| 2024-02-19 |
Felix | examples : clean up common code (whisper/1871) |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | whisper : fix external encoder (whisper/1860) |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | ggml : resolve merge conflicts (#0) |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | common : add IQ1_S (#0) |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | sync : llama.cpp |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | ci : enable -Werror for CUDA builds (llama/5579) |
commit | commitdiff | tree |
| 2024-02-19 |
slaren | cuda, metal : fix nans in soft_max (llama/5574) |
commit | commitdiff | tree |
| 2024-02-19 |
bmwl | ggml : android and old glibc NUMA incompatibility bugfi... |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | ggml : restore vec dot stride arg names (llama/5453) |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | ci : fix wikitext url + compile warnings (llama/5569) |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | metal : fix unused warnings (llama/0) |
commit | commitdiff | tree |
| 2024-02-19 |
Herman Semenov | ggml, common, examples, tests : fixed type arguments... |
commit | commitdiff | tree |
| 2024-02-19 |
Kawrakow | 1.5 bit quantization (llama/5453) |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | ggml : add ALiBi support for ggml_soft_max_ext (llama... |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | sync : llama.cpp |
commit | commitdiff | tree |
| 2024-02-19 |
Ananta Bastola | ci : add an option to fail on compile warning (llama... |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | cmake : fix VULKAN and ROCm builds (llama/5525) |
commit | commitdiff | tree |
| 2024-02-19 |
bmwl | ggml : add numa options (llama/5377) |
commit | commitdiff | tree |
| 2024-02-19 |
slaren | cuda : print message when initialization fails (llama... |
commit | commitdiff | tree |
| 2024-02-19 |
Neuman Vong | vulkan: Find optimal memory type but with fallback... |
commit | commitdiff | tree |
| 2024-02-19 |
AT | Early return for zero size calls to get_tensor. (llama... |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | tests : disable moe test (llama/5473) |
commit | commitdiff | tree |
| 2024-02-19 |
Kawrakow | ggml-quants : fix compiler warnings (shadow variable... |
commit | commitdiff | tree |
| 2024-02-19 |
Abhilash Majumder | ggml-sycl: Replace 3d ops with macro (llama/5458) |
commit | commitdiff | tree |
| 2024-02-19 |
Georgi Gerganov | cmake : update CBLAS build flags (#0) |
commit | commitdiff | tree |
| 2024-02-12 |
slaren | ggml-alloc : allocate all leafs as if they were inputs... |
commit | commitdiff | tree |
| 2024-02-12 |
Georgi Gerganov | sync : whisper.cpp |
commit | commitdiff | tree |
| 2024-02-12 |
dscripka | examples : added audio_ctx argument to main and server... |
commit | commitdiff | tree |
| 2024-02-12 |
Didzis Gosko | metal : option to embed MSL source into compiled binary... |
commit | commitdiff | tree |
| 2024-02-12 |
Georgi Gerganov | examples : initialize context params properly (whisper... |
commit | commitdiff | tree |
| 2024-02-12 |
Georgi Gerganov | sync : llama.cpp |
commit | commitdiff | tree |
| 2024-02-12 |
Georgi Gerganov | ggml-backend : sync remnant |
commit | commitdiff | tree |
| 2024-02-12 |
Johannes Gäßler | CUDA: mul_mat_vec_q tiling, refactor mul mat logic... |
commit | commitdiff | tree |
| 2024-02-12 |
Sergio López | vulkan: only use M-sized matmul on Apple GPUs (llama... |
commit | commitdiff | tree |
| 2024-02-12 |
Georgi Gerganov | ggml : fix compile warnings (unused vars) (llama/4966) |
commit | commitdiff | tree |
| 2024-02-12 |
snadampal | ggml : add mmla kernels for quantized GEMM (llama/4966) |
commit | commitdiff | tree |
| 2024-02-12 |
Ian Bull | metal : use autoreleasepool to avoid memory leaks ... |
commit | commitdiff | tree |
| 2024-02-11 |
slaren | ggml-alloc : v3 (#727) |
commit | commitdiff | tree |
| 2024-02-10 |
Georgi Gerganov | examples : remove old stuff (#728) |
commit | commitdiff | tree |
| 2024-02-10 |
Georgi Gerganov | sync : whisper.cpp |
commit | commitdiff | tree |
| 2024-02-10 |
Didzis Gosko | whisper : expose CUDA device setting in public API... |
commit | commitdiff | tree |
| 2024-02-10 |
Georgi Gerganov | sync : ggml (whisper/0) |
commit | commitdiff | tree |
| 2024-02-10 |
Georgi Gerganov | src : relocate new backend sources |
commit | commitdiff | tree |
| 2024-02-10 |
Georgi Gerganov | sync : llama.cpp |
commit | commitdiff | tree |
| 2024-02-10 |
Georgi Gerganov | ci : fix mpt test |
commit | commitdiff | tree |
| 2024-02-10 |
Georgi Gerganov | tests : fix im2col usage |
commit | commitdiff | tree |
| 2024-02-10 |
Michael Podvitskiy | ggml : fix `error C2078: too many initializers` for... |
commit | commitdiff | tree |
| 2024-02-10 |
0cc4m | Fix Vulkan crash on APUs with very little device memory... |
commit | commitdiff | tree |
| 2024-02-10 |
Johannes Gäßler | CUDA: more warps for mmvq on NVIDIA (llama/5394) |
commit | commitdiff | tree |
| 2024-02-10 |
Abhilash Majumder | Fix f16_sycl cpy call from Arc (llama/5411) |
commit | commitdiff | tree |
| 2024-02-10 |
Johannes Gäßler | CUDA: fixed mmvq kernel for bs 2,3,4 and -sm row (llama... |
commit | commitdiff | tree |
| 2024-02-10 |
0cc4m | Basic Vulkan Multi-GPU implementation (llama/5321) |
commit | commitdiff | tree |
| 2024-02-10 |
Johannes Gäßler | CUDA: mul_mat_vec_q max. batch size 8 -> 4 (llama/5370) |
commit | commitdiff | tree |
| 2024-02-10 |
Kawrakow | Slight quantization improvement for Q4_K and Q5_K ... |
commit | commitdiff | tree |
| 2024-02-10 |
Johannes Gäßler | CUDA: mul_mat_vec_q for batch sizes > 1 (llama/5351) |
commit | commitdiff | tree |
| 2024-02-10 |
Kawrakow | ggml : make use of ggml-quants.h possible in C++ code... |
commit | commitdiff | tree |
| 2024-02-10 |
Dr. Tom Murphy... | ggml : avoid duplicating function calls using MIN/MAX... |
commit | commitdiff | tree |
| 2024-02-10 |
Kawrakow | iq2_xxs: tune quantization (llama/5320) |
commit | commitdiff | tree |
| 2024-02-10 |
AidanBeltonS | Fix cpy with dims of 3 (llama/5289) |
commit | commitdiff | tree |
| 2024-02-10 |
0cc4m | Vulkan Intel Fixes, Optimizations and Debugging Flags... |
commit | commitdiff | tree |
| 2024-02-10 |
AidanBeltonS | Fix im2col with 32fp (llama/5286) |
commit | commitdiff | tree |
| 2024-02-10 |
AidanBeltonS | Tidy ggml-sycl (llama/5261) |
commit | commitdiff | tree |
| 2024-02-10 |
Meng, Hengyu | get MAX_MEM_ALLOC from device property (llama/5270) |
commit | commitdiff | tree |
| 2024-02-10 |
Neo Zhang Jianyu | add --no-mmap in llama-bench (llama/5257) |
commit | commitdiff | tree |
| 2024-02-10 |
0cc4m | Vulkan Phi Fix for AMD Proprietary Drivers (llama/5260) |
commit | commitdiff | tree |
| 2024-02-10 |
slaren | cuda : fix LLAMA_CUDA_F16 (llama/5262) |
commit | commitdiff | tree |
| 2024-02-10 |
Georgi Gerganov | metal : add im2col F32 dst support (llama/5132) |
commit | commitdiff | tree |
| 2024-02-10 |
JidongZhang-THU | llava : add MobileVLM support (llama/5132) |
commit | commitdiff | tree |
| 2024-02-10 |
Neo Zhang Jianyu | format license text, restore apache license by legal... |
commit | commitdiff | tree |
| 2024-02-10 |
slaren | ggml : limit n_threads to the max n_tasks (llama/5238) |
commit | commitdiff | tree |
| 2024-02-10 |
0cc4m | Vulkan Fixes (llama/5223) |
commit | commitdiff | tree |
| 2024-02-10 |
Jared Van Bortel | kompute : llama-bench support and ggml_cpu_has_kompute... |
commit | commitdiff | tree |
| 2024-02-09 |
Michael Podvitskiy | ggml : add abort_callback for cpu backend (#725) |
commit | commitdiff | tree |
| 2024-01-30 |
Georgi Gerganov | sync : whisper.cpp |
commit | commitdiff | tree |
| 2024-01-30 |
JacobLinCool | common : fix wav buffer detection (whisper/1819) |
commit | commitdiff | tree |
| 2024-01-30 |
Georgi Gerganov | sync : llama.cpp |
commit | commitdiff | tree |
| 2024-01-30 |
Kawrakow | ggml : fix IQ3_XXS on Metal (llama/5219) |
commit | commitdiff | tree |
| 2024-01-30 |
Georgi Gerganov | sync : ggml (llama/0) |
commit | commitdiff | tree |
| 2024-01-30 |
Kawrakow | Faster AVX2 dot product for IQ2_XS (llama/5187) |
commit | commitdiff | tree |
| 2024-01-30 |
Kawrakow | SOTA 3-bit quants (llama/5196) |
commit | commitdiff | tree |
| 2024-01-30 |
0cc4m | Vulkan Windows APU Memory Handling (llama/5199) |
commit | commitdiff | tree |
| 2024-01-30 |
Paul Tsochantaris | ggml alloc: Fix for null dereference on alloc failure... |
commit | commitdiff | tree |
| 2024-01-30 |
Jared Van Bortel | Nomic Vulkan backend (llama/4456) |
commit | commitdiff | tree |
| 2024-01-30 |
slaren | ggml : add max buffer sizes to opencl and metal backend... |
commit | commitdiff | tree |
| 2024-01-30 |
Paul Tsochantaris | metal : free metal objects (llama/5161) |
commit | commitdiff | tree |
| 2024-01-29 |
Georgi Gerganov | gguf : fix comparison (#715) |
commit | commitdiff | tree |
| 2024-01-29 |
John Balis | `ggml_cuda_cpy` support for 4d tensors and float16... |
commit | commitdiff | tree |
| 2024-01-29 |
Georgi Gerganov | gguf : add input validation, prevent integer overflows... |
commit | commitdiff | tree |
| 2024-01-29 |
Georgi Gerganov | ci : fix yolo URLs + fix metal capture (#712) |
commit | commitdiff | tree |
| 2024-01-29 |
Jack Mousseau | metal : add debug capture backend function (#694) |
commit | commitdiff | tree |
| 2024-01-28 |
Georgi Gerganov | sync : llama.cpp |
commit | commitdiff | tree |
| 2024-01-28 |
Georgi Gerganov | sync : whisper.cpp |
commit | commitdiff | tree |
| 2024-01-28 |
Georgi Gerganov | sync : llama.cpp |
commit | commitdiff | tree |
| 2024-01-28 |
0cc4m | ggml : add Vulkan backend (llama/2059) |
commit | commitdiff | tree |
| 2024-01-28 |
Abhilash Majumder | ggml : add unified SYCL backend for Intel GPUs (llama... |
commit | commitdiff | tree |
| 2024-01-28 |
Georgi Gerganov | ggml : minor type fix (int64_t -> size_t) |
commit | commitdiff | tree |
| 2024-01-27 |
Georgi Gerganov | sync : whisper.cpp |
commit | commitdiff | tree |
| 2024-01-27 |
Georgi Gerganov | common : fix input buffer check (whisper/1812) |
commit | commitdiff | tree |
| 2024-01-27 |
Ryan Hitchman | server : implement "verbose_json" format with token... |
commit | commitdiff | tree |
| 2024-01-27 |
Georgi Gerganov | sync : llama.cpp |
commit | commitdiff | tree |
| next |
overview / pkg / ggml / sources / ggml / shortlog