]> git.djapps.eu Git - pkg/ggml/sources/ggml/shortlog
pkg/ggml/sources/ggml
2025-02-15 Mathieu BaudierMerge branch 'debian/latest' into debian/bullseye debian/bullseye
2025-02-15 Mathieu BaudierUpdate upstream web site
2025-02-15 Mathieu BaudierMerge tag 'upstream/0.0.1722' into debian/latest
2025-02-12 Georgi Gerganovsync : llama.cpp upstream/0.0.1722
2025-02-12 uvosHIP: Switch to std::vector in rocblas version check...
2025-02-12 bandoticleanup: fix compile warnings associated with gnu_print...
2025-02-12 Richardggml : fix multi-threaded clamp_f32 (llama/11824)
2025-02-12 Weizhao Ouyangggml-cpu: Fix duplicate MATMUL_INT8 (llama/11817)
2025-02-12 Johannes GäßlerCUDA: fix CUDART_VERSION checks (llama/11821)
2025-02-12 Sheldon RobinsonFix #11802: Compile bug - RegQueryValueExA changed...
2025-02-12 Johannes GäßlerCUDA: use arch list for compatibility check (llama...
2025-02-12 Maxim Evtushfix: typos in documentation files (llama/11791)
2025-02-12 Danny Milosavljevicvulkan: Make Vulkan optional at runtime (#11493). ...
2025-02-12 Wagner Brunavulkan: add environment variable GGML_VK_PREFER_HOST_ME...
2025-02-12 Jeff Bolzvulkan: account for lookup tables when checking shared...
2025-02-12 Karol Kontnyggml: Fix data race in ggml threadpool (llama/11736)
2025-02-12 Johannes GäßlerCUDA: fix min. version for movmatrix (llama/11751)
2025-02-12 Jeff Bolzvulkan: print shared memory size (llama/11719)
2025-02-12 Akarshan BiswasSYCL: remove XMX info from print devices (llama/11712)
2025-02-12 Jinyang Heggml : optimize and build warning fix for LoongArch...
2025-02-12 Akarshan BiswasSYCL: Adjust support condition for norm operators ...
2025-02-12 junchao-zhaoggml : fix LoongArch compile error with 128-bit SIMD...
2025-02-12 Jeff Bolzvulkan: optimize coopmat2 iq2/iq3 callbacks (llama...
2025-02-12 Rémy Ovulkan: initial support for IQ4_XS quantization (llama...
2025-02-12 Jeff Bolzvulkan: use smaller combined allocations to avoid fragm...
2025-02-12 Charles Duffymetal : avoid breaking build when metal API predates...
2025-02-12 Georgi Gerganovmetal : adjust support conditions for norm operators...
2025-02-12 Johannes GäßlerCUDA: support for mat. mul. with ne03 != ne13 (llama...
2025-02-12 Johannes GäßlerCUDA: non-contiguous (RMS) norm support (llama/11659)
2025-02-12 fxzjshmHIP: force max threads per block to be 1024 (llama...
2025-02-12 Jhen-Jie Hongmetal : use residency set for other platforms (llama...
2025-02-09 Shang Yuanchuncmake : fix CPU detection on loongarch64 in tests ...
2025-02-08 Georgi Gerganovreadme : remove transfer notice (#1107)
2025-02-07 Shawn yangfix a bug in examples/simple/simple-backend (#1078)
2025-02-06 Patrick Pengrpc: fix known RCE in rpc-server (#1103)
2025-02-06 Mathieu BaudierDisable Vulkan (glslc not supported in bullseye) and...
2025-02-05 Mathieu BaudierMerge tag 'upstream/0.0.1689' into debian/latest
2025-02-05 Mathieu BaudierUse upcoming versioning scheme
2025-02-04 Georgi Gerganovreadme : add info about repository transfer
2025-02-04 Georgi Gerganovauthors : update upstream/0.0.1689
2025-02-04 Georgi Gerganovsync : whisper.cpp
2025-02-03 Christian Kastnercmake: Add ability to pass in GGML_BUILD_NUMBER (#1096)
2025-02-03 Georgi Gerganovsync : llama.cpp
2025-02-03 Johannes GäßlerCUDA: fix Volta FlashAttention logic (llama/11615)
2025-02-03 Georgi Gerganovsync : llama.cpp
2025-02-03 Johannes GäßlerHIP: fix flash_attn_stream_k_fixup warning (llama/11604)
2025-02-03 uvosCUDA/HIP: add support for selectable warp size to mmv...
2025-02-03 uvosHIP: add GGML_CUDA_CC_IS_* for amd familys as increasin...
2025-02-03 Johannes GäßlerCUDA: use mma PTX instructions for FlashAttention ...
2025-02-03 Olivier Chafik`ci`: use sccache on windows instead of ccache (llama...
2025-02-03 uvosHIP: require at least HIP 5.5
2025-02-03 uvosHIP: Prepare reduction operators for wave 64
2025-02-03 uvosCUDA/HIP: add warp_size to cuda_device_info
2025-02-03 Rémy Oudomphengvulkan: implement initial support for IQ2 and IQ3 quant...
2025-02-03 Jeff Bolzvulkan: Catch pipeline creation failure and print an...
2025-01-29 Georgi Gerganovcmake : sync new file
2025-01-29 Georgi Gerganovscripts : sync cmake
2025-01-29 Georgi Gerganovsync : llama.cpp
2025-01-29 uvosHIP: Supress transformation warning in softmax.cu
2025-01-29 Nikita SarychevHIP: Only call rocblas_initialize on rocblas versions...
2025-01-29 someone13574cmake : don't fail on `GGML_CPU=OFF` (llama/11457)
2025-01-29 Akarshan BiswasSYCL : SOFTMAX F16 mask support and other fixes (llama...
2025-01-29 Haus1AMD: parse the architecture as supplied by gcnArchName...
2025-01-29 Ihar Hrachyshkametal: Handle null returned from MTLCreateSystemDefault...
2025-01-29 Georgi Gerganovmetal : use residency sets (llama/11427)
2025-01-29 bandoticmake: add ggml find package (llama/11369)
2025-01-29 Jeff Bolzvulkan: compile shaders on-demand (llama/11406)
2025-01-29 uvosHip: disable VMM on hip as it seams that it dosent...
2025-01-29 uvoship : Add hipGraph and VMM support to ROCM (llama/11362)
2025-01-29 Johannes GäßlerCUDA: fix FP16 cuBLAS GEMM (llama/11396)
2025-01-29 uvosrocBLAS: Avoid fp32->fp16->fp32 conversion on cdna...
2025-01-29 Johannes GäßlerCPU/CUDA: fix (GQA) mul mat back, add CUDA support...
2025-01-29 Bernhard M... cmake : avoid -march=native when reproducible build...
2025-01-29 Jeff Bolztests: fix some mul_mat test gaps (llama/11375)
2025-01-29 amd-dwangVulkan-run-test: fix mmq_wg_denoms (llama/11343)
2025-01-29 Jeff Bolzvulkan: sort shaders for more deterministic binary...
2025-01-29 Jeff Bolzvulkan: fix diag_mask_inf (llama/11323)
2025-01-29 Radoslav Gerganovrpc : better caching of the base buffer pointer (llama...
2025-01-29 Georgi Gerganovmetal : fix out-of-bounds write (llama/11314)
2025-01-29 Jeff Bolzvulkan: fix coopmat2 validation failures (llama/11284)
2025-01-29 Nicolò ScipioneSYCL: Introducing memory host pool (llama/11251)
2025-01-29 Georgi Gerganovcmake : add sanitizer flags for llama.cpp (llama/11279)
2025-01-29 Jeff Bolzvulkan: fix coopmat2 flash attention for non-contiguous...
2025-01-29 Radoslav Gerganovrpc : early register backend devices (llama/11262)
2025-01-29 Jeff Bolzvulkan: support copy from f32 to q4_0/q4_1/q5_0/q5_1...
2025-01-29 Jeff Bolzvulkan: optimize coopmat2 q4_k/q5_k dequant functions...
2025-01-29 Jeff Bolzvulkan: optimize coopmat2 q2_k dequant function (llama...
2025-01-29 Johannes GäßlerCUDA: backwards pass for misc. ops, add tests (llama...
2025-01-29 fj-y-saitoggml: aarch64: implement SVE kernels for q4_K_q8_K...
2025-01-29 Evevulkan: scale caching for k quants + misc fixes (llama...
2025-01-29 Junil Kimfix: ggml: fix vulkan-shaders-gen build (llama/10448)
2025-01-29 Johannes GäßlerRoPE: fix back, CUDA support for back + noncont. (llama...
2025-01-29 Akarshan BiswasSYCL: Add gated linear attention kernel (llama/11175)
2025-01-25 Andrii Ryzhkovcmake : fix build tests on arm (#1084)
2025-01-23 William Tambelliniggml : add option to not print stack on abort (#1081)
2025-01-22 Mathieu BaudierAdd a link in the Debian package to a portable CPU...
2025-01-21 Mathieu BaudierSimplify debian packaging now that GGML backends are...
2025-01-21 Mathieu BaudierInstall GGML backends in libexec
2025-01-21 Mathieu BaudierDebian development package only depends on base libraries
2025-01-21 Mathieu BaudierImprove debian build based on lintian
next