deepspeech-arm

Its a deepspeech docker container for arm.

Installation

Install Raspberry Pi OS [Debian version: 10 (buster), 21.02.2023] on your Raspberry Pi.

It must be an armv7! Check the version with uname -a. Raspberry Pi OS 11 is already on armv8!

Install git:

sudo apt install git wget -y

git clone https://github.com/raci0n/deepspeech-arm.git 

cd deepspeech-arm

Get your models:

wget -O ./model/deepspeech.scorer https://github.com/mozilla/DeepSpeech/releases/download/v0.9.3/deepspeech-0.9.3-models.scorer
wget -O ./model/deepspeech.tflite https://github.com/mozilla/DeepSpeech/releases/download/v0.9.3/deepspeech-0.9.3-models.tflite

Create the docker container:

sudo chmod +x ./install.sh
./install.sh

Start the docker container:

sudo chmod +x ./start.sh

The Docker container is now available via http://localhost:8082.

Example

import requests

url = 'http://localhost:8082/stt'
fname = 'file.wav'

# Open the wav-file 
with open(fname, 'rb') as f:
    data = f.read() # read

# http request to 'http://localhost:8082/stt'
req = requests.post(url, data=data)
print(req.text) # print the result