COMMITS
August 26, 2026
A
ui: Move Settings and MCP Servers routes to dialog-based views (#27744)
Aleksander Grygier committed
A
ui: Update Dialog component styling (#27743)
Aleksander Grygier committed
R
P
Implemented vulkan cross_entropy_loss and cross_entropy_loss_back (#27216)
Pranav Uttarkar committed
R
rpc : implement event and async backend APIs (#18626)
Radoslav Gerganov committed
A
ci: Clean up UI builds from releases (#27706)
Aleksander Grygier committed
D
cuda: unblock mmq for MoE on sm_60 (#26264)
David Friehs committed
S
ci : make cache bucket public (#27728)
Sigbjørn Skjæret committed
D
conversion : fix Nemotron 3.5 Lightning layers (#27729)
Daniel Bevenius committed
D
ggml-metal: add chunked SSD MMA for Mamba-2 prefill optimization (#26647)
Dominik Pantaleoni committed
M
ggml-meta: propagate buffer usage and call init on the new tensors (#27586)
Max Krasnyansky committed
August 25, 2026
J
kleidiai: Rework KleidiAI Build System/Integration (#26077)
Jonathan Clohessy committed
M
ci : update OS used for ROCM to Ubuntu 24.04 (#27681)
Mario Limonciello committed
R
rpc: support apple RDMA as an RPC transport (#26421)
Ryan C committed
K
devops: use GGML_NATIVE=OFF for OpenVINO (#27338)
krzsztf committed
K
A
chat : scope qwen3-coder workarounds (#27679)
Aldehir Rojas committed
S
ci : store ccache on HF buckets (test with cuda-ubuntu for now) (#27699)
Sigbjørn Skjæret committed
A
ui: ESLint config updates (#27700)
Aleksander Grygier committed
Y
metal : null-check buffer alloc to fix OOM crash (#25371)
Yuri Khrustalev committed
G
llama.cpp : bump version to 0.3.0 (#27696)
Georgi Gerganov committed
G
sync : ggml
Georgi Gerganov committed
G
ggml : bump version to 0.22.0 (ggml/1607)
Georgi Gerganov committed
S
grammar : parse \- in char classes as literal hyphen (#27591)
Saad Ali committed
N
sycl : mark tq2_0 as not supported (#27660)
Neo Zhang committed
F
webgpu : fix handling of infinity values during ARGSORT and TOP_K (#27538)
fairydreaming committed
August 24, 2026
Y
metal : per-device tuned (Q, NE) for flash-attn vec (#26570)
YiChen Lv committed
Y
metal: per-op source split + parallel compile (#26561)
YiChen Lv committed
T
misc : read repetition_penalty from generation_config.json (#27659)
Tarek Dakhran committed
F
tests : disable DOTS3NOTE arch test for WebGPU (#27654)
fairydreaming committed