PaddlePaddle / Paddle

PArallel Distributed Deep LEarning: Machine Learning Framework from Industrial Practice (『飞桨』核心框架,深度学习&机器学习高性能单机、分布式训练和跨平台部署)
http://www.paddlepaddle.org/
Apache License 2.0
22.29k stars 5.61k forks source link

Bug fix amp #69694

Closed zhanghonggeng closed 17 hours ago

zhanghonggeng commented 1 day ago

PR Category

Performance Optimization

PR Types

Performance

Description

Fix bug caused by AMP before moving to CINN. The output should not be converted from int64 to float16 for the clip arithmetic. image

image

Pcard-67164

paddle-bot[bot] commented 1 day ago

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.