The fastai deep learning library
tok_text_col
(#3208), thanks to @muellerzr
download_images()
on Windows (#3196)AttributeError: Can only use .cat accessor with a 'category' dtype
(#3165), thanks to @dreamflasher
cont_cat_split
does not support pandas types (#3156)DataBlock.dataloaders
does not support the advertised "shuffle" argument (#3133)nf
in create_head
based on concat_pool
(#3115), thanks to @muellerzr
NativeMixedPrecision
to default MixedPrecision
(and similar for Learner.to_fp16
); old MixedPrecision
is now called NonNativeMixedPrecision
(#3127)
GradientClip
callback instead of the clip
parameter to use gradient clippingCallback
which has the same name as an attribute no longer raises an exception (#3109)RNNCallback
, but does not require RNNRegularizer
; out
and raw_out
have moved to RNNRegularizer
(#3108)
rnn_cbs
to get all callbacks needed for RNN training, optionally with regularizationrun_after
with order
; do not run after
cbs on exception (#3101)GradientClip
callback (#3107)Flatten
cast to TensorBase
to simplify type compatibility (#3106)TensorBase.register_func
to register types for __torch_function__
(#3097)dynamic
flag for controlling dynamic loss scaling in NativeMixedPrecision
(#3096)to_native_fp32
before predict
; set skipped
in NativeMixedPrecision after NaN from dynamic loss scaling (#3095)nf
in create_head
based on concat_pool
(#3115) thanks to @muellerzrNoneType object has no attribute append
in fastbook chapter 6 BIWI example (#3091)before_batch
to perform the data augmentation. See fastai.callback.mixup
2.1.8
negate_func
->not_
2.1.7
TextDataLoaders.from_df()
returns TypeError: 'float' object is not iterable
(#2978)2.1.6
download_images
(#2983), thanks to @mess-lelouch
config
in create_cnn
and instead pass kwargs directly (#2966), thanks to @borisdayma
TextDataLoaders.from_df()
returns TypeError: 'float' object is not iterable
(#2978)