Some modification on my side

[bobosui]

1. deepvoice3_pytorch/init.py
   from .version import version
   this line has error
   version.py is not provided.

2. deepvoice3_pytorch/builder.py
   deepvoice3_multispeaker
   inconsistent with hparams.py

3. deepvoice3_pytorch/deepvoice3.py
   line 474, (done>0.5).all()
   maybe done.data is better

[r9y9]

Thank you for your feedback. Comments follow:

deepvoice3_pytorch/init.py
from .version import version
this line has error
version.py is not provided.

version.py is generated during installation process (python setup.py develop). See

deepvoice3_pytorch/setup.py

Lines 31 to 36 in d990533

	def create_version_file():
	global version, cwd
	print('-- Building version ' + version)
	version_path = os.path.join(cwd, 'deepvoice3_pytorch', 'version.py')
	with open(version_path, 'w') as f:
	f.write("__version__ = '{}'\n".format(version))

deepvoice3_pytorch/builder.py
deepvoice3_multispeaker
inconsistent with hparams.py

Do you mean deepvoice3_vctk is better than deepvoice3_multispeaker? I choose deepvoice3_multispeaker for the builder name (not preset name) in case we can use multi-speaker dataset other than VCTK.

deepvoice3_pytorch/deepvoice3.py
line 474, (done>0.5).all()
maybe done.data is better

I'm not sure if it makes any difference. Do you think it's better?

[bobosui]

Thanks for the quick feedback.
2. deepvoice3_pytorch/builder.py
It is not a big issue for me, I can change the hyper-parameters from "builder=deepvoice3_vctk" to "builder=deepvoice3_multispeaker" to fix it.

3. (done>0.5).all()
   I got an error message here. I guess this is probably because I am using a lower version of pytorch, since is defined with Variable (which has no all() function in my pytorch version).

Thank you very much！