feat(kimi-k3): Kimi K3 support: MXFP4 to BF16 conversion and inference graph - #58

Open

worthant wants to merge 148 commits into

feat/kimi-k3-support

Commits on Jul 15, 2026

sycl : support kernel type fp16 for conv2d_dw (#25653)
arthw
authored
sycl : fix get_rows Q2_K, Q4_K, Q5_K (#25656)
arthw
authored
DeepseekV4: reduce graph splits (#25702)
am17an
authored
ggml: add f16 out_prod support for CPU and out_prod op for Vulkan (#23997)
Lamothe
authored
metal: fuse snake activation (mul, sin, sqr, mul, add) (#25459)
ServeurpersoCom
authored
ci : add HF_TOKEN to self-hosted workflows (#25706)
ggerganov
authored
cuda : relax tensor contiguity requirements for quantized concat (#25678)

fairydreaming
and
sszymczy
authored
ui: fix thinking menu never appearing in single-model mode (#25637)
ServeurpersoCom
authored
server: fix read_file append_loc space breaking edit_file match (#25705)
ServeurpersoCom
authored
Fix crash with draft-simple (#25720)
gaugarg-nv
authored
CUDA: tighter MMQ src1 buffer size for native fp4 (#25613)
leonardHONG
authored
opencl: fix two issues on flash attention for Adreno a7x (#25697)
wanghqc
authored
tokenize : drop --stdin mutual-exclusion check (#25672)
angt
authored
cuda : CUDA GGML_OP_LIGHTNING_INDEXER implementation (generic vector kernel + wmma kernel) (#25545)

fairydreaming
and
sszymczy
authored
ui: Agentic Content UX improvements (#25450)
allozaur
authored
opencl: exclude some moe kernels on Adreno a7x (#25698)
wanghqc
authored

Commits on Jul 16, 2026

Commits on Jul 17, 2026

Commits on Jul 18, 2026

Commits on Jul 20, 2026

Commits on Jul 21, 2026

Commits on Jul 22, 2026

Commits on Jul 24, 2026

Commits on Jul 25, 2026

Commits on Jul 26, 2026

Commits on Jul 27, 2026

Commits on Jul 29, 2026

Commits on Jul 30, 2026

fix(kimi-k3): label the 93-layer config A104B, not A50B
worthant
committed