You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This file contains benchmark AUC performance for predicting each pathlogy on 20% of each dataset.
To reproduce these results use the script scripts/model_calibrate.py. This script will take a long time so it will cache the results so you can work with them after the compute is done. The script also calculates the calibration for a model but those outputs can be ignored for this benchmarking. Below is an example of running different models on the PadChest pc dataset and writing the output as markdown.
python model_calibrate.py pc resnet50-res512-all -mdtable
python model_calibrate.py pc chexpert -mdtable
python model_calibrate.py pc jfhealthcare -mdtable
python model_calibrate.py pc densenet121-res224-all -mdtable