Regularize Callback

Perform Group Regularization in fastai Callback system

from fasterai.core.criteria import *
from fasterai.regularize.all import *
from fastai.vision.all import *

Get your data

path = untar_data(URLs.PETS)
files = get_image_files(path/"images")

def label_func(f): return f[0].isupper()

dls = ImageDataLoaders.from_name_func(path, files, label_func, item_tfms=Resize(64))

Train a model without Regularization as a baseline

learn = vision_learner(dls, resnet18, metrics=accuracy)
learn.unfreeze()

learn.fit_one_cycle(3)

epoch	train_loss	valid_loss	accuracy	time
0	0.675158	0.390553	0.845061	00:06
1	0.334588	0.219738	0.901894	00:03
2	0.178833	0.194565	0.919486	00:04

Create the RegularizeCallback

reg_cb = RegularizeCallback('filter', wd=0.0001)

learn = vision_learner(dls, resnet18, metrics=accuracy)
learn.unfreeze()

learn.fit_one_cycle(3, cbs=reg_cb)

epoch	train_loss	valid_loss	accuracy	time
0	16.783606	15.932514	0.855210	00:04
1	14.619355	13.426625	0.922192	00:03
2	13.077019	12.727697	0.921516	00:04