Implement serialization of Summarizer #1293
Closed
Conversation
This file contains hidden or 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
|
This pull request was exported from Phabricator. Differential Revision: D58227186 |
|
This pull request was exported from Phabricator. Differential Revision: D58227186 |
yucu
pushed a commit
to yucu/captum
that referenced
this pull request
Jun 7, 2024
Summary: Pull Request resolved: pytorch#1293 To save a checkpoint of a Feature Importance job, we will need to persistent below fields in https://www.internalfb.com/code/fbsource/[5e3ebcd68e0c]/fbcode/model_understanding/feature_importance/summarizer.py?lines=89-99 in order save the state of the job: ``` # Initialize summarizers self.input_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) self.input_coverage_summarizer = Summarizer([Mean()]) self.derived_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) self.derived_coverage_summarizer = Summarizer([Mean()]) self.neuron_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) # Initialize batch counts self.input_batch_count = 0 self.derived_batch_count = 0 self.neuron_batch_count = 0 ``` This diff implements (de)serialization of Summarizer. ### future diffs include: - save FI checkpoints to manifold - add the hook to Mast graceful preemption to save the checkpoint of Feature Importance Job - load checkpoint when initializing the Feature Importance Job Reviewed By: MarcioPorto Differential Revision: D58227186
|
@lurunming has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator. |
|
This pull request was exported from Phabricator. Differential Revision: D58227186 |
yucu
pushed a commit
to yucu/captum
that referenced
this pull request
Jun 10, 2024
Summary: To save a checkpoint of a Feature Importance job, we will need to persistent below fields in https://www.internalfb.com/code/fbsource/[5e3ebcd68e0c]/fbcode/model_understanding/feature_importance/summarizer.py?lines=89-99 in order save the state of the job: ``` # Initialize summarizers self.input_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) self.input_coverage_summarizer = Summarizer([Mean()]) self.derived_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) self.derived_coverage_summarizer = Summarizer([Mean()]) self.neuron_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) # Initialize batch counts self.input_batch_count = 0 self.derived_batch_count = 0 self.neuron_batch_count = 0 ``` This diff implements (de)serialization of Summarizer. ### future diffs include: - save FI checkpoints to manifold - add the hook to Mast graceful preemption to save the checkpoint of Feature Importance Job - load checkpoint when initializing the Feature Importance Job Pull Request resolved: pytorch#1293 Reviewed By: MarcioPorto Differential Revision: D58227186 Pulled By: lurunming
|
This pull request was exported from Phabricator. Differential Revision: D58227186 |
yucu
pushed a commit
to yucu/captum
that referenced
this pull request
Jun 10, 2024
Summary: To save a checkpoint of a Feature Importance job, we will need to persistent below fields in https://www.internalfb.com/code/fbsource/[5e3ebcd68e0c]/fbcode/model_understanding/feature_importance/summarizer.py?lines=89-99 in order save the state of the job: ``` # Initialize summarizers self.input_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) self.input_coverage_summarizer = Summarizer([Mean()]) self.derived_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) self.derived_coverage_summarizer = Summarizer([Mean()]) self.neuron_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) # Initialize batch counts self.input_batch_count = 0 self.derived_batch_count = 0 self.neuron_batch_count = 0 ``` This diff implements (de)serialization of Summarizer. ### future diffs include: - save FI checkpoints to manifold - add the hook to Mast graceful preemption to save the checkpoint of Feature Importance Job - load checkpoint when initializing the Feature Importance Job Pull Request resolved: pytorch#1293 Reviewed By: MarcioPorto Differential Revision: D58227186 Pulled By: lurunming
Summary: To save a checkpoint of a Feature Importance job, we will need to persistent below fields in https://www.internalfb.com/code/fbsource/[5e3ebcd68e0c]/fbcode/model_understanding/feature_importance/summarizer.py?lines=89-99 in order save the state of the job: ``` # Initialize summarizers self.input_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) self.input_coverage_summarizer = Summarizer([Mean()]) self.derived_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) self.derived_coverage_summarizer = Summarizer([Mean()]) self.neuron_attr_summarizer = Summarizer([Mean(), StdDev(order=0)]) # Initialize batch counts self.input_batch_count = 0 self.derived_batch_count = 0 self.neuron_batch_count = 0 ``` This diff implements (de)serialization of Summarizer. ### future diffs include: - save FI checkpoints to manifold - add the hook to Mast graceful preemption to save the checkpoint of Feature Importance Job - load checkpoint when initializing the Feature Importance Job Pull Request resolved: pytorch#1293 Reviewed By: MarcioPorto Differential Revision: D58227186 Pulled By: lurunming
|
This pull request was exported from Phabricator. Differential Revision: D58227186 |
|
@lurunming merged this pull request in 3f0cd93. |
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.
Summary:
To save a checkpoint of a Feature Importance job, we will need to persistent below fields in https://www.internalfb.com/code/fbsource/[5e3ebcd68e0c]/fbcode/model_understanding/feature_importance/summarizer.py?lines=89-99 in order save the state of the job:
This diff implements (de)serialization of Summarizer.
future diffs include:
Differential Revision: D58227186