This file stores a representative set of saved models from the manuscript "The Effects of Nonlinear Signal on Expression-Based Prediction Performance". To simplify the uploading process, the file was split into chunks using the `split` utility in Linux. They can be joined back together with the command `cat model_weights* > model_weights.tar.gz` These saved files correspond to the weights and optimizer state of models trained on various biological tasks. They can be "rehydrated" using the `load_model` function of the associated model class from this repo: https://github.com/greenelab/linear_signal