Preprocessed datasets for MAF experiments

Preprocessed datasets used in the experiments of the paper:

• G. Papamakarios, T. Pavlakou, I. Murray. Masked Autoregressive Flow for Density Estimation. Advances in Neural Information Processing Systems, 2017. [pdf] [bibtex]

Code that runs the experiments can be found at:

https://github.com/gpapamak/maf

How to use the datasets:

1. Downdload the code from the github repository.
2. Download the datasets and unpack.
3. Place the unpacked datasets in the same folder as the code.
4. Make sure the code reads from the location the datasets are saved at.
5. Run the code as described on the github repository.

All datasets uploaded here are preprocessed versions of public datasets. The original datasets are:

• POWER:
  http://archive.ics.uci.edu/ml/datasets/Individual+household+electric+power+consumption

• GAS
  http://archive.ics.uci.edu/ml/datasets/Gas+sensor+array+under+dynamic+gas+mixtures

• HEPMASS
  http://archive.ics.uci.edu/ml/datasets/HEPMASS

• MINIBOONE
  http://archive.ics.uci.edu/ml/datasets/MiniBooNE+particle+identification

• BSDS300
  https://www2.eecs.berkeley.edu/Research/Projects/CS/vision/bsds/

• MNIST
  http://yann.lecun.com/exdb/mnist/

• CIFAR-10
  https://www.cs.toronto.edu/~kriz/cifar.html

None of the above datasets belongs to the creators of this upload. This upload exists to make the experiments in Masked Autoregressive Flow for Density Estimation easy to reproduce.