A simple project on the MNIST dataset using deeplearning
- DNN
- Lenet
- Vggnet
- Resnet
To train the model, please first make sure you have create two folders called datasets and checkpoints. Maybe using the following command in jupyter:
!mkdir checkpoints datasets
The following packages are necessary:
- pytorch
- torchvision
- TensorboardX
If you want to run this project with you GPU(s), cudatoolkit is also needed.
you can first run the example script in jupyter:
!python train.py --model lenet --lr 0.0001 --epoch 20
To know more about this script using:
!python train.py -h