Issues

• trim & improve: reduce knowledge contribution cost and generated-file churn

  area:knowledgekind:taskstatus:proposed
  InferMatrixLab/InferRSI#355 · Filter by author Dong1017 opened on Nov 5, 2026
  3

• acceptance: DeepSeek V4.1 Flash 910C 8K/1K TPOT ≤25 ms/token at C1/C4/C16

  area:workbenchkind:acceptance
  InferMatrixLab/InferRSI#353 · Filter by author jiangkuaixue123 opened on Nov 4, 2026
  26

  

• task: distill Ascend ModelAgent (model-agent) into source-grounded candidate skills

  area:knowledgekind:taskstatus:proposed
  InferMatrixLab/InferRSI#352 · Filter by author wuhang2014 opened on Nov 4, 2026
  1

• task: 细化 vLLM Ascend 知识蒸馏与知识库整理

  area:knowledgekind:taskstatus:proposed
  InferMatrixLab/InferRSI#350 · Filter by author jiangkuaixue123 opened on Nov 4, 2026
  9

• task: anchor L1 weight demand against the checkpoint in performance-ceiling-analysis

  area:knowledgekind:taskstatus:proposed
  InferMatrixLab/InferRSI#348 · Filter by author fake0fan opened on Nov 4, 2026
  1

• task: Dashboard 增强：每轮改动说明、动态 workflow 与耗时、测试进度

  area:usabilityarea:workbenchkind:taskstatus:proposed
  InferMatrixLab/InferRSI#345 · Filter by author jiangkuaixue123 opened on Nov 4, 2026
  1

• task: 支持每个优化迭代的知识库命中归因统计

  area:knowledgearea:workbenchkind:taskstatus:proposed
  InferMatrixLab/InferRSI#344 · Filter by author jiangkuaixue123 opened on Nov 4, 2026
  2

• task: connect engineering tasks to existing Agent execution and evaluation

  area:workbenchkind:taskstatus:proposed
  InferMatrixLab/InferRSI#337 · Filter by author Bounty-hunter opened on Nov 3, 2026
  5

• tracking: grow reusable engineering optimization tasks

  area:workbenchbot-reviewedkind:taskpriority:P2status:proposed
  InferMatrixLab/InferRSI#336 · Filter by author Bounty-hunter opened on Nov 3, 2026
  5

• discussion: what trace evidence should InferRSI persist beyond per-run artifacts?

  area:processarea:usabilityarea:workbenchbot-reviewedhelp:wantedkind:rfcpriority:P2status:proposed
  InferMatrixLab/InferRSI#333 · Filter by author Dong1017 opened on Nov 3, 2026
  7

• [Task] Deploy Minimax H3 in 5070 and enhance the performance using InferRSI

  area:workbenchbot-reviewedhelp:wantedkind:taskpriority:P2status:proposed
  InferMatrixLab/InferRSI#331 · Filter by author congw729 opened on Nov 3, 2026
  5

• [Campaign] MiniMax-H3 Ref2VA lossless optimization on 4x910B1

  area:workbenchbot-reviewedkind:acceptancestatus:proposed
  InferMatrixLab/InferRSI#320 · Filter by author amy-why-3459 opened on Nov 2, 2026
  4

• [Roadmap] InferRSI: Optimization, Evaluation, and Engineering

  area:processbot-reviewedkind:roadmap
  InferMatrixLab/InferRSI#316 · Filter by author david6666666 opened on Nov 2, 2026
  1

• [Task] 0 Day Model Support: Reusable Integration Experience from Transformers to GPU and NPU Inference Engines

  area:knowledgebot-reviewedkind:taskstatus:proposed
  InferMatrixLab/InferRSI#315 · Filter by author david6666666 opened on Nov 2, 2026
  1

  

• [Task] RSI Dashboard: Optimization Workflow Visualization and Trace Capture

  area:usabilityarea:workbenchbot-reviewedkind:rfckind:taskstatus:proposed
  InferMatrixLab/InferRSI#314 · Filter by author david6666666 opened on Nov 2, 2026
  3

  

• [RFC] Per-task SOL modeling and a KDA-flow gap metric for fused operators

  area:knowledgekind:rfcstatus:proposed
  InferMatrixLab/InferRSI#303 · Filter by author fake0fan opened on Nov 2, 2026
  4

  

• [Campaign] MiniMax-H3 Ref2VA on 910C: 4 logical devices, 5s/15s × lossless/lossy optimization

  area:workbenchkind:acceptance
  InferMatrixLab/InferRSI#297 · Filter by author Dong1017 opened on Nov 1, 2026
  1

  

• [Draft RFC / Experimental] Async Workbench：独立 manager 与 loop worker 的上下文和生命周期

  area:workbenchkind:rfcstatus:proposed
  InferMatrixLab/InferRSI#295 · Filter by author Dong1017 opened on Nov 1, 2026
  1

• [Campaign] MiniMax-H3 Ref2VA Ascend NPU latency optimization with video/audio quality gates

  area:workbenchkind:acceptance
  InferMatrixLab/InferRSI#293 · Filter by author amy-why-3459 opened on Oct 31, 2026
  2

• [RFC] Integrate AISBench as an optional benchmark tool

  area:workbenchkind:rfcstatus:proposed
  InferMatrixLab/InferRSI#289 · Filter by author wuhang2014 opened on Oct 31, 2026
  3

  

• [Status 03] InferRSI evolution: 48e0d589 → 68108700 (2026-10-31 catch-up)

  area:processbot-reviewedkind:status
  InferMatrixLab/InferRSI#286 · Filter by author Dong1017 opened on Oct 31, 2026
  6

• [RFC] Online quantization knowledge, skills and Skill Indexer selection

  area:knowledgekind:rfcstatus:proposed
  InferMatrixLab/InferRSI#254 · Filter by author zengchuang-hw opened on Oct 29, 2026
  1

• [Tracking] Minimal engineering benchmarks shared by Workbench, SkillEval and SearchSpace experiments

  area:workbenchbot-reviewedhelp:wantedkind:trackingpriority:P2status:proposed
  InferMatrixLab/InferRSI#252 · Filter by author Dong1017 opened on Oct 29, 2026
  14

  

• task: parallel optimization rounds with orthogonal lanes for multi-agent harnesses

  area:processarea:workbenchhelp:wantedkind:task
  InferMatrixLab/InferRSI#249 · Filter by author david6666666 opened on Oct 29, 2026
  6

  

• task: measure whether InferRSI helps Claude Code and Codex optimize inference tasks

  area:knowledgearea:processarea:workbenchhelp:wantedkind:task
  InferMatrixLab/InferRSI#248 · Filter by author david6666666 opened on Oct 29, 2026
  11