PyTorch Classifiers
PyTorch implementation of the classic classifiers network.
Implemented Classifiers
Results
Accuracy on MNIST
Model |
Parameters |
Batch Size |
Iterations |
Accuracy |
LeNet-5 |
0.0371M |
64 |
10k |
97.8404% |
AlexNet |
58.2991M |
64 |
10k |
99.0048% |
VGG16 |
134.3004M |
16 |
2k |
97.6600% |
Accuracy on CIFAR-10
Model |
Parameters |
Batch Size |
Iterations |
Accuracy |
LeNet-5 |
0.0371M |
64 |
10k |
52.1994% |
AlexNet |
58.3223M |
64 |
10k |
80.5434% |
VGG16 |
134.3015M |
16 |
2k |
54.5800% |
Accuracy on CIFAR-100
Model |
Parameters |
Batch Size |
Iterations |
Accuracy |
LeNet-5 |
0.0448M |
64 |
10k |
15.8340% |
AlexNet |
58.6910M |
64 |
10k |
48.8256% |
VGG16 |
134.6702M |
16 |
10k |
43.1100% |
GoogLeNet |
10.6183M |
32 |
10k |
44.8582% |
GoogLeNet |
10.6183M |
32 |
20k |
53.0152% |
GoogLeNet |
10.6183M |
64 |
10k |
50.1393% |
ResNet-18 |
11.2278M |
64 |
10k |
53.4833% |
Architectures
LeNet
Image Credit
AlexNet
Image Credit
VGG16
Image Credit
GoogLeNet
Image Credit
ResNet
Image Credit
ResNet-18
Image Credit
ResNet-34
Image Credit
Others
This repository is for practice purposes only.