Recent changes
3 hours ago ggml : handle graph buffer reservation failure (#26070) (911f6cdc8)
3 hours ago vulkan: add IQ3_S MMQ matmul kernels (#28822) (bbd488c42)
4 hours ago vocab : add ufakzeka pre-tokenizer (#29033) (dc85f89c7)
4 hours ago cmake : fix build when GGML_CPU=OFF and GGML_CUDA=ON (#29026) (8ed1a55ef)
4 hours ago gguf-py: fix Q8_1 block size in GGML_QUANT_SIZES (2+2+32) (#29036) (bb11ebb68)
4 hours ago ci : disable GHA cache for copilot (#29068) (f03cf3e9b)
5 hours ago ci : bump android-actions/setup-android to 4.0.4 (#29065) (bdcbaaf6e)
5 hours ago vulkan: raise the hoisted row-id limit for mul_mat_id from 256 to 512 experts (#28501) (5c53396b8)
19 hours ago ci : add missing evict-old-files (#29041) (972d2313b)
21 hours ago rpc : skip ACCEL devices (#29020) (c77ae695c)
24 hours ago model : skip gate_up_exps if TENSOR_SKIP is set (#29014) (b49650adb)
24 hours ago model : extend Nemotron MTP support (#29018) (707618048)
25 hours ago openvino : Update OpenVINO to 2026.4;fix clangd,MSVC warnings; (#29009) (ebbb18522)
27 hours ago ui: fix removed reasoning menu in single model mode on desktop (#27985) (4ff829ec2)
28 hours ago vulkan: split buffers and debug code into separate files, add shared headers (#28732) (f172be756)
28 hours ago ci : add API/ABI check to make-release workflow [no ci] (#28947) (87f9c82f2)
29 hours ago chat : add message delimiters to the DeepSeek V3.2/V4 parser (#29008) (7f6f0c2a9)
29 hours ago gguf : align the data section relative to the GGUF start, not the file (#28993) (81aeaeb74)
29 hours ago sycl : fix the B70 mem allocate error when >19.3GB (#28953) (c9a5eeeb3)
29 hours ago vulkan: skip unneeded MoE work in mul_mm coopmat1 path (#25483) (7490357f2)
29 hours ago sycl: ssm_conv: fuse the SiLU epilogue into the ssm_conv kernel (#28929) (817e5f83e)
29 hours ago opencl: fix various warnings (#28984) (c57da6fd8)
30 hours ago docs: remove JG as CODEOWNER for test-llama-archs (#29003) (79bfc1d43)
31 hours ago vulkan: fix buffer_reference alignment in im2col shaders (#28996) (05f2dcfdb)
32 hours ago vulkan: support qwen4exp hc ops (#28988) (35822afe5)
34 hours ago [SYCL] Fix function signature for `ggml_backend_sycl_split_buffer_type` (#28981) (aa39d7a3e)
2 days ago vulkan: work around NV bug with argsort_large.comp (#28975) (4bc272fd7)
2 days ago TP: fix split state and granularity for fused QKV gemma4, qwen35 (#28965) (fb27a525d)
2 days ago ci: switch fast jobs back to github (#28959) (c6824a9e4)
2 days ago Enable CUDA graph for MTP draft (#28549) (2f3fd0252)