-
Notifications
You must be signed in to change notification settings - Fork 147
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
Testing STFT/Magnitude against STFTTflite/MagnitudeTflite #142
Comments
That used a DC signal above. If I use a more realistic signal, the difference is significantly reduced. I'm not sure if there's anything to be added to attempt to close the gap. |
hi. there have been some discussion on it. in general, it's possible to make them close enough with parameter tuned, though it may not be apparent. one thing: better to compare it without the decibel scaling, as it exaggerates the numerical difference under https://groups.google.com/a/ismir.net/g/community/c/LiVRv4I7asw/m/H6Ag-MxGAQAJ https://colab.research.google.com/drive/1ptS1UkpHa-dW8w7WEf8xTE63mEQg8NQZ |
I'm testing STFT to Magnitude layers and comparing them against the TFlite versions and get failures for the same set of arguments. Am I doing something incorrectly here or is this expected for some reason?
The text was updated successfully, but these errors were encountered: