how to obtain models/waveglow_256channels_v4.pt for the Inference demo? about flowtron HOT 7 CLOSED

It seems that tecotron2 repo is using old waveglow.

cd tacotron2/waveglow
git checkout master

And then do the inference without conversion (just original waveglow_256channels_v4.pt you posted).

from flowtron.

Please pull from master and try again with the converted waveglow checkpoint. We updated the tacotron2 repo in this repo.

from flowtron.

Use the script below to convert the model https://github.com/NVIDIA/waveglow/blob/master/convert_model.py

from flowtron.

@rafaelvalle This conversion executed successfully, but the same results on the inference demo:

python convert_model.py ..\flowtron\models\waveglow_256channels_v4.pt ..\flowtron\models\waveglow_256channels_v4_new.pt
F:\Users\Erik.DESKTOP-E5E1V83\anaconda3\lib\site-packages\torch\serialization.py:657: SourceChangeWarning: source code of class 'torch.nn.modules.conv.ConvTranspose1d' has changed. you can retrieve the original source code by accessing the object's source attribute or set torch.nn.Module.dump_patches = True and use the patch tool to revert the changes.
warnings.warn(msg, SourceChangeWarning)
F:\Users\Erik.DESKTOP-E5E1V83\anaconda3\lib\site-packages\torch\serialization.py:657: SourceChangeWarning: source code of class 'torch.nn.modules.container.ModuleList' has changed. you can retrieve the original source code by accessing the object's source attribute or set torch.nn.Module.dump_patches = True and use the patch tool to revert the changes.
warnings.warn(msg, SourceChangeWarning)
f class 'torch.nn.modules.conv.Conv1d' has changed. you can retrieve the original source code by accessing the object's warnings.warn(msg, SourceChangeWarning)

Loaded checkpoint 'models/flowtron_ljs.pt')
Number of speakers : 1
Traceback (most recent call last):
File "inference.py", line 122, in
args.n_frames, args.sigma, args.seed)
File "inference.py", line 80, in infer
audio = waveglow.infer(mels.half(), sigma=0.8).float()
File "tacotron2/waveglow\glow.py", line 276, in infer
output = self.WN[k]((audio_0, spect))
File "F:\Users\Erik.DESKTOP-E5E1V83\anaconda3\lib\site-packages\torch\nn\modules\module.py", line 550, in call
result = self.forward(*input, **kwargs)
File "tacotron2/waveglow\glow.py", line 161, in forward
self.cond_layersi,
File "F:\Users\Erik.DESKTOP-E5E1V83\anaconda3\lib\site-packages\torch\nn\modules\module.py", line 594, in getattr
type(self).name, name))
AttributeError: 'WN' object has no attribute 'cond_layers'

from flowtron.

Convert and set the waveglow path to the converted model

from flowtron.

yes, did try,

my experience is similar to this other issue. i tried v2 and it does execute

NVIDIA/tacotron2#333 (comment)

from flowtron.

this is resolved now

from flowtron.