-
Notifications
You must be signed in to change notification settings - Fork 215
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
Create interfaces for OpenFL components with methods that can be implemented #802
Labels
Comments
openfl team 2 is working on this |
danhe1
added a commit
to danhe1/openfl
that referenced
this issue
Apr 28, 2023
…ementedCreate interfaces for OpenFL components with methods that can be implemented Create AggregatorInterface and CollaboratorInterface in openflinterface/interactive_api/experiment.py Support passing the aggregator object and collaborator object in the method FLExperiment.start() If aggregator object or collaborator object is None, the default aggregator and collaborator objects will be created at director and envoy services. Known issue: Since the AggregatorInterface and CollaboratorInterface are for ad-hoc algorithm development. We should modify the definitions of the model, data loader and task interfaces, since we are going to support some algorithms other than Keras, TensorFlow or PyTorch. Fixes securefederatedai#802 Signed-off-by: He, Dan H <[email protected]> Signed-off-by: Jiang, Jiaqiu <[email protected]> Signed-off-by: Li, Qingqing <[email protected]> Signed-off-by: Wang, Le <[email protected]> Signed-off-by: Wu, Caili <[email protected]> Signed-off-by: He, Dan H <[email protected]>
danhe1
added a commit
to danhe1/openfl
that referenced
this issue
Apr 29, 2023
…ementedCreate interfaces for OpenFL components with methods that can be implemented Create AggregatorInterface and CollaboratorInterface in openflinterface/interactive_api/experiment.py Support passing the aggregator object and collaborator object in the method FLExperiment.start() If aggregator object or collaborator object is None, the default aggregator and collaborator objects will be created at director and envoy services. Known issue: Since the AggregatorInterface and CollaboratorInterface are for ad-hoc algorithm development. We should modify the definitions of the model, data loader and task interfaces, since we are going to support some algorithms other than Keras, TensorFlow or PyTorch. Fixes securefederatedai#802 Signed-off-by: He, Dan H <[email protected]> Signed-off-by: Jiang, Jiaqiu <[email protected]> Signed-off-by: Li, Qingqing <[email protected]> Signed-off-by: Wang, Le <[email protected]> Signed-off-by: Wu, Caili <[email protected]> Signed-off-by: He, Dan H <[email protected]> Signed-off-by: He, Dan H <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Is your feature request related to a problem? Please describe.
All OpenFL components like aggregator, collaborator, aggregation algorithms are directly implemented which makes it difficult to write new s/w that would want to customize a few of these components. For ex; Add/test new algos, a new add-on based on OpenFL.
Either the base OpenFL code has to be tweaked or inherit these classes and re-implement the code to suit the requirement
Describe the solution you'd like
If OpenFL components had interfaces defined, these interfaces can be implemented by someone looking to re-use OpenFL.
An option to directly use OpenFL's components should also be provided
Describe alternatives you've considered
TBD
The text was updated successfully, but these errors were encountered: