Neural LCS(Phoneme/Word)

Overview

This repository contains a project that involves training and inference for phoneme and word-level alignment models. The project is structured to handle both phoneme-level and word-level data, with separate directories and models for each level.

Project Structure

• phn_lcs/: Phoneme-level alignment directory.
  • data_phn/: Contains phoneme-level data.
  • model/: You should create a folder named "model" here, and you can download the pretrained phoneme aligner model weight file in the link .
  • inference.ipynb: Jupyter notebook for phoneme-level inference.
  • train.py: Script for training phoneme alignment models.
• requirements.txt: List of required Python dependencies.
• simulation/: Contains scripts for phoneme and word-level simulation.
  • generator_phn.ipynb: Jupyter notebook for generating phoneme-level data.
  • generator_word.ipynb: Jupyter notebook for generating word-level data.

How to Use

Phoneme-Level Inference

1. Navigate to the phn_lcs/ directory.
2. Open the inference.ipynb Jupyter notebook.
3. Follow the steps outlined in the notebook to perform phoneme-level alignment and inference.

We are updating the word level Neural LCS model, which will be released soon

Training a Model

To train a new model for phoneme alignment:

1. Run the train.py script.
2. The script will use the data in data_phn/ to train the model.

Requirements

Install the required dependencies by running:

pip install -r requirements.txt

License

This project is licensed under the MIT License.