xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2025-12-13 23:55:44 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Bugfix] Fix the issue where the model name is empty string, causing no response with the model name. (#15938)

Browse Source 
Signed-off-by: chaunceyjiang <chaunceyjiang@gmail.com>
This commit is contained in:
Chauncey 2025-04-02 21:33:52 +08:00 committed by GitHub
parent 44f990515b
commit 594a8b9030
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 29 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

31
tests/entrypoints/openai/test_chat.py
View File

@ -11,7 +11,7 @@ import pytest
import pytest_asyncio import pytest_asyncio
import requests import requests
import torch import torch
from openai import BadRequestError from openai import BadRequestError, OpenAI
from ...utils import RemoteOpenAIServer from ...utils import RemoteOpenAIServer
from .test_completion import zephyr_lora_added_tokens_files # noqa: F401 from .test_completion import zephyr_lora_added_tokens_files # noqa: F401
@ -1054,7 +1054,7 @@ async def test_long_seed(client: openai.AsyncOpenAI):
@pytest.mark.asyncio @pytest.mark.asyncio
async def test_http_chat_wo_model_name(server: RemoteOpenAIServer): async def test_http_chat_no_model_name_with_curl(server: RemoteOpenAIServer):
url = f"http://localhost:{server.port}/v1/chat/completions" url = f"http://localhost:{server.port}/v1/chat/completions"
headers = { headers = {
"Content-Type": "application/json", "Content-Type": "application/json",
@ -1075,10 +1075,35 @@ async def test_http_chat_wo_model_name(server: RemoteOpenAIServer):
response = requests.post(url, headers=headers, json=data) response = requests.post(url, headers=headers, json=data)
response_data = response.json() response_data = response.json()
print(response_data) print(response_data)
assert response_data.get("model") == MODEL_NAME
choice = response_data.get("choices")[0] choice = response_data.get("choices")[0]
message = choice.get("message") message = choice.get("message")
assert message is not None assert message is not None
content = message.get("content") content = message.get("content")
assert content is not None assert content is not None
assert len(content) > 0 assert len(content) > 0
@pytest.mark.asyncio
@pytest.mark.parametrize("model_name", [MODEL_NAME, ""])
async def test_http_chat_no_model_name_with_openai(server: RemoteOpenAIServer,
model_name: str):
openai_api_key = "EMPTY"
openai_api_base = f"http://localhost:{server.port}/v1"
client = OpenAI(
api_key=openai_api_key,
base_url=openai_api_base,
)
messages = [
{
"role": "user",
"content": "Hello, vLLM!"
},
]
response = client.chat.completions.create(
model="", # empty string
messages=messages,
)
assert response.model == MODEL_NAME

2
vllm/entrypoints/openai/serving_engine.py
View File

@ -537,7 +537,7 @@ class OpenAIServing:
lora_request: Optional[LoRARequest] = None) -> str: lora_request: Optional[LoRARequest] = None) -> str:
if lora_request: if lora_request:
return lora_request.lora_name return lora_request.lora_name
if model_name is None: if not model_name:
return self.models.base_model_paths[0].name return self.models.base_model_paths[0].name
return model_name return model_name