Spaces:
Sleeping
Sleeping
File size: 1,780 Bytes
8a0a2d0 3fb48f1 8a0a2d0 3fb48f1 ce200fa 8a0a2d0 e8aa27e 3fb48f1 8a0a2d0 3fb48f1 8a0a2d0 1dceb84 8a0a2d0 3fb48f1 8a0a2d0 3fb48f1 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 |
from transformers import pipeline, WhisperModel
import gradio as gr
import pandas as pd
import string
pipe = pipeline(model="matteocirca/whisper-small-it-2",return_timestamps="word")
current_audio = None
segments = {}
def audio2segments(audio,word):
global segments,current_audio
if audio != current_audio or current_audio == None:
segments = pipe(audio)
current_audio = audio
if not word:
if current_audio != None:
return segments["text"],"<html><h1>No Word inserted!</h1></html>"
else:
return "","<html><h1>No Word inserted!</h1></html>"
df = pd.DataFrame(columns=["Occurrence n","Starting TimeStamp","Ending TimeStamp"])
if word:
ranges_list = []
ranges = []
print(segments)
for w in segments['chunks']:
if word == w["text"].translate(str.maketrans('', '', string.punctuation)).replace(" ","").lower() :
ranges_list.append(w["timestamp"])
res = "<table><thead><tr><th>Occurrence n°</th><th>Start</th><th>End</th></tr></thead><tbody>"
for i,r in enumerate(ranges_list):
# ranges_list.append({"Occurrence n":i,"Starting TimeStamp":r[0],"Ending TimeStamp":r[1]})
res += f"<tr><td>{i}</td><td>{r[0]}</td><td>{r[1]}</td></tr>"
res+=" </tbody></table>"
print(res)
return segments["text"],res
def clear():
segments = {}
iface = gr.Interface(
fn=audio2segments,
inputs=[gr.Audio(sources=["upload","microphone"], type="filepath"),"text"],
outputs=["text","html"],
title="Whisper Small Italian",
description="Realtime demo for Italian speech recognition using a fine-tuned Whisper small model.",
)
iface.launch()
|