]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/shortlog
pkg/ggml/sources/llama.cpp
2025-12-29 Georgi Gerganovserver : handle closed connection for tasks (#18459)
2025-12-29 Daniel Beveniusmodel-conversion : add device option to embd run orig...
2025-12-29 Héctor Estrada... retrieval : use at most n_seq_max chunks (#18400)
2025-12-29 o7sicommon: fix return value check for setpriority (#18412)
2025-12-29 Johannes GäßlerCUDA: Blackwell features for non-native builds (#18436)
2025-12-29 Aman Guptacuda: fix race condition in cumsum (#18448)
2025-12-28 Tim Neumannci : re-enable rocm build on amd64 (#18439)
2025-12-28 uvosHIP: Use mmq on MFMA devices for MUL_MAT_ID in cases...
2025-12-28 momongamodel : Plamo3 support (#17304)
2025-12-28 Aman GuptaRevert "ggml-cuda: use CMAKE_CUDA_ARCHITECTURES if...
2025-12-28 o7sirpc: fix segfault on invalid endpoint format (#18387)
2025-12-28 Johannes Gäßlerllama-fit-params: fix step size for last device (#18415)
2025-12-28 Johannes Gäßlergithub: update issue templates [no ci] (#18410)
2025-12-28 Xuan-Son Nguyenmtmd: clarify that we no longer accept AI-generated...
2025-12-28 Boian Berberovcmake: Added more x86_64 CPU backends when building...
2025-12-28 QDeltaggml-cuda: use CMAKE_CUDA_ARCHITECTURES if set when...
2025-12-27 lhezopencl: allow resizing transpose buffers (#18384)
2025-12-27 Johannes Gäßlerllama-fit-params: fix overflow check (#18354)
2025-12-27 Johannes Gäßlerllama: fix magic number of 999 for GPU layers (#18266)
2025-12-27 Aman Guptaggml-cuda: Use same regex for GGML_NATIVE=OFF (#18407)
2025-12-27 Johannes Gäßlerllama_fit_params: return enum for fail vs. error (...
2025-12-27 Johannes Gäßlerllama-fit-params: fix Gemma 3 calculation (#18372)
2025-12-26 Jeff Bolzvulkan: preprocess mul_mat_id experts and discard workg...
2025-12-26 Jeff Bolzvulkan: optimize decodeFuncB in coopmat2 mul_mat_id...
2025-12-26 Jeff Bolzvulkan: Use BK=32 for coopmat2 mul_mat_id (#18332)
2025-12-26 Evevulkan: small dequantization improvements (#18380)
2025-12-26 Jeff Bolzvulkan: Support UPSCALE w/antialias (#18327)
2025-12-26 Jeff Bolzvulkan: handle rope with large number of rows (#18306)
2025-12-26 o7siserver : fix crash when seq_rm fails for hybrid/recurre...
2025-12-26 Francisco Herreradocs: added note for pre SYCL Intel hardware (#18016)
2025-12-26 0MarbleCANN: implement the SSM_CONV operator (#17737)
2025-12-25 Aman Guptaggml-cuda: fix regex for arch list (#18371)
2025-12-25 Aman Guptacuda: optimize cumsum cub path (#18362)
2025-12-25 Aman Guptaggml-cuda: fix blackwell native builds (#18361)
2025-12-25 Penglin CaiCANN: Add support for CONV_TRANSPOSE_1D when kernel...
2025-12-25 Aadeshveer... ggml : optimize cuda cumsum fallback kernel (#18343)
2025-12-24 Xuan-Son Nguyenserver: (router) add stop-timeout option (#18350)
2025-12-24 Xuan-Son Nguyenmodel: support MiMo-V2-Flash (#18328)
2025-12-24 Aadeshveer... fit-params : fix race condition in fit-params output...
2025-12-24 Aman GuptaCUDA: experimental native mxfp4 support for blackwell...
2025-12-24 Saba Fallahmodel : support for LlamaBidirectionalModel architectur...
2025-12-24 Jeff Bolzvulkan: fix command buffer corruption in ggml_backend_v...
2025-12-24 Wang WeixuanCANN : refactor ACL graph cache (#17752)
2025-12-24 Jesse Ikonendocs: Fix typos in SYCL documentation (#18269)
2025-12-24 Ruben Ortlamvulkan: use fewer FA rows for small cache runs (#18280)
2025-12-24 TianHao324CANN: Uses yarn_ramp cache in ROPE (#17725)
2025-12-24 ddh0common: add `LLAMA_ARG_OVERRIDE_TENSOR` env var for...
2025-12-23 Xuan-Son Nguyenserver: return_progress to also report 0% processing...
2025-12-23 Pascalwebui: apply webui_settings on first load (#18223)
2025-12-23 Xuan-Son Nguyenserver: fix crash with model not having BOS/EOS (#18321)
2025-12-23 Daniel Beveniusmodel-conversion : add device option to run-org-model...
2025-12-23 Chris Rohlfrpc : add check for rpc buffer type (#18242)
2025-12-23 nullnameggml-hexagon: create generalized functions for cpu...
2025-12-23 Daniel Beveniusmodel-conversion : add trust_remote_code for embedding...
2025-12-23 Neo Zhang[SYCL] replace llama-cli by llama-completion to rm...
2025-12-23 Alessandro98-gitmodel : fix div-by-zero for Nemotron V2 (#18309)
2025-12-22 Ryan Mangenomodel : Granite Embedding support (#15641)
2025-12-22 compiladegguf-py : do not align the data start offset (#18291)
2025-12-22 Shouyuggml-hexagon: gelu optimization (#18151)
2025-12-22 Xuan-Son Nguyengen-docs: automatically update markdown file (#18294)
2025-12-22 Taimur Ahmadllamafile: add rvv support for sgemm kernels (#18199)
2025-12-22 lhezopencl: unpack q4_0 for adreno in get_tensor (#18278)
2025-12-22 Jeff Bolzvulkan: Extend rope fusions to allow mrope (#18264)
2025-12-22 Xuan-Son Nguyenserver: prevent data race from HTTP threads (#18263)
2025-12-22 Xuan-Son Nguyenserver: fix data race in to_json_anthropic (#18283)
2025-12-22 Matttrelease: update release workflow to store XCFramework...
2025-12-22 Aaron Teoconvert: rework ftype heuristics (#18214)
2025-12-22 Xuan-Son Nguyenserver: (docs) remove mention about extra_args (#18262)
2025-12-22 Johannes Gäßlertool/ex/tests: consistently free ctx, then model (...
2025-12-21 Jeff Bolzvulkan: Implement set_tensor_async and the event interf...
2025-12-21 Johannes Gäßlerllama: fix RPC for -fit on (#18233)
2025-12-21 Xuan-Son Nguyenmove copilot instructions to AGENTS.md (#18259)
2025-12-21 Jeff Bolzvulkan: fix im2col overflowing maxworkgroupcount (...
2025-12-21 Jeff Bolzvulkan/cuda: fix topk_moe with exp_probs_b (#18071)
2025-12-21 Jeff Bolzvulkan: support GGML_UNARY_OP_XIELU (#18062)
2025-12-21 Jeff Bolzvulkan: in graph_optimize, try to group ADD operations...
2025-12-21 lovedheartVulkan: some improvement on mul_mat_iq2_xs (#18031)
2025-12-21 Daniel Beveniusdocs : fix links in parsing.md (#18245)
2025-12-21 Aldehir Rojascommon : reorganize includes to prioritize vendored...
2025-12-21 Xuan-Son Nguyenserver: add auto-sleep after N seconds of idle (#18228)
2025-12-20 Jeff Bolztests: Avoid floating point precision false positives...
2025-12-20 Jeff Bolztest-backend-ops: improve msvc build time (#18209)
2025-12-20 Aadeshveer... Added comments explaining thread block size selection...
2025-12-20 Oleksandr Kuvshynovserver : [easy] fix per round speculative decode loggin...
2025-12-20 Xuan-Son Nguyenserver: support load model on startup, support preset...
2025-12-19 Sigbjørn Skjæretci : remove non-windows zip artifacts (#18201)
2025-12-19 Sigbjørn Skjæretci : only save ccache on master (#18207)
2025-12-19 Alfredggml-hexagon: Implement true Q8_0 quantization on Hexag...
2025-12-19 Pascalarg: fix order to use short form before long form ...
2025-12-19 Julius Tischbeinllama : Changing off_t to size_t for Windows (#18204)
2025-12-19 Aman Guptaserver: friendlier error msg when ctx < input (#18174)
2025-12-19 Xuan-Son Nguyenpresets: refactor, allow cascade presets from different...
2025-12-19 Aleksander... webui: Add editing attachments in user messages (#18147)
2025-12-19 Daniel Beveniusmodel-conversion : add verbose flag in run-org-model...
2025-12-19 Naco Sirenandroid: fix missing screenshots for Android.md (#18156)
2025-12-19 Jeff Bolzvulkan: Add perf logger mode with concurrency (#17944)
2025-12-18 Xuan-Son Nguyenmodel : add ASR support for LFM2-Audio-1.5B (conformer...
2025-12-18 Pascalwebui: display prompt processing stats (#18146)
2025-12-18 Taimur Ahmadggml-cpu: extend support for RVV floating-point kernels...
2025-12-18 Xuan-Son Nguyenarg: fix ASAN error on sampler_type_names empty (#18167)
next