# NVIDIA/tacotron2

Tacotron 2 - PyTorch implementation with faster-than-realtime inference

Repository: https://github.com/NVIDIA/tacotron2
Canonical: https://ross.abutalabs.com/products/tacotron2
Language: Jupyter Notebook
License: BSD-3-Clause
License Family: permissive
Last push: 2024-06-12T18:49:15+00:00

## Health v2 (maintenance only)
Score: 32/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3044, "days_push": 812, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 5296, forks 1410 (observed 2026-08-28T04:09:15.241513+00:00)

## What it is
NVIDIA's PyTorch implementation of the Tacotron 2 text-to-speech model, which synthesizes mel spectrograms from text for vocoder-based audio generation. It supports distributed training, automatic mixed precision, and faster-than-realtime inference when paired with vocoders like WaveGlow.

## Use cases
- convert text to speech audio in Python
- train a custom TTS model on my own dataset
- generate speech from text with a pretrained model
- fine-tune Tacotron 2 on LJSpeech
- run faster-than-realtime neural speech synthesis on GPU
- build a voice synthesis pipeline with PyTorch

## When to choose
- you need a well-tested PyTorch Tacotron 2 implementation with pretrained checkpoints
- you want GPU-accelerated training with mixed precision and multi-GPU support
- you are doing research or experimentation with neural TTS on NVIDIA hardware

## When to avoid
- you need the latest state-of-the-art TTS models (e.g., VITS, modern diffusion-based TTS)
- you have no NVIDIA GPU, since CUDA cuDNN is required
- you want a production-ready TTS service with an API out of the box

## Facets
- artifact type: library
- maturity: maintenance
- function: tts, deep-learning, machine-learning
- domain: speech-processing, deep-learning, artificial-intelligence
- platform: python, windows
- tags: pytorch, tacotron2, speech-synthesis, ljspeech, nvidia, text-to-speech, waveglow, gpu, linux

## Member repositories
- NVIDIA/tacotron2 (main) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:15.241513+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:59:09.583267+00:00, confidence not recorded.
  - readme: https://github.com/NVIDIA/tacotron2 (fetched 2026-08-28T04:09:15.241513+00:00, sha 9ab68102a8d2)
- Data as of 2026-08-30T08:39:29.467469+00:00.