Autoencoder in fastai

Question

I'm trying to build an autoencoder with fast.ai version 1.0.52 and struggling with how to set labels to be equal to original images. I was following this blog post: https://alanbertl.com/autoencoder-with-fast-ai/

I replaced ImageItemList in the original code with ImageList since it was changed in the latest fastai versions.

%reload_ext autoreload
%autoreload 2
%matplotlib inline

from fastai.imports import *
from fastai.vision import *
from fastai.data_block import *
from fastai.basic_train import *

import pandas as pd

x = np.random.randint(256, size=(1000, 16384))
x = x/255
x = x.reshape(-1,128,128)
x = np.stack([x,x,x],1)
x.shape

class ArraysImageList(ImageList,FloatList):
    def __init__(self, items:Iterator, log:bool=False, **kwargs):
        if isinstance(items, ItemList):
            items = items.items
        super(FloatList,self).__init__(items,**kwargs)

    def get(self,i):
        return Tensor(super(FloatList,self).get(i).astype('float32'))

x_il = ArraysImageList(x)
x_ils = x_il.split_by_rand_pct()
lls = x_ils.label_from_lists(x_ils.train, x_ils.valid)

Here's the error message I get.

---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-33-cbada9e18af9> in <module>
----> 1 lls = x_ils.label_from_lists(x_ils.train, x_ils.valid)

~/.local/lib/python3.6/site-packages/fastai/data_block.py in label_from_lists(self, train_labels, valid_labels, label_cls, **kwargs)
    484         self.valid = self.valid._label_list(x=self.valid, y=self.train.y.new(valid_labels, **kwargs))
    485         self.__class__ = LabelLists
--> 486         self.process()
    487         return self
    488 

~/.local/lib/python3.6/site-packages/fastai/data_block.py in process(self)
    520         "Process the inner datasets."
    521         xp,yp = self.get_processors()
--> 522         for ds,n in zip(self.lists, ['train','valid','test']): ds.process(xp, yp, name=n)
    523         #progress_bar clear the outputs so in some case warnings issued during processing disappear.
    524         for ds in self.lists:

~/.local/lib/python3.6/site-packages/fastai/data_block.py in process(self, xp, yp, name)
    683     def process(self, xp:PreProcessor=None, yp:PreProcessor=None, name:str=None):
    684         "Launch the processing on `self.x` and `self.y` with `xp` and `yp`."
--> 685         self.y.process(yp)
    686         if getattr(self.y, 'filter_missing_y', False):
    687             filt = array([o is None for o in self.y.items])

~/.local/lib/python3.6/site-packages/fastai/data_block.py in process(self, processor)
     73         if processor is not None: self.processor = processor
     74         self.processor = listify(self.processor)
---> 75         for p in self.processor: p.process(self)
     76         return self
     77 

~/.local/lib/python3.6/site-packages/fastai/data_block.py in process(self, ds)
    334 
    335     def process(self, ds):
--> 336         if self.classes is None: self.create_classes(self.generate_classes(ds.items))
    337         ds.classes = self.classes
    338         ds.c2i = self.c2i

~/.local/lib/python3.6/site-packages/fastai/data_block.py in generate_classes(self, items)
    391         for c in items: classes = classes.union(set(c))
    392         classes = list(classes)
--> 393         classes.sort()
    394         return classes
    395 

RuntimeError: bool value of Tensor with more than one value is ambiguous

Ultimately, I want to read images using a dataframe with image paths. So I also tried the following:

import sklearn

cv = sklearn.model_selection.GroupKFold(n_splits=5)
train_inds, valid_inds = next(cv.split(iso_image_df.group, groups=iso_image_df.group))

img_lists = (ImageList.from_df(iso_image_df, resized_img_path, cols=0).split_by_idxs(train_inds, valid_inds))
src = img_lists.label_from_lists(img_lists.train, img_lists.valid)

data = (src.databunch(bs = 32).normalize(imagenet_stats))

data.show_batch(rows=3, figsize=(10, 10))

Here I get the following error message:

---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-146-2514de511e64> in <module>
----> 1 data.show_batch(rows=3, figsize=(10, 10))

~/.local/lib/python3.6/site-packages/fastai/basic_data.py in show_batch(self, rows, ds_type, reverse, **kwargs)
    190         #TODO: get rid of has_arg if possible
    191         if has_arg(self.train_ds.y.reconstruct, 'x'):
--> 192             ys = [self.train_ds.y.reconstruct(grab_idx(y, i), x=x) for i,x in enumerate(xs)]
    193         else : ys = [self.train_ds.y.reconstruct(grab_idx(y, i)) for i in range(n_items)]
    194         self.train_ds.x.show_xys(xs, ys, **kwargs)

~/.local/lib/python3.6/site-packages/fastai/basic_data.py in <listcomp>(.0)
    190         #TODO: get rid of has_arg if possible
    191         if has_arg(self.train_ds.y.reconstruct, 'x'):
--> 192             ys = [self.train_ds.y.reconstruct(grab_idx(y, i), x=x) for i,x in enumerate(xs)]
    193         else : ys = [self.train_ds.y.reconstruct(grab_idx(y, i)) for i in range(n_items)]
    194         self.train_ds.x.show_xys(xs, ys, **kwargs)

~/.local/lib/python3.6/site-packages/fastai/data_block.py in reconstruct(self, t, x)
     89     def reconstruct(self, t:Tensor, x:Tensor=None):
     90         "Reconstruct one of the underlying item for its data `t`."
---> 91         return self[0].reconstruct(t,x) if has_arg(self[0].reconstruct, 'x') else self[0].reconstruct(t)
     92 
     93     def new(self, items:Iterator, processor:PreProcessors=None, **kwargs)->'ItemList':

AttributeError: 'Image' object has no attribute 'reconstruct'

Any help is highly appreciated!

Answer 1

The lls are being used to create the databunch.

I've looked at it and given the API change in fastai libs I created the databunch without using the lls that were causing the error:

bs = 64
db = (ImageImageList.from_folder(mnist)
        .split_by_folder()          
        .label_from_func(get_y_fn)
        .databunch(bs=bs,num_workers=4))

EDIT: you'll need the get_y_fn; it is very simply defined

def get_y_fn(x): return x

the lls aren't used for anything else anyway

This should fix your problem, let me know if this worked for you.