Issues

• [Bug] Queued tokens for one request duplicate detokenized output

  sgl-project/mini-sglang#171 · Filter by author HuaTNA opened on Oct 19, 2026

• Scheduler crash (ref_count >= 0 assertion) or corrupted radix cache when a request is aborted while its step is in flight (overlap scheduling)

  sgl-project/mini-sglang#170 · Filter by author taigaoma1997 opened on Oct 18, 2026

• Radix cache dedup frees KV pages that a running request still reads (wrong output with overlap scheduling)

  sgl-project/mini-sglang#168 · Filter by author taigaoma1997 opened on Oct 17, 2026

• [Bug] fused_topk renormalizes routing weights twice

  sgl-project/mini-sglang#167 · Filter by author fangpin opened on Oct 14, 2026

• The text completion API format is not comptible with the Open WebUI

  sgl-project/mini-sglang#153 · Filter by author Jiadalee opened on Oct 5, 2026

• RadixPrefixCache.check_integrity is a no-op

  sgl-project/mini-sglang#149 · Filter by author 12dinitrobenzene opened on Aug 18, 2026
  1

• 5070 SM120 Unsupported architecture with TllmGenFmhaRunner

  sgl-project/mini-sglang#129 · Filter by author khan-yin opened on May 10, 2026
  2

• does mini-sglang support SM_75

  sgl-project/mini-sglang#128 · Filter by author Krank41 opened on May 9, 2026
  3

• Support / Clarify Expert Parallelism (EP) support in mini‑SGLang

  sgl-project/mini-sglang#123 · Filter by author albertimff opened on May 6, 2026
  1

• [Feature Request] Mac/Apple Silicon (MPS) support via fallback attention backend

  sgl-project/mini-sglang#120 · Filter by author llwhub opened on Apr 20, 2026

• Interest in adding AMD GPU (ROCm) support

  sgl-project/mini-sglang#90 · Filter by author jhinpan opened on Mar 1, 2026
  6

• I run the mini-sglang on my spark, does the project need a pr for this ?

  sgl-project/mini-sglang#61 · Filter by author qqwqqw689 opened on Jan 10, 2026
  1

• torch.compile compatibility: significant performance overhead due to guard failures and recompilations

  sgl-project/mini-sglang#60 · Filter by author uv-xiao opened on Jan 5, 2026
  2

• crash to run offline inference

  sgl-project/mini-sglang#58 · Filter by author gameofdimension opened on Dec 30, 2025
  6

• any VLM models to support

  sgl-project/mini-sglang#31 · Filter by author tuanhe opened on Dec 22, 2025
  1

• any diffusion model support plan？

  sgl-project/mini-sglang#28 · Filter by author vigo999 opened on Dec 21, 2025
  2

• CPU support

  sgl-project/mini-sglang#12 · Filter by author hnyls2002 opened on Dec 18, 2025
  2