ggml-org/ggml

Pull Requests Commits

sync : llama.cpp

ggerganov committed 1 year ago

Verified 93c462c2

scripts : update sync (#1161)

ggerganov committed 1 year ago

Verified ab9cc55a

files : remove old wkv6 sources (#0)

ggerganov committed 1 year ago

660def06

sync : llama.cpp

ggerganov committed 1 year ago

c407e147

ggml : sync/merge cmake,riscv,powerpc, add common.cmake (#0)

ggerganov committed 1 year ago

2e727b4b

llamafile : ppc64le MMA implementation for Q4_0. (llama/12489)

amritahs-ibm committed 1 year ago

32aca460

SYCL: implement memset ggml backend buffer interface (llama/12580)

qnixsynapse committed 1 year ago

d3f81d0d

HIP: Add support for RDNA4 targets (llama/12372)

slojosic-amd committed 1 year ago

a85120e7

metal : refactor mat-vec code (llama/12569)

ggerganov committed 1 year ago

7fb60bd4

ggml : fix MUL_MAT_ID repack with Q8_K (llama/12544)

ggerganov committed 1 year ago

16323b2e

ggml-cpu : update KleidiAI to v1.5.0 (llama/12568)

eddnjjn committed 1 year ago

d4a83438

SYCL: disable Q4_0 reorder optimization (llama/12560)

qnixsynapse committed 1 year ago

3456ae24

opencl: simplify kernel embedding logic in cmakefile (llama/12503)

lhez committed 1 year ago

c04c3fdb

CUDA: Fix clang warnings (llama/12540)

yeahdongcn committed 1 year ago

3ee39106

vulkan: fix mul_mat_vec failure in backend tests (llama/12529)

jeffbolznv committed 1 year ago

df3540e5

ggml : fix quantized cpy op (llama/12310)

ggerganov committed 1 year ago

028e983b

musa: refine compute capability (llama/12493)

yeahdongcn committed 1 year ago

de9813f8

vulkan: Optimize mul_mat_vec p021 and nc shaders (llama/12505)

jeffbolznv committed 1 year ago

382326da

Vulkan: RTE rounding for cpy to quant (llama/12480)

stduhpf committed 1 year ago

db0f4eee

vulkan: workaround for AMD Windows driver 16 bit unpack8 bug (llama/12472)

netrunnereve committed 1 year ago

32d1e032

Fix build on Windows when ccache enabled (#9954) (llama/9976)

MakeDecisionWorth committed 1 year ago

1b05528c

sycl: cleanup oneDNN related code (llama/12097)

sgeor255 committed 1 year ago

875a46ed

ggml : block interleaving support for Q4_K quantization for x86 AVX2 architecture (llama/12332)

Srihari-mcw committed 1 year ago

fcbd7b1a

CUDA: Improve flash decoding kernel GPU occupancy for BS=1 case (llama/12183)

gaugarg-nv committed 1 year ago

80772659

vulkan: optimize iq1 coopmat2 dequant functions (llama/12427)

jeffbolznv committed 1 year ago

57d28154

Fix visionOS build and add CI (llama/12415)

guusw committed 1 year ago

29a0b315

vulkan: Submit once enough matmul work has been recorded (llama/12406)

jeffbolznv committed 1 year ago

7c8c6895

opencl: improve profiling (llama/12442)

lhez committed 1 year ago

ae8857fe

musa: override warp_size of musa device to 32 (llama/12445)

yeahdongcn committed 1 year ago

46532b33

SYCL: using graphs is configurable by environment variable and compile option (llama/12371)

lslusarczyk committed 1 year ago

fa4da9c2

Older