Issues

• [Feature] 请求发布支持 C500/MACA 与 Qwen3.6-35B-A3B 、qwen3.8-27b的新版 LMDeploy 镜像

  InternLM/lmdeploy#5026 · Filter by author ls94ls opened on Oct 9, 2026

• [Feature] lmdeploy support FP32xbf16 gemm

  InternLM/lmdeploy#5022 · Filter by author Tsundoku958 opened on Oct 8, 2026

• [Bug] deepseek v3.2/v4 DSML parsing: asserts instead of documented ValueError (silently skipped under -O), batch vs streaming grammar and duplicate-parameter divergences

  InternLM/lmdeploy#5011 · Filter by author ptimizeroracle opened on Sep 30, 2026

• [Bug] Proxy connection warmup can accumulate unbounded PD connection wait tasks and OOM

  InternLM/lmdeploy#4972 · Filter by author JPengLi opened on Sep 15, 2026
  

• [Bug] Mooncake async migration can hang forever and make Decode unavailable

  InternLM/lmdeploy#4971 · Filter by author JPengLi opened on Sep 15, 2026
  2

  

• [Bug] DistServe Proxy requests can leak Prefill scheduler metadata and OOM the Prefill engine

  InternLM/lmdeploy#4967 · Filter by author Charles-know opened on Sep 14, 2026
  

• [Bug] Untrusted migration_request can terminate the DistServe Decode EngineLoop and cause persistent denial of service

  InternLM/lmdeploy#4965 · Filter by author JPengLi opened on Sep 14, 2026
  

• [Enhance] Control token should not be treated as special token in user messages

  InternLM/lmdeploy#4960 · Filter by author HAOCHENYE opened on Sep 13, 2026
  

• [Feature] 现在lmdeploy能支持模型GLM-5.3-Flash-W4A16-MTP运行吗？

  InternLM/lmdeploy#4958 · Filter by author simonjhy opened on Sep 12, 2026
  3

• [Bug] Slower decode speed in v0.17 when compared to v0.14

  InternLM/lmdeploy#4953 · Filter by author Viji2029 opened on Sep 10, 2026
  5

  

• [Bug] TurboMind SIGSEGV crash-loop triggered by session_len truncation path (deterministic libc offset across 3 crashes)

  InternLM/lmdeploy#4944 · Filter by author zambalee opened on Sep 8, 2026
  1

  

• [Bug] input_embeddings are omitted from prefix-cache identity, causing wrong KV reuse

  InternLM/lmdeploy#4933 · Filter by author wildoranges opened on Sep 3, 2026
  2

• [Bug] Mooncake external KV keys omit KV format and weights lineage, with no default tenant isolation

  InternLM/lmdeploy#4930 · Filter by author wildoranges opened on Sep 3, 2026
  2

  

• [Feature] 咱们最新版本的lmdeploy可以加载这个 Qwen3.8-Flash吗？

  InternLM/lmdeploy#4905 · Filter by author simonjhy opened on Aug 26, 2026
  2

• [Bug] PyTorch engine: Qwen3.5 hybrid-GDN + AWQ checkpoint crashes on TP>1 - 'AwqLinear' object has no attribute 'block_size'

  InternLM/lmdeploy#4899 · Filter by author u5084757454-collab opened on Aug 24, 2026

• [Feature] 什么时候turbomind能够开始支持MTP或者DFlash之类的加速功能？

  InternLM/lmdeploy#4889 · Filter by author simonjhy opened on Aug 21, 2026
  1

  

• [Bug] pytorch backend silently hangs under concurrent decoding with qwen3_5_mtp speculative decoding

  InternLM/lmdeploy#4883 · Filter by author matrix72c opened on Aug 19, 2026
  1

  

• [Feature] 当前qwen-next等SSM类模型prefix cache缓存命中率低

  InternLM/lmdeploy#4879 · Filter by author Tsundoku958 opened on Aug 18, 2026
  

• [Feature] sm75 can support awq/gptq/fp8/w8a8 in turbomind

  InternLM/lmdeploy#4865 · Filter by author bltcn opened on Aug 14, 2026
  

• [Feature] 现在最新的0.15.0可以支持最新的deepseek-v4-flash-0731吗？

  InternLM/lmdeploy#4832 · Filter by author simonjhy opened on Aug 6, 2026
  2

• [Feature] DeepSeek-V4 MoE is Blackwell-only via DeepGEMM — is a non-Blackwell path wanted?

  InternLM/lmdeploy#4824 · Filter by author Hakureirm opened on Aug 4, 2026
  9

  

• [Feature] qwen3.5可以turbo后端可以开启prefill cache吗

  InternLM/lmdeploy#4775 · Filter by author janelu9 opened on Jul 23, 2026
  2

  

• [Feature] lmcache support?

  InternLM/lmdeploy#4750 · Filter by author HaiFengZeng opened on Jul 14, 2026
  1

  

• [Bug] 0.14.0版本能否支持对Qwen3.5系列模型的量化？

  InternLM/lmdeploy#4720 · Filter by author legends-7 opened on Jun 30, 2026
  

• [Bug] /update_weights deserializes request-controlled pickle data before validation

  InternLM/lmdeploy#4698 · Filter by author starsalt0124 opened on Jun 22, 2026
  2