auspicious3000 / autovc

AutoVC: Zero-Shot Voice Style Transfer with Only Autoencoder Loss

https://arxiv.org/abs/1905.05879

MIT License

987 stars 205 forks source link

AttributeError: 'numpy.ndarray' object has no attribute 'numpy' #1

Closed ak9250 closed 5 years ago

ak9250 commented 5 years ago

AttributeError Traceback (most recent call last)

in () 15 c = spect[1] 16 print(name) ---> 17 waveform = wavegen(model, c=c) 18 librosa.output.write_wav(name+'.wav', waveform, sr=16000) /content/autovc/synthesis.py in wavegen(model, c, tqdm) 46 47 """ ---> 48 c = c.numpy() 49 50 model.eval() AttributeError: 'numpy.ndarray' object has no attribute 'numpy'

auspicious3000 commented 5 years ago

Error fixed

ak9250 commented 5 years ago

thanks i also added

g_checkpoint = torch.load('autovc.ckpt')

g_checkpoint = torch.load('/content/autovc/autovc.ckpt',map_location='cuda:0') in conversion.ipynb line 18 which was giving a error also how can i test this with my own source and target wav files?

auspicious3000 commented 5 years ago

As authors, we certainly appreciate your concerns, and we thank you for your interest in our method. On the other hand, as researchers from IBM, while we are committed to ensuring that our results can be verified and reproduced, we also must balance those considerations against potential misuse of the technology, per our company's policy. We believe that withholding a small fraction of code is a workable compromise that allows the community to nonetheless easily understand and evaluate the method. If you'd like to discuss further, please contact us at yang.zhang2@ibm.com.