wavenet 常见错误

使用的wavenet的链接：
链接: wavenet
记录自己踩过的坑。

q1：
tf1.51.1 运行Tactron2中的wavenet，出现以下错误。

Traceback (most recent call last):
  File "train.py", line 138, in <module>
    main()
  File "train.py", line 132, in main
    train(args, log_dir, hparams)
  File "train.py", line 83, in train
    checkpoint = wavenet_train(args, log_dir, hparams, input_path)
  File "/root/Tacotron-2/wavenet_vocoder/train.py", line 346, in wavenet_train
    return train(log_dir, args, hparams, input_path)
  File "/root/Tacotron-2/wavenet_vocoder/train.py", line 230, in train
    model, stats = model_train_mode(args, feeder, hparams, global_step)
  File "/root/Tacotron-2/wavenet_vocoder/train.py", line 173, in model_train_mode
    model = create_model(model_name or args.model, hparams, init)
  File "/root/Tacotron-2/wavenet_vocoder/models/__init__.py", line 12, in create_model
    return WaveNet(hparams, init)
  File "/root/Tacotron-2/wavenet_vocoder/models/wavenet.py", line 109, in __init__
    name='input_convolution')
  File "/root/Tacotron-2/wavenet_vocoder/models/modules.py", line 376, in __init__
    name=name, **kwargs
  File "/root/Tacotron-2/wavenet_vocoder/models/modules.py", line 230, in __init__
    self._track_checkpointable(layer, name='layer')
AttributeError: 'Conv1D1x1' object has no attribute '_track_checkpointable'

a1:
self._track_checkpointable 替换成 self._track_trackable

q2:

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) = ] [42350] [y (WaveNet_model/inference/strided_slice_6:0) = ] [30800]
         [[node WaveNet_model/inference/assert_equal_1/Assert/Assert (defined at /usr/local/lib/python3.6/dist-packages/tensorflow_core/python/framework/ops.py:1748) ]]
         [[WaveNet_model/optimizer_1/gradients/WaveNet_model/inference/ResidualConv1DGLU_11/ResidualConv1DGLU_11/residual_block_causal_conv_ResidualConv1DGLU_11/residual_block_causal_conv_ResidualConv1DGLU
  (1) Invalid argument: assertion failed: [Condition x == y did not hold element-wise:] [x (WaveNet_model/inference/strided_slice_5:0) = ] [42350] [y (WaveNet_model/inference/strided_slice_6:0) = ] [30800]
         [[node WaveNet_model/inference/assert_equal_1/Assert/Assert (defined at /usr/local/lib/python3.6/dist-packages/tensorflow_core/python/framework/ops.py:1748) ]]
0 successful operations.
0 derived errors ignored.

a2:
我用的是16kHz的语料库，hop_size用的是200， tacotron2默认用的是 LJSpeech-1.1语料库，采样率是22050Hz， hop_size用的是275。在训练参数中，下面两个参数需要注意
upsample_type = ‘SubPixel’
upsample_scales = [11, 25], #prod(upsample_scales) should be equal to hop_size
改成[8, 25]就可以。