https://github.com/ashispati/MusicPerfAssessment/blob/31dc045785177189be51d082cb3d9938e2ab49f1/src/models/SpectralCRNN.py#L27 The output size of this layer should be 1, not 11.