llguidance : official v0.7.20 release (no actual changes) [noci] (ggm…

…l-org#13594)

minja: sync (qwen3) (ggml-org#13573)

* minja: sync google/minja@f06140f

- google/minja#67 (@grf53)
- google/minja#66 (@taha-yassine)
- google/minja#63 (@grf53)
- google/minja#58

---------

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

gguf : use ggml log system (ggml-org#13571)

* gguf : use ggml log system

* llama : remove unnecessary new lines in exception messages

server : proper error handling for missing elements in messages array…

… (OpenAI compatible backend) (ggml-org#13540)

`common`: add partial regex support (ggml-org#12808)

* move string_find_partial_stop & string_ends_with to common

* add common_regex (supports partial matches)

Co-authored-by: Georgi Gerganov <[email protected]>

* Update common/regex-partial.cpp

Co-authored-by: Georgi Gerganov <[email protected]>

* Update common/regex-partial.cpp

Co-authored-by: Georgi Gerganov <[email protected]>

* Update common/regex-partial.h

Co-authored-by: Georgi Gerganov <[email protected]>

* partial regex: add missing iterator end checks

* string utils: use string_views

* direct throw to avoid ggml.h include

* regex-partial: replace missed ggml_asserts

---------

Co-authored-by: ochafik <[email protected]>
Co-authored-by: Georgi Gerganov <[email protected]>

CUDA: faster Deepseek FA, add Turing support (ggml-org#13435)

SYCL: Refactor and enable FP16 in binary broadcast OPs (ggml-org#12975)

* SYCL: refactor move to a separate file

* Fix binbcast

* Remove duplicates

* fix include formatting

* fix typo

sycl: Support sycl_ext_oneapi_limited_graph (ggml-org#12873)

The current usage of the SYCL-Graph extension checks for
the `sycl_ext_oneapi_graph` device aspect. However, it is also
possible to support `sycl_ext_oneapi_limied_graph` devices that
don't support update

hellaswag: display estimated score confidence interval (ggml-org#12797)

sync: minja (ggml-org#12739)

* sync: minja

google/minja#57

* fix json include