starmountain1997 2cb9f76a0f [CI] Ds32 ep aime2025 (#8496) ...

Backport of #7882 to releases/v0.18.0. Adds aime2025 benchmark test for
DeepSeek-V3.2-W8A8 EP with disaggregated prefill on A3 (4-node, 16 NPUs
per node, accuracy benchmark baseline 66.67%).

Signed-off-by: guozr <guozr1997@hotmail.com>
Co-authored-by: guozr <guozr1997@hotmail.com>

2026-04-21 19:49:06 +08:00

..

e2e

[CI] Ds32 ep aime2025 (#8496)

2026-04-21 19:49:06 +08:00

ut

[Performance] Use forward_native for Conv3dLayer and add UT (#8375)

2026-04-20 17:20:40 +08:00

__init__.py

[SpecDecode] Add spec decode support (#500)

2025-04-17 20:16:32 +08:00