The fastai deep learning library
resize_images
creates folder structure at dest
when recurse=True
(#3692)DistributedDL
(#3637), thanks to @tmabraham
channels_last
cb (#3634), thanks to @tcapelle
self.loss_func
to device if it is an instance on nn.Module (#3395), thanks to @arampacha
load_model
and let LRFind be ran in a distributed setup (#3689), thanks to @muellerzr
Tokenizer.get_lengths
(#3667), thanks to @karotchykau
load_learner
with cpu=False
doesn't respect the current cuda device if model exported on another; fixes #3656 (#3657), thanks to @ohmeow
learner.fine_tune(1)
(#3283)cnn_learner
to vision_learner
since we now support models other than CNNs too (#3625)setup.py
for PyPI (#3604), thanks to @nzw0301
loss_not_reduced
work with DiceLoss (#3583), thanks to @hiromis
lr_find()
at different times (#3295)at_end
feature to SaveModelCallback
(#3296), thanks to @tmabraham