This is a package that includes MATLAB Software for Convolutional Neural Network Training of All Sky Imager (ASI) Images. The software was used to train ASI images for ability to classify the images on the basis of whether they: (a) contain Equatorial Plasma Bubbles (EPBs), (b) do not contain EPBs, (c) are cloudy/noisy. The software used for this training is the one named "pro8_NNTrain.m". In line 2 of the "pro8_NNTrain.m" program is a string variable (named groupDir) that represents the directory where the images/samples for the groups/classes are stored. In our case. the images are stored in 3 different folders (named "Cloudy", "EPB", and "No_EPB", and containing 1000 images each for those classes). These 3 folders are contained in the folder/directory named groupDir. The other 3 files included in this publication are resulting networks obtained from training 3 sub-sets of the ASI images which are composed as detailed in a paper titled "A Bootstrapping Convolutional Neural Network Technique for Optimizing Automated Detection of Equatorial Plasma Bubbles by Optical All‐Sky Imagers", and submitted to the AGU Earth and Space Science journal for publication.