sgl-project/sglang

[Feature] Create benchmark and dataset usage scripts for embedding models

Open

#9,674 opened on Aug 27, 2025

View on GitHub
 (6 comments) (0 reactions) (0 assignees)Python (6,216 forks)auto 404
good first issuehelp wanted

Repository metrics

Stars
 (28,442 stars)
PR merge metrics
 (Avg merge 2d 1h) (1,000 merged PRs in 30d)

Description

Look at the following for related tests, and the benchmark script should be similar to bench_serving Existing Tests:

models/test_embedding_models.py — Backend, 73s — PR: core embedding model implementations (covers srt/models/*_embedding.py, srt/entrypoints/openai/serving_embedding.py).

models/test_encoder_embedding_models.py — Backend, 100s — Post-merge: encoder-based embedding models (BERT/Roberta-style).

models/test_cross_encoder_models.py — Backend, 100s — Post-merge: cross-encoder/reranker models (related to embedding-based tasks but a different architecture).

openai_server/basic/test_serving_embedding.py — Frontend/unit, 10s — Run on all PRs: embedding serving layer (validates the serving path).

openai_server/basic/test_openai_embedding.py — Frontend, 141s — Frontend (PR/post-merge as listed): end-to-end OpenAI embedding API endpoints.

test_input_embeddings.py — Backend, 38s — Post-merge: tests the input embedding handling/path.

Contributor guide

Research direction
Review the existing `bench serving` script and the embedding test files (e.g., test embedding models.py) to understand the pattern for creating benchmark and dataset usage scripts. Focus on replicating the benchmarking structure for embedding models.
Tech stack
python
Domain
backendtesting
Issue type
Feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
3
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
1-2 days
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
Pythonpytest
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
50