Making a non image dataset compatible with Fuel

[oema...@uncg.edu]

Hello, 

I have been reading the documentation on making the data compatible with both h5py and with fuel and I was wondering about the "Image features" listed in the documentation. Is it possible to get a dataset that does not have image features in Fuel? (For example, datasets that contain network traffic or weather patterns.) If not, is there a work around like having NaN values for all of the image features values for the dataset?