Hello,
I am using fastai for image segmentation task I have the following snippet that works perfectly fine
src = (SegmentationItemList.from_folder(path_img)
# Load in x data from folder
.split_by_rand_pct(0.2, seed=13)
# Split data into training and validation set
.label_from_func(get_y_fn, classes=codes)
# Label data using the get_y_fn function
)
data = (src.transform(get_transforms(), size=size, tfm_y=True)
.databunch(bs=bs, num_workers=0)
.normalize(imagenet_stats))
I am having errors when I try to add test set in this kind of process(obviously I don’t know),
Another question that I have is what if I have a trained model and want to predict a new set of images to predict, how can I efficiently do that? do I need to create a databunch or what?
any help will be appreciated. Thank you in advance.