Closed meetps closed 8 years ago
The CRNN takes a 96*1366 spectrogram image as input. However the weights (for input batch normalization layer) provided for Tensorflow for CRNN have the shape of the
96*1366
bn_0_freq_beta_1 (1440,) bn_0_freq_gamma_1 (1440,) bn_0_freq_running_mean_1 (1440,) bn_0_freq_running_std_1 (1440,)
which is incompatible with the input image size.
It is compatible because there is zero padding.
The CRNN takes a
96*1366
spectrogram image as input. However the weights (for input batch normalization layer) provided for Tensorflow for CRNN have the shape of thewhich is incompatible with the input image size.