id=request_id,
|
||||
lora_request=lora_request,
|
||||
prompt_adapter_request=prompt_adapter_request,
|
||||
|
||||
Reference in New Issue
Block a user