PyTorch implementation for Score-Based Generative Modeling through Stochastic Differential Equations (ICLR 2021, Oral)

Overview

Score-Based Generative Modeling through Stochastic Differential Equations

PWC

This repo contains a PyTorch implementation for the paper Score-Based Generative Modeling through Stochastic Differential Equations

by Yang Song, Jascha Sohl-Dickstein, Diederik P. Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole


We propose a unified framework that generalizes and improves previous work on score-based generative models through the lens of stochastic differential equations (SDEs). In particular, we can transform data to a simple noise distribution with a continuous-time stochastic process described by an SDE. This SDE can be reversed for sample generation if we know the score of the marginal distributions at each intermediate time step, which can be estimated with score matching. The basic idea is captured in the figure below:

schematic

Our work enables a better understanding of existing approaches, new sampling algorithms, exact likelihood computation, uniquely identifiable encoding, latent code manipulation, and brings new conditional generation abilities (including but not limited to class-conditional generation, inpainting and colorization) to the family of score-based generative models.

All combined, we achieved an FID of 2.20 and an Inception score of 9.89 for unconditional generation on CIFAR-10, as well as high-fidelity generation of 1024px Celeba-HQ images (samples below). In addition, we obtained a likelihood value of 2.99 bits/dim on uniformly dequantized CIFAR-10 images.

FFHQ samples

What does this code do?

Aside from the NCSN++ and DDPM++ models in our paper, this codebase also re-implements many previous score-based models in one place, including NCSN from Generative Modeling by Estimating Gradients of the Data Distribution, NCSNv2 from Improved Techniques for Training Score-Based Generative Models, and DDPM from Denoising Diffusion Probabilistic Models.

It supports training new models, evaluating the sample quality and likelihoods of existing models. We carefully designed the code to be modular and easily extensible to new SDEs, predictors, or correctors.

JAX version

Please find a JAX implementation here, which additionally supports class-conditional generation with a pre-trained classifier, and resuming an evalution process after pre-emption.

JAX vs. PyTorch

In general, this PyTorch version consumes less memory but runs slower than JAX. Here is a benchmark on training an NCSN++ cont. model with VE SDE. Hardware is 4x Nvidia Tesla V100 GPUs (32GB)

Framework Time (second per step) Memory usage in total (GB)
PyTorch 0.56 20.6
JAX (n_jitted_steps=1) 0.30 29.7
JAX (n_jitted_steps=5) 0.20 74.8

How to run the code

Dependencies

Run the following to install a subset of necessary python packages for our code

pip install -r requirements.txt

Stats files for quantitative evaluation

We provide the stats file for CIFAR-10. You can download cifar10_stats.npz and save it to assets/stats/. Check out #5 on how to compute this stats file for new datasets.

Usage

Train and evaluate our models through main.py.

main.py:
  --config: Training configuration.
    (default: 'None')
  --eval_folder: The folder name for storing evaluation results
    (default: 'eval')
  --mode: <train|eval>: Running mode: train or eval
  --workdir: Working directory
  • config is the path to the config file. Our prescribed config files are provided in configs/. They are formatted according to ml_collections and should be quite self-explanatory.

    Naming conventions of config files: the path of a config file is a combination of the following dimensions:

    • dataset: One of cifar10, celeba, celebahq, celebahq_256, ffhq_256, celebahq, ffhq.
    • model: One of ncsn, ncsnv2, ncsnpp, ddpm, ddpmpp.
    • continuous: train the model with continuously sampled time steps.
  • workdir is the path that stores all artifacts of one experiment, like checkpoints, samples, and evaluation results.

  • eval_folder is the name of a subfolder in workdir that stores all artifacts of the evaluation process, like meta checkpoints for pre-emption prevention, image samples, and numpy dumps of quantitative results.

  • mode is either "train" or "eval". When set to "train", it starts the training of a new model, or resumes the training of an old model if its meta-checkpoints (for resuming running after pre-emption in a cloud environment) exist in workdir/checkpoints-meta . When set to "eval", it can do an arbitrary combination of the following

    • Evaluate the loss function on the test / validation dataset.

    • Generate a fixed number of samples and compute its Inception score, FID, or KID. Prior to evaluation, stats files must have already been downloaded/computed and stored in assets/stats.

    • Compute the log-likelihood on the training or test dataset.

    These functionalities can be configured through config files, or more conveniently, through the command-line support of the ml_collections package. For example, to generate samples and evaluate sample quality, supply the --config.eval.enable_sampling flag; to compute log-likelihoods, supply the --config.eval.enable_bpd flag, and specify --config.eval.dataset=train/test to indicate whether to compute the likelihoods on the training or test dataset.

How to extend the code

  • New SDEs: inherent the sde_lib.SDE abstract class and implement all abstract methods. The discretize() method is optional and the default is Euler-Maruyama discretization. Existing sampling methods and likelihood computation will automatically work for this new SDE.
  • New predictors: inherent the sampling.Predictor abstract class, implement the update_fn abstract method, and register its name with @register_predictor. The new predictor can be directly used in sampling.get_pc_sampler for Predictor-Corrector sampling, and all other controllable generation methods in controllable_generation.py.
  • New correctors: inherent the sampling.Corrector abstract class, implement the update_fn abstract method, and register its name with @register_corrector. The new corrector can be directly used in sampling.get_pc_sampler, and all other controllable generation methods in controllable_generation.py.

Pretrained checkpoints

All checkpoints are provided in this Google drive.

Instructions: You may find two checkpoints for some models. The first checkpoint (with a smaller number) is the one that we reported FID scores in our paper's Table 3 (also corresponding to the FID and IS columns in the table below). The second checkpoint (with a larger number) is the one that we reported likelihood values and FIDs of black-box ODE samplers in our paper's Table 2 (also FID(ODE) and NNL (bits/dim) columns in the table below). The former corresponds to the smallest FID during the course of training (every 50k iterations). The later is the last checkpoint during training.

Per Google's policy, we cannot release our original CelebA and CelebA-HQ checkpoints. That said, I have re-trained models on FFHQ 1024px, FFHQ 256px and CelebA-HQ 256px with personal resources, and they achieved similar performance to our internal checkpoints.

Here is a detailed list of checkpoints and their results reported in the paper. FID (ODE) corresponds to the sample quality of black-box ODE solver applied to the probability flow ODE.

Checkpoint path FID IS FID (ODE) NNL (bits/dim)
ve/cifar10_ncsnpp/ 2.45 9.73 - -
ve/cifar10_ncsnpp_continuous/ 2.38 9.83 - -
ve/cifar10_ncsnpp_deep_continuous/ 2.20 9.89 - -
vp/cifar10_ddpm/ 3.24 - 3.37 3.28
vp/cifar10_ddpm_continuous - - 3.69 3.21
vp/cifar10_ddpmpp 2.78 9.64 - -
vp/cifar10_ddpmpp_continuous 2.55 9.58 3.93 3.16
vp/cifar10_ddpmpp_deep_continuous 2.41 9.68 3.08 3.13
subvp/cifar10_ddpm_continuous - - 3.56 3.05
subvp/cifar10_ddpmpp_continuous 2.61 9.56 3.16 3.02
subvp/cifar10_ddpmpp_deep_continuous 2.41 9.57 2.92 2.99
Checkpoint path Samples
ve/bedroom_ncsnpp_continuous bedroom_samples
ve/church_ncsnpp_continuous church_samples
ve/ffhq_1024_ncsnpp_continuous ffhq_1024
ve/ffhq_256_ncsnpp_continuous ffhq_256_samples
ve/celebahq_256_ncsnpp_continuous celebahq_256_samples

Demonstrations and tutorials

Link Description
Open In Colab Load our pretrained checkpoints and play with sampling, likelihood computation, and controllable synthesis (JAX + FLAX)
Open In Colab Load our pretrained checkpoints and play with sampling, likelihood computation, and controllable synthesis (PyTorch)
Open In Colab Tutorial of score-based generative models in JAX + FLAX
Open In Colab Tutorial of score-based generative models in PyTorch

Tips

  • When using the JAX codebase, you can jit multiple training steps together to improve training speed at the cost of more memory usage. This can be set via config.training.n_jitted_steps. For CIFAR-10, we recommend using config.training.n_jitted_steps=5 when your GPU/TPU has sufficient memory; otherwise we recommend using config.training.n_jitted_steps=1. Our current implementation requires config.training.log_freq to be dividable by n_jitted_steps for logging and checkpointing to work normally.
  • The snr (signal-to-noise ratio) parameter of LangevinCorrector somewhat behaves like a temperature parameter. Larger snr typically results in smoother samples, while smaller snr gives more diverse but lower quality samples. Typical values of snr is 0.05 - 0.2, and it requires tuning to strike the sweet spot.
  • For VE SDEs, we recommend choosing config.model.sigma_max to be the maximum pairwise distance between data samples in the training dataset.

References

If you find the code useful for your research, please consider citing

@inproceedings{
  song2021scorebased,
  title={Score-Based Generative Modeling through Stochastic Differential Equations},
  author={Yang Song and Jascha Sohl-Dickstein and Diederik P Kingma and Abhishek Kumar and Stefano Ermon and Ben Poole},
  booktitle={International Conference on Learning Representations},
  year={2021},
  url={https://openreview.net/forum?id=PxTIG12RRHS}
}

This work is built upon some previous papers which might also interest you:

  • Song, Yang, and Stefano Ermon. "Generative Modeling by Estimating Gradients of the Data Distribution." Proceedings of the 33rd Annual Conference on Neural Information Processing Systems. 2019.
  • Song, Yang, and Stefano Ermon. "Improved techniques for training score-based generative models." Proceedings of the 34th Annual Conference on Neural Information Processing Systems. 2020.
  • Ho, Jonathan, Ajay Jain, and Pieter Abbeel. "Denoising diffusion probabilistic models." Proceedings of the 34th Annual Conference on Neural Information Processing Systems. 2020.
Owner
Yang Song
PhD Candidate in Stanford AI Lab
Yang Song
Users can free try their models on SIDD dataset based on this code

SIDD benchmark 1 Train python train.py If you want to train your network, just modify the yaml in the options folder. 2 Validation python validation.p

Yuzhi ZHAO 2 May 20, 2022
Self-Supervised Collision Handling via Generative 3D Garment Models for Virtual Try-On

Self-Supervised Collision Handling via Generative 3D Garment Models for Virtual Try-On [Project website] [Dataset] [Video] Abstract We propose a new g

71 Dec 24, 2022
Object tracking implemented with YOLOv4, DeepSort, and TensorFlow.

Object tracking implemented with YOLOv4, DeepSort, and TensorFlow. YOLOv4 is a state of the art algorithm that uses deep convolutional neural networks to perform object detections. We can take the ou

The AI Guy 1.1k Dec 29, 2022
Replication Code for "Self-Supervised Bug Detection and Repair" NeurIPS 2021

Self-Supervised Bug Detection and Repair This is the reference code to replicate the research in Self-Supervised Bug Detection and Repair in NeurIPS 2

Microsoft 85 Dec 24, 2022
Py-FEAT: Python Facial Expression Analysis Toolbox

Py-FEAT is a suite for facial expressions (FEX) research written in Python. This package includes tools to detect faces, extract emotional facial expressions (e.g., happiness, sadness, anger), facial

Computational Social Affective Neuroscience Laboratory 147 Jan 06, 2023
Implementation of Kalman Filter in Python

Kalman Filter in Python This is a basic example of how Kalman filter works in Python. I do plan on refactoring and expanding this repo in the future.

Enoch Kan 35 Sep 11, 2022
SelfRemaster: SSL Speech Restoration

SelfRemaster: Self-Supervised Speech Restoration Official implementation of SelfRemaster: Self-Supervised Speech Restoration with Analysis-by-Synthesi

Takaaki Saeki 46 Jan 07, 2023
A curated list and survey of awesome Vision Transformers.

English | 简体中文 A curated list and survey of awesome Vision Transformers. You can use mind mapping software to open the mind mapping source file. You c

OpenMMLab 281 Dec 21, 2022
Second-Order Neural ODE Optimizer, NeurIPS 2021 spotlight

Second-order Neural ODE Optimizer (NeurIPS 2021 Spotlight) [arXiv] ✔️ faster convergence in wall-clock time | ✔️ O(1) memory cost | ✔️ better test-tim

Guan-Horng Liu 39 Oct 22, 2022
Tutorial to set up TensorFlow Object Detection API on the Raspberry Pi

A tutorial showing how to set up TensorFlow's Object Detection API on the Raspberry Pi

Evan 1.1k Dec 26, 2022
MultiTaskLearning - Multi Task Learning for 3D segmentation

Multi Task Learning for 3D segmentation Perception stack of an Autonomous Drivin

2 Sep 22, 2022
Omnidirectional Scene Text Detection with Sequential-free Box Discretization (IJCAI 2019). Including competition model, online demo, etc.

Box_Discretization_Network This repository is built on the pytorch [maskrcnn_benchmark]. The method is the foundation of our ReCTs-competition method

Yuliang Liu 266 Nov 24, 2022
A collection of implementations of deep domain adaptation algorithms

Deep Transfer Learning on PyTorch This is a PyTorch library for deep transfer learning. We divide the code into two aspects: Single-source Unsupervise

Yongchun Zhu 647 Jan 03, 2023
Official MegEngine implementation of CREStereo(CVPR 2022 Oral).

[CVPR 2022] Practical Stereo Matching via Cascaded Recurrent Network with Adaptive Correlation This repository contains MegEngine implementation of ou

MEGVII Research 309 Dec 30, 2022
Official implementation of the paper Momentum Capsule Networks (MoCapsNet)

Momentum Capsule Network Official implementation of the paper Momentum Capsule Networks (MoCapsNet). Abstract Capsule networks are a class of neural n

8 Oct 20, 2022
The official implementation of paper Siamese Transformer Pyramid Networks for Real-Time UAV Tracking, accepted by WACV22

SiamTPN Introduction This is the official implementation of the SiamTPN (WACV2022). The tracker intergrates pyramid feature network and transformer in

Robotics and Intelligent Systems Control @ NYUAD 29 Jan 08, 2023
Mahadi-Now - This Is Pakistani Just Now Login Tools

PAKISTANI JUST NOW LOGIN TOOLS Install apt update apt upgrade apt install python

MAHADI HASAN AFRIDI 19 Apr 06, 2022
[CVPR 2021] Involution: Inverting the Inherence of Convolution for Visual Recognition, a brand new neural operator

involution Official implementation of a neural operator as described in Involution: Inverting the Inherence of Convolution for Visual Recognition (CVP

Duo Li 1.3k Dec 28, 2022
[CVPR 2022] Pytorch implementation of "Templates for 3D Object Pose Estimation Revisited: Generalization to New objects and Robustness to Occlusions" paper

template-pose Pytorch implementation of "Templates for 3D Object Pose Estimation Revisited: Generalization to New objects and Robustness to Occlusions

Van Nguyen Nguyen 92 Dec 28, 2022
Yet Another Reinforcement Learning Tutorial

This repo contains self-contained RL implementations

Sungjoon 65 Dec 10, 2022