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

Separate trainer and updater #1319

Closed
helinwang opened this issue Feb 13, 2017 · 1 comment
Closed

Separate trainer and updater #1319

helinwang opened this issue Feb 13, 2017 · 1 comment

Comments

@helinwang
Copy link
Contributor

helinwang commented Feb 13, 2017

Before we are planning to use a single concept trainer to combine trainer and updater.
E.g.,

paddle.trainer.Momentum
paddle.trainer.MomentumParallel
paddle.trainer.AdagradParallel
paddle.trainer.Adagrad
paddle.trainer.AdagradParallel
paddle.trainer.Adam
paddle.trainer.AdamParallel

The number-of-class will be number-of-trainer x number-of-update-algorithm.

To make number of class fewer, maybe we can separate trainer and updater. And update is a C++ interface with a pure virtual function update:

paddle.trainer.SGD
paddle.trainer.SGDParallel
paddle.updater.Adam
paddle.updater.Adagrad
paddle.updater.Momentum
@wangkuiyi
Copy link
Collaborator

I agree. And I am adding this point into the API design.

wangkuiyi added a commit to wangkuiyi/Paddle that referenced this issue Feb 13, 2017
wangxicoding pushed a commit to wangxicoding/Paddle that referenced this issue Dec 9, 2021
* Update Taskflow REAME.md and enable download logs

* Enable download log for Taskflow
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

No branches or pull requests

2 participants