merge from upstream#73

Merged

l3utterfly merged 73 commits intolayla-buildfrom

Jul 6, 2025

Commits on Jul 1, 2025

Commits on Jul 2, 2025

ggml : Callback before abort (#14481)

ScaledLizard
and
slaren
authored
github : add OpenCL backend to issue templates (#14492)
EZForever
authored
ci : add OpenCL to labeler workflow (#14496)
CISC
authored
opencl : update upscale to support align corners (#14488)
lhez
authored
opencl : skip empty nodes on cgraph compute (#14491)
EZForever
authored
simple-chat : fix context-exceeded condition (#14494)
ggerganov
authored
opencl : fix possible buffer overflow in dump_tensor (#14490)
jeffzhou2000
authored
ggml : support bcast ggml_soft_max_ext, ggml_flash_attn_ext (#14435)
ggerganov
committed
vulkan: support softmax/FA batch and broadcast (#14449)

jeffbolznv
authored and
ggerganov
committed
CUDA: broadcasting for FlashAttention mask (#14500)

JohannesGaessler
authored and
ggerganov
committed
CUDA: add softmax broadcast (#14475)

am17an
authored and
ggerganov
committed
Set RPATH to "@loader_path" / "$ORIGIN" to ensure executables and dynamic libraries search for dependencies in their origin directory. (#14309)
rotemdan
authored
ggml : add version function to get lib version (ggml/1286)

danbev
and
ggerganov
committed
sync : ggml
ggerganov
committed
llama : initial Mamba-2 support (#9126)
compilade
authored
gguf-py : add support for chat template jinja files (#14508)
CISC
authored
CUDA: add dynamic shared mem to softmax, refactor general usage (#14497)
am17an
authored

Commits on Jul 3, 2025

Commits on Jul 6, 2025