This includes the new models needed for the Tensorflow 2 models.
They are named like openl3_audio_linear_music-v0_4_0.h5.gz following the format of the v0.2.0 files. I added a script tests/package_weights.py that essentially inverts the gzip decompression code from setup.py.
The only difference between these weights and the previous version is that the spectrogram layers were deleted from it (using tests/migration/remove_layers.py)
This includes the new models needed for the Tensorflow 2 models.
They are named like
openl3_audio_linear_music-v0_4_0.h5.gz
following the format of thev0.2.0
files. I added a script tests/package_weights.py that essentially inverts the gzip decompression code fromsetup.py
.The only difference between these weights and the previous version is that the spectrogram layers were deleted from it (using tests/migration/remove_layers.py)