This repository is the implementation of some classical Convolutional Neural Networks for Fashion-MNIST in PyTorch. Please feel free to ask anything!
- LeNet, Paper Link
- AlexNet, Paper Link
- VGGNet, Paper Link
- InceptionNet, Paper Link
- ResNet, Paper Link
python 3.9.11
cuda 11.0.3
cudnn 8.0.5
torch 1.7.1
- Install pytorch:
pip install torch==1.7.1+cu110 torchvision==0.8.2+cu110 torchaudio==0.7.2 -f https://download.pytorch.org/whl/torch_stable.html
- Install dependencies:
pip install -r requirements.txt
- Download the Fashion-MNIST then put the unzipped file into the data folder
- Download the best.pth.tar from corresponding pre-trained model parameter folder (seeing the table in Working Record) and put the best.pth.tar file into the model folder
- Choose the model you need in Line125-129 of train.py and Check the hyper parameters
.
|-- config/
|-- data/
|-- model/
|-- resources/
|-- train.py
|-- requirements.txt
|-- .gitignore
|-- LICENSE
|-- README.md
State | Network | Train Epochs / Time | Best Test Acc | Model Param |
---|---|---|---|---|
☑ 09 June, 2022 | LeNet | 30 / 18m-29s | 0.9016 | LeNet-best.pth.tar |
☑ 11 June, 2022 | AlexNet1 | 30 / 35m-39s | 0.9219 | AlexNet-best.pth.tar |
☑ 13 June, 2022 | VGGNet2 | 30 / 30m-35s | 0.9135 | VGGNet-best.pth.tar |
☑ 22 June, 2022 | InceptionNet | 30 / 416m-54s | 0.9274 | InceptionNet-best.pth.tar |
☑ 16 June, 2022 | ResNet | 30 / 82m-24s | 0.9340 | ResNet-best.pth.tar |
- Device on GPU: NVIDIA GeForce GTX 1070, CPU: Intel i7-7700K, RAM: 32GB and Win10 System.
- Training loss curves can be seen in Loss Curves folder.
- 1 Considering the image size of Fashion-MNIST, here in AlexNet has some tiny differences with the original AlexNet Framework, and we don't take the seperate group and LRN structure as well.
- 2 Considering the image size of Fashion-MNIST, here in VGGNet16, we delete the last two block of convolutional layers, which are layer eight to thirteen.
- LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324.
- Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems, 25.
- Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.
- Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., ... & Rabinovich, A. (2015). Going deeper with convolutions. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 1-9).
- He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778).