Fabrice-TIERCELIN commited on
Commit
fa7808c
1 Parent(s): 559b00c
Files changed (1) hide show
  1. app.py +4 -10
app.py CHANGED
@@ -1,8 +1,6 @@
1
  import gradio as gr
2
  import json
3
  import torch
4
- import wavio
5
- import spaces
6
 
7
  from tqdm import tqdm
8
  from huggingface_hub import snapshot_download
@@ -81,15 +79,10 @@ tango = Tango(device = "cpu")
81
  tango.vae.to(device_type)
82
  tango.stft.to(device_type)
83
  tango.model.to(device_type)
84
-
85
- @spaces.GPU(duration = 60)
86
  def gradio_generate(prompt, steps, guidance):
87
  output_wave = tango.generate(prompt, steps, guidance)
88
- # output_filename = f"{prompt.replace(' ', '_')}_{steps}_{guidance}"[:250] + ".wav"
89
- output_filename = "temp.wav"
90
- wavio.write(output_filename, output_wave, rate = 16000, sampwidth = 2)
91
-
92
- return output_filename
93
 
94
  description_text = """
95
  <p style="text-align: center;">
@@ -111,10 +104,11 @@ description_text = """
111
  """
112
  # Gradio input and output components
113
  input_text = gr.Textbox(label = "Prompt", value = "Snort of a horse", lines = 2, autofocus = True)
114
- output_audio = gr.Audio(label = "Generated Audio", type = "filepath")
115
  denoising_steps = gr.Slider(label = "Steps", minimum = 100, maximum = 200, value = 100, step = 1, interactive = True)
116
  guidance_scale = gr.Slider(label = "Guidance Scale", minimum = 1, maximum = 10, value = 3, step = 0.1, interactive = True)
117
 
 
 
118
  # Gradio interface
119
  gr_interface = gr.Interface(
120
  fn = gradio_generate,
 
1
  import gradio as gr
2
  import json
3
  import torch
 
 
4
 
5
  from tqdm import tqdm
6
  from huggingface_hub import snapshot_download
 
79
  tango.vae.to(device_type)
80
  tango.stft.to(device_type)
81
  tango.model.to(device_type)
82
+
 
83
  def gradio_generate(prompt, steps, guidance):
84
  output_wave = tango.generate(prompt, steps, guidance)
85
+ return gr.make_waveform((16000, output_wave))
 
 
 
 
86
 
87
  description_text = """
88
  <p style="text-align: center;">
 
104
  """
105
  # Gradio input and output components
106
  input_text = gr.Textbox(label = "Prompt", value = "Snort of a horse", lines = 2, autofocus = True)
 
107
  denoising_steps = gr.Slider(label = "Steps", minimum = 100, maximum = 200, value = 100, step = 1, interactive = True)
108
  guidance_scale = gr.Slider(label = "Guidance Scale", minimum = 1, maximum = 10, value = 3, step = 0.1, interactive = True)
109
 
110
+ output_audio = gr.Audio(label = "Generated Audio")
111
+
112
  # Gradio interface
113
  gr_interface = gr.Interface(
114
  fn = gradio_generate,