Skip to content

[https://nvbugs/5625743][fix] Cherry-pick support for n>1 with pytorch backend for openai completion and add tests. #9471

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
SimengLiu-nv wants to merge 4 commits into from
Closed

[https://nvbugs/5625743][fix] Cherry-pick support for n>1 with pytorch backend for openai completion and add tests. #9471

Changes from 1 commit
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
338a78c
[TRTLLM-8598][feat] enable n > 1 in OpenAI API with PyTorch backend (…
ixlmar Nov 8, 2025
c362159
[https://nvbugs/5625743][fix] Add tests for completion with n>1 and s…
SimengLiu-nv Nov 26, 2025
7a5fa3f
Address comments.
SimengLiu-nv Dec 3, 2025
7a0c328
Fix the output corruption with n > 1 and stream=True.
SimengLiu-nv Dec 5, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
10 changes: 7 additions & 3 deletions tests/unittest/llmapi/apps/_test_openai_completions.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@

import openai
import pytest
from utils.util import similar

from ..test_llm import get_model_path
from .openai_server import RemoteOpenAIServer
Expand Down Expand Up @@ -206,7 +207,7 @@ async def test_batch_completions_streaming(async_client: openai.AsyncOpenAI,


@pytest.mark.asyncio(loop_scope="module")
@pytest.mark.parametrize("prompts", [["Hello, my name is", "What is the AI?"]])
@pytest.mark.parametrize("prompts", [["Hello, my name is"] * 2])
async def test_batch_completions_with_option_n_streaming(
async_client: openai.AsyncOpenAI, model_name, prompts):
# test beam search with streaming
Expand All @@ -215,7 +216,7 @@ async def test_batch_completions_with_option_n_streaming(
prompt=prompts,
n=3, # number of completions to generate for each prompt.
max_tokens=5,
temperature=0.0,
temperature=0.0001,
stream=True,
)
texts = [""] * 6 # 2 prompts × 3 generations per prompt = 6 choices
Expand All @@ -224,7 +225,10 @@ async def test_batch_completions_with_option_n_streaming(
choice = chunk.choices[0]
texts[choice.index] += choice.text

assert [""] not in texts # Assert all the generations are not empty
assert "" not in texts # Assert all the generations are not empty

for i, j in zip(texts[:3], texts[3:]):
assert similar(i, j, threshold=0.8)


@pytest.mark.asyncio(loop_scope="module")
Expand Down