ggml-org/ggml

Pull Requests Commits

sync : whisper.cpp

ggerganov committed 2 years ago

Verified 03c1ad89

examples : added audio_ctx argument to main and server (whisper/1857)

dscripka committed 2 years ago

Verified 3faad1ef

metal : option to embed MSL source into compiled binary (whisper/1842)

didzis committed 2 years ago

Verified a05a28fa

examples : initialize context params properly (whisper/1852)

ggerganov committed 2 years ago

Verified 055b4c29

sync : llama.cpp

ggerganov committed 2 years ago

Verified 15438356

ggml-backend : sync remnant

ggerganov committed 2 years ago

Verified e69a1c51

CUDA: mul_mat_vec_q tiling, refactor mul mat logic (llama/5434)

JohannesGaessler committed 2 years ago

Verified 5f60268a

vulkan: only use M-sized matmul on Apple GPUs (llama/5412)

slp committed 2 years ago

Verified adc3925d

ggml : fix compile warnings (unused vars) (llama/4966)

ggerganov committed 2 years ago

Verified 1a781c5a

ggml : add mmla kernels for quantized GEMM (llama/4966)

snadampal committed 2 years ago

Verified 3655bf83

metal : use autoreleasepool to avoid memory leaks (llama/5437)

irbull committed 2 years ago

Verified 04b3bf5d

ggml-alloc : v3 (#727)

slaren committed 2 years ago

Verified 5070f078

examples : remove old stuff (#728)

ggerganov committed 2 years ago

Verified ee07c17b

sync : whisper.cpp

ggerganov committed 2 years ago

Verified 82248c97

whisper : expose CUDA device setting in public API (whisper/1840)

didzis committed 2 years ago

Verified 80d5565e

sync : ggml (whisper/0)

ggerganov committed 2 years ago

Verified d12ec3b5

src : relocate new backend sources

ggerganov committed 2 years ago

Verified 14b0a66d

sync : llama.cpp

ggerganov committed 2 years ago

Verified d253ec6a

ci : fix mpt test

ggerganov committed 2 years ago

Verified 0578c5bc

tests : fix im2col usage

ggerganov committed 2 years ago

Verified 3c327016

ggml : fix `error C2078: too many initializers` for MSVC ARM64 (llama/5404)

Xarbirus committed 2 years ago

Verified 4bbcf566

Fix Vulkan crash on APUs with very little device memory (llama/5424)

0cc4m committed 2 years ago

Verified d04625d1

CUDA: more warps for mmvq on NVIDIA (llama/5394)

JohannesGaessler committed 2 years ago

Verified 09aa31e9

Fix f16_sycl cpy call from Arc (llama/5411)

abhilash1910 committed 2 years ago

Verified c8efd654

CUDA: fixed mmvq kernel for bs 2,3,4 and -sm row (llama/5386)

JohannesGaessler committed 2 years ago

Verified 3fc39d6f

Basic Vulkan Multi-GPU implementation (llama/5321)

0cc4m committed 2 years ago

Verified 1c8791cd

CUDA: mul_mat_vec_q max. batch size 8 -> 4 (llama/5370)

JohannesGaessler committed 2 years ago

Verified 14541b44

Slight quantization improvement for Q4_K and Q5_K (llama/5361)

ikawrakow committed 2 years ago

Verified 82389b52

CUDA: mul_mat_vec_q for batch sizes > 1 (llama/5351)

JohannesGaessler committed 2 years ago

Verified 8f5bb5dc

ggml : make use of ggml-quants.h possible in C++ code (llama/5338)

ikawrakow committed 2 years ago

Verified 63d8fce8

Older