b1663

CUDA: Faster Mixtral prompt processing (#4538)

* CUDA: make MoE tensors contiguous for batch size>1

* Update ggml-cuda.cu

Co-authored-by: slaren <[email protected]>

---------

Co-authored-by: slaren <[email protected]>

b1662

ggml : fixed check for _MSC_VER (#4535)

Co-authored-by: Eric Sommerlade <[email protected]>

b1661

ggml-cuda: Fix HIP build (#4528)

regression of #4490
Adds defines for two new datatypes
cublasComputeType_t, cudaDataType_t.

Currently using deprecated hipblasDatatype_t since newer ones very recent.

b1657

llama : fix try_override for bool_value which always return true (#4519)

b1656

decode : fix logits_valid for legacy API (#4516)

b1644

Merge pull request #2 from ggerganov/master

[pull] master from ggerganov:master

b1641

Merge branch 'ggerganov:master' into master

b1640

ggml : remove n_dims from ggml_tensor (#4469)

ggml-ci

b1635

Merge pull request #1 from ggerganov/master

[pull] master from ggerganov:master