Hey guys,
I am trying to build a classifier.
the images are in folder and id’s in csv file.
Its multi column multi category classification i think.
- a small look at the above image for csv format
label_cols = ['class_A','class_B','class_C','class_D']
tfms = get_transforms(do_flip=True,max_rotate=0.1,max_lighting=0.15)
data = (ImageList.from_df(df,path,folder='train',suffix='.jpeg',cols='id')
.split_by_rand_pct(0.2)
.label_from_df(cols=label_cols , label_cls=MultiCategoryList , one_hot =True)
.transform(tfms)
.databunch(bs=128)
.normalize(imagenet_stats))
Am I doing something wrong in forming the databunch
But the it’s not working outputs and prediction are something of 0,0,0.
Thanks in advance.