Installing and Running pre-trained DeepSpeech Model

Installing DeepSpeech and executing a sample audio file on the Mozilla’s pre-trained deepspeech model in Ubuntu.

1. Setup python environment.
2. Install virtualenv package.
   virtual environment is a tool to create isolated python environments.
3. Download the DeepSpeech github repository
   $ git clone https://github.com/mozilla/DeepSpeech
4. Download the pre-trained model
   $ wget -O — https://github.com/mozilla/DeepSpeech/releases/download/v0.3.0-models.tar.gz | tar xvfz -
5. Create the Virtual Environment
   $ virtualenv -p python3 $HOME/tmp/deepspeech-venv/
6. Activate the virtual environment
   $ source $HOME/tmp/deepspeech-venv/bin/activate
7. Install DeepSpeech python binding
   $ pip3 install deepspeech
8. Within the DeepSpeech directory create an audio file to test
   $ arecord my_audio_file.wav , record the audio and ctrl+c
9. Install Sox for processing the audio files
   $ sudo apt-get install sox
10. Within the DeepSpeech directory, calling and executing the deepspeech model
    $ deepspeech --model models/output_graph.pbmm --alphabet models/alphabt.txt --lm models/lm.binary --trie models/trie --audio my_audio_file.wav
11. The inferred text is displayed.