Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Cherry-pick][Paddle Inference] fix mixed precision diff #49477

Merged

Conversation

yuanlehome
Copy link
Contributor

@yuanlehome yuanlehome commented Dec 30, 2022

PR types

Others

PR changes

Others

Describe

  • scale op的scale和bias属性值可能会影响float16计算精度,handle it in amp pass.
  • fused_fc_elementwise_layernorm kernel的float16计算有精度问题,fix it.
  • fc_kernel的float16计算有精度问题,fix it.
  • 避免amp pass插入冗余的cast op并修复一些函数和变量命名不当问题。
  • 将一些无精度问题的op移除黑名单。
  • 将gpu_ernie_half_test之前由于fc kernel的精度问题而禁用的fc_fuse_pass开启并修改绝对误差容忍值。

@paddle-bot
Copy link

paddle-bot bot commented Dec 30, 2022

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.

@yuanlehome yuanlehome changed the title Cp fix mixed precision diff [Cherry-pick][Paddle Inference] fix mixed precision diff Dec 30, 2022
@yuanlehome yuanlehome force-pushed the cp_fix_mixed_precision_diff branch 2 times, most recently from 246b74c to cffd8fa Compare December 31, 2022 16:11
@qingqing01 qingqing01 merged commit 1d25c66 into PaddlePaddle:release/2.4 Jan 4, 2023
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants