Skip to content

Latest commit

 

History

History
13 lines (13 loc) · 2.49 KB

File metadata and controls

13 lines (13 loc) · 2.49 KB
Field Expected Datatype Descriptions
input-path str The path to the Dataset
output-path str Location to flush all outputs
negative-dir list[str] Directory Location to Control Data
positive-dir list[str] Directory Location to Cancer Patients
epoch int Amount of Epochs to train the model
lr float or list[float] Learning Rate, or a List of Learning Rates
change-lr-at float or list[float] Epochs to change Learning Rate at. This should be the same datatype as "lr"
train-split float The proportion of data to be served as the training data
bag-accummulate-loss int The amount of patients to incur a step down the gradient
l2-penalty float Amount of Weight Decay
encoding str The symbolic encoding method