New project: deepspeech websocket server & client

lissyx · January 10, 2020, 11:08am

Read data/lm/, it’s already documented.

utunga · January 13, 2020, 2:33am

Yes, for what it’s worth i have the streaming API working with our custom (te reo Māori) model right now. I’ll also just mention, in case it’s helpful, that working with the pyAudio library was driving me a bit bonkers so I re-purposed @daanzu’s great pyAudio based client/server websocket demo by adding a ‘sounddevice’ based version, which is here in case its of any use to someone else (“device id” hell was driving me a bit bonkers on ubuntu desktop)

github.com

utunga/deepspeech-websocket-server/blob/master/sound_device_audio.py

import argparse
import math
import shutil
import numpy as np
import pyaudio
import threading, collections, queue, os, os.path
import samplerate as sr
import webrtcvad
import sys
from halo import Halo
import sounddevice as sd
import logging
from lomond import WebSocket, events

logger = logging.getLogger(__name__)
logging.basicConfig(level=30,
    format="%(asctime)s.%(msecs)03d: %(name)s: %(levelname)s: %(funcName)s(): %(message)s",
    datefmt="%Y-%m-%d %p %I:%M:%S",
    )
logging.getLogger('lomond').setLevel(30)

This file has been truncated. show original

deep_learning · January 17, 2020, 12:39pm

Thank you @lissyx . Finally, I succeeded to create a language model.

Hassaan_Saleem · September 16, 2020, 4:12pm

i am getting same error how did you solve it?