Skip to content

Pull requests: kvcache-ai/ktransformers

New pull request New
42 Open 704 Closed
42 Open 704 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

[feat](kt-kernel): --kt-mmap-experts-dir to keep CPU expert weights in a file, not anonymous RAM
#2190 opened Sep 7, 2026 by tim-odonnell Loading…
[feat] Add native RAWINT4 routed-expert LoRA SFT for Kimi K2.5/K2.6
#2186 opened Sep 4, 2026 by yyj6666667 Collaborator Loading…
[feat](kt-kernel): add native MXFP4 routed-expert LoRA SFT
#2185 opened Sep 3, 2026 by yyj6666667 Collaborator Loading…
2 tasks done
[fix](kt-sft): scale chunked_prefill_size by the per-device batch size
#2180 opened Aug 28, 2026 by devangpratap Contributor Loading…
2 tasks done
1
1
[fix](kt-kernel): keep CPU expert buffers alive during CUDA graph capture
#2174 opened Aug 27, 2026 by 05yuki Contributor Loading…
1 of 2 tasks
chore: update third_party/sglang to support_telechat4
#2168 opened Aug 19, 2026 by 157933996 • Draft
[build] Sync sglang-kt submodule (v0.7.0.post1) automated dependencies
#2166 opened Aug 19, 2026 by github-actions Bot Loading…
fix(kt-kernel): AMXMoEWrapper cpu_save branch crashes on load_experts keys/prefix (#2151)
#2161 opened Aug 18, 2026 by Anai-Guo Contributor Loading…
[feat]: add AMXINT4 SMART 3-dtype wrapper-conversion MoE routing
#2155 opened Aug 14, 2026 by NukePeace Loading…
1 of 2 tasks
12
[WIP][feat](kt-kernel): Add Ascend NPU CPU-MoE backend + llama.cpp b3173 patch support
#2153 opened Aug 13, 2026 by Nikonz • Draft
2 tasks done
1
[fix](cli): do not compare a git SHA1 against a local SHA256 when verifying
#2142 opened Aug 5, 2026 by adityasingh2400 Loading…
1
docs: add API benchmark example
#2138 opened Aug 5, 2026 by zcxGGmu Loading…
5
docs: document Qwen3-Next balance serve build
#2137 opened Aug 5, 2026 by zcxGGmu Loading…
4
docs: add Ascend NPU example links
#2135 opened Aug 4, 2026 by zcxGGmu Loading…
1
docs: clarify CPU-only runtime path
#2134 opened Aug 4, 2026 by zcxGGmu Loading…
4
docs: add flash-attn import troubleshooting
#2133 opened Aug 4, 2026 by zcxGGmu Loading…
3
docs: document build parallel limit
#2132 opened Aug 4, 2026 by zcxGGmu Loading…
1
docs: clarify Qwen3 model metadata path
#2131 opened Aug 4, 2026 by zcxGGmu Loading…
3
docs: pin transformers for MiniMax M2.5
#2130 opened Aug 4, 2026 by zcxGGmu Loading…
3
docs: document CUDA architecture override
#2129 opened Aug 4, 2026 by zcxGGmu Loading…
3
docs: add model provenance note
#2126 opened Aug 3, 2026 by zcxGGmu Loading…
1
docs: add CUDA host compiler troubleshooting
#2125 opened Aug 3, 2026 by zcxGGmu Loading…
4
docs: add submodule install troubleshooting
#2124 opened Aug 3, 2026 by zcxGGmu Loading…
1
docs: clarify kt-kernel is not AMX only
#2123 opened Aug 3, 2026 by zcxGGmu Loading…
3
docs: answer CPU and GPU weights FAQ
#2122 opened Aug 3, 2026 by zcxGGmu Loading…
1
ProTip! no:milestone will show everything without a milestone.