app.py

import gradio as gr
from pyannote.audio import Pipeline

def diarization(audio):
    pipeline = Pipeline.from_pretrained("pyannote/speaker-segmentation")
    #parameters = {"onset":0.7,"offset":0.3,"min_duration_on":0.0,"min_duration_off":0.0}
    #pipeline.instantiate(parameters)
    output = pipeline(audio)
    answer = ""
    for turn, _, speaker in output.itertracks(yield_label=True):
        answer += "{} said something starting from {:.2f} and ends on {:.2f}\n".format(speaker,turn.start,turn.end)
    return answer

app = gr.Interface(fn=diarization,
                inputs=gr.inputs.Audio(source="upload", type="filepath", label="audio"),
                outputs="text")
app.launch()