Combine Qwen hybrid prompt-cache fixes on current upstream#3

Closed

Regrad wants to merge 73 commits into

fix/qwen-hybrid-cache-radical

Commits on Jun 3, 2026

Commits on Jun 4, 2026

sycl : Improve SYCL doc (#23025)
malsbat
authored
ggml-cpu: extend RVV quantization vec dot to higher VLENs (#22754)

authored
ggml-webgpu: FlashAttention refactor + standardize quantization support (#23834)
reeselevine
authored
metal : reduce rset heartbeat from 500ms -> 5ms (#24074)
ggerganov
authored
tests : refactor test-save-load-state to accept token input (#24073)
ggerganov
authored
readme : add status badges (#24104)
ggerganov
authored
fix(mtmd): handle Gemma 4 audio projector embedding size (#24091)

abetlen
and
ngxson
authored
cmake: skip cvector-generator and export-lora when CPU backend is disabled (#24053)
arichiardi
authored
server : add header to tools/server/server-http.h (#24089)
abawany
authored
build : use umbrella Headers directory for XCFramework module map (#23974)
gmarzjr
authored
webui: fix tool selector toggle/counter, key tools by stable identity (#24065)
ServeurpersoCom
authored
agents: refactor, include more guidelines (#24111)
ngxson
authored
server: avoid unnecessary checkpoint restore when new tokens are present (#24110)

Abioy
and
ggerganov
authored
ggml: vectorize ggml_vec_dot_q4_1_q8_1 with WASM SIMD128 (#22209)
sirohikartik
authored
convert: Fix Gemma 4 Unified conversion (#24118)
pcuenca
authored
return filter to save memory (#24125)

forforever73
and
lvyichen
authored
ui: added single line reasoning preview (#23601)

gugugiyu
and
allozaur
authored
ui: Fixed packages (#24119)
allozaur
authored
Move duplicated imatrix code into single common imatrix-loader.cpp (#22445)
bartowski1182
authored
webui: [a11y] fix keyboard navigation issues in chat interface and sidebar (#23132)
vignesh191
authored
arg: fix double mtp downloads (#24128)
ngxson
authored
server : disable on-device spec checkpoints (#24108)
ggerganov
authored
server : do not consume prompt-cache entry on load
alainnothere
committed
server : test that prompt-cache entries survive cross-conversation load
alainnothere
committed

Commits on Jun 5, 2026

Commits on Jun 6, 2026