]> git.djapps.eu Git - pkg/ggml/sources/llama.cpp/shortlog
pkg/ggml/sources/llama.cpp
2026-02-13 Adrien Gallouëtllama : remove deprecated codecvt (#19565)
2026-02-13 Adrien Gallouëtvendor : update BoringSSL to 0.20260211.0 (#19562)
2026-02-13 Georgi Gerganovmemory : fix kv cache size for hybrid models (#19559)
2026-02-13 Georgi Gerganovmetal : improve concurrency (#19555)
2026-02-13 Georgi Gerganovmetal : support GGML_OP_SET (#19548)
2026-02-12 Shupei Fanhexagon: fix typo in vtcm_needs_release (#19545)
2026-02-12 lhezopencl: add basic support for q4_1 (#19534)
2026-02-12 Georgi Gerganovargs : add -kvu to llama-parallel (#19577)
2026-02-12 Aleksander... webui: Add switcher to Chat Message UI to show raw...
2026-02-12 Adrien Gallouëtvendor : update cpp-httplib (#19537)
2026-02-12 Christian Schmitzllama : update outdated comment in llama.h (#19428)
2026-02-12 Aleksander... (webui) FEATURE: Enable adding or injecting System...
2026-02-12 Daniel Beveniusscripts : add support for forks in pr2wt.sh (#19540)
2026-02-12 Aleksander... (webui) REFACTOR: UI primitives and polish (#19551)
2026-02-12 Aleksander... WebUI Architecture Cleanup (#19541)
2026-02-12 Georgi Gerganovmetal : update sum_rows kernel to support float4 (...
2026-02-12 Mario LimoncielloAdd a workaround for compilation with ROCWMMA_FATTN...
2026-02-12 RichardScottOZserver : fix typo in README.md for features list (...
2026-02-12 TriDefenderdocs : update path in snapdragon README.md (#19533)
2026-02-12 Max Krasnyanskyhexagon: further optimization and tuning of matmul...
2026-02-12 Adrien Gallouëtcommon : replace deprecated codecvt using parse_utf8_co...
2026-02-11 lhezopencl: add general Q6_K mm and Q4_K mv (#19347)
2026-02-11 Georgi Gerganovggml : unary ops support non-cont src0 + metal F16...
2026-02-11 Daniel Beveniuscommon : remove unused token util functions (#19506)
2026-02-11 AesSedaimodel: Add Kimi-K2.5 support (#19170)
2026-02-11 Daniel Beveniusbuild : fix case in dSYMs path for build-macos [no...
2026-02-11 Georgi Gerganovmetal : extend l2_norm support for non-cont src0 (...
2026-02-11 Johannes Gäßlerdocs: ban AI for issues and discussions [no CI] (#19512)
2026-02-11 Adrien Gallouëtcommon : improve download error reporting (#19491)
2026-02-11 Max Krasnyanskyhexagon: Add ARGSORT, DIV, SQR, SQRT, SUM_ROWS, GEGLU...
2026-02-11 thecaptain789llama : correct typos 'occured' and 'occurences' (...
2026-02-11 Georgi Gerganovmodel : fix wavtokenizer embedding notions (#19479)
2026-02-11 Georgi Gerganovggml : extend bin bcast for permuted src1 (#19484)
2026-02-11 Georgi Gerganovmetal : consolidate unary ops (#19490)
2026-02-11 Daniel Beveniusllama : refactor sampling_info to use buffer_view templ...
2026-02-10 Oliver SimonsCUDA : Update CCCL-tag for 3.2 to final release from...
2026-02-10 Nikhil Jain[WebGPU] Plug memory leaks and free resources on shutdo...
2026-02-10 JJJYmmmmodels : support qwen3.5 series (#19468)
2026-02-10 Xuan-Son Nguyentest: fix IMROPE perf test case (#19465)
2026-02-10 Alberto Cabrera... ggml-cpu: arm64: q6_K repack gemm and gemv (and generic...
2026-02-10 k4ss4nggml : use noexcept overload for is_regular_file in...
2026-02-10 Piotr Wilkin... convert : move experts permutation from Qwen2MoeModel...
2026-02-10 Daniel Beveniustts : fix typos in README.md [no ci] (#19463)
2026-02-10 Raul TorresCANN: Remove unnecessary wrapper for `gml_backend_buft_...
2026-02-10 hipuddingCANN: implement quantized MUL_MAT_ID for MoE models...
2026-02-10 Georgi Gerganovcuda : extend GGML_OP_PAD to work with non-cont src0...
2026-02-09 Xuan-Son Nguyenchat: fix case where template accepts type content...
2026-02-09 Tarek Dakhranmtmd: Implement tiling for LFM2-VL (#19454)
2026-02-09 손희준Server: log when converting requests to chat completion...
2026-02-09 Sascha Rogmannspec : remove check rate (#19377)
2026-02-09 Georgi Gerganovci : add metal server workflows (#19293)
2026-02-09 Georgi Gerganovrevert : "[Model] Qwen3.5 dense and MoE support (no...
2026-02-09 Kevin Pougetggml-virtgpu: add backend documentation (#19354)
2026-02-09 Hugocmake : add variable to skip installing tests (#19370)
2026-02-08 Piotr Wilkin... [Model] Qwen3.5 dense and MoE support (no vision) ...
2026-02-08 Oliver SimonsCUDA: Fix non-contig rope (#19338)
2026-02-08 Adrien Gallouëtrpc : update from common.cpp (#19400)
2026-02-08 Georgi Gerganovserver : improve context checkpoint logic (#19408)
2026-02-08 ddh0llama-quantize : cleanup `--help` output (#19317)
2026-02-08 Sigbjørn Skjæretci : remove server job from webui and move slow test...
2026-02-07 Georgi Gerganovci : use -j param correctly when building with sanitize...
2026-02-07 Georgi Gerganovmetal : consolidate bin kernels (#19390)
2026-02-07 Georgi Gerganovmetal : fix event synchronization in cpy_tensor_async...
2026-02-06 forforever73model : support Step3.5-Flash (#19283)
2026-02-06 Alex Trottagguf-py : bump sentencepiece version (#19319)
2026-02-06 Abhijit Rameshggml-webgpu: JIT compile binary operators and handle...
2026-02-06 Nechama Krashinskisycl: add F16 support for GGML_OP_CEIL (#19306)
2026-02-06 Jeff Bolztests: reduce number of FA test permutations (#19381)
2026-02-06 Georgi Gerganovcommon : add common_speculative_is_compat() (#19270)
2026-02-06 Lasse Lauwerysunicode : MSVC regex fix (#19340)
2026-02-06 ymckiKimi-Linear support (backend agnostic + MLA KV cache...
2026-02-06 Jeff Bolzvulkan: For coopmat2 FA, use fp16 accumulators for...
2026-02-06 Jeff Bolzvulkan: make FA mask/softcap enables spec constants...
2026-02-06 Georgi Gerganovmetal : skip loading all-zero mask (#19337)
2026-02-06 Daniel Beveniusllama : rename llama-sampling to llama-sampler (#19363)
2026-02-06 Georgi Gerganovcuda : cuda graphs now compare all node params (#19383)
2026-02-05 Georgi Gerganovmetal : adaptive CPU/GPU interleave based on number...
2026-02-05 Jeff Bolzvulkan: Preprocess FA mask to detect all-neg-inf and...
2026-02-05 Georgi Gerganovbenches : update models + numbers (#19359)
2026-02-05 Sigbjørn Skjæretdocker : fix vulkan build (#19352)
2026-02-05 Adrien Gallouëtvendor : update BoringSSL to 0.20260204.0 (#19333)
2026-02-05 Georgi Gerganovmetal : add diag (#19330)
2026-02-05 Oleksandr Kuvshynovvulkan: fix GPU deduplication logic. (#19222)
2026-02-05 Jeff Bolzvulkan: Set k_load_shmem to false when K is too large...
2026-02-05 Jeff Bolzvulkan: fix non-contig rope (#19299)
2026-02-05 will-lmsmetal : add missing includes (#19348)
2026-02-05 Sigbjørn Skjæretvendor : add missing llama_add_compile_flags (#19322)
2026-02-04 Aaron Teovendor: update cpp-httplib version (#19313)
2026-02-04 Daniel Beveniuscodeowners : add danbev for examples/debug (#19332)
2026-02-04 Xuan-Son Nguyendebug: make common_debug_print_tensor readable (#19331)
2026-02-04 Georgi Gerganovci : fix sanitize workflow to enable ggml sanitizers...
2026-02-04 Xuan-Son Nguyenmodel: (qwen3next) correct vectorized key_gdiff calcula...
2026-02-04 Georgi Gerganovtests : add non-cont, inplace rope tests (#19296)
2026-02-04 Daniel Beveniusmodel-conversion : add tensor-info.py utility (#18954)
2026-02-04 Georgi Gerganovspec : fix the check-rate logic of ngram-simple (#19261)
2026-02-04 Daniel Beveniuscompletion : simplify batch (embd) processing (#19286)
2026-02-04 Kevin Pougetggml-virtgpu: make the code thread safe (#19204)
2026-02-04 Aman Guptaggml-cpu: use LUT for converting e8->f32 scales on...
2026-02-03 Georgi Gerganovmetal : add solve_tri (#19302)
2026-02-03 Georgi Gerganovci : add sanitizer runs for server (#19291)
next