DCGAN LSGAN WGAN-GP DRAGAN PyTorch

Last update: Nov 30, 2022

Overview

Recommendation

Our GAN based work for facial attribute editing - AttGAN.

News

8 April 2019: We re-implement these GANs by Tensorflow 2! The old version is here: v1 or in the "v1" directory.
PyTorch Version

GANs - Tensorflow 2

Tensorflow 2 implementations of DCGAN, LSGAN, WGAN-GP and DRAGAN.

Exemplar results

Fashion-MNIST

DCGAN	LSGAN	WGAN-GP	DRAGAN

CelebA

DCGAN	LSGAN

WGAN-GP	DRAGAN

Anime

WGAN-GP	DRAGAN

Usage

Environment
- Python 3.6
- TensorFlow 2.2, TensorFlow Addons 0.10.0
- OpenCV, scikit-image, tqdm, oyaml
- we recommend Anaconda or Miniconda, then you can create the TensorFlow 2.2 environment with commands below
```
conda create -n tensorflow-2.2 python=3.6

source activate tensorflow-2.2

conda install scikit-image tqdm tensorflow-gpu=2.2

conda install -c conda-forge oyaml

pip install tensorflow-addons==0.10.0
```
- NOTICE: if you create a new conda environment, remember to activate it before any other command
```
source activate tensorflow-2.2
```
Datasets
- Fashion-MNIST will be automatically downloaded
- CelebA should be prepared by yourself in ./data/img_align_celeba/*.jpg
  - dataset link (find "img_align_celeba.zip"):
    - Baidu Netdisk (password rp0s) or
    - Google Drive
- the Anime dataset should be prepared by yourself in ./data/faces/*.jpg
  - dataset link: https://www.kaggle.com/splcher/animefacedataset

Examples of training

Fashion-MNIST DCGAN

CUDA_VISIBLE_DEVICES=0 python train.py --dataset=fashion_mnist --epoch=25 --adversarial_loss_mode=gan

CelebA DRAGAN

CUDA_VISIBLE_DEVICES=0 python train.py --dataset=celeba --epoch=25 --adversarial_loss_mode=gan --gradient_penalty_mode=dragan

Anime WGAN-GP

CUDA_VISIBLE_DEVICES=0 python train.py --dataset=anime --epoch=200 --adversarial_loss_mode=wgan --gradient_penalty_mode=wgan-gp --n_d=5

see more training exampls in commands.sh

tensorboard for loss visualization

tensorboard --logdir ./output/fashion_mnist_gan/summaries --port 6006

Comments

GPU is full

Hello, when the code runs, the memory is full. What happened? My python version is 3.6, tensorflow version is 1.11, my GPU is 1080ti, thanks! the error is as follow: An error ocurred while starting the kernel 2019󈚥󈚦 08:40:45.229298: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 2019󈚥󈚦 08:40:45.601632: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1411] Found device 0 with properties: name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582 pciBusID: 0000:65:00.0 totalMemory: 11.00GiB freeMemory: 9.10GiB 2019󈚥󈚦 08:40:45.603929: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1490] Adding visible gpu devices: 0 2019󈚥󈚦 08:40:46.556261: I tensorflow/core/common_runtime/gpu/gpu_device.cc:971] Device interconnect StreamExecutor with strength 1 edge matrix: 2019󈚥󈚦 08:40:46.558110: I tensorflow/core/common_runtime/gpu/gpu_device.cc:977] 0 2019󈚥󈚦 08:40:46.558407: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990] 0: N 2019󈚥󈚦 08:40:46.558836: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1103] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 8789 MB memory) ‑> physical GPU (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:65:00.0, compute capability: 6.1)

opened by lixingbao 9
About problem with generating image size

Hello,can your code only generate 64×64 images? Can I generate an image of the specified size? For example: 256 × 256, if you can, what parameters need to be modified?thank you!

opened by lixingbao 6
WGAN-GP does not work!!!

I have updated the code from TensorFlow 2.0-alpha to TensorFlow 2.0, everything works well except for WGAN-GP (it works in tf2.0-alpha). In tf2.0, The gradient penalty seems very unstable, but I cannot find out the problem. Does anybody help? I will be grateful.
help wanted

opened by LynnHo 3
how about 3D data?

Hi!

cartoon faces original size is [96, 96, 3],the number 3 means 3 channel RGB data. But if I have grayscale data with 3 slices, i.e the size is [121,145,3], Can I simply use this code? If not, what should I change based on this code?

Thanks for your work! Look forward to your response.

opened by KakaVlasic 3
c_iter isn't used

c_iter is defined but not used in all of the WGAN files. What is the correct behaviour? i.e should the critic be optimised heavily initially or not?

Also, can you confirm that you use a learning rate parameter of 0.0002, regardless of whether RMSProp or Adam is used as the optimiser?

opened by davidADSP 3
About n_critic = 5

i use the code whitch be used to train cartoon pictures with WGAN-GP. i don't know what the mean of n_critic = 5 , and why do you to set it. thanks.

opened by tuoniaoren 3
Error while using celeba dataset

I am getting this error while running train.py TypeError: Input 'filename' of 'ReadFile' Op has type float32 that does not match expected type of string. Please help with this. Thanks in advance

opened by yksolanki9 2
Where do you freeze the gradient descent?

Hello, I am confused about how do you freeze the gradient descent to the other model. When training d_step, I suppose the generator should be freezed, as f_logit is based on generator and used in d_loss; similarly, when training g_step, I suppose the discriminator should be freezed, as f_logit depens on discriminator.

However, I do not see where you stop those gradients flowing to the unwanted part, either generator or discriminator. Would you please provide some hints for me? Thank you.

opened by ybsave 2
do you try to use Resnet in wgan-gp?

Have you compared the difference between the network structure of DCGAN and the structure of Resnet in WGAN-GP?Is the effect of Resnet will be better than the structure of DCGAN.

opened by tuoniaoren 2
running question

do you meet the program stop without mistakes when the code running for some time ,and the GPU stops work.I changed the value of num_threads。（from 16 to 10）。i run it again.i don't know Is it because the value is too high。

opened by tuoniaoren 2
License

Hi Zhenliang He, I wonder whether you would be willing to please license this code under an open source license? If so please add a license, or if not please just close this request. Thanks, Connelly

opened by connellybarnes 2
A problem for your DCGAN architecture

Hi, - Your work is really interesting. But I have found there is a problem for your DCGAN that I didn't understand. You generate noise twice when train discriminator and generator for each iteration, like the blue lines in the following picture. In soumith code (includes some official DCGAN code), he only generate noise once: https://github.com/soumith/dcgan.torch. Could you please tell me the reason?

opened by RayGuo-C 1
NameError: name 'shape' is not defined

Traceback (most recent call last): File "D:/github/DCGAN-LSGAN-WGAN-GP-DRAGAN-Tensorflow-2-master/DCGAN-LSGAN-WGAN-GP-DRAGAN-Tensorflow-2-master/train.py", line 91, in G = module.ConvGenerator(input_shape=(1, 1, args.z_dim), output_channels=shape[-1], n_upsamplings=n_G_upsamplings, name='G_%s' % args.dataset) NameError: name 'shape' is not defined please tell me why

opened by Tonyztj 0

Releases(v1)

v1(Apr 6, 2019)
DCGAN-LSGAN-WGAN-WGAN-GP-Tensorflow-r1.2

tensorflow r1.2

python 2.7

Source code(tar.gz)
Source code(zip)

Owner

Zhenliang He

GitHub Repository

[ICCV'2021] Image Inpainting via Conditional Texture and Structure Dual Generation

122 Dec 11, 2022

PyTorch implementation of the R2Plus1D convolution based ResNet architecture described in the paper "A Closer Look at Spatiotemporal Convolutions for Action Recognition"

R2Plus1D-PyTorch PyTorch implementation of the R2Plus1D convolution based ResNet architecture described in the paper "A Closer Look at Spatiotemporal

342 Dec 16, 2022

Honours project, on creating a depth estimation map from two stereo images of featureless regions

image-processing This module generates depth maps for shape-blocked-out images Install If working with anaconda, then from the root directory: conda e

2 Oct 17, 2022

SlotRefine: A Fast Non-Autoregressive Model forJoint Intent Detection and Slot Filling

SlotRefine: A Fast Non-Autoregressive Model for Joint Intent Detection and Slot Filling Reference Main paper to be cited (Di Wu et al., 2020) @article

34 Nov 03, 2022

DNA sequence classification by Deep Neural Network

DNA sequence classification by Deep Neural Network: Project Overview worked on the DNA sequence classification problem where the input is the DNA sequ

0 Aug 02, 2022

PyTorch code for 'Efficient Single Image Super-Resolution Using Dual Path Connections with Multiple Scale Learning'

Efficient Single Image Super-Resolution Using Dual Path Connections with Multiple Scale Learning This repository is for EMSRDPN introduced in the foll

7 Feb 10, 2022

Code implementation of Data Efficient Stagewise Knowledge Distillation paper.

Data Efficient Stagewise Knowledge Distillation Table of Contents Data Efficient Stagewise Knowledge Distillation Table of Contents Requirements Image

112 Dec 02, 2022

Pytorch modules for paralel models with same architecture. Ideal for multi agent-based systems

WideLinears Pytorch parallel Neural Networks A package of pytorch modules for fast paralellization of separate deep neural networks. Ideal for agent-b

1 Dec 17, 2021

PyTorch implementation for NED. It can be used to manipulate the facial emotions of actors in videos based on emotion labels or reference styles.

Neural Emotion Director (NED) - Official Pytorch Implementation Example video of facial emotion manipulation while retaining the original mouth motion

89 Dec 23, 2022

A Neural Net Training Interface on TensorFlow, with focus on speed + flexibility

Tensorpack is a neural network training interface based on TensorFlow. Features: It's Yet Another TF high-level API, with speed, and flexibility built

6.2k Jan 09, 2023

[NeurIPS2021] Code Release of K-Net: Towards Unified Image Segmentation

K-Net: Towards Unified Image Segmentation Introduction This is an official release of the paper K-Net:Towards Unified Image Segmentation. K-Net will a

423 Jan 02, 2023

Super-Fast-Adversarial-Training - A PyTorch Implementation code for developing super fast adversarial training

Super-Fast-Adversarial-Training This is a PyTorch Implementation code for develo

26 Dec 02, 2022

Vit-ImageClassification - Pytorch ViT for Image classification on the CIFAR10 dataset

Vit-ImageClassification Introduction This project uses ViT to perform image clas

4 Jun 01, 2022

Attention for PyTorch with Linear Memory Footprint

Attention for PyTorch with Linear Memory Footprint Unofficially implements https://arxiv.org/abs/2112.05682 to get Linear Memory Cost on Attention (+

11 Jan 09, 2022

Unofficial TensorFlow implementation of the Keyword Spotting Transformer model

Keyword Spotting Transformer This is the unofficial TensorFlow implementation of the Keyword Spotting Transformer model. This model is used to train o

8 May 11, 2022

PyTorch implementation of Hierarchical Multi-label Text Classification: An Attention-based Recurrent Network

hierarchical-multi-label-text-classification-pytorch Hierarchical Multi-label Text Classification: An Attention-based Recurrent Network Approach This

17 Dec 13, 2022

The source code for the Cutoff data augmentation approach proposed in this paper: "A Simple but Tough-to-Beat Data Augmentation Approach for Natural Language Understanding and Generation".

Cutoff: A Simple Data Augmentation Approach for Natural Language This repository contains source code necessary to reproduce the results presented in

49 Dec 22, 2022