ET backend generalization and performance uplift#8

Open

marty1885 wants to merge 455 commits intoaifoundry-org:etfrom

marty1885:backend-dev-2

This pull request is big! We're only showing the most recent 250 commits

Commits on Mar 30, 2026

Commits on Mar 31, 2026

Commits on Apr 1, 2026

Commits on Apr 2, 2026

Commits on Apr 3, 2026

Commits on Apr 6, 2026

Commits on Apr 7, 2026

Commits on Apr 8, 2026

Commits on Apr 9, 2026

common : skip non-primary GGUF split files when selecting model (#21633)
angt
authored
vulkan: unify type macros to use Vx instead of _VECx (#21605)
0cc4m
authored
ci: drop v5 `all:` composition from labeler.yml (#21627)

Marxist-Leninist
and
noreply
authored
sycl : add flash-attn support for head size 512 (#21654)
qnixsynapse
authored
webui: Add option to pre-encode conversation for faster next turns (#21034)
allozaur
authored
server : fix grammar commandline args (#21543)
AUTOMATIC1111
authored
fix: Model Selector choice sync (#21628)
allozaur
authored
metal : add missing mm-id specializations for q1_0 (#21662)
ggerganov
authored
jinja : support ensure_ascii=true, string repetition and int/float self-filtering (#21623)

kwajiehao
and
CISC
authored
vocab: add gemma4 tokenizer tests, fix edge case (#21534)
pwilkin
authored
mtmd: support dots.ocr (#17575)
ngxson
authored
model: fix multimodal padding token for gemma3n/gemma4 (#21625)
ngxson
authored
common : simplify autoparser tagged parser rules (#21216)
aldehir
authored
common : fix ambiguous grammar rule in gemma4 (#21661)
aldehir
authored
webui: add "Send message on Enter" setting (#21577)
mourix
authored
requirements : update transformers to 5.5.1 (#21617)
danbev
authored
ggml : check return value of CUB calls used in argsort and top-k (they all return cudaError_t) (#21676)

fairydreaming
and
sszymczy
authored
ggml: backend-agnostic tensor parallelism (experimental) (#19378)

authored
HIP: add CDNA4 (gfx950) architecture support for MI350X/MI355X (#21570)
andyluo7
authored

Commits on Apr 10, 2026

Commits on Apr 12, 2026