KeyError: '__getstate__'

[mingri-lt]

When I tried to train my own model, I ran into this error

Traceback (most recent call last):
  File "E:\VocieData\tacotron2\melgan-0.3-alpha\utils\train.py", line 60, in train
    validate(hp, args, model_g, model_d, valloader, writer, step)
  File "E:\VocieData\tacotron2\melgan-0.3-alpha\utils\validation.py", line 13, in validate
    for mel, audio in loader:
  File "D:\Software\Anaconda3\envs\melgan\lib\site-packages\tqdm\std.py", line 1195, in __iter__
    for obj in iterable:
  File "D:\Software\Anaconda3\envs\melgan\lib\site-packages\torch\utils\data\dataloader.py", line 359, in __iter__
    return self._get_iterator()
  File "D:\Software\Anaconda3\envs\melgan\lib\site-packages\torch\utils\data\dataloader.py", line 305, in _get_iterator
    return _MultiProcessingDataLoaderIter(self)
  File "D:\Software\Anaconda3\envs\melgan\lib\site-packages\torch\utils\data\dataloader.py", line 918, in __init__
    w.start()
  File "D:\Software\Anaconda3\envs\melgan\lib\multiprocessing\process.py", line 112, in start
    self._popen = self._Popen(self)
  File "D:\Software\Anaconda3\envs\melgan\lib\multiprocessing\context.py", line 223, in _Popen
    return _default_context.get_context().Process._Popen(process_obj)
  File "D:\Software\Anaconda3\envs\melgan\lib\multiprocessing\context.py", line 322, in _Popen
    return Popen(process_obj)
  File "D:\Software\Anaconda3\envs\melgan\lib\multiprocessing\popen_spawn_win32.py", line 89, in __init__
    reduction.dump(process_obj, to_child)
  File "D:\Software\Anaconda3\envs\melgan\lib\multiprocessing\reduction.py", line 60, in dump
    ForkingPickler(file, protocol).dump(obj)
KeyError: '__getstate__'

Then I made a few changes：
loader = tqdm.tqdm(valloader, desc='Validation loop')
loader = tqdm.tqdm(valloader.dataset, desc='Validation loop')

and then：

Validation loop:   0%|                                                                                  | 0/1000 [00:02<?, ?it/s]
2022-10-06 09:59:39,152 - INFO - Exiting due to exception: Expected 3-dimensional input for 3-dimensional weight [512, 80, 7], but got 2-dimensional input of size [80, 173] instead
Traceback (most recent call last):
  File "E:\VocieData\tacotron2\melgan-0.3-alpha\utils\train.py", line 60, in train
    validate(hp, args, model_g, model_d, valloader, writer, step)
  File "E:\VocieData\tacotron2\melgan-0.3-alpha\utils\validation.py", line 18, in validate
    fake_audio = generator(mel)
  File "D:\Software\Anaconda3\envs\melgan\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
    return forward_call(*input, **kwargs)
  File "E:\VocieData\tacotron2\melgan-0.3-alpha\model\generator.py", line 48, in forward
    return self.generator(mel)
  File "D:\Software\Anaconda3\envs\melgan\lib\site-packages\torch\nn\modules\module.py", line 1102, in _call_impl
    return forward_call(*input, **kwargs)
  File "D:\Software\Anaconda3\envs\melgan\lib\site-packages\torch\nn\modules\container.py", line 141, in forward
    input = module(input)
  File "D:\Software\Anaconda3\envs\melgan\lib\site-packages\torch\nn\modules\module.py", line 1120, in _call_impl
    result = forward_call(*input, **kwargs)
  File "D:\Software\Anaconda3\envs\melgan\lib\site-packages\torch\nn\modules\conv.py", line 301, in forward
    return self._conv_forward(input, self.weight, self.bias)
  File "D:\Software\Anaconda3\envs\melgan\lib\site-packages\torch\nn\modules\conv.py", line 298, in _conv_forward
    self.padding, self.dilation, self.groups)
RuntimeError: Expected 3-dimensional input for 3-dimensional weight [512, 80, 7], but got 2-dimensional input of size [80, 173] instead

How do I fix it please?

[mingri-lt]

I have fix this problem and use biaobei tts dataset to train success.

I have made some changes to the code in order to smooth out the training.

in validation.py :

def validate(hp, args, generator, discriminator, valloader, writer, step):
    ... ...
    loader = tqdm.tqdm(valloader.dataset, desc='Validation loop')
    ... ...
    
    
    for mel, audio in loader:
        mel = mel.reshape(1, mel.size(0), mel.size(1))
        audio = audio.reshape(1, 1, audio.size(1))
        ... ...

in train.py :

            ... ...
            loader = tqdm.tqdm(trainloader.dataset, desc='Loading train data')
            
            for (melG, audioG), (melD, audioD) in loader:
                ... ...
                melG = melG.reshape(1, melG.size(0), melG.size(1))
                audioG = audioG.reshape(1, audioG.size(0), audioG.size(1))
                melD = melD.reshape(1, melD.size(0), melD.size(1))
                audioD = audioD.reshape(1, audioD.size(0), audioD.size(1))
                ... ...