while wavenet training, 0 derived errors ignored.

[ghost]

Hi guys.

I'm trying to train wavenet model, but got following error.

Generated 64 train batches of size 8 in 0.393 sec Exiting due to exception: 2 root error(s) found. (0) Invalid argument: assertion failed: [] [Condition x == y did not hold element-wise:] [x (WaveNet_model/inference/strided_slice_5:0) = ] [15125] [y (WaveNet_model/inference/strided_slice_6:0) = ] [11000] [[node WaveNet_model/inference/assert_equal/Assert/Assert (defined at /Tacotron-2/wavenet_vocoder/models/wavenet.py:699) ]] [[WaveNet_model/clip_by_value_147/_2571]] (1) Invalid argument: assertion failed: [] [Condition x == y did not hold element-wise:] [x (WaveNet_model/inference/strided_slice_5:0) = ] [15125] [y (WaveNet_model/inference/strided_slice_6:0) = ] [11000] [[node WaveNet_model/inference/assert_equal/Assert/Assert (defined at /Tacotron-2/wavenet_vocoder/models/wavenet.py:699) ]] 0 successful operations. 0 derived errors ignored.

Machine spec: Ubuntu 18.04 GPU: Tesla T4 Python 3.6.9 Cuda 10.0, Cudnn 7.6.5 Tensorflow=gpu 1.14

Thanksful any help for this.