akhaliq HF staff commited on
Commit
4e45c79
1 Parent(s): bcb66b7

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +5 -2
app.py CHANGED
@@ -1,6 +1,7 @@
1
  import torch
2
  import torchaudio
3
  import gradio as gr
 
4
 
5
  device="cpu"
6
  bundle = torchaudio.pipelines.TACOTRON2_WAVERNN_PHONE_LJSPEECH
@@ -36,15 +37,17 @@ def inference(text):
36
  spec, _, _ = tacotron2.infer(processed, lengths)
37
 
38
  plt.imshow(spec[0].cpu().detach())
 
 
39
 
40
  with torch.no_grad():
41
  waveforms = waveglow.infer(spec)
42
 
43
  torchaudio.save("output_waveglow.wav", waveforms[0:1].cpu(), sample_rate=22050)
44
- return "output_waveglow.wav",plt
45
 
46
  title="TACOTRON 2"
47
  description="Gradio demo for TACOTRON 2: The Tacotron 2 model for generating mel spectrograms from text. To use it, simply add you text or click on one of the examples to load them. Read more at the links below."
48
  article = "<p style='text-align: center'><a href='https://arxiv.org/abs/1712.05884' target='_blank'>Natural TTS Synthesis by Conditioning WaveNet on Mel Spectrogram Predictions</a> | <a href='https://github.com/NVIDIA/DeepLearningExamples/tree/master/PyTorch/SpeechSynthesis/Tacotron2' target='_blank'>Github Repo</a></p>"
49
  examples=[["life is like a box of chocolates"]]
50
- gr.Interface(inference,"text",[gr.outputs.Audio(type="file"),gr.outputs.Image(type="plot",label="Spectrogram")],title=title,description=description,article=article,examples=examples).launch(enable_queue=True)
 
1
  import torch
2
  import torchaudio
3
  import gradio as gr
4
+ import matplotlib as plt
5
 
6
  device="cpu"
7
  bundle = torchaudio.pipelines.TACOTRON2_WAVERNN_PHONE_LJSPEECH
 
37
  spec, _, _ = tacotron2.infer(processed, lengths)
38
 
39
  plt.imshow(spec[0].cpu().detach())
40
+ plt.axis('off')
41
+ plt.savefig("test.png", bbox_inches='tight')
42
 
43
  with torch.no_grad():
44
  waveforms = waveglow.infer(spec)
45
 
46
  torchaudio.save("output_waveglow.wav", waveforms[0:1].cpu(), sample_rate=22050)
47
+ return "output_waveglow.wav","test.png"
48
 
49
  title="TACOTRON 2"
50
  description="Gradio demo for TACOTRON 2: The Tacotron 2 model for generating mel spectrograms from text. To use it, simply add you text or click on one of the examples to load them. Read more at the links below."
51
  article = "<p style='text-align: center'><a href='https://arxiv.org/abs/1712.05884' target='_blank'>Natural TTS Synthesis by Conditioning WaveNet on Mel Spectrogram Predictions</a> | <a href='https://github.com/NVIDIA/DeepLearningExamples/tree/master/PyTorch/SpeechSynthesis/Tacotron2' target='_blank'>Github Repo</a></p>"
52
  examples=[["life is like a box of chocolates"]]
53
+ gr.Interface(inference,"text",[gr.outputs.Audio(type="file"),gr.outputs.Image(type="file",label="Spectrogram")],title=title,description=description,article=article,examples=examples).launch(enable_queue=True)