Rayhane-mamah/Tacotron-2

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2322 stars · 898 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

A TensorFlow implementation of DeepMind's Tacotron-2 neural text-to-speech architecture, including both the Tacotron spectrogram predictor and WaveNet vocoder. It supports training on LJSpeech and M-AILABS-style datasets and provides pretrained models and audio samples.

Use cases

• train a Tacotron-2 text-to-speech model on LJSpeech
• synthesize speech from text with a WaveNet vocoder
• reproduce the results of the Tacotron-2 paper in TensorFlow
• fine-tune a TTS model on M-AILABS speech data
• generate mel spectrograms and audio from a trained checkpoint

When to choose

• you want a faithful, well-starred TensorFlow reproduction of the Tacotron-2 paper
• your dataset is LJSpeech or M-AILABS format
• you need both Tacotron and WaveNet training pipelines in one repo

When to avoid

• you need a production-ready or actively maintained TTS system
• you prefer PyTorch or modern TTS frameworks
• your dataset is not LJSpeech-like and you don't want to write custom preprocessing

Facets

library · maturity maintenance

tts machine-learning deep-learning speech-processing deep-learning machine-learning python cross-platform tacotron-2 wavenet tensorflow speech-synthesis text-to-speech research-implementation gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="Rayhane-mamah/Tacotron-2")