-
Notifications
You must be signed in to change notification settings - Fork 1.5k
Pull requests: kvcache-ai/ktransformers
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
Add MiniMax-M3 and MiniMax-M2.7 to the kt-cli model registry
#2112
opened Jul 31, 2026 by
octo-patch
Loading…
fix(kt-kernel): move write_weight_scale_to_buffer helpers to BaseMoEWrapper (#2108)
#2111
opened Jul 30, 2026 by
Anai-Guo
Contributor
Loading…
fix(cli): match model files by relative path when verifying
#2104
opened Jul 23, 2026 by
ebarkhordar
Loading…
2 tasks done
fix(kt-kernel): pick default CUDA archs by nvcc version (fixes #2098)
#2103
opened Jul 23, 2026 by
Anai-Guo
Contributor
Loading…
fix(kt-kernel): harden qwen3.5 build script for setuptools Path / rglob
#2097
opened Jul 20, 2026 by
fooSynaptic
Loading…
[perf](kt-kernel): blocked accumulators and opt-in packed INT4 weights for the AVX-VNNI-256 RAWINT4 backend
#2095
opened Jul 19, 2026 by
blazingphoenix7
Contributor
Loading…
2 tasks done
[feat](kt-kernel): add co-activation-aware and online-EMA GPU expert placement
#2093
opened Jul 18, 2026 by
neerajdad123-byte
Loading…
2 tasks done
[fix](kt-kernel): fix AMX BF16 Full-FT expert gradients
#2086
opened Jul 13, 2026 by
Illumination111
Contributor
Loading…
2 tasks done
[feat](kt-kernel): opt-in skip of CPU copies for GPU-resident experts
#2085
opened Jul 12, 2026 by
gvr13n
Loading…
[build] Sync sglang-kt submodule (v0.6.4)
automated
dependencies
#2069
opened Jul 4, 2026 by
github-actions
Bot
Loading…
feat(kt-kernel): add score-aware expert placement
#2064
opened Jun 25, 2026 by
githubshaurya
Loading…
fix(balance_serve): bind scheduler RPC to loopback to close pre-auth …
#2043
opened Jun 16, 2026 by
AAtomical
Loading…
2 tasks
[fix(docker)]: build fine-tune env from source instead of the unpublished release wheel
run-ci
#2028
opened May 31, 2026 by
devangpratap
Contributor
Loading…
1 of 2 tasks
[Feature] Add MESH expert residency with io_uring direct I/O for MoE inference
#2003
opened May 13, 2026 by
RaQiu
Loading…
1 of 2 tasks
ProTip!
Find all pull requests that aren't related to any open issues with -linked:issue.