Hello,
I’m trying to load in some extra transforms for an object detection problem using image points, and I’m getting an error when trying to make my databunch:
tfms = get_transforms(flip_vert=False, max_rotate=10, max_zoom=1, max_warp=0.05, xtra_tfms=[dihedral_affine, flip_lr])
data = (PointsItemList.from_folder(path)
.split_by_folder()
.label_from_func(get_ctr)
.transform(tfms, tfm_y=True)
.databunch(bs=64).normalize(imagenet_stats)
)
The error:
You can deactivate this warning by passing no_check=True.
/home/waydegg/anaconda3/envs/fastai-course-v3/lib/python3.7/site-packages/fastai/basic_data.py:205: UserWarning: There seems to be something wrong with your dataset, can’t access self.train_ds[i] for all i in [440, 544, 636, 488, 767, 499, 426, 158, 46, 83, 518, 662, 653, 143, 79, 216, 539, 226, 685, 296, 747, 642, 427, 3, 597, 259, 58, 441, 50, 262, 505, 728, 552, 467, 745, 640, 10, 763, 771, 335, 59, 711, 163, 694, 299, 279, 615, 738, 785, 180, 210, 410, 566, 141, 82, 203, 118, 592, 34, 302, 162, 489, 315, 202]
warn(f"There seems to be something wrong with your dataset, can’t access self.train_ds[i] for all i in {idx}")
Not sure where to look to fix this problem, any insight would be appreciated.
Thanks,
Wayde