Conversation
The init method now allows specifying several common options such as - bandpassing (crop spec) - mel scale - rescaling - changing lower db limit - spectrogram parameters - image resizing I added documentation of how to customize preprocessing as examples in SpectrogramPreprocessor.save_onnx() docstring
…t, one value per audio file Fixes #1071
…fig()) Fixes #1171 also add wandb tests with pytest.skip option for both SpectrogramClassifier and LightningSpectrogramClassifier
Collaborator
Author
|
I checked that onnx model outputs match pytorch model up to small differences. I didn't add a test for this as it would require adding onnx and onnxruntime to the development environment - this might be worth doing in the future? |
Collaborator
Author
|
needs tests for ONNX model export |
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Only review the diff after we merge #1174 , since this branch has
spec_refactorbranch as the baseImplements TorchSpectrogramPreprocessor and save CNNs to onnx