codecat0 / cv

本仓库将使用Pytorch框架实现经典的图像分类网络、目标检测网络、图像分割网络，图像生成网络等，并会持续更新！！！

CVers 集合!!!

Image Classification

1. AlexNet

• 论文地址：https://arxiv.org/abs/1404.5997
• 论文详解：https://blog.csdn.net/qq_42735631/article/details/115800315
• 模型代码: https://github.com/codecat0/CV/blob/main/Image_Classification/models/alexnet.py

2. VGGNet

• 论文地址: https://arxiv.org/pdf/1409.1556.pdf
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/116071166
• 模型代码: https://github.com/codecat0/CV/blob/main/Image_Classification/models/vggnet.py

3. GoogLeNet

• 论文地址: http://arxiv.org/abs/1409.4842
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/116404840
• 模型代码: https://github.com/codecat0/CV/blob/main/Image_Classification/models/googlenet.py

4. ResNet

• 论文地址: https://arxiv.org/pdf/1512.03385.pdf
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/116803066
• 模型代码: https://github.com/codecat0/CV/blob/main/Image_Classification/models/resnet.py

5. DenseNet

• 论文地址: https://arxiv.org/pdf/1707.06990.pdf
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/116861074
• 模型代码: https://github.com/codecat0/CV/blob/main/Image_Classification/models/densenet.py

6. MobileNet系列

6.1 MobileNet V1

• 论文地址: https://arxiv.org/abs/1704.04861
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121053364
• 模型代码

6.2 MobileNet V2

• 论文地址: https://arxiv.org/abs/1801.04381
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121053364
• 模型代码: https://github.com/codecat0/CV/blob/main/Image_Classification/models/mobilenetv2.py

6.3 MobileNet V3

• 论文地址: https://arxiv.org/abs/1905.02244
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121053364
• 模型代码: https://github.com/codecat0/CV/blob/main/Image_Classification/models/mobilenetv3.py

7. ShuffleNet系列

7.1 ShuffleNet V1

• 论文地址: https://arxiv.org/abs/1707.01083
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121064514
• 模型代码: https://github.com/codecat0/CV/blob/main/Image_Classification/models/shufflenetv1.py

7.2 ShuffleNet V2

• 论文地址: https://arxiv.org/abs/1807.11164
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121064514
• 模型代码: https://github.com/codecat0/CV/blob/main/Image_Classification/models/shufflenetv2.py

8. GhostNet

• 论文地址: https://arxiv.org/abs/1911.11907
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121065063
• 模型代码: https://github.com/codecat0/CV/blob/main/Image_Classification/models/ghostnet.py

Object Detection

1. R-CNN

• 论文地址: https://arxiv.org/abs/1311.2524
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121147838
• 模型代码

2. Fast R-CNN

• 论文地址: https://arxiv.org/abs/1504.08083
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121151584
• 模型代码

3. Faster R-CNN

• 论文地址: https://arxiv.org/abs/1506.01497
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121164268
• 模型代码: https://github.com/codecat0/CV/tree/main/Object_Detection/Faster_RCNN

4. SSD

• 论文地址: https://arxiv.org/abs/1512.02325
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121331064
• 模型代码: https://github.com/codecat0/CV/tree/main/Object_Detection/SSD

5. YOLO系列

5.1 YOLOV1

• 论文地址: https://arxiv.org/abs/1506.02640
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121445432
• 模型代码

5.2 YOLOV2

• 论文地址: https://arxiv.org/abs/1612.08242
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121456856
• 模型代码

5.3 YOLOV3

• 论文地址: https://arxiv.org/abs/1804.02767
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121479035
• 模型代码: https://github.com/codecat0/CV/tree/main/Object_Detection/YOLOV3

5.3 YOLOV4

• 论文地址: https://arxiv.org/abs/2004.10934
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/121621972
• 模型代码: https://github.com/codecat0/CV/tree/main/Object_Detection/YOLOV4

6. FPN

• 论文地址: https://arxiv.org/abs/1612.03144
• 论文详解
• 模型代码

7. Mask R-CNN

• 论文地址
• 论文详解
• 模型代码

Semantic Segmentation

1. DeepLab系列

• 论文地址
• 论文详解
• 模型代码

2. U-Net

• 论文地址: https://arxiv.org/abs/1505.04597
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/122182266?spm=1001.2014.3001.5501
• 模型代码: https://github.com/codecat0/CV/tree/main/Semantic_Segmentation/UNet

3. FCN

• 论文地址: https://www.cv-foundation.org/openaccess/content_cvpr_2015/papers/Long_Fully_Convolutional_Networks_2015_CVPR_paper.pdf
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/122247211?spm=1001.2014.3001.5501
• 模型代码: https://github.com/codecat0/CV/tree/main/Semantic_Segmentation/FCN

4. RCF

• 论文地址
• 论文详解
• 模型代码

5. SegNet

• 论文地址: https://arxiv.org/abs/1511.00561
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/122252894?spm=1001.2014.3001.5501
• 模型代码: https://github.com/codecat0/CV/tree/main/Semantic_Segmentation/SegNet

6. PSPNet

• 论文地址：https://arxiv.org/abs/1612.01105
• 论文详解: https://blog.csdn.net/qq_42735631/article/details/122069409
• 模型代码: https://github.com/codecat0/CV/tree/main/Semantic_Segmentation/PSPNet

Generative Models

1. AutoEncoder

• 论文地址
• 论文详解
• 模型代码

2. VAE

• 论文地址
• 论文详解
• 模型代码

3. GAN

• 论文地址
• 论文详解
• 模型代码

4. DCGAN

• 论文地址
• 论文详解
• 模型代码

5. CycleGAN

• 论文地址
• 论文详解
• 模型代码

6. WGAN

• 论文地址
• 论文详解
• 模型代码

7. StyleGAN

• 论文地址
• 论文详解
• 模型代码

cv's Issues

准确率是咋显示的？

[train epoch 0] loss: 0.339386, acc: 7.600000: 100%|
[valid epoch 0] loss: 0.001, acc 1.000: 100%|
[train epoch 1] loss: 0.001157, acc: 8.000000: 100%
这准确率是多少啊？100%？还是说没显示？

SegNet模型保存

您好，我在model_path导入ep050.pth的文件，但是他出现错误：KeyError: 'module.encoder.encode1.0.weight'
同时，
ep050.pth在训练集与测试集loss均达0.05，但是，我用它predict的时候，只有一种颜色输出，甚至用训练集里的照片也是这样的结果
想请教一下：
请问是模型保存的出现问题了吗？
或者是torch.nn.DataParallel的问题吗？（我的cuda定义是True，但是我只有一张显卡；可是我cuda定义是False时是cpu训练，很慢）

SegNet 网络结构

SegNet的网络如果根据网络结构图来看，Encoder345和Decoder123是三层卷积+BN+ReLU吧。

环境配置冲突

您好，在pycharm依照配置文件安装环境时出现torch、torchvision和opencv无法安装成功的情况。用anaconda控制台安装显示pytorch和torchvision与其他环境有冲突？

backbone训练

博主 请问如果要选择其中之一作为backbone进行训练得到预训练权重文件，是不是只要把输出通道数和分类的代码改掉就可以了？