LLM inference in C/C++
C++ ggml
Commits on master
Revert "common: share thread pools when `n_threads` differ (#27138)" (#27337)
3e73446 Niklas Wenzel committed 3 months ago
model : GraniteSWAForCausalLM / GraniteMoeSWAForCausalLM (#25505)
7221e24 Gabe Goodhart committed 3 months ago
sycl: report zero devices instead of aborting when the host has none (#27291)
6cc504a Titaniumtown committed 3 months ago
ggml: add ggml_rope_set_offset (+ metal support) (#27120)
fe8156f Xuan-Son Nguyen committed 3 months ago
server: add dedup-cache-models preset option (#27346)
ee4c505 Xuan-Son Nguyen committed 3 months ago
vulkan: tiled transpose for 0<->2 permuted CONT (#26585)
98d1e92 Kevin Hopper committed 3 months ago
ggml-webgpu: add mulmat with overlapping src0/src1 (e.g., for minimax-01) (#27321)
5112b97 Masashi Yoshimura committed 3 months ago
ci : add attestation for signed release artifacts (#25933)
0adcc3b Jeremie Miller committed 3 months ago
tests: skip test-unicode build on win32/BUILD_SHARED_LIBS (#27336)
79fe799 Jeff Bolz committed 3 months ago
server: (cosmetic) do not print cmd_child_to_router messages [no release] (#27347)
6d05498 Xuan-Son Nguyen committed 3 months ago
opencl: fix WAR race in the generic FA tile kernels when the WG spans subgroups (#26434)
60adddd Hongqiang Wang committed 3 months ago
mtmd: deepseek-ocr SAM ggml_conv_2d with the im2col kept in F32 (#26727)
9731ad3 Saba Fallah committed 3 months ago
RPC: populate use_count to enable fusion inside backends (#27142)
af51726 Aman Gupta committed 3 months ago
common: share thread pools when `n_threads` differ (#27138)
04b5691 Niklas Wenzel committed 3 months ago
quant : Optimise memory usage by evicting weights after processing each layer (#22877)
0596704 Ed Addario committed 3 months ago
unicode : include '~' in collapsed symbol class (#26972)
afd439d Thiago Padilha committed 3 months ago
readme : update status badges + regen AUTHORS (#27317)
169e4a7 Georgi Gerganov committed 3 months ago
ci : Update OpenVINO to 2026.3, skip nemotron-h rollback test (#27292)
9d77fa1 Zijun Yu committed 3 months ago
ci: add Windows ARM64 CUDA support to the manual workflow (#27300)
c029602 shivamkumard-ctrl committed 3 months ago