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

Add an accuracy threshold to stop the trainings #15

Open
eyp opened this issue Nov 13, 2020 · 0 comments
Open

Add an accuracy threshold to stop the trainings #15

eyp opened this issue Nov 13, 2020 · 0 comments
Assignees
Labels
backend Related to the backend client enhancement New feature or request frontend Related to the frontend server
Milestone

Comments

@eyp
Copy link
Owner

eyp commented Nov 13, 2020

Add another hyperparam, min_accuracy or accurracy_threshold to stop the training rounds when this value of accuracy is reached. This is not simple, because are the client nodes the ones that validates the training, so this accuracy value is not shared yet, probably the last accuracy value it should be returned by the clients along the model weights, then the server could do an average of the accuracy and use this value to check if the threshold has been reached.

@eyp eyp added enhancement New feature or request frontend Related to the frontend backend Related to the backend client server labels Nov 13, 2020
@eyp eyp added this to the Version 1.0 milestone Nov 13, 2020
@eyp eyp self-assigned this Nov 13, 2020
@eyp eyp modified the milestones: Version 1.0, Version 1.1 Dec 6, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
backend Related to the backend client enhancement New feature or request frontend Related to the frontend server
Projects
None yet
Development

No branches or pull requests

1 participant