COMMITS
March 31, 2026
A
ggml-webgpu: port all AOT operators to JIT (#20728)
Abhijit Ramesh committed
A
fix: Use lower-case proxy headers naming (#21235)
Aleksander Grygier committed
A
common : cleanup logs and modernize the progress bar (#21215)
Adrien Gallouët committed
H
CANN: fix multi-thread set_tensor race conditions (#20151)
hipudding committed
X
server: (webui) no more gzip compression (#21073)
Xuan-Son Nguyen committed
A
common : gpt-oss handle builtin and unsolicited tool calls (#21213)
Aldehir Rojas committed
L
fix: correct misspellings in code comments (#21217)
lainon1 committed
S
CI: Enable CPU and Vulkan ARM64 Release (#21207)
Seungmin Kim committed
G
sync : ggml
Georgi Gerganov committed
A
common : move up common_init() and fix Windows UTF-8 logs (#21176)
Adrien Gallouët committed
N
sycl : enhance fattn perf (#21185)
Neo Zhang committed
M
S
fix: include API key in CORS proxy requests for MCP connections (#21193)
SATISH K C committed
P
server/webui: cleanup dual representation approach, simplify to openai-compat (#21090)
Piotr Wilkin (ilintar) committed
A
vendor : update BoringSSL to 0.20260327.0 (#21211)
Adrien Gallouët committed
G
March 30, 2026
G
ggml : bump version to 0.9.9 (ggml/1449)
Georgi Gerganov committed
S
opencl: add q4_K gemm and gemv kernels for Adreno (#20919)
shaofeiqi committed
S
CI : Enable CUDA and Vulkan ARM64 runners and fix CI/CD (#21122)
Seungmin Kim committed
Z
jinja : handle empty expressions correctly (#20913)
Zhihao "Zephyr" Yao committed
O
CUDA : Fix CUB's argsort when nrows % block_size == 0 CCCL < 3.1 (#21181)
Oliver Simons committed
R
rpc : fix misleading error log (#21184)
Radoslav Gerganov committed
A
webui: Fix branching logic on edit message (#21175)
Aleksander Grygier committed
A
llama-model-loader: print warning when using overrides with mmap (#20978)
Aman Gupta committed
S
ci : bump ty to 0.0.26 (#21156)
Sigbjørn Skjæret committed
X
server: wrap headers for mcp proxy (#21072)
Xuan-Son Nguyen committed
March 29, 2026
S
add missing ROPE_FACTORS_LONG/SHORT for MiniCPM (#21150)
Sigbjørn Skjæret committed
G
Optimize MOE GEMV kernel for BS > 1. (#20905)
Gaurav Garg committed
M
hexagon: dma optimizations (mostly fixing regressions) (#21137)
Max Krasnyansky committed
D
devops: including compute-runtime for intel.Dockerfile (#21076)
Davi Henrique Linhares committed