UFLDL Exercises

This is my solution to the Unsupervised Feature Learning and Deep Learning (UFLDL) exercises.

Datasets

To minimize the size of the code, I remove all the datasets from the package. Before running the code, you should download the following datasets and unpack to the corresponding directories.

1. Dowload MNIST dataset by running the following command

   ./MNIST/get_mnist.sh
   

2. Download A subset of the STL10 Dataset (stlSubset.zip) unpack to './cnn_exercise'

   ./cnn_exercise/get_stl.sh
   

3. Sampled 8x8 patches from the STL-10 dataset (stl10_patches_100k.zip) unpack to './linear_decoder_exercise'

4. All .mat file should be downloaded from the tutorial site