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

[fix] parse compression pipeline from plan.yaml during fx model save #1152

Merged

Conversation

kta-intel
Copy link
Collaborator

Resolves #1142

This PR implements logic to parse the compression pipeline from plan.yaml in order to instantiate the correct pipeline to be used to deconstruct the model proto when calling fx model save

Originally, a model proto is constructed with the best/last models using some specified compression pipeline. These pipelines come with a backward() pass to deconstruct the proto and extract the model. However, the default pipeline that is invoked with fx model save is the NoCompressionPipeline(), which will raise an error if the original proto was constructed with a different pipeline. This PR ensures that the same pipeline is used in both construction and deconstruction of the model proto

@MasterSkepticista MasterSkepticista merged commit a977fb3 into securefederatedai:develop Nov 16, 2024
29 checks passed
@kta-intel kta-intel deleted the compression/model-save branch November 16, 2024 14:05
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

Successfully merging this pull request may close these issues.

ValueError: buffer size must be a multiple of element size
2 participants