NVIDIA/NeMo-text-processing

NeMo text processing for ASR and TTS

What it solves

It provides a way to handle text normalization and inverse text normalization, which is essential for converting text into a format suitable for speech synthesis or processing.

How it works

The package uses a combination of techniques, including Weighted Finite-State Transducers (WFST) for grammar customization and a hybrid approach that can integrate with PyTorch for advanced normalization.

Who it’s for

Developers and researchers working on NLP and speech-related AI, specifically those needing to prepare text for downstream tasks like text-to-speech (TTS) systems.

Highlights

  • Text normalization: Converting written text (e.g., "$10") into a spoken form (e.g., "ten dollars").
  • Inverse text normalization: Converting spoken forms back into written forms.
  • Support for grammar customization via WFST.
  • Hybrid normalization options using PyTorch.

Related

  • Project
    wenet-e2e/WeTextProcessing

    A production-ready text processing toolkit for text normalization and inverse text normalization, supporting multiple languages for speech-related AI applications.

  • Project
    nateshmbhat/pyttsx3

    A Python library for offline text-to-speech conversion that utilizes native system engines to generate voice output without an internet connection.

  • Project
    stephenhky/PyShortTextCategorization

    A Python package for supervised and unsupervised short text categorization, providing tools for text representation, topic modeling, and neural network classification.

  • Project
    VOICEVOX/voicevox_engine

    An HTTP server engine for the VOICEVOX text-to-speech system that enables high-quality speech and singing synthesis via a REST API.

  • Project
    explosion/spaCy

    An industrial-strength NLP library for Python and Cython that provides pretrained pipelines and neural network models for advanced text processing across 70+ languages.