ggerganov/whisper.cpp

Pull Requests Commits

ci : fix windows builds to use 2019

ggerganov committed 1 year ago

Verified b0aeef2d

whisper : use backend registry (#0)

ggerganov committed 1 year ago

37c88027

ggml/sched : do not skip views in pre-assignments

slaren committed 1 year ago

9db070a3

whisper : adapt to new ggml (wip)

ggerganov committed 1 year ago

7fd8d9c2

talk-llama : sync llama.cpp

ggerganov committed 1 year ago

06e059b8

ggerganov committed 1 year ago

c9f49d5f

ggml : sync resolve (skip) (#0)

ggerganov committed 1 year ago

f4c1d7df

Add required ggml-base and backend libs to cmake pkg (llama/10407)

mtmcp committed 1 year ago

339b8e55

cuda : fix CUDA_FLAGS not being applied (llama/10403)

slaren committed 1 year ago

5f6d6919

sycl : Add option to set the SYCL architecture for all targets (llama/10266)

Rbiessy committed 1 year ago

8ee76773

vulkan: Optimize soft_max (llama/10301)

jeffbolznv committed 1 year ago

45f1f914

sycl: Revert MUL_MAT_OP support changes (llama/10385)

Alberto Cabrera Pérez committed 1 year ago

53589c8f

cuda : only use native when supported by cmake (llama/10389)

slaren committed 1 year ago

7ac2f17f

vulkan: remove use of null initializer (llama/10372)

jeffbolznv committed 1 year ago

48862c7b

metal : fox offset integer overflows in im2col (ggml/1015)

pminev committed 1 year ago

44f7d9f4

Vulkan: Fix device info output format specifiers (llama/10366)

0cc4m committed 1 year ago

fd123025

metal : add `GGML_UNARY_OP_ELU` kernel (ggml/1018)

PABannier committed 1 year ago

f80bef46

CUDA: fix MMV kernel being used for FP16 src1 (llama/10357)

JohannesGaessler committed 1 year ago

161b4435

CMake: fix typo in comment [no ci] (llama/10360)

JohannesGaessler committed 1 year ago

ef7fbe1c

llama : only use default buffer types for the KV cache (llama/10358)

slaren committed 1 year ago

0879d359

metal : refactor kernel args into structs (llama/10238)

ggerganov committed 1 year ago

2a444dc5

ggml : fix undefined reference to 'getcpu' (llama/10354)

FirstTimeEZ committed 1 year ago

45cf1634

CUDA: remove DMMV, consolidate F16 mult mat vec (llama/10318)

JohannesGaessler committed 1 year ago

dcb2922d

CMake: default to -arch=native for CUDA build (llama/10320)

JohannesGaessler committed 1 year ago

3c5c7511

ggml : fix possible buffer use after free in sched reserve (llama/9930)

slaren committed 1 year ago

24ad19d0

ggml : inttypes.h -> cinttypes (llama/0)

ggerganov committed 1 year ago

bd574b05

ggml : adapt AMX to tensor->grad removal (llama/0)

ggerganov committed 1 year ago

7e0eafcb

ggml : fix compile warnings (llama/0)

ggerganov committed 1 year ago

75670ae6

llamafile : fix include path (llama/0)

ggerganov committed 1 year ago

d4fcdf60

vulkan: Optimize some mat-vec mul quant shaders (llama/10296)

jeffbolznv committed 1 year ago

1bebb1a1

Older