Thank you for sharing the codes for Pathomic Fusion.
I went through the Mobadersany et al dataset (pnas_splits.csv) and saw that each of the splits only contains the "train" and "test" labels (no "validation" labels). So I was wondering what Pathomic fusion uses as validation data.
In the train_cv.py, within the train function (line 45), there is a call to the test function (line 88) for validating the model, which seems to be the same as the call for testing the model (line 49). So I was wondering if the data used for model validation is same as the data used for testing the model. I might be mistaken but please share your thoughts.
Hi Richard,
Thank you for sharing the codes for Pathomic Fusion.
I went through the Mobadersany et al dataset (pnas_splits.csv) and saw that each of the splits only contains the "train" and "test" labels (no "validation" labels). So I was wondering what Pathomic fusion uses as validation data.
In the train_cv.py, within the train function (line 45), there is a call to the test function (line 88) for validating the model, which seems to be the same as the call for testing the model (line 49). So I was wondering if the data used for model validation is same as the data used for testing the model. I might be mistaken but please share your thoughts.
Thank you.