AttributeError: 'Learner' object has no attribute 'pred'

[turgut090]

Hi. I am getting the following error:

from upit.data.unpaired import *
from upit.models.cyclegan import *
from upit.train.cyclegan import *
from fastai.vision.all import *

horse2zebra = untar_data('https://people.eecs.berkeley.edu/~taesung_park/CycleGAN/datasets/horse2zebra.zip')

folders = horse2zebra.ls().sorted()
trainA_path = folders[2]
trainB_path = folders[3]
testA_path = folders[0]
testB_path = folders[1]

dls = get_dls(trainA_path, trainB_path,num_A=100)
cycle_gan = CycleGAN(3,3,64)
learn = cycle_learner(dls, cycle_gan,show_img_interval=1)
learn.show_training_loop()
learn.lr_find()

AttributeError: 'Learner' object has no attribute 'pred'

[tmabraham]

Thank you for your interest in my package. I tested the code in a Colab notebook (installing UPIT in the first cell), and I don't get any error. What version of fastai are you using (check fastai.__version__)?

[turgut090]

>>> fastai.__version__
'2.0.16'

[tmabraham]

@henry090 Very odd... I have the same version of fastai and it works.

Please send over a full traceback of the error.

Also, check this colab notebook and see if you can find any meaningful differences between what you do and what I did. I created the notebook using the same code you provided.

[turgut090]

Traceback (most recent call last):
  File "<string>", line 2, in <module>
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/callback/schedule.py", line 228, in lr_find
    with self.no_logging(): self.fit(n_epoch, cbs=cb)
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastcore/logargs.py", line 56, in _f
    return inst if to_return else f(*args, **kwargs)
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 207, in fit
    self._with_events(self._do_fit, 'fit', CancelFitException, self._end_cleanup)
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 155, in _with_events
    try:       self(f'before_{event_type}')       ;f()
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 197, in _do_fit
    self._with_events(self._do_epoch, 'epoch', CancelEpochException)
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 155, in _with_events
    try:       self(f'before_{event_type}')       ;f()
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 191, in _do_epoch
    self._do_epoch_train()
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 183, in _do_epoch_train
    self._with_events(self.all_batches, 'train', CancelTrainException)
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 155, in _with_events
    try:       self(f'before_{event_type}')       ;f()
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 161, in all_batches
    for o in enumerate(self.dl): self.one_batch(*o)
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 179, in one_batch
    self._with_events(self._do_one_batch, 'batch', CancelBatchException)
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 157, in _with_events
    finally:   self(f'after_{event_type}')        ;final()
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 133, in __call__
    def __call__(self, event_name): L(event_name).map(self._call_one)
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastcore/foundation.py", line 342, in map
    def map(self, f, *args, gen=False, **kwargs): return self._new(map_ex(self, f, *args, gen=gen, **kwargs))
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastcore/foundation.py", line 202, in map_ex
    return list(res)
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastcore/foundation.py", line 185, in __call__
    return self.fn(*fargs, **kwargs)
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 137, in _call_one
    [cb(event_name) for cb in sort_by_run(self.cbs)]
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/learner.py", line 137, in <listcomp>
    [cb(event_name) for cb in sort_by_run(self.cbs)]
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/fastai/callback/core.py", line 44, in __call__
    if self.run and _run: res = getattr(self, event_name, noop)()
  File "/home/turgut/.local/share/r-miniconda/envs/r-reticulate/lib/python3.6/site-packages/upit/train/cyclegan.py", line 122, in after_batch
    fake_A, fake_B = self.learn.pred[0].detach(), self.learn.pred[1].detach()
AttributeError: 'Learner' object has no attribute 'pred'

[turgut090]

A new environment with fresh installation still throws this error. Interesting. However, it works in Colab.

[tmabraham]

@henry090 Have you tried running regular fastai code and checked if any error comes? For example, the metric functionality of fastai also uses learn.pred (see here) so I would expect a similar error if you use simple metrics and training as well.

[turgut090]

Works fine:

>>> from fastai.vision.all import *
>>> from fastai.vision.gan import *
>>> 
>>> path = 'oxford-iiit-pet'
>>> path_anno = 'oxford-iiit-pet/annotations'
>>> path_img = 'oxford-iiit-pet/images'
>>> fnames = get_image_files(path_img)
>>> fnames
(#7390) [Path('oxford-iiit-pet/images/newfoundland_90.jpg'),Path('oxford-iiit-pet/images/american_pit_bull_terrier_129.jpg'),Path('oxford-iiit-pet/images/wheaten_terrier_140.jpg'),Path('oxford-iiit-pet/images/havanese_1.jpg'),Path('oxford-iiit-pet/images/Birman_182.jpg'),Path('oxford-iiit-pet/images/english_setter_174.jpg'),Path('oxford-iiit-pet/images/pug_17.jpg'),Path('oxford-iiit-pet/images/scottish_terrier_185.jpg'),Path('oxford-iiit-pet/images/boxer_191.jpg'),Path('oxford-iiit-pet/images/english_cocker_spaniel_40.jpg')...]
>>> 
>>> dls = ImageDataLoaders.from_name_re(
...     path, fnames, pat=r'(.+)_\d+.jpg$', item_tfms=Resize(460), bs=20,
...     batch_tfms=[*aug_transforms(size=224, min_scale=0.75), Normalize.from_stats(*imagenet_stats)],
...     device='cuda')
>>> 
>>> cnn = cnn_learner(dls, resnet18, metrics=[accuracy,error_rate])
>>> cnn.fit_one_cycle(2)
epoch     train_loss  valid_loss  accuracy  error_rate  time    
0         0.869473    0.373598    0.878214  0.121786    00:23     
1         0.542021    0.312157    0.895805  0.104195    00:22 

[turgut090]

I found out why. It is weird. But a load size fixed the problem.

 get_dls(trainA_path, trainB_path, num_A = 100, load_size = 100,crop_size = 100,bs=4) # load size and crop size

epoch   train_loss   id_loss_A   id_loss_B   gen_loss_A   gen_loss_B   cyc_loss_A   cyc_loss_B   D_A_loss   D_B_loss   time  
------  -----------  ----------  ----------  -----------  -----------  -----------  -----------  ---------  ---------  ------
0       11.219095    1.689572    1.781705    0.425382     0.400838     3.470019     3.671166     0.367610   0.367610   00:08 
1       10.067577    1.352973    1.465599    0.354002     0.349632     2.821623     3.051174     0.254647   0.254647   00:08 
2       9.473677     1.291322    1.370914    0.364927     0.374129     2.676564     2.848548     0.236779   0.236779   00:08 
3       9.007480     1.182500    1.285101    0.384088     0.418259     2.485179     2.705824     0.222265   0.222265   00:08 
4       8.644550     1.099371    1.273532    0.385339     0.413444     2.304044     2.691824     0.217611   0.217611   00:08 
5       8.281004     1.058014    1.214808    0.372960     0.412336     2.192252     2.531816     0.239461   0.239461   00:08 
6       7.982881     0.992101    1.175483    0.412752     0.418220     2.028669     2.515495     0.220087   0.220087   00:08 
7       7.642418     0.916930    1.150054    0.372994     0.443032     1.892262     2.369661     0.205539   0.205539   00:08 
8       7.336694     0.873163    1.128663    0.369428     0.436099     1.799056     2.310677     0.203636   0.203636   00:08 
9       7.096091     0.841627    1.099711    0.394596     0.417606     1.700073     2.295223     0.205812   0.205812   00:08 

Instead of CUDA out of memory, it threw another strange error. Thanks for your time! I am closing the issue.