ggerganov/llama.cpp

Pull Requests Commits

llama : fix K-shift with quantized K (wip)

slaren committed 1 year ago

5271c756

llama : fix session save/load with quantized KV (#5649)

slaren committed 1 year ago

Verified 7fe4678b

gemma : allow offloading the output tensor (#5646)

slaren committed 1 year ago

Verified ba2135cc

examples : do not assume BOS when shifting context (#5622)

cebtenzzre committed 1 year ago

Verified 89febfed

ggerganov committed 1 year ago

Verified 5022cf24

server: health: fix race condition on slots data using tasks queue (#5634)

phymbert committed 1 year ago

Verified 1ecea255

readme : add LocalAI to the availables UI (#5629)

mudler committed 1 year ago

Verified a00a35ce

sync : ggml (#5633)

ggerganov committed 1 year ago

Verified eccd7a26

readme : update hot topics

ggerganov committed 1 year ago

Verified c14f72db

llava : add --skip-unknown to 1.6 convert.py (#5632)

danbev committed 1 year ago

Verified cc6cac08

llama : add `gemma` model (#5631)

postmasters committed 1 year ago

Verified 580111d4

[SYCL] conext add name (#5624)

airMeng committed 1 year ago

Verified 88c46cbd

IQ4_NL: 4-bit non-linear quants with blocks of 32 (#5590)

ikawrakow committed 1 year ago

Verified a14679cc

server : support llava 1.6 (#5553)

cjpais committed 1 year ago

Verified 6560bed3

make : fix debug build with CUDA (#5616)

slaren committed 1 year ago

Verified 06bf2cf8

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

danbev committed 1 year ago

Verified 4ed8e4fb

Server: use llama_chat_apply_template (#5593)

ngxson committed 1 year ago

Verified 9c405c9f

readme : update UI list (#5605)

danemadsen committed 1 year ago

Verified 5207b3fb

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

tonyfettes committed 1 year ago

Verified 8dbbd757

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

phymbert committed 1 year ago

Verified c0a8c6db

Update ggml_sycl_op_mul_mat_vec_q (#5502)

AidanBeltonS committed 1 year ago

Verified b9111bd2

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

dokterbob committed 1 year ago

633782b8

Enable Vulkan MacOS CI

0cc4m committed 1 year ago

22f83f0c

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

0cc4m committed 1 year ago

bb9dcd56

Add check for VK_KHR_portability_enumeration for MoltenVK support

0cc4m committed 1 year ago

f50db6ae

Add preprocessor checks for Apple devices.

dokterbob committed 1 year ago

d8c05451

Resolve ErrorIncompatibleDriver with Vulkan on MacOS.

dokterbob committed 1 year ago

42f664a3

Allow for Vulkan build with Accelerate.

dokterbob committed 1 year ago

5dde5408

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

slaren committed 1 year ago

Verified 40c3a6c1

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

cebtenzzre committed 1 year ago

Verified f24ed14e

Older