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] fix the error of fake_quant_dequant op name (#32866) #32879

Merged
merged 1 commit into from
May 17, 2021

Conversation

juncaipeng
Copy link
Contributor

@juncaipeng juncaipeng commented May 12, 2021

PR types

Bug fixes

PR changes

Others

Describe

fix the error of fake_quant_dequant op name

动态图量化训练:

  • 如果权重使用channel_wise_abs_max方式,产出的量化模型中原本量化的op会有skip_quant属性,导致部署时所有量化op依旧执行fp32 kernel;
  • 如果权重使用abs_max方式,不会有上述问题。
  • 权重使用abs_max和channel_wise_abs_max方式,精度基本相同。

在动态图量化训练功能,本PR修复判断的op名字,修复上述问题。

@paddle-bot-old
Copy link

Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

@XiaoguangHu01 XiaoguangHu01 merged commit 4831e37 into PaddlePaddle:release/2.1 May 17, 2021
@juncaipeng juncaipeng deleted the release/2.1-ch11 branch May 17, 2021 08:46
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.

3 participants