Implementation of Deep Speech 2 Automatic Speech Recognition model from Scratch

About • Training results • Installation • How To Use • Credits • License

About

This repository contains a implementation of Deep Speech 2 Automatic Speech Recognition model made form scratch with PyTorch. Deep Speech 2 is an end-to-end deep learning model designed for Automatic Speech Recognition (ASR). It uses a combination of convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to process raw audio features (e.g., spectrograms) and predict corresponding text transcriptions. The model is based on Baidu DeepSpeech2 paper and follows this architecture:

Training results

Tables with model's outputs during training and evaluation can be viewed in report. Overall model achieved WER of 0.44 and CER of 0.12 on average during evaluation

Installation

Follow these steps to install the project:

0. (Optional) Create and activate new environment using conda or venv (+pyenv).

   a. conda version:

   # create env
   conda create -n project_env python=PYTHON_VERSION
   
   # activate env
   conda activate project_env

   b. venv (+pyenv) version:

   # create env
   ~/.pyenv/versions/PYTHON_VERSION/bin/python3 -m venv project_env
   
   # alternatively, using default python version
   python3 -m venv project_env
   
   # activate env
   source project_env

1. Install all required packages

   pip install -r requirements.txt

2. Install pre-commit:

   pre-commit install

How To Use

To train a model, run the following command:

python3 train.py -cn=CONFIG_NAME HYDRA_CONFIG_ARGUMENTS

Where CONFIG_NAME is a config from src/configs and HYDRA_CONFIG_ARGUMENTS are optional arguments.

To run inference (evaluate the model or save predictions):

python3 inference.py HYDRA_CONFIG_ARGUMENTS

Credits

This repository is based on a PyTorch Project Template.

License