ggerganov/llama.cpp

Pull Requests Commits

Using dp4a ptx intrinsics for an improved Mul8MAT perf [By Alcpz]

OuadiElfarouki committed 2 years ago

eab4a882

cuda : organize vendor-specific headers into vendors directory (#8746)

yeahdongcn committed 2 years ago

Verified 439b3fc7

[SYCL] add conv support (#8688)

airMeng committed 2 years ago

Verified 0832de72

cmake: use 1 more thread for non-ggml in CI (#8740)

JohannesGaessler committed 2 years ago

Verified 6eeaeba1

chore : Fix vulkan related compiler warnings, add help text, improve CLI options (#8477)

teleprint-me committed 2 years ago

Verified 4730faca

llama : refactor session file management (#8699)

compilade committed 2 years ago

Verified 4c676c85

feat: Support Moore Threads GPU (#8383)

yeahdongcn committed 2 years ago

Verified e54c35e4

scripts : sync vulkan-shaders (#0)

ggerganov committed 2 years ago

Verified 5e2727fe

scripts : sync ggml-aarch64 sources

ggerganov committed 2 years ago

Verified 56f20aa2

ggml : add missing semicolon (#0)

ggerganov committed 2 years ago

345c8c0c

ggerganov committed 2 years ago

ae7985cd

ggml : loop tiling optimizations for scalar path (ggml/898)

heshpdx committed 2 years ago

a05ca936

ggml: add support for float16 input tensors in pooling operations (ggml/895)

vanaka11 committed 2 years ago

9f77d899

vulkan : initialize vk_buffer_struct members to VK_NULL_HANDLE (ggml/893)

neobrain committed 2 years ago

203b7f15

cmake : only enable GGML_NATIVE and x86 flags if not crosscompiling (ggml/885)

iboB committed 2 years ago

d2b851bf

ggml : remove unnecessary UNUSED macro call (ggml/880)

danbev committed 2 years ago

c12b6e8e

llama : add support for llama 3.1 rope scaling factors (#8676)

jmorganca committed 2 years ago

Verified b5e95468

llama : add function for model-based max number of graph nodes (#8622)

ggerganov committed 2 years ago

Verified 92090eca

common : add --no-warmup option for main/llama-cli (#8712)

danbev committed 2 years ago

Verified 9d03d085

cann: Fix Multi-NPU execution error (#8710)

wangshuai09 committed 2 years ago

Verified bfb4c749

ggml : reduce hash table reset cost (#8698)

slaren committed 2 years ago

Verified 2b1f616b

llama : fix order of parameters (#8706)

foldl committed 2 years ago

Verified 01245f5b

server : add Speech Recognition & Synthesis to UI (#8679)

ElYaiko committed 2 years ago

Verified 01aec4a6

examples : export-lora : fix issue with quantized base models (#8687)

ngxson committed 2 years ago

Verified 41cd47ca

ggml: handle ggml_init failure to fix NULL pointer deref (#8692)

DavidKorczynski committed 2 years ago

Verified 49ce0ab6

llama : fix build + fix fabs compile warnings (#8683)

ggerganov committed 2 years ago

Verified 4226a8d1

ggml : fix build on Windows with Snapdragon X (#8531)

AndreasKunar committed 2 years ago

Verified bf5a81df

tests : fix printfs (#8068)

ggerganov committed 2 years ago

Verified 88954f7f

[SYCL] fix multi-gpu issue on sycl (#8554)

ClarkChin08 committed 2 years ago

Verified ed67bcb2

ggml : add and use ggml_cpu_has_llamafile() (#8664)

ggerganov committed 2 years ago

Verified eddcb523

Older