ggerganov/llama.cpp

Pull Requests Commits

server : health endpoint configurable failure on no slot (#5594)

phymbert committed 2 years ago

Verified c0a8c6db

Update ggml_sycl_op_mul_mat_vec_q (#5502)

AidanBeltonS committed 2 years ago

Verified b9111bd2

nix: now that we can do so, allow MacOS to build Vulkan binaries

dokterbob committed 2 years ago

633782b8

Enable Vulkan MacOS CI

0cc4m committed 2 years ago

22f83f0c

Refactor validation and enumeration platform checks into functions to clean up ggml_vk_instance_init()

0cc4m committed 2 years ago

bb9dcd56

Add check for VK_KHR_portability_enumeration for MoltenVK support

0cc4m committed 2 years ago

f50db6ae

Add preprocessor checks for Apple devices.

dokterbob committed 2 years ago

d8c05451

Resolve ErrorIncompatibleDriver with Vulkan on MacOS.

dokterbob committed 2 years ago

42f664a3

Allow for Vulkan build with Accelerate.

dokterbob committed 2 years ago

5dde5408

cuda : ignore peer access already enabled errors (#5597)

slaren committed 2 years ago

Verified 40c3a6c1

make : pass CPPFLAGS directly to nvcc, not via -Xcompiler (#5598)

cebtenzzre committed 2 years ago

Verified f24ed14e

examples : support minItems/maxItems in JSON grammar converter (#5039)

nopperl committed 2 years ago

Verified 9d679f0f

llava : remove extra cont (#5587)

ggerganov committed 2 years ago

Verified 1387cf60

llava : replace ggml_cpy with ggml_cont

slaren committed 2 years ago

6fd41379

ggerganov committed 2 years ago

337c9cbd

ggml-alloc : apply ggml/731

ggerganov committed 2 years ago

a3145bdc

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

didzis committed 2 years ago

890559ab

ci : enable -Werror for CUDA builds (#5579)

ggerganov committed 2 years ago

Verified d0e3ce51

make : fix CUDA build (#5580)

ggerganov committed 2 years ago

Verified 68a6b98b

readme : fix typo in README-sycl.md (#5353)

valiray committed 2 years ago

Verified 70d45af0

cmake : remove obsolete sycl compile flags (#5581)

abhilash1910 committed 2 years ago

Verified 13e2c771

minor : fix trailing whitespace (#5538)

ggerganov committed 2 years ago

Verified f53119ce

llava : avoid changing the original BakLLaVA model (#5577)

danbev committed 2 years ago

Verified 70847553

baby-llama : allocate graphs in ggml_context (#5573)

NawafAlansari committed 2 years ago

Verified 4480542b

llama : add llama_chat_apply_template() (#5538)

ngxson committed 2 years ago

Verified 11b12de3

cuda, metal : fix nans in soft_max (#5574)

slaren committed 2 years ago

Verified 3a9cb4ca

readme : update (#5572)

Mirko185 committed 2 years ago

Verified 769a716e

ggml : android and old glibc NUMA incompatibility bugfixes (#5557)

bmtwl committed 2 years ago

Verified f0d1fafc

build : pass all warning flags to nvcc via -Xcompiler (#5570)

cebtenzzre committed 2 years ago

Verified a0c2dad9

ggml : restore vec dot stride arg names (#5453)

ggerganov committed 2 years ago

Verified 14278f55

Newer Older