]> git.djapps.eu Git - pkg/ggml/sources/whisper.cpp/shortlog
pkg/ggml/sources/whisper.cpp
2024-12-08 Adrien Gallouëtggml-cpu: replace AArch64 NEON assembly with intrinsics...
2024-12-08 Evevulkan: Dynamic subgroup size support for Q6_K mat_vec...
2024-12-08 Georgi Gerganovggml : fix I8MM Q4_1 scaling factor conversion (llama...
2024-12-08 Shupei Fanggml-cpu: fix typo in gemv/gemm iq4_nl_4_4 (llama/10580)
2024-12-08 Alberto Cabrera... sycl : offload of get_rows set to 0 (llama/10432)
2024-12-08 Alberto Cabrera... sycl : Reroute permuted mul_mats through oneMKL (llama...
2024-12-08 Chenguang LiCANN: RoPE operator optimization (llama/10563)
2024-12-08 Jeff Bolzvulkan: get the first command buffer submitted sooner...
2024-12-08 Georgi Gerganovggml : remove redundant copyright notice + update authors
2024-12-08 Georgi Gerganovggml : fix row condition for i8mm kernels (llama/10561)
2024-12-08 Georgi Gerganovcmake : fix ARM feature detection (llama/10543)
2024-12-08 Shupei Fanggml-cpu: support IQ4_NL_4_4 by runtime repack (llama...
2024-12-08 Sergio Lópezkompute : improve backend to pass test_backend_ops...
2024-12-08 leo-ponyCANN: Fix SOC_TYPE compile bug (llama/10519)
2024-12-08 Chenguang LiCANN: ROPE operator optimization (llama/10540)
2024-12-08 uvosAdd some minimal optimizations for CDNA (llama/10498)
2024-12-08 Georgi Gerganovmetal : fix group_norm support condition (llama/0)
2024-12-08 Jeff Bolzvulkan: define all quant data structures in types.comp...
2024-12-08 Jeff Bolzvulkan: Handle GPUs with less shared memory (llama...
2024-12-08 Jeff Bolzvulkan: further optimize q5_k mul_mat_vec (llama/10479)
2024-12-08 Jeff Bolzvulkan: skip integer div/mod in get_offsets for batch_i...
2024-12-08 Jeff Bolzvulkan: optimize Q2_K and Q3_K mul_mat_vec (llama/10459)
2024-12-08 R0CKSTARmtgpu: Add MUSA_DOCKER_ARCH in Dockerfiles && update...
2024-12-08 Jeff Bolzvulkan: fix group_norm (llama/10496)
2024-12-08 Georgi Gerganovcmake : enable warnings in llama (llama/10474)
2024-12-08 Charles Xuggml-cpu: cmake add arm64 cpu feature check for macos...
2024-12-08 Shanshan ShenCANN: Improve the Inferencing Performance for Ascend...
2024-12-08 Chenguang LiCANN: RoPE and CANCAT operator optimization (llama...
2024-12-08 Junil Kimvulkan: Fix a vulkan-shaders-gen arugment parsing error...
2024-12-08 Georgi Gerganovmetal : enable mat-vec kernels for bs <= 4 (llama/10491)
2024-12-08 Diego Devesallama : accept a list of devices to use to offload...
2024-12-08 Diego Devesaggml : add support for dynamic loading of backends...
2024-12-08 Georgi Gerganovmetal : minor code formatting
2024-12-08 Diego Devesaggml : do not use ARM features not included in the...
2024-12-08 leo-ponyCANN: Support Ascend310P to accelerate F32 and F16...
2024-12-08 Diego Devesacuda : optimize argmax (llama/10441)
2024-12-08 Jeff Bolzvulkan: predicate max operation in soft_max shaders...
2024-12-08 Jeff Bolzvulkan: copy iq4_nl LUT into shared memory (llama/10409)
2024-12-08 Jeff Bolzvulkan: further optimize mul_mat_vec using larger loads...
2024-12-08 haopengadd cmake rvv support (llama/10411)
2024-12-08 mahorozteCUDA: remove unnecessary warp reduce in FA (ggml/1032)
2024-12-08 PABfeat: add `GGML_UNARY_OP_ARGMAX` Metal kernel (ggml...
2024-12-08 PABmetal : add `GGML_OP_CONV_TRANSPOSE_1D` kernels (ggml...
2024-12-08 Frankie RobertsonDo not include arm_neon.h when compiling CUDA code...
2024-12-08 Johannes Gäßlerggml-opt: fix data corruption (ggml/1022)
2024-11-28 KITAITI Makotoruby : Add low-level methods to transcribe (#2585)
2024-11-28 Michael Rienstramodels : add `q8_0` models to `download-ggml-model...
2024-11-21 KITAITI Makotoruby : Follow source tree change (#2580)
2024-11-20 Georgi Gerganovwhisper : use backend registry (#0)
2024-11-20 slarenggml/sched : do not skip views in pre-assignments
2024-11-20 Georgi Gerganovwhisper : adapt to new ggml (wip)
2024-11-20 Georgi Gerganovtalk-llama : sync llama.cpp
2024-11-20 Georgi Gerganovsync : ggml
2024-11-20 Georgi Gerganovggml : sync resolve (skip) (#0)
2024-11-20 bandotiAdd required ggml-base and backend libs to cmake pkg...
2024-11-20 Diego Devesacuda : fix CUDA_FLAGS not being applied (llama/10403)
2024-11-20 Romain Biessysycl : Add option to set the SYCL architecture for...
2024-11-20 Jeff Bolzvulkan: Optimize soft_max (llama/10301)
2024-11-20 Alberto Cabrera... sycl: Revert MUL_MAT_OP support changes (llama/10385)
2024-11-20 Diego Devesacuda : only use native when supported by cmake (llama...
2024-11-20 Jeff Bolzvulkan: remove use of null initializer (llama/10372)
2024-11-20 Plamen Minevmetal : fox offset integer overflows in im2col (ggml...
2024-11-20 0cc4mVulkan: Fix device info output format specifiers (llama...
2024-11-20 PABmetal : add `GGML_UNARY_OP_ELU` kernel (ggml/1018)
2024-11-20 Johannes GäßlerCUDA: fix MMV kernel being used for FP16 src1 (llama...
2024-11-20 Johannes GäßlerCMake: fix typo in comment [no ci] (llama/10360)
2024-11-20 Diego Devesallama : only use default buffer types for the KV cache...
2024-11-20 Georgi Gerganovmetal : refactor kernel args into structs (llama/10238)
2024-11-20 FirstTimeEZggml : fix undefined reference to 'getcpu' (llama/10354)
2024-11-20 Johannes GäßlerCUDA: remove DMMV, consolidate F16 mult mat vec (llama...
2024-11-20 Johannes GäßlerCMake: default to -arch=native for CUDA build (llama...
2024-11-20 Diego Devesaggml : fix possible buffer use after free in sched...
2024-11-20 Georgi Gerganovggml : inttypes.h -> cinttypes (llama/0)
2024-11-20 Georgi Gerganovggml : adapt AMX to tensor->grad removal (llama/0)
2024-11-20 Georgi Gerganovggml : fix compile warnings (llama/0)
2024-11-20 Georgi Gerganovllamafile : fix include path (llama/0)
2024-11-20 Jeff Bolzvulkan: Optimize some mat-vec mul quant shaders (llama...
2024-11-20 Dan Johanssonggml : optimize Q4_0 into Q4_0_X_Y repack (llama/10324)
2024-11-20 Srihari-mcwMake updates to fix issues with clang-cl builds while...
2024-11-20 Johannes Gäßlerggml: new optimization interface (ggml/988)
2024-11-20 Georgi Gerganovggml : remove duplicated sources from the last sync...
2024-11-20 slarenggml : fix some build issues
2024-11-20 Georgi Gerganovsync : leftovers (ggml/0)
2024-11-20 Georgi Gerganovcmake : restore CMakeLists.txt (llama/10256)
2024-11-20 EveAVX BF16 and single scale quant optimizations (llama...
2024-11-20 Romain Biessysycl: Use syclcompat::dp4a (llama/10267)
2024-11-20 Charles Xubackend cpu: add online flow for aarch64 Q4_0 GEMV...
2024-11-20 Diego Devesaggml : build backends as libraries (llama/10256)
2024-11-20 Georgi Gerganovscripts : update sync
2024-11-19 Georgi Gerganovrelease : v1.7.2
2024-11-18 Stefan Sydowsycl: fix example build (#2570)
2024-11-16 Georgi Gerganovci : use local ggml in Android build (#2567)
2024-11-16 Georgi Gerganovggml : tmp workaround for whisper.cpp (skip) (#2565)
2024-11-15 Georgi Gerganovupdate : readme
2024-11-15 Georgi Gerganovscripts : fix sync path
2024-11-15 Jhen-Jie Hongwhisper.swiftui : switch Mac dest to Mac (Designed...
2024-11-15 Georgi Gerganovcmake : fix ppc64 check (#0)
2024-11-15 Georgi Gerganovwhisper : include ggml-cpu.h (#0)
2024-11-15 Georgi Gerganovbuild : fixes
2024-11-15 Georgi Gerganovtalk-llama : sync llama.cpp
next