VoiceCraft: Zero-Shot Speech Editing and Text-to-Speech in the Wild

TL;DR

VoiceCraft is a token infilling neural codec language model, that achieves state-of-the-art performance on both speech editing and zero-shot text-to-speech (TTS) on in-the-wild data including audiobooks, internet videos, and podcasts.

To clone or edit an unseen voice, VoiceCraft needs only a few seconds of reference.

News

:star: 03/28/2024: Model weights are up on HuggingFace🤗 here!

TODO

Codebase upload
Environment setup
Inference demo for speech editing and TTS
Training guidance
RealEdit dataset and training manifest
Model weights (both 330M and 830M, the former seems to be just as good)
Write colab notebooks for better hands-on experience
HuggingFace Spaces demo
Better guidance on training/finetuning

How to run TTS inference

There are two ways:

with docker. see quickstart
without docker. see envrionment setup

When you are inside the docker image or you have installed all dependencies, Checkout inference_tts.ipynb.

If you want to do model development such as training/finetuning, I recommend following envrionment setup and training.

QuickStart

:star: To try out TTS inference with VoiceCraft, the best way is using docker. Thank @ubergarm and @jayc88 for making this happen.

Tested on Linux and Windows and should work with any host with docker installed.

# 1. clone the repo on in a directory on a drive with plenty of free space
git clone git@github.com:jasonppy/VoiceCraft.git
cd VoiceCraft

# 2. assumes you have docker installed with nvidia container container-toolkit (windows has this built into the driver)
# https://docs.nvidia.com/datacenter/cloud-native/container-toolkit/1.13.5/install-guide.html
# sudo apt-get install -y nvidia-container-toolkit-base || yay -Syu nvidia-container-toolkit || echo etc...

# 3. Try to start an existing container otherwise create a new one passing in all GPUs
./start-jupyter.sh  # linux
start-jupyter.bat   # windows

# 4. now open a webpage on the host box to the URL shown at the bottom of:
docker logs jupyter

# 5. optionally look inside from another terminal
docker exec -it jupyter /bin/bash
export USER=(your_linux_username_used_above)
export HOME=/home/$USER
sudo apt-get update

# 6. confirm video card(s) are visible inside container
nvidia-smi

# 7. Now in browser, open inference_tts.ipynb and work through one cell at a time
echo GOOD LUCK

Environment setup

conda create -n voicecraft python=3.9.16
conda activate voicecraft

pip install -e git+https://github.com/facebookresearch/audiocraft.git@c5157b5bf14bf83449c17ea1eeb66c19fb4bc7f0#egg=audiocraft
pip install xformers==0.0.22
pip install torchaudio==2.0.2 torch==2.0.1 # this assumes your system is compatible with CUDA 11.7, otherwise checkout https://pytorch.org/get-started/previous-versions/#v201
apt-get install ffmpeg # if you don't already have ffmpeg installed
apt-get install espeak-ng # backend for the phonemizer installed below
pip install tensorboard==2.16.2
pip install phonemizer==3.2.1
pip install datasets==2.16.0
pip install torchmetrics==0.11.1
# install MFA for getting forced-alignment, this could take a few minutes
conda install -c conda-forge montreal-forced-aligner=2.2.17 openfst=1.8.2 kaldi=5.5.1068
# conda install pocl # above gives an warning for installing pocl, not sure if really need this

# to run ipynb
conda install -n voicecraft ipykernel --no-deps --force-reinstall

If you have encountered version issues when running things, checkout environment.yml for exact matching.

Inference Examples

Checkout inference_speech_editing.ipynb and inference_tts.ipynb

Training

To train an VoiceCraft model, you need to prepare the following parts:

utterances and their transcripts
encode the utterances into codes using e.g. Encodec
convert transcripts into phoneme sequence, and a phoneme set (we named it vocab.txt)
manifest (i.e. metadata)

Step 1,2,3 are handled in ./data/phonemize_encodec_encode_hf.py, where

Gigaspeech is downloaded through HuggingFace. Note that you need to sign an agreement in order to download the dataset (it needs your auth token)
phoneme sequence and encodec codes are also extracted using the script.

An example run:

conda activate voicecraft
export CUDA_VISIBLE_DEVICES=0
cd ./data
python phonemize_encodec_encode_hf.py \
--dataset_size xs \
--download_to path/to/store_huggingface_downloads \
--save_dir path/to/store_extracted_codes_and_phonemes \
--encodec_model_path path/to/encodec_model \
--mega_batch_size 120 \
--batch_size 32 \
--max_len 30000

where encodec_model_path is avaliable here. This model is trained on Gigaspeech XL, it has 56M parameters, 4 codebooks, each codebook has 2048 codes. Details are described in our paper. If you encounter OOM during extraction, try decrease the batch_size and/or max_len. The extracted codes, phonemes, and vocab.txt will be stored at path/to/store_extracted_codes_and_phonemes/${dataset_size}/{encodec_16khz_4codebooks,phonemes,vocab.txt}.

As for manifest, please download train.txt and validation.txt from here, and put them under path/to/store_extracted_codes_and_phonemes/manifest/. Please also download vocab.txt from here if you want to use our pretrained VoiceCraft model (so that the phoneme-to-token matching is the same).

Now, you are good to start training!

conda activate voicecraft
cd ./z_scripts
bash e830M.sh

License

The codebase is under CC BY-NC-SA 4.0 (LICENSE-CODE), and the model weights are under Coqui Public Model License 1.0.0 (LICENSE-MODEL). Note that we use some of the code from other repository that are under different licenses: ./models/codebooks_patterns.py is under MIT license; ./models/modules, ./steps/optim.py, data/tokenizer.py are under Apache License, Version 2.0; the phonemizer we used is under GNU 3.0 License.

Acknowledgement

We thank Feiteng for his VALL-E reproduction, and we thank audiocraft team for open-sourcing encodec.

Citation

@article{peng2024voicecraft,
  author    = {Peng, Puyuan and Huang, Po-Yao and Li, Daniel and Mohamed, Abdelrahman and Harwath, David},
  title     = {VoiceCraft: Zero-Shot Speech Editing and Text-to-Speech in the Wild},
  journal   = {arXiv},
  year      = {2024},
}

Disclaimer

Any organization or individual is prohibited from using any technology mentioned in this paper to generate or edit someone's speech without his/her consent, including but not limited to government leaders, political figures, and celebrities. If you do not comply with this item, you could be in violation of copyright laws.