Applying transforms to fastai v2 vision

Question

In fastai v2 i am trying to add image augmentations

So

tfms = aug_transforms(do_flip = True,
                                 flip_vert=True, 
                                 max_lighting=0.1, 
                                 )
data = ImageDataLoaders.from_df(df,bs=5,item_tfms=tfms,folder=path_to_data)

this give output

Could not do one pass in your dataloader, there is something wrong in it

And when i do

data.show_batch()

it give

RuntimeError: "check_uniform_bounds" not implemented for 'Byte'

How to resolve

Answer 1

I didn't try the do_flip transformation, but what worked for me was to apply them not as item_tfms but as batch_tfms:

item_tfms = [ Resize((200, 150), method='squish')]
 
batch_tfms = [Brightness(max_lighting = 0.3, p = 0.4),
    Contrast(max_lighting = 0.6, p = 0.4),
    Saturation(max_lighting = 0.75, p = 0.4)]
 
db = DataBlock(blocks = (ImageBlock, CategoryBlock),
                  get_items = get_image_files,
                  splitter = RandomSplitter(valid_pct=0.2, seed=42),
                  item_tfms=item_tfms,
                  batch_tfms=batch_tfms,
                  get_y = parent_label)

You can then feed the DataBlock into a DataLoader like in the fastbook tutorial

Answer 2

I got this from fastai docs. Adding related to question stuff you can check everything here Related to augs

    class AlbumentationsTransform(RandTransform):
        "A transform handler for multiple `Albumentation` transforms"
        split_idx,order=None,2
        def __init__(self, train_aug, valid_aug): store_attr()
        
        def before_call(self, b, split_idx):
            self.idx = split_idx
        
        def encodes(self, img: PILImage):
            if self.idx == 0:
                aug_img = self.train_aug(image=np.array(img))['image']
            else:
                aug_img = self.valid_aug(image=np.array(img))['image']
            return PILImage.create(aug_img)

Basically this is all you need

For setting diff augs use

    def get_train_aug(): return albumentations.Compose([
                albumentations.RandomResizedCrop(224,224),
                albumentations.Transpose(p=0.5),
    ])
    
    def get_valid_aug(): return albumentations.Compose([
        albumentations.CenterCrop(224,224, p=1.),
        albumentations.Resize(224,224)
    ], p=1.)

And then

    item_tfms = [Resize(256), AlbumentationsTransform(get_train_aug(), get_valid_aug())]

    dls = ImageDataLoaders.from_name_func(
        path, get_image_files(path), valid_pct=0.2, seed=42,
        label_func=is_cat, item_tfms=item_tfms)
    dls.train.show_batch(max_n=4)

Enjoy