text-generation-webui/extensions/whisper_stt/script.py
EliasVincent 7a03d0bda3 cleanup
2023-03-09 20:33:00 +01:00

37 lines
961 B
Python

import gradio as gr
import speech_recognition as sr
input_hijack = {
'state': False,
'value': ["", ""]
}
def input_modifier(string):
return string
def do_stt():
transcription = ""
r = sr.Recognizer()
with sr.Microphone() as source:
r.adjust_for_ambient_noise(source)
audio = r.listen(source)
try:
transcription = r.recognize_whisper(audio, language="english", model="tiny.en")
except sr.UnknownValueError:
print("Whisper could not understand audio")
except sr.RequestError as e:
print("Could not request results from Whisper", e)
input_hijack.update({"state": True, "value": [transcription, transcription]})
return transcription
def ui():
speech_button = gr.Button(value="🎙️")
output_transcription = gr.Textbox(label="STT-Preview", placeholder="Speech Preview. Click \"Generate\" to send")
speech_button.click(do_stt, outputs=[output_transcription])