Pytorch Taskrunner Workspace with Keras 3 #1333
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Signed-off-by: yes <[email protected]>
Signed-off-by: yes <[email protected]>
Signed-off-by: yes <[email protected]>
Signed-off-by: yes <[email protected]>
Signed-off-by: yes <[email protected]>
Signed-off-by: yes <[email protected]>
Signed-off-by: yes <[email protected]>
tanwarsh
requested review from
MasterSkepticista,
kta-intel,
psfoley and
teoparvanov
as code owners
kta-intel
reviewed
Feb 3, 2025
There was a problem hiding this comment.
Thanks @tanwarsh ! Some open ended comments/questions:
- We should consider having this one directory down, either under
keras/pytorchworkspace or underpytorch/keras. This would simply be a way to further indicate that is either (1) an extended functionality of the native keras taskrunner or (2) a special case for pytorch that uses the keras taskrunner. - Is there a reason you decided to go with a customized
fit()function as opposed to a custom training loop? This way is not wrong, of course, but the custom training loop seems like a more natural addition to the task runner since a dev would be expected to modify thetrain_andvalidatefunctions, whereas inheritingkeras.Modeland then modifyingtrain_stepandtest_stepmay be confusing. If we stick to this, I suggest we move the model definition to a separate file then import it intaskrunner.py. Also, even if we stick with this, I think we should also add in the a workspace that takes advantage of the custom training loop (it can be added as backlog item to be added in a separate PR)
kta-intel
reviewed
Feb 3, 2025
|
|
||
| compression_pipeline : | ||
| defaults : plan/defaults/compression_pipeline.yaml | ||
| # To use different Compression Pipeline, uncomment the following lines |
There was a problem hiding this comment.
We an remove these comments since we have dedicated workspaces / docs for compression pipeline
| """ | ||
| super().__init__(batch_size, **kwargs) | ||
|
|
||
| # TODO: We should be downloading the dataset shard into a directory |
There was a problem hiding this comment.
We can remove these TODOs and add it as a backlog item. FYI, I don't think it's entirely necessary to download the dataset shard into a directory for the --template workspaces, but automatic sharding based on rank / size of collab list could be interesting
|
|
||
| from openfl.federated import KerasTaskRunner | ||
|
|
||
| class CNNModel(keras.Model): |
There was a problem hiding this comment.
from my previous comment, we should move this to a separate file then import it
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Pytorch Taskrunner Workspace with Keras 3
This PR is blocked due to keras-team/keras#20847.