基于pytorch构建cyclegan示例

Related tags

Deep Learningcyclegan-demo
Overview

cyclegan-demo

基于Pytorch构建CycleGAN示例

如何运行

准备数据集

将数据集整理成4个文件，分别命名为

  • trainA, trainB：训练集，A、B代表两类图片
  • testA, testB：测试集，A、B代表两类图片

例如

D:\CODE\CYCLEGAN-DEMO\DATA\SUMMER2WINTER
├─testA
├─testB
├─trainA
└─trainB

之后在main.py中将root设为数据集的路径。

参数设置

见main.py中的初始化参数

# 初始化参数
# seed: 随机种子
# root: 数据集路径
# output_model_root: 模型的输出路径
# image_size: 图片尺寸
# batch_size: 一次喂入的数据量
# lr: 学习率
# betas: 一阶和二阶动量
# epochs: 训练总次数
# historical_epochs: 历史训练次数
# - 0表示不沿用历史模型
# - >0表示对应训练次数的模型
# - -1表示最后一次训练的模型
# save_every: 保存频率
# loss_range: Loss的显示范围
seed = 123
data_root = 'D:/code/cyclegan-demo/data/summer2winter'
output_model_root = 'output/model'
image_size = 64
batch_size = 16
lr = 2e-4
betas = (.5, .999)
epochs = 100
historical_epochs = -1
save_every = 1
loss_range = 1000

安装和运行

  1. 安装依赖
pip install -r requirements.txt
  1. 打开命令行，运行Visdom
python -m visdom.server
  1. 运行主程序
python main.py

训练过程的可视化展示

访问地址http://localhost:8097即可进入Visdom可视化页面，页面中将展示：

  • A类真实图片 -【A2B生成器】 -> B类虚假图片 -【B2A生成器】 -> A类重构图片
  • B类真实图片 -【B2A生成器】 -> A类虚假图片 -【A2B生成器】 -> B类重构图片
  • 判别器A、B以及生成器的Loss曲线

一些可视化的具体用法可见Visdom的使用方法。

测试

TODO

介绍

目录结构

  • dataset.py 数据集
  • discriminator.py 判别器
  • generater.py 生成器
  • main.py 主程序
  • replay_buffer.py 缓冲区
  • resblk.py 残差块
  • util.py 工具方法

原理介绍

残差块是生成器的组成部分，其结构如下

Resblk(
  (main): Sequential(
    (0): ReflectionPad2d((1, 1, 1, 1))
    (1): Conv2d(3, 3, kernel_size=(3, 3), stride=(1, 1))
    (2): InstanceNorm2d(3, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
    (3): ReLU(inplace=True)
    (4): ReflectionPad2d((1, 1, 1, 1))
    (5): Conv2d(3, 3, kernel_size=(3, 3), stride=(1, 1))
    (6): InstanceNorm2d(3, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
  )
)

生成器结构如下，由于采用全卷积结构，事实上其结构与图片尺寸无关

Generater(
  (input): Sequential(
    (0): ReflectionPad2d((3, 3, 3, 3))
    (1): Conv2d(3, 64, kernel_size=(7, 7), stride=(1, 1))
    (2): InstanceNorm2d(64, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
    (3): ReLU(inplace=True)
  )
  (downsampling): Sequential(
    (0): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
    (1): InstanceNorm2d(128, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
    (2): ReLU(inplace=True)
    (3): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1))
    (4): InstanceNorm2d(256, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
    (5): ReLU(inplace=True)
  )
  (resnet): Sequential(
    (0): Resblk
    (1): Resblk
    (2): Resblk
    (3): Resblk
    (4): Resblk
    (5): Resblk
    (6): Resblk
    (7): Resblk
    (8): Resblk
  )

  (upsampling): Sequential(
    (0): ConvTranspose2d(256, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), output_padding=(1, 1))
    (1): InstanceNorm2d(128, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
    (2): ReLU(inplace=True)
    (3): ConvTranspose2d(128, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), output_padding=(1, 1))
    (4): InstanceNorm2d(64, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
    (5): ReLU(inplace=True)
  )
  (output): Sequential(
    (0): ReplicationPad2d((3, 3, 3, 3))
    (1): Conv2d(64, 3, kernel_size=(7, 7), stride=(1, 1))
    (2): Tanh()
  )
)

判别器结构如下，池化层具体尺寸由图片尺寸决定，64x64的图片对应池化层为6x6

Discriminator(
  (main): Sequential(
    (0): Conv2d(3, 64, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
    (1): LeakyReLU(negative_slope=0.2, inplace=True)
    (2): Conv2d(64, 128, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
    (3): InstanceNorm2d(128, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
    (4): LeakyReLU(negative_slope=0.2, inplace=True)
    (5): Conv2d(128, 256, kernel_size=(4, 4), stride=(2, 2), padding=(1, 1))
    (6): InstanceNorm2d(256, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
    (7): LeakyReLU(negative_slope=0.2, inplace=True)
    (8): Conv2d(256, 512, kernel_size=(4, 4), stride=(1, 1), padding=(1, 1))
    (9): InstanceNorm2d(512, eps=1e-05, momentum=0.1, affine=False, track_running_stats=False)
    (10): LeakyReLU(negative_slope=0.2, inplace=True)
    (11): Conv2d(512, 1, kernel_size=(4, 4), stride=(1, 1), padding=(1, 1))
  )
  (output): Sequential(
    (0): AvgPool2d(kernel_size=torch.Size([6, 6]), stride=torch.Size([6, 6]), padding=0)
    (1): Flatten(start_dim=1, end_dim=-1)
  )
)

训练共有三个优化器，分别负责生成器、判别器A、判别器B的优化。

损失有三种类型：

  • 一致性损失：A(B)类真实图片与经生成器生成的图片的误差，该损失使得生成后的风格与原图更接近，采用L1Loss
  • 对抗损失：A(B)类图片经生成器得到B(A)类图片，再经判别器判别的错误率，采用MSELoss
  • 循环损失：A(B)类图片经生成器得到B(A)类图片，再经生成器得到A(B)类的重建图片，原图和重建图片的误差，采用L1Loss

生成器的训练过程：

  1. 将A(B)类真实图片送入生成器，得到生成的图片，计算生成图片与原图的一致性损失
  2. 将A(B)类真实图片送入生成器得到虚假图片，再送入判别器得到判别结果，计算判别结果与真实标签1的对抗损失(虚假图片应能被判别器判别为真实图片，即生成器能骗过判别器)
  3. 将A(B)类虚假图片送入生成器，得到重建图片，计算重建图片与原图的循环损失
  4. 计算、更新梯度

判别器A的训练过程：

  1. 将A类真实图片送入判别器A，得到判别结果，计算判别结果与真实标签1的对抗损失(判别器应将真实图片判别为真实)
  2. 将A类虚假图片送入判别器A，得到判别结果，计算判别结果与虚假标签0的对抗损失(判别器应将虚假图片判别为虚假)
  3. 计算、更新梯度
Owner
Koorye
学习？学个屁
Koorye
GitHub Repository
Balancing Principle for Unsupervised Domain Adaptation

Blancing Principle for Domain Adaptation NeurIPS 2021 Paper Abstract We address the unsolved algorithm design problem of choosing a justified regulari

Marius-Constantin Dinu 4 Dec 15, 2022
Adapter-BERT: Parameter-Efficient Transfer Learning for NLP.

Adapter-BERT: Parameter-Efficient Transfer Learning for NLP.

Google Research 340 Jan 03, 2023
Implementation of the Transformer variant proposed in "Transformer Quality in Linear Time"

FLASH - Pytorch Implementation of the Transformer variant proposed in the paper Transformer Quality in Linear Time Install $ pip install FLASH-pytorch

Phil Wang 209 Dec 28, 2022
Real-time VIBE: Frame by Frame Inference of VIBE (Video Inference for Human Body Pose and Shape Estimation)

Real-time VIBE Inference VIBE frame-by-frame. Overview This is a frame-by-frame inference fork of VIBE at [https://github.com/mkocabas/VIBE]. Usage: i

23 Jul 02, 2022
Deep motion transfer

animation-with-keypoint-mask Paper The right most square is the final result. Softmax mask (circles): \ Heatmap mask: \ conda env create -f environmen

9 Nov 01, 2022
A configurable, tunable, and reproducible library for CTR prediction

FuxiCTR This repo is the community dev version of the official release at huawei-noah/benchmark/FuxiCTR. Click-through rate (CTR) prediction is an cri

XUEPAI 397 Dec 30, 2022
BDDM: Bilateral Denoising Diffusion Models for Fast and High-Quality Speech Synthesis

Bilateral Denoising Diffusion Models (BDDMs) This is the official PyTorch implementation of the following paper: BDDM: BILATERAL DENOISING DIFFUSION M

172 Dec 23, 2022
OneShot Learning-based hotword detection.

EfficientWord-Net Hotword detection based on one-shot learning Home assistants require special phrases called hotwords to get activated (eg:"ok google

ANT-BRaiN 102 Dec 25, 2022
[BMVC2021] The official implementation of "DomainMix: Learning Generalizable Person Re-Identification Without Human Annotations"

DomainMix [BMVC2021] The official implementation of "DomainMix: Learning Generalizable Person Re-Identification Without Human Annotations" [paper] [de

Wenhao Wang 17 Dec 20, 2022
code from "Tensor decomposition of higher-order correlations by nonlinear Hebbian plasticity"

Code associated with the paper "Tensor decomposition of higher-order correlations by nonlinear Hebbian learning," Ocker & Buice, Neurips 2021. "plot_f

Gabriel Koch Ocker 4 Oct 16, 2022
Source code for our EMNLP'21 paper 《Raise a Child in Large Language Model: Towards Effective and Generalizable Fine-tuning》

Child-Tuning Source code for EMNLP 2021 Long paper: Raise a Child in Large Language Model: Towards Effective and Generalizable Fine-tuning. 1. Environ

46 Dec 12, 2022
An end-to-end implementation of intent prediction with Metaflow and other cool tools

You Don't Need a Bigger Boat An end-to-end (Metaflow-based) implementation of an intent prediction flow for kids who can't MLOps good and wanna learn

Jacopo Tagliabue 614 Dec 31, 2022
Face recognition project by matching the features extracted using SIFT.

MV_FaceDetectionWithSIFT Face recognition project by matching the features extracted using SIFT. By : Aria Radmehr Professor : Ali Amiri Dependencies

Aria Radmehr 4 May 31, 2022
Package for extracting emotions from social media text. Tailored for financial data.

EmTract: Extracting Emotions from Social Media Text Tailored for Financial Contexts EmTract is a tool that extracts emotions from social media text. I

13 Nov 17, 2022
Code and real data for the paper "Counterfactual Temporal Point Processes", available at arXiv.

counterfactual-tpp This is a repository containing code and real data for the paper Counterfactual Temporal Point Processes. Pre-requisites This code

Networks Learning 11 Dec 09, 2022
Vis2Mesh: Efficient Mesh Reconstruction from Unstructured Point Clouds of Large Scenes with Learned Virtual View Visibility ICCV2021

Vis2Mesh This is the offical repository of the paper: Vis2Mesh: Efficient Mesh Reconstruction from Unstructured Point Clouds of Large Scenes with Lear

71 Dec 25, 2022
Semantic Image Synthesis with SPADE

Semantic Image Synthesis with SPADE New implementation available at imaginaire repository We have a reimplementation of the SPADE method that is more

NVIDIA Research Projects 7.3k Jan 07, 2023
LaBERT - A length-controllable and non-autoregressive image captioning model.

Length-Controllable Image Captioning (ECCV2020) This repo provides the implemetation of the paper Length-Controllable Image Captioning. Install conda

bearcatt 53 Nov 13, 2022
Convenient tool for speeding up the intern/officer review process.

icpc-app-screen Convenient tool for speeding up the intern/officer applicant review process. Eliminates the pain from reading application responses of

1 Oct 30, 2021
A foreign language learning aid using a neural network to predict probability of translating foreign words

Langy Langy is a reading-focused foreign language learning aid orientated towards young children. Reading is an activity that every child knows. It is

Shona Lowden 6 Nov 17, 2021