DeepLearning

介绍

个人学习项目，用于复现论文代码，深入理解算法原理。

注意：每个项目都可独立运行。若要运行某个项目，你需要将该项目作为根目录，以便找到对应模块。

每个项目都有对应论文解读，解读详情搜 [知乎] - 琪小钧

• paper read: 知乎 论文解读（有些博主对于某些论文已经作了很深刻的理解，因此有些算法直接引用了他们的知乎文章。如有处理不当的地方，请联系我。）
• code:对应项目代码

classification

• [x] mnist paper read / code
• [x] vggNet paper read / code
• [x] resnet paper read / code
• [x] resNeXt paper read / code
• [x] resNeSt paper read / code
• [x] coatNet paper read / code
• [x] convNext paper read / code
• [x] seNet paper read / code
• [x] skNet paper read / code
• [x] TransFG paper read / code
• [x] RepVGG paper read / code
• [x] efficientNet paper read / code
• [x] shuffleNet paper read / code
• [x] vision_transformer paper read / code
• [x] swin-transformer paper read / code
• [x] googleNet paper read / code
• [ ] inception
• [ ] denseNet
• [ ] CBAM
• [ ] mobileNet
• [ ] Xception
• [ ] SqueezeNet

detection

• [x] FPN(实现resnet50 + fpn) paper read / code
• [x] Faster-rcnn paper read / code
• [x] yolov7 paper read / code
• [x] RetinaNet (包含focal_loss) code
• [x] YOLOV5 V5.0 (实现注释，更新pt->onnx代码) code
• [x] yolox (修改了voc数据读取方式) code
• [x] FCOS code
• [ ] yoloF
• [ ] yoloR
• [ ] detr
• [ ] ssd
• [ ] Mask-rcnn
• [ ] Cascade-rcnn
• [ ] SPPNet
• [ ] CenterNet
• [ ] RepPoints
• [ ] OTA
• [ ] ATSS

segmentation

• [x] U-Net paper read / code
• [x] HR-Net-Seg paper read / code
• [x] DeepLabv3 paper read / code
• [x] DeepLabv3Plus paper read / code
• [x] few_shot segmentation code
• [x] FCN code
• [ ] Mask-rcnn
• [ ] Cascade-rcnn
• [ ] UNet++
• [ ] PSPNet
• [ ] Segmenter

metric_learning

• [x] BDB (用于图像检索) code
• [x] Happy-Whale (鲸鱼竞赛检索baseline) code

self-supervised

• [x] MAE (实现VIT+MAE) paper read / code
• [x] SupCon (实现自对比学习+t-SNE可视化+swa) paper read / code
• [ ] MoCo
• [ ] SwAV

deep_stereo

• [x] Real_time_self_adaptive_deep_stereo (实时双目里立体匹配，细节待完善) code

other

• [x] label_convert (三种不同标注文件之间的转换以及box可视化) paper read / code
• [x] normalization (BN、LN、IN、GN、SN图解) paper read / code
• [x] DDP (模型分布式计算) paper read / code
• [x] tensorboard test (可视化网络，图片，训练过程以及卷积核) / code
• [x] load weights test (权重部分加载) / code
• [x] visual weights map test (特征图、卷积核可视化分析) / code
• [x] class_Activation_Map_Visual (可视化CNN的类激活图) / code
• [x] deploy (pytorch模型转onnx，支持自定义算子 示例) / code