b5579

server : disable speculative decoding for SWA models (#13970)

* server : use swa-full fo draft context

ggml-ci

* server : disable speculative decoding for SWA models

b5577

gemma : more consistent attention scaling for v2 and v3 (#13951)

* gemma : fix attn scale for 27B

* cont : apply scale before attn

* cont : consistent attention scaling

b5576

`server`: update deepseek reasoning format (pass reasoning_content as…

b5575

mtmd : fix memory leak in mtmd_helper_eval_chunk_single (#13961)

* mtmd : fix memory in mtmd_helper_eval_chunk_single

* mtmd-cli : fix mem leak

* Update tools/mtmd/mtmd-cli.cpp

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

---------

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

b5574

cmake : Handle mixed-case 'Power' strings in POWER CPU detection (#13…

b5573

sycl: quantize and reorder the input to q8_1 when reorder is enabled …

b5572

gguf: fix failure on version == 0 (#13956)

b5571

convert : fix nomic-bert-moe mask token (#13757)

b5569

ggml: check if non-native endian model is being loaded (#13943)

* gguf: prevent non-native endian models from being loaded

Signed-off-by: Aaron Teo <[email protected]>

* gguf: update error message

Signed-off-by: Aaron Teo <[email protected]>

* gguf: make the non-native endian check more verbose

Signed-off-by: Aaron Teo <[email protected]>

* ggml: move ggml_assert location

Signed-off-by: Aaron Teo <[email protected]>

* ggml: reword the endianness check error message

Signed-off-by: Aaron Teo <[email protected]>

---------

Signed-off-by: Aaron Teo <[email protected]>

b5568

sync : ggml

ggml-ci