app.py

import gradio as gr
from transformers import pipeline
from helper import load_image_from_url, render_results_in_image
from helper import summarize_predictions_natural_language

od_pipe = pipeline("object-detection", model="facebook/detr-resnet-50")
tts_pipe = pipeline("text-to-speech",
                    model="kakao-enterprise/vits-ljs")

def get_pipeline_prediction(pil_image):
    
    pipeline_output = od_pipe(pil_image)
    text = summarize_predictions_natural_language(pipeline_output)
    #text = "Hello, my name is Ratha"
    gen_audio = tts_pipe(text)
    processed_image = render_results_in_image(pil_image,
                                            pipeline_output)
    rate= gen_audio["sampling_rate"]
    return processed_image, text, (rate, gen_audio["audio"][0])

demo = gr.Interface(
  fn=get_pipeline_prediction,
  inputs=gr.Image(label="Input image", 
                  type="pil"),
  outputs= [
          gr.Image(label="Output image with predicted instances", type="pil"),
          gr.Textbox(label="Prediction Summary"),
          gr.Audio(label="Generated Speech")]
)

demo.launch()
#text = itt_pipe(input)


#tts_pipe = pipeline("text-to-speech",
#                    model="kakao-enterprise/vits-ljs")


#narrated_text = tts_pipe(tts_pipe[0]['generated_text'])

#def launch(text):
#    out = tts_pipe(text)
#    audio = IPythonAudio(out["audio"][0],
#             rate=out["sampling_rate"])
#    return audio
    
#iface = gr.Interface(launch,
#                     inputs=gr.Image(type='pil'),
#                     outputs="text")

    
#iface.launch()