ImageDataBunch.from_folder does not see my test set

Question

I’m working on the Food-101 dataset and I’m trying to supply ImageDataBunch.from_folder() with a separate test set but for some reason, it doesn’t recognize it.

ImageDataBunch;

Train: LabelList (60600 items)
x: ImageList
Image (3, 224, 224),Image (3, 224, 224),Image (3, 224, 224),Image (3, 224, 224),Image (3, 224, 224)
y: CategoryList
train,train,train,train,train
Path: images/train;

Valid: LabelList (15150 items)
x: ImageList
Image (3, 224, 224),Image (3, 224, 224),Image (3, 224, 224),Image (3, 224, 224),Image (3, 224, 224)
y: CategoryList
train,train,train,train,train
Path: images/train;

Test: None

When I had this problem previously, I was told that I had to keep my train and test images on the same level. This is my file structure atm:

Food-101
      images
            train
               ...75750 train images
            test
               ...25250 test images
      meta
            classes.txt
            labes.txt
            test.json
            test.txt
            train.json
            train.txt
      README.txt
      license_agreement.txt

And this is basically the code I’m using:

train_image_path = Path('images/train/')
test_image_path = Path('images/test/')
path = Path('../Food-101')

food_names = get_image_files(train_image_path)

file_parse = r'/([^/]+)_\d+\.(png|jpg|jpeg)$'

data = ImageDataBunch.from_folder(train_image_path, test_image_path, valid_pct=0.2, ds_tfms=get_transforms(), size=224)
data.normalize(imagenet_stats)

Answer 1

Specify the test parameter as the name of the folder that contains the tests

e.g.

data = ImageDataBunch.from_folder(path=path, train = "Train", test = "Test", ds_tfms=get_transforms(),size=224,valid_pct=0.3)