intel/auto-round

Pull Requests Commits

ark: document sparse qtile256 experiments and results

yiliu30 committed 91 days ago

6a9906a9

ark: add focused sparse-dev benchmark modes

yiliu30 committed 91 days ago

48bd8b7f

ark: extend sparse prefill runtime for decoupled rows

yiliu30 committed 91 days ago

5f55ff30

Add sparse kernel profiling modes and dev bench path

yiliu30 committed 93 days ago

99549bc6

Checkpoint sparse Sage kernel baseline

yiliu30 committed 95 days ago

c22378a2

Add opt-in Triton CUDA-port preprocess backend

yiliu30 committed 96 days ago

cefff6f0

Add top-level reproduce guide

yiliu30 committed 96 days ago

6aecab18

Document reproduce commands for Triton-XPU sparse work

yiliu30 committed 96 days ago

827a6e4b

Add Triton-XPU sparse prefill backend

yiliu30 committed 96 days ago

cde63c16

yiliu30 committed 104 days ago

b38e2b5e

[ARK] Clean sparse Python glue module

yiliu30 committed 110 days ago

62a62177

[ARK] Split sparse Sage outer wrapper

yiliu30 committed 111 days ago

b3f722c8

[ARK] Document sparse attention clean branch

yiliu30 committed 111 days ago

77ea766e

[ARK] Document current sparse attention profiling status

yiliu30 committed 111 days ago

855d12a1

add nhd integration

yiliu30 committed 111 days ago

98b05279

[ARK] Update sparse kernel status and prefetch notes

yiliu30 committed 111 days ago

924dcd61

[ARK] Use copy-based sparse Sage mainloop baseline

yiliu30 committed 111 days ago

e2d6e7c1

[ARK] Add XPU sparse kernel benchmark harness

yiliu30 committed 111 days ago

b65f12cd

[ARK] Add Wan and Flux sparse attention demos

yiliu30 committed 111 days ago

15e78c8d

[ARK] Add sparse attention design and status docs

yiliu30 committed 111 days ago

c6b58f0f

[ARK] Add sparse attention tests and benchmark

yiliu30 committed 111 days ago

9d9c77dd

[ARK] Add sparse Sage kernel and preprocess path

yiliu30 committed 111 days ago

59a8c26c

fix random rotation and update rotation doc. (#1884)

lkk12014402 committed 117 days ago

Verified 4e2ad772

fix performance regression (#1886)

wenhuach21 committed 122 days ago

Verified 30f72bda

Suppress misleading warning when detecting model type for GGUF export directories (#1887)

lvliang-intel committed 123 days ago

Verified fa48924e

[ARK] Add bf16 + NHD layout support, refactor sage_dynamic_quant (#1882)

luoyu-intel committed 123 days ago

Verified 08c5d6d4

gguf format add support for mtp quantization (#1866)

n1ck-guo committed 123 days ago

Verified f7e77e03

Fix SGL CI dependency conflict (#1877)

WeiweiZhang1 committed 125 days ago

Verified 77720aea

Bump build version to v0.14.0 (#1875)

chensuyue committed 126 days ago

Verified 15e22ede

fix segmentation failure (#1871)

xin3he committed 126 days ago

Verified 22b0fa1b

Older