sglang
https://github.com/sgl-project/sglang
Python
SGLang is a high-performance serving framework for large language models and multimodal models.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Python not yet supported1 Subscribers
Add a CodeTriage badge to sglang
Help out
- Issues
- [Perf] [AMD] kimi_k3: Row-shard the latent MoE up_proj over TP in prefill
- [Bug] [InferHunt] OpenAI validation errors expose server source paths and endpoint locations
- [Perf] [AMD] gemm: Split-K bf16 GEMV for single-row decode on gfx1151
- [Fix] moe: Apply swiglu_limit in the FlashInfer CuTe DSL NVFP4 MoE runner
- [Perf] spec_v2: Eliminate per-step seq_lens_sum D2H host syncs in Frozen-KV MTP
- [Fix] speculative: Preserve NaN argmax semantics in verify kernels
- [Fix] environ: Register the batch-invariant matmul fallback flag in Envs
- [Feat] RL: Report prompt KV weight versions in generation metadata
- [Feat] kv-canary: Pass through canary info when the KV-canary is enabled
- [Feat] RL: Hash raw received weights by name in check_weights
- Docs
- Python not yet supported