Repository Issues
sgl-project/sglang
SGLang is a high-performance serving framework for large language models and multimodal models.
Issues
Closed
[Feature] Support encoder models for flash_infer backend
good first issue
No assignee yetHas a beginner-friendly labelContributing guide available
2 comments1 reaction0 assignees
Open
[Feature] Add KV cache usage prometheus metrics
good first issue
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
11 comments5 reactions0 assignees
Open
[Feature] Support more multi-modal input for VLM
Multi-modalfeaturegood first issuehelp wanted
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
5 comments5 reactions1 assignee
Open
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
3 comments0 reactions0 assignees
Closed
[Feature] Support PDL on norm in sgl-kernel
good first issuesgl-kernel
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
4 comments2 reactions1 assignee
Open
Extensive benchmarking of reasoning models including variance
good first issue
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
3 comments0 reactions1 assignee
Open
[Feature] Automatically truncate when the maximum tokens are exceeded instead of throwing an error
good first issue
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
7 comments0 reactions1 assignee
Open
[Feature] support `gather` instead of `all_gather` when gathering the logits
good first issuehelp wanted
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
9 comments0 reactions1 assignee
Open
[Feature] Add examples for running SGLang on Slurm
documentationgood first issuehelp wanted
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
5 comments0 reactions1 assignee
Open
[Bug] Decode Throughput Inconsistency Between bench_serving and Engine Logs
good first issuehelp wanted
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
16 comments0 reactions1 assignee
Open
[Feature] Support bitsandbytes in QWen2 VL
good first issuehelp wanted
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
4 comments0 reactions0 assignees
Open
[Feature] Serving VLM VILA
good first issuehelp wanted
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
2 comments1 reaction0 assignees
Closed
[CI] Print nightly evaluation results to GITHUB_STEP_SUMMARY
good first issuehelp wanted
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
2 comments0 reactions0 assignees
Open
[Kernel] cuDNN attention backend
enhancementgood first issuehelp wantedhigh priorityinactive
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
5 comments1 reaction2 assignees
Open
[Kernel] Optimize triton decoding kernels for long context
good first issuehelp wantedhigh priorityinactive
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
7 comments1 reaction1 assignee
Closed
[Feature] support torchao for qwen2 models
good first issuehelp wanted
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
11 comments0 reactions0 assignees
Closed
[Bug] disk cache io error when simultaneously loading lots of sglang offline engine
good first issuehelp wanted
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
13 comments0 reactions0 assignees
Open
[Feature] Save cache from requests and load
good first issue
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
2 comments0 reactions0 assignees
Open
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
9 comments0 reactions0 assignees
Open
[Feature] Cascade attention kernels
good first issuehigh priority
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
8 comments1 reaction0 assignees