# AI CUP 紀錄
[toc]

## 資料前處理
問題
- 圖片太大 -> 裁剪 or resize
- 準心不在目標上 -> 整張圖做 resize

0925
- 將圖片 resize 為 288*288
- 選擇長寬較短的一邊作為邊長
- 等比例縮放

## build model
https://pytorch.org/vision/stable/models.html#id35

CNN
[link](https://pytorch.org/tutorials/beginner/blitz/cifar10_tutorial.html?highlight=image%20classifier)

VGG
[link
](https://pytorch.org/vision/main/models/vgg.html)

VGG16
[link
](https://wanghao.blog.csdn.net/article/details/117753156?spm=1001.2101.3001.6650.4&utm_medium=distribute.pc_relevant.none-task-blog-2%7Edefault%7ECTRLIST%7ERate-4-117753156-blog-126231856.t5_download_0_7w&depth_1-utm_source=distribute.pc_relevant.none-task-blog-2%7Edefault%7ECTRLIST%7ERate-4-117753156-blog-126231856.t5_download_0_7w&utm_relevant_index=9)

[【小白学PyTorch】13 EfficientNet详解及PyTorch实现](https://blog.csdn.net/qq_34107425/article/details/108791151?spm=1001.2101.3001.6650.5&utm_medium=distribute.pc_relevant.none-task-blog-2%7Edefault%7EBlogCommendFromBaidu%7ERate-5-108791151-blog-104210354.pc_relevant_multi_platform_whitelistv4&depth_1-utm_source=distribute.pc_relevant.none-task-blog-2%7Edefault%7EBlogCommendFromBaidu%7ERate-5-108791151-blog-104210354.pc_relevant_multi_platform_whitelistv4&utm_relevant_index=8)

[一文读懂LeNet、AlexNet、VGG、GoogleNet、ResNet到底是什么？](https://zhuanlan.zhihu.com/p/162881214)

[无需数学背景，读懂 ResNet、Inception 和 Xception 三大变革性架构](https://www.jiqizhixin.com/articles/2017-08-19-4)

[pytorch image classification](https://github.com/bentrevett/pytorch-image-classification)