mmha lightVersion

[YKTian-x2b]

PR Category

Performance Optimization

PR Types

Performance

Description

mmha lightVersion, split mmha_kernel to two kernel, which will lead a more clear impl, but a little perf loss

[paddle-bot[bot]]

你的PR提交成功，感谢你对开源项目的贡献! 请关注后续CI自动化测试结果，详情请参考Paddle-CI手册。 Your PR has been submitted. Thanks for your contribution! Please wait for the result of CI firstly. See Paddle CI Manual for details.

[paddle-ci-bot[bot]]

Sorry to inform you that 98ef7db's CIs have passed for more than 7 days. To prevent PR conflicts, you need to re-run all CIs manually.