I am using ImageDataBunch class to create my datasets. All of my code is working fine. Now i have been told that i have to implement KFold Cross validation on my datasets and then train the model accordingly. I am giving below my directory structure.
Two folders namely Train and Valid: Two subfolders under each folder for two classes.
np.random.seed(41)
data = ImageDataBunch.from_folder(path, train=“train”, valid =“Valid”,
ds_tfms=get_transforms(), size=(256,256), bs=32, num_workers=4).normalize()
Please note there are no error messages, my deep learning model is working fine on existing datasets.
All i have to do is to train my model by applying KFold Cross Validation on my datasets. So that each and every piece of my image will act as both training and validation datasets.
Thanks
Kannadasan