Current issue state, recent activity, and per-issue timelines from the indexed issue data.
| Date | Opened | Closed | Comments | Events | Open Backlog |
|---|---|---|---|---|---|
| 2026-09-09 | 0 | 0 | 0 | 0 | 1 |
| 2026-09-08 | 1 | 0 | 1 | 1 | 346 |
| 2026-09-07 | 0 | 0 | 0 | 0 | 0 |
| 2026-09-06 | 0 | 0 | 0 | 0 | 0 |
| 2026-09-05 | 0 | 0 | 0 | 0 | 0 |
| 2026-09-04 | 0 | 0 | 0 | 0 | 0 |
| 2026-09-03 | 2 | 0 | 0 | 0 | 0 |
| 2026-09-02 | 0 | 0 | 0 | 0 | 0 |
| 2026-09-01 | 0 | 0 | 0 | 0 | 0 |
| 2026-08-31 | 0 | 1 | 0 | 0 | 0 |
| 2026-08-30 | 0 | 0 | 0 | 0 | 0 |
| 2026-08-29 | 1 | 0 | 0 | 0 | 0 |
| 2026-08-28 | 0 | 0 | 0 | 0 | 0 |
| 2026-08-27 | 0 | 0 | 0 | 0 | 0 |
Opened: 3
Closed: 0
Comments: 1
Events: 1
| Issue | Author | State | Labels | Comments | Reactions | Updated |
|---|---|---|---|---|---|---|
#4944 [Bug] TurboMind SIGSEGV crash-loop triggered by session_len truncation path (deterministic libc offset across 3 crashes) Opened 3 days ago | zambalee | open | No labels | 1 | 0 | 3 days ago |
#4933 [Bug] input_embeddings are omitted from prefix-cache identity, causing wrong KV reuse Opened 7 days ago | wildoranges | open | No labels | 0 | 0 | 7 days ago |
#4930 [Bug] Mooncake external KV keys omit KV format and weights lineage, with no default tenant isolation Opened 7 days ago | wildoranges | open | No labels | 2 | 0 | 7 days ago |
#4824 [Feature] DeepSeek-V4 MoE is Blackwell-only via DeepGEMM — is a non-Blackwell path wanted? Opened 1 month ago | Hakureirm | open | No labels | 9 | 0 | 10 days ago |
#4530 [Feature] support DFlash: Block Diffusion for Flash Speculative Decoding Opened 5 months ago | hicofeng | open | planned feature | 3 | 0 | 10 days ago |
#4905 [Feature] 咱们最新版本的lmdeploy可以加载这个 Qwen3.8-Flash吗？ Opened 15 days ago | simonjhy | open | No labels | 1 | 0 | 10 days ago |
#4889 [Feature] 什么时候turbomind能够开始支持MTP或者DFlash之类的加速功能？ Opened 21 days ago | simonjhy | open | No labels | 1 | 0 | 10 days ago |
#4917 [Bug] Prefix-cache KV is not invalidated by /update_weights: hot weight update silently reuses stale KV (wrong outputs, cached_tokens still reported) Opened 12 days ago | wildoranges | closed - completed | No labels | 1 | 0 | 11 days ago |
#4899 [Bug] PyTorch engine: Qwen3.5 hybrid-GDN + AWQ checkpoint crashes on TP>1 - 'AwqLinear' object has no attribute 'block_size' Opened 17 days ago | u5084757454-collab | open | No labels | 0 | 4 | 17 days ago |
#4781 [Bug] legacy /generate's non-streaming disconnect branch discards its 400 response, returns 200 null instead Opened 2 months ago | AmirF194 | closed - completed | No labels | 1 | 0 | 18 days ago |
#4883 [Bug] pytorch backend silently hangs under concurrent decoding with qwen3_5_mtp speculative decoding Opened 22 days ago | matrix72c | open | No labels | 1 | 0 | 21 days ago |
#4879 [Feature] 当前qwen-next等SSM类模型prefix cache缓存命中率低 Opened 23 days ago | Tsundoku958 | open | No labels | 0 | 0 | 22 days ago |
#4863 [Bug] turbomind can not support fp8 in SM75 Opened 27 days ago | bltcn | closed - completed | No labels | 2 | 0 | 22 days ago |
#4870 [Feature] lmdeploy 框架的 API 层:默认不支持思维深度设置为 xhigh 吗？ Opened 25 days ago | simonjhy | closed - completed | No labels | 0 | 0 | 25 days ago |
#4864 [Feature] 什么时候能够开始支持 https://huggingface.co/Qwen/Qwen3.8-27B？ Opened 27 days ago | simonjhy | closed - completed | No labels | 1 | 0 | 27 days ago |
#4865 [Feature] sm75 can support awq/gptq/fp8/w8a8 in turbomind Opened 27 days ago | bltcn | open | No labels | 0 | 0 | 27 days ago |
#4849 [Bug] int4 KV cache: the quantization range is stretched to include 0 when the packed head width is not a power of two Opened 29 days ago | truong-v | closed - completed | No labels | 0 | 0 | 28 days ago |
#4804 [Bug] Unauthenticated pickle deserialization via ZMQ in disaggregated serving → RCE Opened 1 month ago | AAtomical | closed - completed | No labels | 1 | 0 | 29 days ago |
#4832 [Feature] 现在最新的0.15.0可以支持最新的deepseek-v4-flash-0731吗？ Opened 1 month ago | simonjhy | open | No labels | 2 | 0 | 30 days ago |
#4329 [Bug] 基于https://github.com/InternLM/lmdeploy/pull/4320 构建Docker Image，启动GLM-4.7-Flash依然报错 Opened 7 months ago | simonjhy | closed - completed | No labels | 9 | 0 | 1 month ago |
#4400 [Bug] 在lmdeploy中加载glm-4.7-flash参数如何配置？ Opened 6 months ago | simonjhy | closed - completed | No labels | 6 | 0 | 1 month ago |
#3781 [Bug] fill_kv_cache 实现有bug Opened 1 year ago | zxy1123 | closed - completed | No labels | 1 | 0 | 1 month ago |
#4715 [Bug] 在使用qwen3.5-122B的时候，模型是多模态模型，但是在使用lmdeploy收到请求的时候，报错了 Opened 2 months ago | simonjhy | closed - not_planned | awaiting response Stale | 3 | 0 | 1 month ago |
#4743 [Bug] INT4 KV cache (--quant-policy 4) breaks qwen3coder tool call parser for Qwen3-Coder-30B Opened 2 months ago | zambalee | closed - not_planned | awaiting response Stale | 3 | 0 | 1 month ago |
#4776 [Bug] legacy /v1/completions disconnect-abort calls nonexistent AsyncEngine.stop_session Opened 2 months ago | AmirF194 | closed - completed | No labels | 1 | 0 | 1 month ago |
