ngxson/llama.cpp

Pull Requests Commits

ngxson committed 2 years ago

f6b2e1d8

server: only check model template if there is no custom tmpl

ngxson committed 2 years ago

2ab9cb96

server: add AlphaMonarch to test chat template

ngxson committed 2 years ago

10d86733

add new chat template

ngxson committed 2 years ago

b599545e

server: fallback to chatml

ngxson committed 2 years ago

9576df90

server : support llava 1.6 (#5553)

cjpais committed 2 years ago

Verified 6560bed3

make : fix debug build with CUDA (#5616)

slaren committed 2 years ago

Verified 06bf2cf8

llava : add explicit instructions for llava-1.6 (#5611)

danbev committed 2 years ago

Verified 4ed8e4fb

Server: use llama_chat_apply_template (#5593)

ngxson committed 2 years ago

Verified 9c405c9f

readme : update UI list (#5605)

danemadsen committed 2 years ago

Verified 5207b3fb

metal : add build system support for embedded metal library (#5604)

tonyfettes committed 2 years ago

Verified 8dbbd757

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

Older