cann : add Ascend NPU instructions (#2410)

cmake: Fix libdir value in pkgconfig file (#2407)

revert : cmake : set MSVC to use UTF-8 on source files (#2346)

sync : ggml

ggml: fix ggml_graph_cpy undefined behavior (ggml/943)

cann : fix doxy (ggml/0)

vulkan : fix build (llama/0)

cuda : mark BF16 CONT as unsupported

ggml : fix cont with transposed tensors when one dimension is 1 (ggml/934)

cmake : set MSVC to use UTF-8 on source files (#2346)

readme : remove invalid flag from Python example (#2396)

readme : fix link (#2394)

go : add beamsize/entropythold/maxcontext to context interface (#2350)

talk-llama : sync llama.cpp

whisper : update FA call

sync : ggml

sync : vulkan (skip) (llama/0)

ggml : do not crash when quantizing q4_x_x with an imatrix (llama/9192)

metal : separate scale and mask from QKT in FA kernel (llama/9189)

ggml : add SSM Metal kernels (llama/8546)

metal : gemma2 flash attention support (llama/9159)

CPU/CUDA: Gemma 2 FlashAttention support (llama/8542)

Add a space to supress a cmake warning (llama/9133)

Add oneDNN primitive support (llama/9091)

llama : simplify Mamba with advanced batch splits (llama/8526)

fallback mmvq (llama/9088)

Fix SYCL `im2col` and `convert` Overflow with Large Dims (llama/9052)

rpc : print error message when failed to connect endpoint (llama/9042)

rpc : prevent crashes on invalid input (llama/9040)

ggml : dynamic ggml_sched_max_splits based on graph_size (llama/9047)

cmake : remove unused option GGML_CURL (llama/9011)

ggml : move rope type enum to ggml.h (llama/8949)

ggml: fix div-by-zero (llama/9003)

Optimize Vulkan backend for better CPU performance and less GPU synchronization overhead. (llama/8943)

feat: ref. cross entropy, add CUDA, fix grad test (ggml/929)

ggml: remove bad assert (ggml/928)

examples: add MNIST training + missing ops

models : add support for wget2 for fedora (#2387)

readme : update the path to bench.py (#2386)

readme : fix typo (#2383)

readme : fix broken links in implementation details section (#2382)

whisper : fix compile warning for unused params

sync : ggml vulkan (ggml/0)

yolo : add backend support (ggml/924)

ggml : fix typo in ggml-quants.c comment (ggml/922)

feat: add new `sin` and `cos` operators (ggml/919)

readme : fix broken links (#2358)

examples : use colorblind friendly TTY color scheme (#2360)

sync : ggml

ggml : support forward pass broadcasting in ggml_sub (ggml/914)