sgl-project/sglang issues and pull requests

#1746 - Fix sliding window attention and gemma-2 unit tests in CI

Pull Request - State: closed - Opened by merrymercy 19 days ago

#1745 - Introducing SGLang Guru on Gurubase.io

Pull Request - State: open - Opened by kursataktas 20 days ago - 4 comments

#1744 - [Bug] Issue in latest sglang docker image

Issue - State: closed - Opened by shubhamgajbhiye1994 20 days ago - 1 comment

#1743 - Fix prefill oom

Pull Request - State: closed - Opened by hnyls2002 20 days ago

#1742 - [Bug] when sglang received over 16 concurrency request(means i create 16 thread to call the service all the time), it will return abnormal result, and in the log, will occur NaN

Issue - State: closed - Opened by GGBond8488 20 days ago - 5 comments

#1741 - Maintain seq_lens_sum to make more FlashInfer operations non-blocking

Pull Request - State: closed - Opened by merrymercy 20 days ago

#1740 - Make token mapping non-blocking in the overlapped mode

Pull Request - State: closed - Opened by merrymercy 20 days ago

#1739 - [Bug] Prefill OOM!

Issue - State: closed - Opened by yichuan520030910320 20 days ago - 2 comments

#1738 - Faster overlap mode scheduler

Pull Request - State: closed - Opened by merrymercy 20 days ago

#1737 - misc: add CODEOWNERS

Pull Request - State: closed - Opened by zhyncs 20 days ago

#1736 - Add GLM-4 TextGeneration Model support for SGLang

Pull Request - State: closed - Opened by sixsixcoder 20 days ago

#1735 - Simplify batch result resolution

Pull Request - State: closed - Opened by merrymercy 20 days ago

#1734 - Simplify the usage of device

Pull Request - State: closed - Opened by merrymercy 20 days ago

#1733 - Add documentations for Installation

Pull Request - State: closed - Opened by zhaochenyang20 20 days ago - 3 comments

#1732 - [Feature] Cache-aware Data Parallel Router

Issue - State: open - Opened by ByronHsu 20 days ago
Labels: feature

#1731 - Optimize ZMQ receive operations to reduce idle CPU usage

Pull Request - State: closed - Opened by zyearw1024 21 days ago - 1 comment

#1730 - [Bug] 100% CPU Usage When Idle in sglang

Issue - State: closed - Opened by zyearw1024 21 days ago - 1 comment

#1729 - [Bug][minimal reproducible demo] High variability across batch inference runs

Issue - State: open - Opened by FredericOdermatt 21 days ago - 6 comments
Labels: bug

#1728 - [LoRA, Performance] Add gemm expand triton kernel for multi-LoRA

Pull Request - State: open - Opened by Ying1123 21 days ago

#1727 - [Bugfix] qwen2vl forward_extend

Pull Request - State: closed - Opened by yizhang2077 21 days ago

#1726 - Split the overlapped version of TpModelWorkerClient into a separate file

Pull Request - State: closed - Opened by merrymercy 21 days ago

#1725 - Temporarily skip the test_mixed_batch for QWen2VL

Pull Request - State: closed - Opened by merrymercy 21 days ago - 1 comment

#1724 - Unify the memory pool api and tp worker API

Pull Request - State: closed - Opened by merrymercy 21 days ago

#1723 - docs: fix README

Pull Request - State: closed - Opened by zhyncs 21 days ago - 1 comment

#1722 - Update README.md

Pull Request - State: closed - Opened by Ying1123 21 days ago

#1721 - Support qwen2 vl model

Pull Request - State: closed - Opened by zhyncs 21 days ago - 5 comments

#1720 - Update vllm to 0.6.3 (#1711)

Pull Request - State: closed - Opened by zhyncs 21 days ago

#1719 - CPU Inference

Issue - State: closed - Opened by JocelynPanPan 21 days ago - 1 comment

#1718 - Simplify the interface of tp_worker

Pull Request - State: closed - Opened by merrymercy 21 days ago

#1717 - Created SECURITY.md

Pull Request - State: closed - Opened by NishantRana07 21 days ago

#1716 - Update readme and workflow

Pull Request - State: closed - Opened by merrymercy 21 days ago