test-ok / resnet18

Resnet18

This is a homework for AI.

*main.py: The main resnet model and the codes of training.
*resnet18_10000_rate_0_001.pth: The trained model by training 10000 iterations (batchsize=100, epoch=20).
*filter_view.py: Python file of filter visualization.
*feature_map_view.py: Python file of feature map visualization without reconstruction.
*feature_map_reconv.py: Python file of feature map visualization of the layer ‘Conv1’ with reconstruction.
*feature_map_reconv5.py: Python file of feature map visualization of the layer ‘Conv5’ with reconstruction.
*data: the datasets of Cifar10, you can download from the website https://www.cs.toronto.edu/~kriz/cifar.html
*feature_reconv: the images of feature map visualization of the layer ‘Conv1’ with reconstruction.
*feature_reconv5: the images of feature map visualization of the layer ‘Conv5’ with reconstruction.
*runs: the logs of tensorboard