Implementation of Google Brain's WaveGrad high-fidelity vocoder

Overview

alt-text-1

WaveGrad

Implementation (PyTorch) of Google Brain's high-fidelity WaveGrad vocoder (paper). First implementation on GitHub with high-quality generation for 6-iterations.

Status

  • Documented API.
  • High-fidelity generation.
  • Multi-iteration inference support (stable for low iterations).
  • Stable and fast training with mixed-precision support.
  • Distributed training support.
  • Training also successfully runs on a single 12GB GPU with batch size 96.
  • CLI inference support.
  • Flexible architecture configuration for your own data.
  • Estimated RTF on popular GPU and CPU devices (see below).
  • 100- and lower-iteration inferences are faster than real-time on RTX 2080 Ti. 6-iteration inference is faster than one reported in the paper.
  • Parallel grid search for the best noise schedule.
  • Uploaded generated samples for different number of iterations (see generated_samples folder).
  • Pretrained checkpoint on 22KHz LJSpeech dataset with noise schedules.

Real-time factor (RTF)

Number of parameters: 15.810.401

Model Stable RTX 2080 Ti Tesla K80 Intel Xeon 2.3GHz*
1000 iterations + 9.59 - -
100 iterations + 0.94 5.85 -
50 iterations + 0.45 2.92 -
25 iterations + 0.22 1.45 -
12 iterations + 0.10 0.69 4.55
6 iterations + 0.04 0.33 2.09

*Note: Used an old version of Intel Xeon CPU.


About

WaveGrad is a conditional model for waveform generation through estimating gradients of the data density with WaveNet-similar sampling quality. This vocoder is neither GAN, nor Normalizing Flow, nor classical autoregressive model. The main concept of vocoder is based on Denoising Diffusion Probabilistic Models (DDPM), which utilize Langevin dynamics and score matching frameworks. Furthemore, comparing to classic DDPM, WaveGrad achieves super-fast convergence (6 iterations and probably lower) w.r.t. Langevin dynamics iterative sampling scheme.


Installation

  1. Clone this repo:
git clone https://github.com/ivanvovk/WaveGrad.git
cd WaveGrad
  1. Install requirements:
pip install -r requirements.txt

Training

1 Preparing data

  1. Make train and test filelists of your audio data like ones included into filelists folder.
  2. Make a configuration file* in configs folder.

*Note: if you are going to change hop_length for STFT, then make sure that the product of your upsampling factors in config is equal to your new hop_length.

2 Single and Distributed GPU training

  1. Open runs/train.sh script and specify visible GPU devices and path to your configuration file. If you specify more than one GPU the training will run in distributed mode.
  2. Run sh runs/train.sh

3 Tensorboard and logging

To track your training process run tensorboard by tensorboard --logdir=logs/YOUR_LOGDIR_FOLDER. All logging information and checkpoints will be stored in logs/YOUR_LOGDIR_FOLDER. logdir is specified in config file.

4 Noise schedule grid search

Once model is trained, grid search for the best schedule* for a needed number of iterations in notebooks/inference.ipynb. The code supports parallelism, so you can specify more than one number of jobs to accelerate the search.

*Note: grid search is necessary just for a small number of iterations (like 6 or 7). For larger number just try Fibonacci sequence benchmark.fibonacci(...) initialization: I used it for 25 iteration and it works well. From good 25-iteration schedule, for example, you can build a higher-order schedule by copying elements.

Noise schedules for pretrained model
  • 6-iteration schedule was obtained using grid search. After, based on obtained scheme, by hand, I found a slightly better approximation.
  • 7-iteration schedule was obtained in the same way.
  • 12-iteration schedule was obtained in the same way.
  • 25-iteration schedule was obtained using Fibonacci sequence benchmark.fibonacci(...).
  • 50-iteration schedule was obtained by repeating elements from 25-iteration scheme.
  • 100-iteration schedule was obtained in the same way.
  • 1000-iteration schedule was obtained in the same way.

Inference

CLI

Put your mel-spectrograms in some folder. Make a filelist. Then run this command with your own arguments:

sh runs/inference.sh -c <your-config> -ch <your-checkpoint> -ns <your-noise-schedule> -m <your-mel-filelist> -v "yes"

Jupyter Notebook

More inference details are provided in notebooks/inference.ipynb. There you can also find how to set a noise schedule for the model and make grid search for the best scheme.


Other

Generated audios

Examples of generated audios are provided in generated_samples folder. Quality degradation between 1000-iteration and 6-iteration inferences is not noticeable if found the best schedule for the latter.

Pretrained checkpoints

You can find a pretrained checkpoint file* on LJSpeech (22KHz) via this Google Drive link.

*Note: uploaded checkpoint is a dict with a single key 'model'.


Important details, issues and comments

  • During training WaveGrad uses a default noise schedule with 1000 iterations and linear scale betas from range (1e-6, 0.01). For inference you can set another schedule with less iterations. Tune betas carefully, the output quality really highly depends on it.
  • By default model runs in a mixed-precision way. Batch size is modified compared to the paper (256 -> 96) since authors trained their model on TPU.
  • After ~10k training iterations (1-2 hours) on a single GPU the model performs good generation for 50-iteration inference. Total training time is about 1-2 days (for absolute convergence).
  • At some point training might start to behave weird and crazy (loss explodes), so I have introduced learning rate (LR) scheduling and gradient clipping. If loss explodes for your data, then try to decrease LR scheduler gamma a bit. It should help.
  • By default hop length of your STFT is equal 300 (thus total upsampling factor). Other cases are not tested, but you can try. Remember, that total upsampling factor should be still equal to your new hop length.

History of updates

  • (NEW: 10/24/2020) Huge update. Distributed training and mixed-precision support. More correct positional encoding. CLI support for inference. Parallel grid search. Model size significantly decreased.
  • New RTF info for NVIDIA Tesla K80 GPU card (popular in Google Colab service) and CPU Intel Xeon 2.3GHz.
  • Huge update. New 6-iteration well generated sample example. New noise schedule setting API. Added the best schedule grid search code.
  • Improved training by introducing smarter learning rate scheduler. Obtained high-fidelity synthesis.
  • Stable training and multi-iteration inference. 6-iteration noise scheduling is supported.
  • Stable training and fixed-iteration inference with significant background static noise left. All positional encoding issues are solved.
  • Stable training of 25-, 50- and 1000-fixed-iteration models. Found no linear scaling (C=5000 from paper) of positional encoding (bug).
  • Stable training of 25-, 50- and 1000-fixed-iteration models. Fixed positional encoding downscaling. Parallel segment sampling is replaced by full-mel sampling.
  • (RELEASE, first on GitHub). Parallel segment sampling and broken positional encoding downscaling. Bad quality with clicks from concatenation from parallel-segment generation.

References

Owner
Ivan Vovk
• Mathematics • Machine Learning • Speech technologies
Ivan Vovk
Credit fraud detection in Python using a Jupyter Notebook

Credit-Fraud-Detection - Credit fraud detection in Python using a Jupyter Notebook , using three classification models (Random Forest, Gaussian Naive Bayes, Logistic Regression) from the sklearn libr

Ali Akram 4 Dec 28, 2021
A multi-mode modulator for multi-domain few-shot classification (ICCV)

A multi-mode modulator for multi-domain few-shot classification (ICCV)

Yanbin Liu 8 Apr 28, 2022
Meta-meta-learning with evolution and plasticity

Evolve plastic networks to be able to automatically acquire novel cognitive (meta-learning) tasks

5 Jun 28, 2022
PyKaldi GOP-DNN on Epa-DB

PyKaldi GOP-DNN on Epa-DB This repository has the tools to run a PyKaldi GOP-DNN algorithm on Epa-DB, a database of non-native English speech by Spani

18 Dec 14, 2022
A Closer Look at Structured Pruning for Neural Network Compression

A Closer Look at Structured Pruning for Neural Network Compression Code used to reproduce experiments in https://arxiv.org/abs/1810.04622. To prune, w

Bayesian and Neural Systems Group 140 Dec 05, 2022
Compute descriptors for 3D point cloud registration using a multi scale sparse voxel architecture

MS-SVConv : 3D Point Cloud Registration with Multi-Scale Architecture and Self-supervised Fine-tuning Compute features for 3D point cloud registration

42 Jul 25, 2022
Official implementation of the paper Label-Efficient Semantic Segmentation with Diffusion Models

Label-Efficient Semantic Segmentation with Diffusion Models Official implementation of the paper Label-Efficient Semantic Segmentation with Diffusion

Yandex Research 355 Jan 06, 2023
Wider or Deeper: Revisiting the ResNet Model for Visual Recognition

ademxapp Visual applications by the University of Adelaide In designing our Model A, we did not over-optimize its structure for efficiency unless it w

Zifeng Wu 338 Dec 12, 2022
PyTorch implementation for Convolutional Networks with Adaptive Inference Graphs

Convolutional Networks with Adaptive Inference Graphs (ConvNet-AIG) This repository contains a PyTorch implementation of the paper Convolutional Netwo

Andreas Veit 176 Dec 07, 2022
Auxiliary Raw Net (ARawNet) is a ASVSpoof detection model taking both raw waveform and handcrafted features as inputs, to balance the trade-off between performance and model complexity.

Overview This repository is an implementation of the Auxiliary Raw Net (ARawNet), which is ASVSpoof detection system taking both raw waveform and hand

6 Jul 08, 2022
Get a Grip! - A robotic system for remote clinical environments.

Get a Grip! Within clinical environments, sterilization is an essential procedure for disinfecting surgical and medical instruments. For our engineeri

Jay Sharma 1 Jan 05, 2022
LieTransformer: Equivariant Self-Attention for Lie Groups

LieTransformer This repository contains the implementation of the LieTransformer used for experiments in the paper LieTransformer: Equivariant Self-At

OxCSML (Oxford Computational Statistics and Machine Learning) 50 Dec 28, 2022
The official repo for OC-SORT: Observation-Centric SORT on video Multi-Object Tracking. OC-SORT is simple, online and robust to occlusion/non-linear motion.

OC-SORT Observation-Centric SORT (OC-SORT) is a pure motion-model-based multi-object tracker. It aims to improve tracking robustness in crowded scenes

Jinkun Cao 325 Jan 05, 2023
This is the face keypoint train code of project face-detection-project

face-key-point-pytorch 1. Data structure The structure of landmarks_jpg is like below: |--landmarks_jpg |----AFW |------AFW_134212_1_0.jpg |------AFW_

I‘m X 3 Nov 27, 2022
Versatile Generative Language Model

Versatile Generative Language Model This is the implementation of the paper: Exploring Versatile Generative Language Model Via Parameter-Efficient Tra

Zhaojiang Lin 17 Dec 02, 2022
Code for our ACL 2021 paper - ConSERT: A Contrastive Framework for Self-Supervised Sentence Representation Transfer

ConSERT Code for our ACL 2021 paper - ConSERT: A Contrastive Framework for Self-Supervised Sentence Representation Transfer Requirements torch==1.6.0

Yan Yuanmeng 478 Dec 25, 2022
This demo showcase the use of onnxruntime-rs with a GPU on CUDA 11 to run Bert in a data pipeline with Rust.

Demo BERT ONNX pipeline written in rust This demo showcase the use of onnxruntime-rs with a GPU on CUDA 11 to run Bert in a data pipeline with Rust. R

Xavier Tao 14 Dec 17, 2022
Jupyter Dock is a set of Jupyter Notebooks for performing molecular docking protocols interactively, as well as visualizing, converting file formats and analyzing the results.

Molecular Docking integrated in Jupyter Notebooks Description | Citation | Installation | Examples | Limitations | License Table of content Descriptio

Angel J. Ruiz Moreno 173 Dec 25, 2022
Lane follower: Lane-detector (OpenCV) + Object-detector (YOLO5) + CAN-bus

Lane Follower This code is for the lane follower, including perception and control, as shown below. Environment Hardware Industrial Camera Intel-NUC(1

Siqi Fan 3 Jul 07, 2022
Hierarchical Attentive Recurrent Tracking

Hierarchical Attentive Recurrent Tracking This is an official Tensorflow implementation of single object tracking in videos by using hierarchical atte

Adam Kosiorek 147 Aug 07, 2021