app.py

# import gradio as gr
# from googletrans import Translator
# import torch
# # Initialize Translator
# from transformers import pipeline
# translator = Translator()
# MODEL_NAME = "openai/whisper-base"
# device = 0 if torch.cuda.is_available() else "cpu"
# pipe = pipeline(
#     task="automatic-speech-recognition",
#     model=MODEL_NAME,
#     chunk_length_s=30,
#     device=device,
# )


# def transcribe_audio(audio):
#    text = pipe(audio)["text"]
#    return text
#     # return translated_text

# audio_record = gr.inputs.Audio(source='microphone', label='Record Audio')
# output_text = gr.outputs.Textbox(label='Transcription')

# interface = gr.Interface(fn=transcribe_audio, inputs=audio_record, outputs=output_text)
# interface.launch()

import gradio as gr
from transformers import pipeline

modelo = pipeline("automatic-speech-recognition", model="openai/whisper-base")

def transcribe(audio):
    text = modelo(audio)["text"]
    return text

# Criar a interface Gradio
gr.Interface(
    fn=transcribe,
    inputs=[gr.Audio(source="microphone", type="filepath")],
    outputs=["textbox"]
).launch(share=True)  # Adicionar o parâmetro share=True para criar um link público