Speech

A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognitio

What is Speech?

A scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech)

Overview of Speech

What it is

NVIDIA NeMo Speech is a scalable generative AI framework built for researchers and developers working on Large Language Models, Multimodal, and Speech AI (Automatic Speech Recognition and Text-to-Speech). It is designed to help users efficiently create, customize, and deploy new AI models by leveraging existing code and pre-trained model checkpoints. The framework supports Speech models including Automatic Speech Recognition (ASR), Text to Speech (TTS), and Speech LLMs.

Key features

  • Supports Automatic Speech Recognition (ASR), Text to Speech (TTS), and Speech LLMs.
  • Provides pre-trained model checkpoints for various speech tasks.
  • Includes streaming and offline inference capabilities.
  • Supports multiple languages (e.g., 40 languages for Nemotron-3.5-ASR-Streaming-0.6B, 9 languages for MagpieTTS).
  • Built on cache-aware Fastconformer architecture for streaming ASR.
  • Offers full-duplex, natural, interruptible voice chat with low latency (Nemotron 3 VoiceChat in Early Access).
  • Works with Python 3.12+, PyTorch 2.7+, and NVIDIA GPU with CUDA.

Use cases

Researchers and PyTorch developers working on speech models can use NeMo Speech to create, customize, and deploy AI models for automatic speech recognition, text-to-speech, and speech-based large language models. It is suitable for both offline and streaming inference scenarios, and supports multilingual applications.

Getting started

Install from source with uv (recommended): bash git clone https://github.com/NVIDIA-NeMo/Speech.git cd Speech uv sync --extra all --extra cu13 Or use Docker: bash docker pull nvcr.io/nvidia/nemo-speech:26.07.00 docker run --rm -it --gpus all -v $PWD:/workspace nvcr.io/nvidia/nemo-speech:26.07.00 bash Alternatively, install from PyPI with pip: bash pip install 'nemo-toolkit[asr,tts]'

License and pricing

NeMo Speech is licensed under the Apache License 2.0. It is open source.

Based on the project's public README and repository metadata.

Add this badge to your README

Featured on YunoDB Badge styles & help

Maintain this project? Show that it is featured on YunoDB.

Markdown[![Featured on YunoDB](https://yunodb.com/badge/nvidia-nemospeech.svg)](https://yunodb.com/tools/nvidia-nemospeech?utm_source=github&utm_medium=badge)
HTML<a href="https://yunodb.com/tools/nvidia-nemospeech?utm_source=github&utm_medium=badge"><img src="https://yunodb.com/badge/nvidia-nemospeech.svg" alt="Featured on YunoDB" height="20"></a>