Annotated, understandable, and visually interpretable PyTorch implementations of: VAE, BIRVAE, NSGAN, MMGAN, WGAN, WGANGP, LSGAN, DRAGAN, BEGAN, RaGAN, InfoGAN, fGAN, FisherGAN

Overview

Overview

PyTorch 0.4.1 | Python 3.6.5

Annotated implementations with comparative introductions for minimax, non-saturating, wasserstein, wasserstein gradient penalty, least squares, deep regret analytic, bounded equilibrium, relativistic, f-divergence, Fisher, and information generative adversarial networks (GANs), and standard, variational, and bounded information rate variational autoencoders (VAEs).

Paper links are supplied at the beginning of each file with a short summary of the paper. See src folder for files to run via terminal, or notebooks folder for Jupyter notebook visualizations via your local browser. The main file changes can be see in the train, train_D, and train_G of the Trainer class, although changes are not completely limited to only these two areas (e.g. Wasserstein GAN clamps weight in the train function, BEGAN gives multiple outputs from train_D, fGAN has a slight modification in viz_loss function to indicate method used in title).

All code in this repository operates in a generative, unsupervised manner on binary (black and white) MNIST. The architectures are compatible with a variety of datatypes (1D, 2D, square 3D images). Plotting functions work with binary/RGB images. If a GPU is detected, the models use it. Otherwise, they default to CPU. VAE Trainer classes contain methods to visualize latent space representations (see make_all function).

Usage

To initialize an environment:

python -m venv env  
. env/bin/activate  
pip install -r requirements.txt  

For playing around in Jupyer notebooks:

jupyter notebook

To run from Terminal:

cd src
python bir_vae.py

New Models

One of the primary purposes of this repository is to make implementing deep generative model (i.e., GAN/VAE) variants as easy as possible. This is possible because, typically but not always (e.g. BIRVAE), the proposed modifications only apply to the way loss is computed for backpropagation. Thus, the core training class is structured in such a way that most new implementations should only require edits to the train_D and train_G functions of GAN Trainer classes, and the compute_batch function of VAE Trainer classes.

Suppose we have a non-saturating GAN and we wanted to implement a least-squares GAN. To do this, all we have to do is change two lines:

Original (NSGAN)

def train_D(self, images):
  ...
  D_loss = -torch.mean(torch.log(DX_score + 1e-8) + torch.log(1 - DG_score + 1e-8))

  return D_loss
def train_G(self, images):
  ...
  G_loss = -torch.mean(torch.log(DG_score + 1e-8))

  return G_loss

New (LSGAN)

def train_D(self, images):
  ...
  D_loss = (0.50 * torch.mean((DX_score - 1.)**2)) + (0.50 * torch.mean((DG_score - 0.)**2))

  return D_loss
def train_G(self, images):
  ...
  G_loss = 0.50 * torch.mean((DG_score - 1.)**2)

  return G_loss

Model Architecture

The architecture chosen in these implementations for both the generator (G) and discriminator (D) consists of a simple, two-layer feedforward network. While this will give sensible output for MNIST, in practice it is recommended to use deep convolutional architectures (i.e. DCGANs) to get nicer outputs. This can be done by editing the Generator and Discriminator classes for GANs, or the Encoder and Decoder classes for VAEs.

Visualization

All models were trained for 25 epochs with hidden dimension 400, latent dimension 20. Other implementation specifics are as close to the respective original paper (linked) as possible.

Model Epoch 1 Epoch 25 Progress Loss
MMGAN
NSGAN
WGAN
WGPGAN
DRAGAN
BEGAN
LSGAN
RaNSGAN
FisherGAN
InfoGAN
f-TVGAN
f-PearsonGAN
f-JSGAN
f-ForwGAN
f-RevGAN
f-HellingerGAN
VAE
BIRVAE

To Do

Models: CVAE, denoising VAE, adversarial autoencoder | Bayesian GAN, Self-attention GAN, Primal-Dual Wasserstein GAN
Architectures: Add DCGAN option
Datasets: Beyond MNIST

Owner
Shayne O'Brien
NLP / Machine Learning / Network Science. Moved from MIT to Apple 06/2019
Shayne O'Brien
Unsupervised Video Interpolation using Cycle Consistency

Unsupervised Video Interpolation using Cycle Consistency Project | Paper | YouTube Unsupervised Video Interpolation using Cycle Consistency Fitsum A.

NVIDIA Corporation 100 Nov 30, 2022
TransMIL: Transformer based Correlated Multiple Instance Learning for Whole Slide Image Classification

TransMIL: Transformer based Correlated Multiple Instance Learning for Whole Slide Image Classification [NeurIPS 2021] Abstract Multiple instance learn

132 Dec 30, 2022
SW components and demos for visual kinship recognition. An emphasis is put on the FIW dataset-- data loaders, benchmarks, results in summary.

FIW Data Development Kit Table of Contents Introduction Families In the Wild Database Publications Organization To Do License Getting Involved Introdu

Joseph P. Robinson 12 Jun 04, 2022
The Most Efficient Temporal Difference Learning Framework for 2048

moporgic/TDL2048+ TDL2048+ is a highly optimized temporal difference (TD) learning framework for 2048. Features Many common methods related to 2048 ar

Hung Guei 5 Nov 23, 2022
The official codes of "Semi-supervised Models are Strong Unsupervised Domain Adaptation Learners".

SSL models are Strong UDA learners Introduction This is the official code of paper "Semi-supervised Models are Strong Unsupervised Domain Adaptation L

Yabin Zhang 26 Dec 26, 2022
This project is for a Twitter bot that monitors a bird feeder in my backyard. Any detected birds are identified and posted to Twitter.

Backyard Birdbot Introduction This is a silly hobby project to use existing ML models to: Detect any birds sighted by a webcam Identify whic

Chi Young Moon 71 Dec 25, 2022
Implementation of Stochastic Image-to-Video Synthesis using cINNs.

Stochastic Image-to-Video Synthesis using cINNs Official PyTorch implementation of Stochastic Image-to-Video Synthesis using cINNs accepted to CVPR202

CompVis Heidelberg 135 Dec 28, 2022
DaReCzech is a dataset for text relevance ranking in Czech

Dataset DaReCzech is a dataset for text relevance ranking in Czech. The dataset consists of more than 1.6M annotated query-documents pairs,

Seznam.cz a.s. 8 Jul 26, 2022
Autonomous racing with the Anki Overdrive

Anki Autonomous Racing Autonomous racing with the Anki Overdrive. Using the Overdrive-Python API (https://github.com/xerodotc/overdrive-python) develo

3 Dec 11, 2022
OCR-D wrapper for detectron2 based segmentation models

ocrd_detectron2 OCR-D wrapper for detectron2 based segmentation models Introduction Installation Usage OCR-D processor interface ocrd-detectron2-segm

Robert Sachunsky 13 Dec 06, 2022
The PyTorch improved version of TPAMI 2017 paper: Face Alignment in Full Pose Range: A 3D Total Solution.

Face Alignment in Full Pose Range: A 3D Total Solution By Jianzhu Guo. [Updates] 2020.8.30: The pre-trained model and code of ECCV-20 are made public

Jianzhu Guo 3.4k Jan 02, 2023
Implementations of polygamma, lgamma, and beta functions for PyTorch

lgamma Implementations of polygamma, lgamma, and beta functions for PyTorch. It's very hacky, but that's usually ok for research use. To build, run: .

Rachit Singh 24 Nov 09, 2021
Taichi Course Homework Template

太极图形课S1-标题部分 这个作业未来或将是你的开源项目,标题的内容可以来自作业中的核心关键词,让读者一眼看出你所完成的工作/做出的好玩demo 如果暂时未想好,起名时可以参考“太极图形课S1-xxx作业” 如下是作业(项目)展开说明的方法,可以帮大家理清思路,并且也对读者非常友好,请小伙伴们多多参

TaichiCourse 30 Nov 19, 2022
Contrastive Learning for Compact Single Image Dehazing, CVPR2021

AECR-Net Contrastive Learning for Compact Single Image Dehazing, CVPR2021. Official Pytorch based implementation. Paper arxiv Pytorch Version TODO: mo

glassy 253 Jan 01, 2023
[CVPR 2022] Unsupervised Image-to-Image Translation with Generative Prior

GP-UNIT - Official PyTorch Implementation This repository provides the official PyTorch implementation for the following paper: Unsupervised Image-to-

Shuai Yang 125 Jan 03, 2023
RuDOLPH: One Hyper-Modal Transformer can be creative as DALL-E and smart as CLIP

[Paper] [Хабр] [Model Card] [Colab] [Kaggle] RuDOLPH 🦌 🎄 ☃️ One Hyper-Modal Transformer can be creative as DALL-E and smart as CLIP Russian Diffusio

AI Forever 232 Jan 04, 2023
Where2Act: From Pixels to Actions for Articulated 3D Objects

Where2Act: From Pixels to Actions for Articulated 3D Objects The Proposed Where2Act Task. Given as input an articulated 3D object, we learn to propose

Kaichun Mo 69 Nov 28, 2022
Neighborhood Reconstructing Autoencoders

Neighborhood Reconstructing Autoencoders The official repository for Neighborhood Reconstructing Autoencoders (Lee, Kwon, and Park, NeurIPS 2021). T

Yonghyeon Lee 24 Dec 14, 2022
Rank 3 : Source code for OPPO 6G Data Generation Challenge

OPPO 6G Data Generation with an E2E Framework Homepage of OPPO 6G Data Generation Challenge Datasets H1_32T4R.mat H2_32T4R.mat Please put the original

Sen Pei 97 Jan 07, 2023
Pseudo lidar - (CVPR 2019) Pseudo-LiDAR from Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Driving

Pseudo-LiDAR from Visual Depth Estimation: Bridging the Gap in 3D Object Detection for Autonomous Driving This paper has been accpeted by Conference o

Yan Wang 881 Dec 27, 2022