Image-Classification

Here are repository of Image Classification with state-of-the-art techniques (Fourier Neural Operator, FNO/ U-Net, etc) orororor Transfer Learning Model

I' m now tackling on any renowned algorithm to modeling MNIST, cifar10, food101 classifier. (ImageNet)

Hope you enjoy my works !

Benchmarks on Food 101 datasets: https://tensorboard.dev/experiment/MVceQhnBTXC6d2rGEh9G9w/#scalars

Benchmarks on MNIST classifier

Model	Train Accuracy	Train Loss	Test Accuracy	Test Loss
Fully connected network - After 5 Epochs	0.8923	0.2935	0.8731	0.2432
Convolutional network - After 5 Epochs	0.8860	0.3094	0.9048	0.1954
Residual network - After 5 Epochs	0.9064	0.2610	0.8713	0.3398
Fourier Neural Operator - After 5 Epochs	0.9486	0.1622	0.9455	0.1806
U-Net - After 5 Epochs	0.9867	0.0431	0.9873	0.0436

Demonstration in MNIST:

Demonstration in Cifar10:

---

Citation

@code{Image Classification, 
           author = "Kozak Hou"
           email = "kozak20010716@g.ncu.edu.tw"
           Tel : +886-905804898
           Affiliation = "Department of Space Science and Engineering, National Central University"
     }