]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/shortlog
pkg/ggml/sources/llama.cpp
2025-12-11 PascalAdd a search field on model selector / improve mobile...
2025-12-11 Piotr Wilkin... SOLVE_TRI extension to more dimensions (#17793)
2025-12-11 Georgi Gerganovggml-alloc : fix reuse-parent logic for misaligned...
2025-12-11 Georgi Gerganovbatch : fix sequence id ownership (#17915)
2025-12-11 Yuichiro Utsumidocs: use port 8080 in Docker examples (#17903)
2025-12-10 nullnameggml-hexagon: fix `rope` failure at `test-backend-ops...
2025-12-10 Sigbjørn Skjæretci: fix riscv64-native build (#17916)
2025-12-10 Xuan-Son Nguyenmtmd: some small clean up (#17909)
2025-12-10 Xuan-Son Nguyencli: enable jinja by default (#17911)
2025-12-10 Pascalserver: add presets (config) when using multiple models...
2025-12-10 Max KrasnyanskyFix race conditions in threadpool when dealing with...
2025-12-10 Georgi Gerganovggml : remove GGML_KQ_MASK_PAD constant (#17910)
2025-12-10 Sigbjørn Skjæretcuda : add missing support check for xielu (#17895)
2025-12-10 Xuan-Son Nguyencli: new CLI experience (#17824)
2025-12-10 Eric Zhangmodel : Qwen3-Next-80B-A3B has 48 layers (#17898)
2025-12-10 lhezdocs : update opencl ops (#17904)
2025-12-10 Johannes GäßlerCUDA: fix unpadded strides in MMA FA kernel (#17891)
2025-12-10 Xuan-Son Nguyenconvert: allow using quantized Mistral weight (#17889)
2025-12-10 Neo Zhang Jianyufix softmax for iGPU (#17838)
2025-12-09 Aldehir Rojascommon : add parser for ministral/mistral large 3/devst...
2025-12-09 Sigbjørn Skjæretdocs : update cpu and cuda ops (#17890)
2025-12-09 Gabe Goodhartmetal: SSM kernel improvements (#17876)
2025-12-09 Piotr Wilkin... Add DIAG for CUDA (#17873)
2025-12-09 Johannes Gäßlerdocs: clarify that CPU support should be first (#17886)
2025-12-09 Gabe Goodhartggml : Provide macos-specific backtrace printing to...
2025-12-09 Georgi Gerganovmetal : print node names for debugging (#17882)
2025-12-09 Sigbjørn Skjæretggml : allow fill node alloc inplace (#17870)
2025-12-09 Rhys-Tcmake: fix Mach-O current version number (#17877)
2025-12-09 Sigbjørn Skjæretmodel : nit, DeepSeek V1 MoE is 16B and GigaChat is...
2025-12-09 Xuan-Son Nguyenconsole: allow using arrow left/right, home/end keys...
2025-12-09 Chenguang LiCANN: add support for partial RoPE and Vision mode...
2025-12-09 Johannes GäßlerCUDA: fix FP16 overflow in tile FA kernel (#17875)
2025-12-09 Aldehir Rojasllama : add token matching support to llama-grammar...
2025-12-09 philip-essentialmodel : support Rnj-1 (#17811)
2025-12-08 Sigbjørn Skjæretgraph : use fill instead of scale_bias in grouped exper...
2025-12-08 Daniel Beveniusmodel-conversion : add token ids to prompt token output...
2025-12-08 Xuan-Son Nguyenserver: delegate result_state creation to server_task...
2025-12-08 Neo Zhangci : support bfloat16 SYCL release package (#17855)
2025-12-08 Xuan-Son Nguyenserver: improve speed of speculative decoding (#17808)
2025-12-08 Piotr Wilkin... Make graph_max_nodes vary by ubatch size (#17794)
2025-12-08 hksdpc255Fix Kimi-K2 tool-call parsing issues (#17376)
2025-12-08 Jay Zenithcuda : add FILL op support (#17851)
2025-12-08 Xuan-Son Nguyenserver : add development documentation (#17760)
2025-12-08 Georgi Gerganovserver : make cache_reuse configurable per request...
2025-12-08 wsbagnsv1cuda: optimize SOLVE_TRI using registers and FMAF ...
2025-12-08 ixgbeggml-cpu: add ggml_thread_cpu_relax with Zihintpause...
2025-12-07 Xuan-Son Nguyenmodel: add llama 4 scaling for mistral-large (deepseek...
2025-12-07 lovedheartVulkan: improve mul_mat_vec_iq1_m (#16907)
2025-12-07 Sigbjørn Skjæretci : add windows-cuda 13.1 release (#17839)
2025-12-07 Sigbjørn Skjæretcommon : change --color to accept on/off/auto, default...
2025-12-07 Law Po Yingsycl: add missing BF16 conversion support for Intel...
2025-12-06 Jeff Bolzvulkan: perf_logger improvements (#17672)
2025-12-06 Vishal Singhggml-zendnn : add ZenDNN backend for AMD CPUs (#17690)
2025-12-06 Xuan-Son Nguyenserver: support multiple generations from one prompt...
2025-12-06 Phylliida Devggml : add circular tiling support to pad, for Vulkan...
2025-12-06 Johannes GäßlerHIP: fix RDNA3 FP16/BF16 matrix multiplication (#17817)
2025-12-06 Aleksander... webui: Stop generation from chat sidebar (#17806)
2025-12-06 Aleksander... webui: Fix context available value in Multi-model Route...
2025-12-06 Aleksander... webui: Per-conversation system message with UI displayi...
2025-12-06 Skyggml : improve error handling for search path existence...
2025-12-06 Daniel Beveniusllama : remove quantization sanity check (#17788)
2025-12-06 Jeff Bolzvulkan: Use one row per workgroup for f32 mmv (#17711)
2025-12-06 Xuan-Son Nguyenconvert: support Mistral 3 Large MoE (#17730)
2025-12-06 Jeff Bolzvulkan: support solve_tri with larger N/K values (...
2025-12-06 Georgi Gerganovcontrib : stale PRs (#17803)
2025-12-06 Georgi Gerganovmetal : fix build(#17799)
2025-12-06 Masato Nakasakavulkan: Replace deprecated VK_EXT_validation_features...
2025-12-06 Masato Nakasakavulkan: Fix mismatch in TOPK_MOE unit test (#17541)
2025-12-05 Jeff Bolzvulkan: add more num_blocks instantiations in rms_norm...
2025-12-05 Jeff Bolzvulkan: fix top_k bug when there are ties in the input...
2025-12-05 Aclyvulkan : support conv-2d with large output size (#17685)
2025-12-05 Reese Levineggml webgpu: unary op suppport, code refactoring, ops...
2025-12-05 Jeff Bolzvulkan: enable mmvq for q2_k on NVIDIA (#17675)
2025-12-05 Jeff Bolzvulkan: set all memory allocations to high priority...
2025-12-05 Georgi Gerganovrpc : fix alloc size logic (#17116)
2025-12-05 Georgi Gerganovmetal : add residency sets keep-alive heartbeat (#17766)
2025-12-05 Johannes GäßlerHIP : fix RDNA4 build (#17792)
2025-12-05 Pascalfix: prevent segfault in tokenizer on highly repetitive...
2025-12-05 Adrien Gallouëtci : fix winget workflow (#17790)
2025-12-05 shalinib-ibmQ4/Q8 Tiled Gemm Optimization. (#16999)
2025-12-05 Piotr Wilkin... Add pwilkin to CODEOWNERS for chat files (#17789)
2025-12-05 Johannes GäßlerCUDA: fix FA VKQ accumulator overflow (#17746)
2025-12-05 Jiacheng (Jason... HIP: enable WMMA-MMQ INT kernels for RDNA 3 (#17576)
2025-12-05 Sigbjørn Skjæretci : transform release binary root dir in tar to llama...
2025-12-04 Gabe Goodhartdocs : update ops.md (Metal, BLAS) (#17768)
2025-12-04 Piotr Wilkin... Add support for CUMSUM and TRI for CUDA. (#17584)
2025-12-04 Gabe Goodhartmetal: TRI, FILL, EXPM1, SOFTPLUS (#16623)
2025-12-04 Xuan-Son Nguyenserver: strip content-length header on proxy (#17734)
2025-12-04 Xuan-Son Nguyenserver: move msg diffs tracking to HTTP thread (#17740)
2025-12-04 Daniel Beveniusexamples : add missing code block end marker [no ci...
2025-12-04 Daniel Beveniuscommon : skip model validation when --help is requested...
2025-12-04 Alberto Cabrera... ggml-cpu : remove asserts always evaluating to false...
2025-12-04 SmartestWashingMachineconvert: use existing local chat_template if mistral...
2025-12-04 Adrien Gallouëtcmake : simplify build info detection using standard...
2025-12-04 Sigbjørn Skjæretci : disable ggml-ci-x64-amd-* (#17753)
2025-12-04 Adrien Gallouëtcommon: use native MultiByteToWideChar (#17738)
2025-12-04 Georgi Gerganovmetal : use params per pipeline instance (#17739)
2025-12-04 Georgi Gerganovllama : fix sanity checks during quantization (#17721)
2025-12-04 Adrien Gallouëtbuild : move _WIN32_WINNT definition to headers (#17736)
2025-12-04 Jeff Bolzbuild: enable parallel builds in msbuild using MTT...
next