Grad-TTS

A fork of the official implementation of the Grad-TTS model based on Diffusion Probabilistic Modelling. This fork cleans up the code to focus on easy installation and inference.

Installation

Inference

See grad_tts_cli.py for how to use the model for inference.

python grad_tts_cli.py \
    --file /PATH/TO/TEXT_FILE \
    --checkpoint /PATH/TO/GRAD_TTS_CHECKPOINT \
    --hifigan_checkpoint /PATH/TO/HIFIGAN_CHECKPOINT \
    --outdir /PATH/TO/OUTPUT_DIR

You can download Grad-TTS and HiFi-GAN checkpoints trained on LJSpeech.

References

• HiFi-GAN model is used as vocoder, official github repository: link.
• Monotonic Alignment Search algorithm is used for unsupervised duration modelling, official github repository: link.
• Phonemization utilizes CMUdict, official github repository: link.