Close
Table 3: Deep learning hyperparameter settings held constant for all experiments