app.py

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


asr_model = nemo_asr.models.ASRModel.from_pretrained(model_name="QuartzNet15x5Base-En")


def transcribe(audio):
   
    transcription = asr_model.transcribe([audio])
    return transcription[0]


iface = gr.Interface(
    fn=transcribe,
    inputs=gr.inputs.Audio(source="microphone", type="filepath"),
    outputs="text",
    title="ASR with NeMo",
    description="Speak into your microphone and get a transcription using NVIDIA NeMo QuartzNet ASR."
)


iface.launch()