Amphion Multi-Speaker TTS Pre-trained Model

Quick Start

We provide the pre-trained checkpoint of VITS, trained on Hi-fi TTS, which consists of a total of 291.6 hours audio contributed by 10 speakers, on an average of 17 hours per speaker. To utilize the pre-trained model, run the following commands:

Step1: Download the checkpoint

git lfs install
git clone https://huggingface.co/amphion/vits_hifitts

Step2: Clone the Amphion's Source Code of GitHub

git clone https://github.com/open-mmlab/Amphion.git

Step3: Specify the checkpoint's path

Use the soft link to specify the downloaded checkpoint in the first step:

cd Amphion
mkdir -p ckpts/tts
ln -s  ../../../vits_hifitts  ckpts/tts/

Step4: Inference

You can follow the inference part of this recipe to generate speech from text. For example, if you want to synthesize a clip of speech with the text of "This is a clip of generated speech with the given text from a TTS model.", just, run:

sh egs/tts/VITS/run.sh --stage 3 --gpu "0" \
    --config ckpts/tts/vits_hifitts/args.json \
    --infer_expt_dir ckpts/tts/vits_hifitts/ \
    --infer_output_dir ckpts/tts/vits_hifitts/result \
    --infer_mode "single" \
    --infer_text "This is a clip of generated speech with the given text from a TTS model." \
    --infer_speaker_name "hifitts_92"

Note: The supported infer_speaker_name values can be seen here.