]> git.djapps.eu Git - pkg/ggml/sources/ggml/shortlog
pkg/ggml/sources/ggml
2024-09-24 Max Krasnyanskythreads: improve ggml_barrier scaling with large number...
2024-09-24 Srihari-mcwggml : AVX512 gemm for Q4_0_8_8 (llama/9532)
2024-09-24 Georgi Gerganovmetal : use F32 prec for K*Q in vec FA (llama/9595)
2024-09-24 Akarshan BiswasRevert "[SYCL] fallback mmvq (#9088)" (llama/9579)
2024-09-24 R0CKSTARmusa: enable building fat binaries, enable unified...
2024-09-24 Molly SophiaFix merge error in #9454 (llama/9589)
2024-09-24 Johannes GäßlerCUDA: enable Gemma FA for HIP/Pascal (llama/9581)
2024-09-24 Molly SophiaRWKV v6: RWKV_WKV op CUDA implementation (llama/9454)
2024-09-24 slarenggml-alloc : fix list of allocated tensors with GGML_AL...
2024-09-24 agray3Update CUDA graph on scale change plus clear nodes...
2024-09-20 Georgi Gerganovexamples : adapt to ggml.h changes (#0)
2024-09-20 Georgi Gerganovsync : llama.cpp
2024-09-20 Georgi Gerganovggml : refactoring (llama/#0)
2024-09-20 Georgi Gerganovggml : fix builds (llama/0)
2024-09-20 Georgi Gerganovggml : fix trailing whitespace (llama/0)
2024-09-20 Johannes GäßlerCUDA: fix sum.cu compilation for CUDA < 11.7 (llama...
2024-09-20 slarenggml : fix n_threads_cur initialization with one thread...
2024-09-20 Max Krasnyanskythreadpool : skip polling for unused threads (llama...
2024-09-20 Michael Podvitskiyggml : link MATH_LIBRARY not by its full path (llama...
2024-09-20 Georgi Gerganovcmake : do not hide GGML options + rename option (llama...
2024-09-20 Eveggml : IQ4_NL sgemm + Q4_0 AVX optimization (llama...
2024-09-20 Georgi Gerganovmetal : handle zero-sized allocs (llama/9466)
2024-09-20 Georgi Gerganovcommon : reimplement logging (llama/9418)
2024-09-20 Michael Podvitskiycmake : correct order of sycl flags (llama/9497)
2024-09-20 Michael Podvitskiycmake : try to fix sycl+intel build (llama/9487)
2024-09-20 Yuri Khrustalevggml : ggml_type_name return "NONE" for invalid values...
2024-09-20 Georgi Gerganovcmake : use list(APPEND ...) instead of set() + dedup...
2024-09-20 Dou Xinpengcann: Add host buffer type for Ascend NPU (llama/9406)
2024-09-20 Ahmad Tameemriscv : modify Makefile and add a RISCV_VECT to print...
2024-09-20 Xinpeng Doucann: Fix error when running a non-exist op (llama...
2024-09-20 Johannes GäßlerCUDA: fix --split-mode row race condition (llama/9413)
2024-09-20 R0CKSTARmusa: remove Clang builtins mapping (llama/9421)
2024-09-20 Alberto Cabrera... sycl : update support conditions (llama/9394)
2024-09-20 Georgi Gerganovmetal : fix compile warning with GGML_METAL_NDEBUG...
2024-09-20 Radoslav Gerganovrpc : fix segfault with nkvo (llama/9389)
2024-09-20 Prashant Vithuleggml : vector length agnostic SVE support (llama/9290)
2024-09-20 Johannes GäßlerCUDA: fix variable name conflict for Windows build...
2024-09-20 Markus TavenrathOverlap cmdbuffer creation and cmdbuffer execution...
2024-09-20 Georgi Gerganovcuda : fix FA Q src index (1 -> 0) (llama/9374)
2024-09-20 Neo Zhang Jianyuadd check malloc result on device (llama/9346)
2024-09-20 Georgi Gerganovscripts : add context to sync-llama-am.sh
2024-09-20 Johannes Gäßlerggml/examples: add backend support for numerical optimi...
2024-09-18 Jeff Bolzbuild: fix msvc build (#960)
2024-09-08 Georgi Gerganovexamples : add null threadpool args where needed (#0)
2024-09-08 Georgi Gerganovsync : llama.cpp
2024-09-08 Georgi Gerganovmetal : update support condition for im2col + fix warni...
2024-09-08 slarenggml : always check bounds on get_rows operations ...
2024-09-08 Xuan Son Nguyenggml : fix missing `cpu_set_t` on emscripten (llama...
2024-09-08 Markus TavenrathImprove Vulkan shader build system (llama/9239)
2024-09-08 compiladeggml-quants : ternary packing for TriLMs and BitNet...
2024-09-08 slarencuda : fix defrag with quantized KV (llama/9319)
2024-09-08 Srihari-mcwggml : AVX2 support for Q4_0_8_8 (llama/8713)
2024-09-08 Ouadie EL FAROUKIFix DMMV dequantization (llama/9279)
2024-09-08 yuri@FreeBSDggml : add pthread includes on FreeBSD (llama/9258)
2024-09-08 Molly Sophiallama : support RWKV v6 models (llama/8980)
2024-09-08 Faisal ZaghloulThreadpool: take 2 (llama/8672)
2024-09-06 Salvatore Mesoracavulkan: fix compilation with GGML_VULKAN_DEBUG=ON ...
2024-09-06 Salvatore Mesoracavulkan: add dryrun support to sin and cos ops (#947)
2024-09-06 Salvatore Mesoracavulkan: correctly report support for OP_CONT (#946)
2024-09-03 Johannes Gäßlertests: add gradient tests for all backends (#932)
2024-08-31 Johannes Gäßlerggml: fix ggml_graph_cpy undefined behavior (#943)
2024-08-28 Georgi Gerganovcann : fix doxy (#0)
2024-08-28 Georgi Gerganovsync : whisper.cpp
2024-08-28 Justine Tunneyexamples : use colorblind friendly TTY color scheme...
2024-08-28 Mengqing Caocann : add Ascend NPU support (whisper/2336)
2024-08-28 Georgi Gerganovsync : llama.cpp
2024-08-28 Georgi Gerganovscripts : sync vulkan shaders generator
2024-08-28 Georgi Gerganovvulkan : fix build (llama/0)
2024-08-28 Georgi Gerganovtests : disable BF16 cont test
2024-08-28 Georgi Gerganovcuda : mark BF16 CONT as unsupported
2024-08-28 Salvatore Mesoracaggml : fix cont with transposed tensors when one dimens...
2024-08-28 Georgi Gerganovci : fix yolo check
2024-08-27 Georgi Gerganovsync : llama.cpp
2024-08-27 Georgi Gerganovci : disable mnist test
2024-08-27 Georgi Gerganovsync : vulkan (skip) (llama/0)
2024-08-27 Georgi Gerganovsync : llama.cpp
2024-08-27 slarenggml : do not crash when quantizing q4_x_x with an...
2024-08-27 Georgi Gerganovmetal : separate scale and mask from QKT in FA kernel...
2024-08-27 Georgi Gerganovggml : add SSM Metal kernels (llama/8546)
2024-08-27 slarenmetal : gemma2 flash attention support (llama/9159)
2024-08-27 Johannes GäßlerCPU/CUDA: Gemma 2 FlashAttention support (llama/8542)
2024-08-27 Akarshan BiswasAdd a space to supress a cmake warning (llama/9133)
2024-08-27 luoyu-intelAdd oneDNN primitive support (llama/9091)
2024-08-27 compiladellama : simplify Mamba with advanced batch splits ...
2024-08-27 Meng, Hengyufallback mmvq (llama/9088)
2024-08-27 zhentaoyuFix SYCL `im2col` and `convert` Overflow with Large...
2024-08-27 Radoslav Gerganovrpc : print error message when failed to connect endpoi...
2024-08-27 Radoslav Gerganovrpc : prevent crashes on invalid input (llama/9040)
2024-08-27 Nico Bosshardggml : dynamic ggml_sched_max_splits based on graph_siz...
2024-08-27 Georgi Gerganovcmake : remove unused option GGML_CURL (llama/9011)
2024-08-27 Daniel Beveniusggml : move rope type enum to ggml.h (llama/8949)
2024-08-27 DavidKorczynskiggml: fix div-by-zero (llama/9003)
2024-08-27 Markus TavenrathOptimize Vulkan backend for better CPU performance...
2024-08-27 Johannes Gäßlerfeat: ref. cross entropy, add CUDA, fix grad test ...
2024-08-27 Salvatore Mesoracatests : fix memory leaks (#936)
2024-08-24 Johannes Gäßlerggml: remove bad assert (#928)
2024-08-24 ucag.liset NULL to ggml_context pointer to pass assert check...
2024-08-22 Johannes Gäßlerexamples: add MNIST training + missing ops
2024-08-19 Radoslav Gerganovyolo : add backend support (#924)
2024-08-15 Daniel Beveniusggml : fix typo in ggml-quants.c comment (#922)
next