Fig. 3
From: Learning curves for drug response prediction in cancer cell lines
Two neural network architectures used in the analysis: a single-input network (sNN, 4.254 million trainable parameters) and b multi-input network (mNN, 4.250 million trainable parameters)