24 lines
666 B
Python
24 lines
666 B
Python
"""vLLM: a high-throughput and memory-efficient inference engine for LLMs"""
|
|
|
|
from vllm.engine.arg_utils import AsyncEngineArgs, EngineArgs
|
|
from vllm.engine.async_llm_engine import AsyncLLMEngine
|
|
from vllm.engine.llm_engine import LLMEngine
|
|
from vllm.engine.ray_utils import initialize_ray_cluster
|
|
from vllm.entrypoints.llm import LLM
|
|
from vllm.outputs import CompletionOutput, RequestOutput
|
|
from vllm.sampling_params import SamplingParams
|
|
|
|
__version__ = "0.4.0"
|
|
|
|
__all__ = [
|
|
"LLM",
|
|
"SamplingParams",
|
|
"RequestOutput",
|
|
"CompletionOutput",
|
|
"LLMEngine",
|
|
"EngineArgs",
|
|
"AsyncLLMEngine",
|
|
"AsyncEngineArgs",
|
|
"initialize_ray_cluster",
|
|
]
|