spring-media/TransformerTTS

Health v2 · maintenance only

Flags: archived no_license

Adoption not part of the score

1161 stars · 223 forks observed · 2026-08-28

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

A TensorFlow 2 implementation of a non-autoregressive Transformer-based neural network for text-to-speech synthesis, based on FastSpeech and FastPitch papers. It includes pre-trained LJSpeech models compatible with vocoders like MelGAN and HiFiGAN, with controllable speed and pitch.

Use cases

• synthesize speech from text with a transformer model
• generate fast non-autoregressive TTS audio
• control speed and pitch of generated speech
• train a FastSpeech-style TTS model on custom datasets
• convert text to mel spectrograms for vocoders like MelGAN or HiFiGAN
• run text-to-speech inference from the command line or Colab

When to choose

• you need a TensorFlow-based non-autoregressive TTS implementation
• you want fast, robust speech synthesis without autoregressive failure modes
• you need controllable pitch and speed in synthesized speech
• you want to train or fine-tune a FastSpeech/FastPitch-style model

When to avoid

• you need a production-ready TTS service with active development
• you prefer PyTorch-based TTS tooling
• you need the latest state-of-the-art TTS quality
• you require Windows support out of the box

Facets

library · maturity maintenance

tts machine-learning deep-learning speech-processing deep-learning python text-to-speech tensorflow transformer non-autoregressive fastspeech ljspeech speech-synthesis vocoder natural-language-processing linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="spring-media/TransformerTTS")