TensorFlow implementation of the paper Supervised Representation Learning: Transfer Learning with Deep Autoencoders.
- Numpy
- Tensorflow >= 1.0
- CIFAR-100 Dataset
Run bash setup.sh in terminal to automatically download the CIFAR-100 dataset.
Enter python TLDA.py in bash for fast with default setting.
Use --cifar_path to specify the path of pretrained VGG Net. By default, the model is located under datasets/cifar-100-python directory
Use --outfile to specify file to save the result. By default, the file is "result.csv".
Use --help to acquire more information.