MaykaGR commited on
Commit
1e3f5f7
verified
1 Parent(s): 2e5161a

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +20 -13
app.py CHANGED
@@ -22,28 +22,35 @@ with gr.Blocks(theme=gr.themes.Ocean(primary_hue="pink", neutral_hue="indigo", f
22
  button = gr.Button("Describir", variant="primary")
23
  clear = gr.Button("Borrar")
24
  output = gr.Textbox(label="Resumen")
 
 
 
 
25
 
26
  def describir(url):
27
  raw_image = Image.open(requests.get(url, stream=True).raw).convert('RGB')
28
  inputs = processor(raw_image, return_tensors="pt").to("cpu")
29
  out = model.generate(**inputs)
30
- models, cfg, task = load_model_ensemble_and_task_from_hf_hub(
31
- "facebook/fastspeech2-en-ljspeech",
32
- arg_overrides={"vocoder": "hifigan", "fp16": False}
33
- )
34
- ble = models[0]
35
- TTSHubInterface.update_cfg_with_data_cfg(cfg, task.data_cfg)
36
- generator = task.build_generator(ble, cfg)
37
-
38
- text = processor.decode(out[0], skip_special_tokens=True)
 
39
 
40
- sample = TTSHubInterface.get_model_input(task, text)
41
- wav, rate = TTSHubInterface.get_prediction(task, ble, generator, sample)
42
- ipd.Audio(wav, rate=rate)
43
 
44
- return processor.decode(out[0], skip_special_tokens=True)
 
 
 
45
 
46
 
47
  button.click(describir, [textbox], output)
 
48
 
49
  demo.launch(debug=True)
 
22
  button = gr.Button("Describir", variant="primary")
23
  clear = gr.Button("Borrar")
24
  output = gr.Textbox(label="Resumen")
25
+ with gr.Row():
26
+ button2 = gr.Button("Leer", variant="primary")
27
+ clear = gr.Button("Borrar")
28
+ output2 = gr.Audio(label="Audio")
29
 
30
  def describir(url):
31
  raw_image = Image.open(requests.get(url, stream=True).raw).convert('RGB')
32
  inputs = processor(raw_image, return_tensors="pt").to("cpu")
33
  out = model.generate(**inputs)
34
+ return processor.decode(out[0], skip_special_tokens=True)
35
+
36
+ def leer(texto):
37
+ models, cfg, task = load_model_ensemble_and_task_from_hf_hub(
38
+ "facebook/fastspeech2-en-ljspeech",
39
+ arg_overrides={"vocoder": "hifigan", "fp16": False}
40
+ )
41
+ model = models[0]
42
+ TTSHubInterface.update_cfg_with_data_cfg(cfg, task.data_cfg)
43
+ generator = task.build_generator(model, cfg)
44
 
45
+ text = texto
 
 
46
 
47
+ sample = TTSHubInterface.get_model_input(task, text)
48
+ wav, rate = TTSHubInterface.get_prediction(task, model, generator, sample)
49
+ ipd.Audio(wav, rate=rate)
50
+ return None
51
 
52
 
53
  button.click(describir, [textbox], output)
54
+ button2.click(leer, [output], output2)
55
 
56
  demo.launch(debug=True)