Language : 🇺🇸 | 🇨🇳
An unofficial PyTorch implementation of VALL-E(Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers).
We can train the VALL-E model on one GPU.
Demo
Broader impacts
Since VALL-E could synthesize speech that maintains speaker identity, it may carry potential risks in misuse of the model, such as spoofing voice identification or impersonating a specific speaker.
To avoid abuse, Well-trained models and services will not be provided.
Install Deps
To get up and running quickly just follow the steps below:
# PyTorch
pip install torch==1.13.1 torchaudio==0.13.1 --extra-index-url https://download.pytorch.org/whl/cu116
pip install torchmetrics==0.11.1
# fbank
pip install librosa==0.8.1
# phonemizer pypinyin
apt-get install espeak-ng
## OSX: brew install espeak
pip install phonemizer==3.2.1 pypinyin==0.48.0
# lhotse update to newest version
# https://github.com/lhotse-speech/lhotse/pull/956
# https://github.com/lhotse-speech/lhotse/pull/960
pip uninstall lhotse
pip uninstall lhotse
pip install git+https://github.com/lhotse-speech/lhotse
# k2
# find the right version in https://huggingface.co/csukuangfj/k2
pip install https://huggingface.co/csukuangfj/k2/resolve/main/cuda/k2-1.23.4.dev20230224+cuda11.6.torch1.13.1-cp310-cp310-linux_x86_64.whl
# icefall
git clone https://github.com/k2-fsa/icefall
cd icefall
pip install -r requirements.txt
export PYTHONPATH=`pwd`/../icefall:$PYTHONPATH
echo "export PYTHONPATH=`pwd`/../icefall:\$PYTHONPATH" >> ~/.zshrc
echo "export PYTHONPATH=`pwd`/../icefall:\$PYTHONPATH" >> ~/.bashrc
cd -
source ~/.zshrc
# valle
git clone https://github.com/lifeiteng/valle.git
cd valle
pip install -e .
Training&Inference
-
examples/libritts/README.md
English example -
examples/aishell1/README.md
Chinese example -
Paper Chapter 5.1 "The average length of the waveform in LibriLight is 60 seconds. During training, we randomly crop the waveform to a random length between 10 seconds and 20 seconds. For the NAR acoustic prompt tokens, we select a random segment waveform of 3 seconds from the same utterance." Prefix Mode 0 1 2 4 for NAR Decoder
- 0: no acoustic prompt tokens
- 1: random prefix of current batched utterances (This is recommended)
- 2: random segment of current batched utterances
- 4: same as the paper (As they randomly crop the long waveform to multiple utterances, so the same utterance means pre or post utterance in the same long waveform.)
# If train NAR Decoders with prefix_mode 4 python3 bin/trainer.py --prefix_mode 4 --dataset libritts --input-strategy PromptedPrecomputedFeatures ...
LibriTTS demo Trained on one GPU with 24G memory
cd examples/libritts
# step1 prepare dataset
bash prepare.sh --stage -1 --stop-stage 3
# step2 train the model on one GPU with 24GB memory
exp_dir=exp/valle
## Train AR model
python3 bin/trainer.py --max-duration 80 --filter-min-duration 0.5 --filter-max-duration 14 --train-stage 1 \
--num-buckets 6 --dtype "bfloat16" --save-every-n 10000 --valid-interval 20000 \
--model-name valle --share-embedding true --norm-first true --add-prenet false \
--decoder-dim 1024 --nhead 16 --num-decoder-layers 12 --prefix-mode 1 \
--base-lr 0.05 --warmup-steps 200 --average-period 0 \
--num-epochs 20 --start-epoch 1 --start-batch 0 --accumulate-grad-steps 4 \
--exp-dir ${exp_dir}
## Train NAR model
cp ${exp_dir}/best-valid-loss.pt ${exp_dir}/epoch-2.pt # --start-epoch 3=2+1
python3 bin/trainer.py --max-duration 40 --filter-min-duration 0.5 --filter-max-duration 14 --train-stage 2 \
--num-buckets 6 --dtype "float32" --save-every-n 10000 --valid-interval 20000 \
--model-name valle --share-embedding true --norm-first true --add-prenet false \
--decoder-dim 1024 --nhead 16 --num-decoder-layers 12 --prefix-mode 1 \
--base-lr 0.05 --warmup-steps 200 --average-period 0 \
--num-epochs 40 --start-epoch 3 --start-batch 0 --accumulate-grad-steps 4 \
--exp-dir ${exp_dir}
# step3 inference
python3 bin/infer.py --output-dir infer/demos \
--checkpoint=${exp_dir}/best-valid-loss.pt \
--text-prompts "KNOT one point one five miles per hour." \
--audio-prompts ./prompts/8463_294825_000043_000000.wav \
--text "To get up and running quickly just follow the steps below." \
# Demo Inference
https://github.com/lifeiteng/lifeiteng.github.com/blob/main/valle/run.sh#L68
Troubleshooting
- SummaryWriter segmentation fault (core dumped)
- LINE
tb_writer = SummaryWriter(log_dir=f"{params.exp_dir}/tensorboard")
- FIX https://github.com/tensorflow/tensorboard/pull/6135/files
file=`python -c 'import site; print(f"{site.getsitepackages()[0]}/tensorboard/summary/writer/event_file_writer.py")'` sed -i 's/import tf/import tensorflow_stub as tf/g' $file
- LINE
Training on a custom dataset?
- prepare the dataset to
lhotse manifests
- There are plenty of references here lhotse/recipes
python3 bin/tokenizer.py ...
python3 bin/trainer.py ...
Contributing
Citing
To cite this repository:
@misc{valle,
author={Feiteng Li},
title={VALL-E: A neural codec language model},
year={2023},
url={http://github.com/lifeiteng/vall-e}
}
@article{VALL-E,
title = {Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers},
author = {Chengyi Wang, Sanyuan Chen, Yu Wu,
Ziqiang Zhang, Long Zhou, Shujie Liu,
Zhuo Chen, Yanqing Liu, Huaming Wang,
Jinyu Li, Lei He, Sheng Zhao, Furu Wei},
year = {2023},
eprint = {2301.02111},
archivePrefix = {arXiv},
volume = {abs/2301.02111},
url = {http://arxiv.org/abs/2301.02111},
}
Sponsor this project 赞助该项目
As an independent developer, I am not employed by any company, which means my projects are freely available and open. However, I do need to pay for GPU cloud services to acquire the computational resources required for my experiments. Therefore, I need to raise funds to sustain these operational expenses. By securing sufficient funding, sponsors will receive comprehensive technical support.
When I gather adequate funding, sponsors will receive full technical support. This means that as a sponsor, you will enjoy unique benefits and privileges, including but not limited to:
-
Priority Support: As a sponsor, your questions and needs will be prioritized and addressed promptly. I will do my best to provide personalized technical support and solutions tailored to your requirements.
-
Custom Development: As a sponsor, you can request specific features and functionalities, and I will strive to meet your personalized needs by customizing the development accordingly.
-
Early Access: As a sponsor, you will have early access to my latest projects and features, allowing you to stay informed about my latest advancements.
-
Sponsor-exclusive Perks: As a sponsor, you will have access to additional special benefits, such as participation in sponsor-exclusive discussion forums, conferences, or seminars, as well as potential collaboration opportunities.
I firmly believe that with your generous sponsorship, I will be able to continue conducting innovative research and projects while providing outstanding technical support and personalized services to my esteemed sponsors. Thank you very much for your support!
作为一名自由独立的开发者,我并没有与任何公司签约,这意味着我的项目是自由开放的。然而,为了购买实验所需的GPU算力资源,我需要支付云服务商不菲的费用。为了确保项目的顺利运行,我需要筹集资金来支持这些开支。
当我筹集到足够的资金时,赞助者们将获得充分的技术支持。这意味着您作为赞助者将享受到独特的待遇和特权,包括但不限于:
-
优先支持:作为赞助者,您的问题和需求将优先考虑和解决。我将尽我所能为您提供个性化的技术支持和解决方案。
-
定制开发:作为赞助者,您可以提出特定的需求和功能要求,我将致力于满足您的个性化需求,并根据您的要求定制开发。
-
早期访问:作为赞助者,您将优先获得我最新项目和功能的早期访问权限,您将第一时间了解到我的最新进展。
-
赞助者专属福利:作为赞助者,您将享受到额外的特殊福利,例如赞助者专属的讨论论坛、会议或研讨会的参与权,以及可能的合作机会等。
我深信,您的慷慨赞助将使我能够继续开展创新的研究和项目,并为广大赞助者提供出色的技术支持和个性化服务。 非常感谢您对我的支持!