llama : add infill sampler (ggml-org#9896)

ggml-ci

ggml : synchronize threads using barriers (ggml-org#7993)

Vulkan Embedding Fix (ggml-org#7360)

* Fix empty Vulkan host buffers

Add fp32 fp16 matmul shader

Fix matmul shader alignment

* Remove deprecated tensor->backend uses

* Fix Vulkan validation errors on embedding models with no offloaded layers

* Fix Vulkan llava segfault when not offloading layers

Merge pull request #15 from ggerganov/master

push update

Revert "move ndk code to a new library (ggml-org#6951)" (ggml-org#7282)

This reverts commit efc8f76.

Merge pull request #12 from JohannesGaessler/server-ngram-4

Server ngram 4

Merge pull request #10 from JohannesGaessler/cuda-fa-no-tc-11

Cuda fa no tc 11

cmake : fix version cmp (ggml-org#7227)

llama-bench : add pp+tg test type (ggml-org#7199)

Comment explaining a decision (ggml-org#6531)