Merge pull request #4 from kabkabm/get_celeba_fix

blackbox.py

@@ -236,13 +236,13 @@ def get_celeba(data_path, test_on_dev=True, orig_data=False):
         images: Images of the dataset.
         labels: Labels of the loaded images.
     """
-    dev_name = 'dev'
+    dev_name = 'val'
     if not test_on_dev:
         dev_name = 'test'
     ds = CelebA(attribute=FLAGS.attribute)
     ds.load()
     ds_test = CelebA(attribute=FLAGS.attribute)
-    ds_test.load(split=dev_name, transform_type=1)
+    ds_test.load(split=dev_name)
     train_labels = ds.labels
     test_labels = ds_test.labels
 

datasets/dataset.py

@@ -137,7 +137,7 @@ def __getitem__(self, index):
         if isinstance(index, int):
             return self._get_image(self.filepaths[index])
         # Case of a slice or array of indices.
-        elif isinstance(index, slice) or isinstance(index, np.ndarray):
+        elif isinstance(index, slice):
             if isinstance(index, slice):
                 if index.start is None:
                     index = range(index.stop)
@@ -146,9 +146,16 @@ def __getitem__(self, index):
                 else:
                     index = range(index.start, index.stop, index.step)
             return np.array(
-                [self._get_image(self.filepaths[i]) for i in index])
+                [self._get_image(self.filepaths[i]) for i in index]
+            )
         else:
-            raise TypeError("Index must be an integer or a slice.")
+            try:
+                inds = [int(i) for i in index]
+                return np.array(
+                    [self._get_image(self.filepaths[i]) for i in inds]
+                )
+            except TypeError:
+                raise TypeError("Index must be an integer, a slice, a container or an integer generator.")
 
     def get_subset(self, indices):
         """Gets a subset of the images
@@ -167,6 +174,14 @@ def get_subset(self, indices):
         else:
             raise TypeError("Index must be an integer or a slice.")
 
+    @property
+    def shape(self):
+        return tuple([None] + list(self._get_image(self.filepaths[0]).shape))
+
+    @property
+    def dtype(self):
+        return self._get_image(self.filepaths[0]).dtype
+
 
 class PickleLazyDataset(LazyDataset):
     """This dataset is a lazy dataset for working with saved pickle files

requirements.txt

@@ -1,7 +1,8 @@
 numpy==1.14.2
 scipy==1.0.1
-tensorflow-gpu==1.7.0
+requests==2.20.0
 keras==2.1.5
 opencv-python==3.4.0.12
 scikit-image==0.13.1
-matplotlib==2.1.2
+matplotlib==2.1.2
+tqdm=4.28.1