app.py

import gradio as gr
import nemo.collections.asr as nemo_asr

asr_model = nemo_asr.models.ASRModel.from_pretrained(
    model_name="nvidia/parakeet-rnnt-110m-da-dk"
)


def transcribe_audio(audio):
    transcription = asr_model.transcribe([audio])
    return transcription[0].text if transcription else "No transcription available."


with gr.Blocks() as demo:
    gr.Markdown(
        """# [Parakeet-RNNT-110M-Danish](https://huggingface.co/nvidia/parakeet-rnnt-110m-da-dk) Speech-to-Text Transcription Demo"""
    )
    audio_input = gr.Audio(type="filepath", label="Record or upload your audio")
    transcribe_button = gr.Button("Transcribe")
    transcription_output = gr.Textbox(label="Transcription")

    transcribe_button.click(
        fn=transcribe_audio, inputs=audio_input, outputs=transcription_output
    )

demo.launch()