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.

Sign up for GitHub

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

Jump to bottom

Modify max_epochs to support a time abstraction #134

Closed
moinnadeem opened this issue Dec 4, 2021 · 1 comment
Closed

Modify max_epochs to support a time abstraction #134

moinnadeem opened this issue Dec 4, 2021 · 1 comment
Labels
bert enhancement New (engineering) enhancements, such as features or API changes.

Comments

@moinnadeem
Copy link
Contributor

🚀 Feature Request

Currently, we support max_epochs to determine how long to train for. This breaks under an abstraction where the user wants to train for a certain number of samples (which could result in ~3.5 epochs). We should update max_epochs to be max_iters, and values can use the time abstraction to be 3ep or 3131ba (as an example).

Motivation

A user wants to train for fractions of an epoch. How do they do this? In other words, they want to train for a fixed number of steps instead of a fixed number of epochs.
@moinnadeem moinnadeem added enhancement New (engineering) enhancements, such as features or API changes. bert labels Dec 4, 2021
@ravi-mosaicml
Copy link
Contributor

Closing for #146

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
bert enhancement New (engineering) enhancements, such as features or API changes.
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@moinnadeem @ravi-mosaicml