Hey guys, I got my dataset from the following kaggle competition:
This file contains .Bmp files as images and when I ran the following piece of code:
data = (ImageList.from_csv(path, 'trainLabels.csv', folder='train', suffix = '.Bmp').split_by_rand_pct(0.2).label_from_df(label_delim=',').databunch().normalize(imagenet_stats))
I get the following error:
/opt/anaconda3/lib/python3.7/site-packages/fastai/basic_data.py:269: UserWarning: It's not possible to collate samples of your dataset together in a batch. Shapes of the inputs/targets: [[torch.Size([3, 58, 30]), torch.Size([3, 37, 38]), torch.Size([3, 62, 42]), torch.Size([3, 33, 25]), torch.Size([3, 83, 76]), torch.Size([3, 102, 67]), torch.Size([3, 104, 69]), torch.Size([3, 62, 30]), torch.Size([3, 77, 46]), torch.Size([3, 73, 62]), torch.Size([3, 47, 38]), torch.Size([3, 20, 25]), torch.Size([3, 69, 27]), torch.Size([3, 42, 25]), torch.Size([3, 51, 30]), torch.Size([3, 65, 24]), torch.Size([3, 49, 43]), torch.Size([3, 63, 44]), torch.Size([3, 377, 202]), torch.Size([3, 109, 67]), torch.Size([3, 66, 50]), torch.Size([3, 50, 35]), torch.Size([3, 105, 73]), torch.Size([3, 131, 129]), torch.Size([3, 17, 16]), torch.Size([3, 105, 71]), torch.Size([3, 41, 49]), torch.Size([3, 31, 6]), torch.Size([3, 94, 42]), torch.Size([3, 70, 32]), torch.Size([3, 84, 34]), torch.Size([3, 69, 50]), torch.Size([3, 101, 27]), torch.Size([3, 22, 18]), torch.Size([3, 89, 111]), torch.Size([3, 71, 82]), torch.Size([3, 43, 33]), torch.Size([3, 31, 26]), torch.Size([3, 89, 35]), torch.Size([3, 116, 80]), torch.Size([3, 24, 15]), torch.Size([3, 128, 61]), torch.Size([3, 23, 18]), torch.Size([3, 162, 93]), torch.Size([3, 161, 64]), torch.Size([3, 53, 27]), torch.Size([3, 74, 35]), torch.Size([3, 126, 112]), torch.Size([3, 88, 51]), torch.Size([3, 68, 83]), torch.Size([3, 18, 10]), torch.Size([3, 78, 62]), torch.Size([3, 103, 118]), torch.Size([3, 51, 46]), torch.Size([3, 81, 47]), torch.Size([3, 108, 113]), torch.Size([3, 50, 38]), torch.Size([3, 51, 55]), torch.Size([3, 24, 15]), torch.Size([3, 55, 54]), torch.Size([3, 100, 58]), torch.Size([3, 69, 57]), torch.Size([3, 97, 75]), torch.Size([3, 56, 46])], [(62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,), (62,)]] warn(message)
Is the dataset just not good to use with fastai?