[Intel HPU] enable tensor_wise_fp8 - #5324

Merged

zoooo0820 merged 9 commits into

PaddlePaddle:developfrom

fmiao2372:develop_hpu_upstream_fp8

Dec 17, 2025

Commits on Dec 16, 2025

[Intel HPU] enable tensor_wise_fp8
fmiao2372
committed
6d027a7
update code based on comments
fmiao2372
committed
b7c405a
fix code style issue
fmiao2372
committed
48e8153
fix bug about RP 5138
fmiao2372
committed
307efee
mv kv_cache modifications to HPU backend

yanfeich
authored and
fmiao2372
committed
082917d
fix FP8 Precision Issues

yanfeich
authored and
fmiao2372
committed
ae4fd32
fix FP8 Precision Issues

yanfeich
authored and
fmiao2372
committed
5aca689
Merge branch 'develop' into develop_hpu_upstream_fp8
EmmonsCurse
authored
b7e8c36
Add quantization UT
yanfeich
committed
cb65b09