Issues

• [Feature]: Add opt-in Muse Spark Open 816B-A42B support

  Niko1221/Strata#2762 · Filter by author BColsey opened on Oct 16, 2026

• Second server start should fail fast: PID lock before model loading

  Niko1221/Strata#2760 · Filter by author ver9118-droid opened on Oct 16, 2026

• Qwen3.6A3B fails with out of memory with recommended configs while Qwen3.8-Flash-Next unsloth Q4 runs fine

  Niko1221/Strata#2747 · Filter by author Coderx7 opened on Oct 16, 2026

• [Bug]: NVFP4 2-GPU layer split + STRATA_PREFILL_HELP=1: deterministic non-finite logits (#879) on 1024..3276-token single-chunk prompts (PeerPrefill omits tail_off scales)

  Niko1221/Strata#2745 · Filter by author sankerXD opened on Oct 16, 2026

• [FEATURE] QAD Model? - Nvidia's Quantization-Aware Distillation

  Niko1221/Strata#2744 · Filter by author SabinStargem opened on Oct 16, 2026

• iq_embed_rows: missing the >65535-row guard - fixed --prefill 65536 fails deterministically at a 64K chunk (invalid argument)

  Niko1221/Strata#2743 · Filter by author daifeiya opened on Oct 16, 2026

• error download model

  Niko1221/Strata#2741 · Filter by author wertrigone opened on Oct 16, 2026
  2

• Why Qwen3.6-35B-A3B run more and more slowly than Qwen3.8-Flash-Next

  Niko1221/Strata#2736 · Filter by author zhangiser opened on Oct 16, 2026
  1

• NVFP4 prompt modes on a 96 GB Blackwell card (every expert in VRAM): w4a4 reads fresh prompts 9-13% faster than w4a8

  Niko1221/Strata#2735 · Filter by author alphastorm opened on Oct 16, 2026

• Hope add support for the LMCache framework

  Niko1221/Strata#2734 · Filter by author chenwei0930 opened on Oct 16, 2026

• Profile Studio: when will the /profile routes from #2661/#2688 ship in a release?

  Niko1221/Strata#2733 · Filter by author ShaggyDragon opened on Oct 16, 2026
  1

• 2x2080Ti 44G 64G RAM UD-IQ4 512K 配置 KV 复用

  Niko1221/Strata#2732 · Filter by author lingopen opened on Oct 16, 2026

• serve: a reply that ends inside its thinking still returns an empty answer on 0.1.44.1 (the #1814 shape)

  Niko1221/Strata#2731 · Filter by author 47Hunter47 opened on Oct 16, 2026

• How to use tools like with Unsloth

  Niko1221/Strata#2730 · Filter by author avtech23 opened on Oct 16, 2026

• [Bug]: gfx1100 (RX 7900 XT 20 GB, 31 GiB RAM): amdkfd SVM restore hangs and kswapd order-0 alloc failures persist with HSA_USERPTR_FOR_PAGED_MEM=0; suspend with the SVM mapping active never resumes

  Niko1221/Strata#2729 · Filter by author Smuggz69 opened on Oct 16, 2026
  1

• [Feature]: Offline conversion of FP4-Local CSF checkpoints

  Niko1221/Strata#2726 · Filter by author hagope opened on Oct 16, 2026

• --prefill-head-alias: 3x slower cold prefill on a 16GB card, and the safety stop aborts the request at the end of prompt read

  Niko1221/Strata#2724 · Filter by author ShaggyDragon opened on Oct 16, 2026

• Not rendering Latex

  Niko1221/Strata#2723 · Filter by author zinahe opened on Oct 16, 2026

• Strata Profile Studio: a page that shows what limits Strata on your PC, and whether a tune helped

  Niko1221/Strata#2722 · Filter by author qiuzan2001 opened on Oct 16, 2026

• [Feature]: an opt-in setting to refuse a request whose model is not the loaded model

  Niko1221/Strata#2717 · Filter by author BoogeAI opened on Oct 16, 2026

• [Feature] Request support for the NEW NVIDIA NVFP4 checkpoint

  Niko1221/Strata#2716 · Filter by author Pragmatism0220 opened on Oct 16, 2026
  1

• --batch on the Metal backend is accepted at boot and exits the engine with code 86 on the first request that runs two sequences together

  Niko1221/Strata#2715 · Filter by author alvarorsouza-arch opened on Oct 16, 2026

• [Feature]: let a coding agent's own heavy jobs have the GPU and RAM, then continue the conversation without re-reading it

  Niko1221/Strata#2714 · Filter by author page1 opened on Oct 16, 2026

• let Cascade Lake CPUs (AVX512F/BW/VL/DQ + VNNI, no VBMI) use the AVX-512 i-quant kernels

  Niko1221/Strata#2713 · Filter by author zhai-hello opened on Oct 16, 2026

• Community report: Strix Halo (gfx1151, Windows): default 15 pool workers vs 6 + STRATA_VERIFY_PROFILE=1, 32.0 -> 41.7 tok/s in one ~120 W budget

  Niko1221/Strata#2711 · Filter by author yoheier opened on Oct 16, 2026