biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[Core] Support Lora lineage and base model metadata management (#6315)

Browse Source
This commit is contained in:
Jiaxin Shan
2024-09-19 23:20:56 -07:00
committed by GitHub
parent 9e5ec35b1f
commit 260d40b5ea
15 changed files with 337 additions and 45 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
tests/entrypoints/openai/test_serving_engine.py
View File

@@ -8,9 +8,10 @@ from vllm.engine.protocol import EngineClient
from vllm.entrypoints.openai.protocol import (ErrorResponse,
LoadLoraAdapterRequest,
UnloadLoraAdapterRequest)
from vllm.entrypoints.openai.serving_engine import OpenAIServing
from vllm.entrypoints.openai.serving_engine import BaseModelPath, OpenAIServing
MODEL_NAME = "meta-llama/Llama-2-7b"
BASE_MODEL_PATHS = [BaseModelPath(name=MODEL_NAME, model_path=MODEL_NAME)]
LORA_LOADING_SUCCESS_MESSAGE = (
"Success: LoRA adapter '{lora_name}' added successfully.")
LORA_UNLOADING_SUCCESS_MESSAGE = (
@@ -25,7 +26,7 @@ async def _async_serving_engine_init():
serving_engine = OpenAIServing(mock_engine_client,
mock_model_config,
served_model_names=[MODEL_NAME],
BASE_MODEL_PATHS,
lora_modules=None,
prompt_adapters=None,
request_logger=None)