I was reading the paper written by Taesup Kim and Yoshua Bengio] (2016) named Deep Directed Generative Models with Energy-Based Probability Estimation. arXiv:1606.03439
I have been working on a project in which we make predictions with caffe for non-image data. The very first step would be save the data in HDF5 file format.
Hierarchical Data Format (HDF) is an open source file format for storing huge amounts of numerical data. Think of HDF as a file system within a file. It lets you organize the data hierarchically and manage large amount of data very efficiently. But HDF5 data doesn’t support data transformation. You have to pre-process your data with desired transformations before feeding them in.