This is a pytorch implementation of the paper Unsupervised Domain Adaptation by Backpropagation

Environment

• Pytorch 1.6
• Python 3.8.5

Network Structure

Dataset

First, download target dataset mnist_m from pan.baidu.com fetch code: kjan or Google Drive, and put mnist_m dataset into dataset/mnist_m, the structure is as follows:

--dataset--mnist_m--mnist_m_train
                 |--mnist_m_test
                 |--mnist_m_train_labels.txt
                 |--mnist_m_test_labels.txt
                 |--.gitkeep

Training

Then, run python main.py

Docker

• build image

docker build -t pytorch_dann .

• run docker container

docker run -it --runtime=nvidia \
  -u $(id -u):$(id -g) \
  -v /YOUR/DANN/PROJECT/dataset:/DANN/dataset \
  -v /YOUR/DANN/PROJECT/models:/DANN/models \
  pytorch_dann:latest \
  python main.py