b4936

model : do not repack if a GPU device is present (#12498)

ggml-ci

b4935

chore : cleanup llama_model_loader::TENSOR_ usage (#12492)

b4934

llama-tts : avoid crashes related to bad model file paths (#12482)

b4933

[SYCL] Fix build on Windows when ccache enabled (#9954) (#9976)

* [SYCL] Fix build on Windows when ccache enabled (#9954)

* take effect only on windows and force it to icl

---------

Co-authored-by: Romain Biessy <[email protected]>

b4932

sycl: cleanup oneDNN related code (#12097)

b4930

llama : make Qwen2MoE QKV bias optional (#12477)

b4929

ggml : block interleaving support for Q4_K quantization for x86 AVX2 …

b4927

context : clear sets containing encoder output sequence ids before st…

b4926

CUDA: Improve flash decoding kernel GPU occupancy for BS=1 case (#12183)

- Find out active blocks per SM using cudaOccupancyMaxActiveBlocksPerMultiprocessor API. Use this value to determine the optimal parallel_blocks value.
- Prefer vector flash attention kernels over MMA kernel for BS=1

Fixes Issue: #12182
---------

Co-authored-by: Johannes Gäßler <[email protected]>

b4925

vulkan: optimize iq1 coopmat2 dequant functions (#12427)