Weight scaling #129
Merged
Weight scaling #129
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
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Outdated suggestions cannot be applied.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.
Related issues
Added feature.
Description
Here a conversion factor (weight scaling) is introduced that scales the weight range of the cross bar in respect to the digital output. For instance during initialization for training, the
weight_scaling_omegafactor can be used to set the absolute max of the weight matrix to some value (e.g. to the maximum of the weight range) and scale the output accordingly. This ensures that tile operates in a proper dynamic regime for a given DNN and usually improves achievable accuracy.Details
See Rasch, Gokmen & Haensch (2019) for algorithmic details.