Re-run formatter

Author: (Ian Stenbit)

examples/layers/preprocessing/cut_mix_demo.py

@@ -21,7 +21,9 @@ def resize(image, label, num_classes=10):
 
 
 def main():
-    data, ds_info = tfds.load("oxford_flowers102", with_info=True, as_supervised=True)
+    data, ds_info = tfds.load(
+        "oxford_flowers102", with_info=True, as_supervised=True
+    )
     train_ds = data["train"]
 
     num_classes = ds_info.features["label"].num_classes

examples/layers/preprocessing/mix_up_demo.py

@@ -21,7 +21,9 @@ def resize(image, label, num_classes=10):
 
 
 def main():
-    data, ds_info = tfds.load("oxford_flowers102", with_info=True, as_supervised=True)
+    data, ds_info = tfds.load(
+        "oxford_flowers102", with_info=True, as_supervised=True
+    )
     train_ds = data["train"]
 
     num_classes = ds_info.features["label"].num_classes

keras_cv/__init__.py

@@ -12,4 +12,6 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
-from keras_cv import layers, metrics, util
+from keras_cv import layers
+from keras_cv import metrics
+from keras_cv import util

keras_cv/layers/preprocessing/cut_mix.py

@@ -87,8 +87,12 @@ def _cutmix(self, images, labels):
             input_shape[2],
         )
 
-        permutation_order = tf.random.shuffle(tf.range(0, batch_size), seed=self.seed)
-        lambda_sample = CutMix._sample_from_beta(self.alpha, self.alpha, (batch_size,))
+        permutation_order = tf.random.shuffle(
+            tf.range(0, batch_size), seed=self.seed
+        )
+        lambda_sample = CutMix._sample_from_beta(
+            self.alpha, self.alpha, (batch_size,)
+        )
 
         ratio = tf.math.sqrt(1 - lambda_sample)
 
@@ -130,7 +134,9 @@ def _update_labels(self, images, labels, lambda_sample, permutation_order):
         cutout_labels = tf.gather(labels, permutation_order)
 
         lambda_sample = tf.reshape(lambda_sample, [-1, 1])
-        labels = lambda_sample * labels_smoothed + (1.0 - lambda_sample) * cutout_labels
+        labels = (
+            lambda_sample * labels_smoothed + (1.0 - lambda_sample) * cutout_labels
+        )
         return images, labels
 
     def _smooth_labels(self, labels):

keras_cv/layers/preprocessing/mix_up.py

@@ -80,9 +80,13 @@ def call(self, images, labels):
 
     def _mixup(self, images, labels):
         batch_size = tf.shape(images)[0]
-        permutation_order = tf.random.shuffle(tf.range(0, batch_size), seed=self.seed)
+        permutation_order = tf.random.shuffle(
+            tf.range(0, batch_size), seed=self.seed
+        )
 
-        lambda_sample = MixUp._sample_from_beta(self.alpha, self.alpha, (batch_size,))
+        lambda_sample = MixUp._sample_from_beta(
+            self.alpha, self.alpha, (batch_size,)
+        )
         lambda_sample = tf.reshape(lambda_sample, [-1, 1, 1, 1])
 
         mixup_images = tf.gather(images, permutation_order)
@@ -96,7 +100,8 @@ def _update_labels(self, images, labels, lambda_sample, permutation_order):
 
         lambda_sample = tf.reshape(lambda_sample, [-1, 1])
         labels = (
-            lambda_sample * labels_smoothed + (1.0 - lambda_sample) * labels_for_mixup
+            lambda_sample * labels_smoothed
+            + (1.0 - lambda_sample) * labels_for_mixup
         )
 
         return images, labels

keras_cv/metrics/coco/base.py

@@ -137,7 +137,9 @@ def update_state(self, y_true, y_pred, sample_weight=None):
                     false_positives = tf.cast(pred_matches == -1, tf.float32)
 
                     true_positives_sum = tf.math.reduce_sum(true_positives, axis=-1)
-                    false_positives_sum = tf.math.reduce_sum(false_positives, axis=-1)
+                    false_positives_sum = tf.math.reduce_sum(
+                        false_positives, axis=-1
+                    )
 
                     true_positives_update = tf.tensor_scatter_nd_add(
                         true_positives_update, [indices], [true_positives_sum]

keras_cv/metrics/coco/base_test.py

@@ -53,7 +53,8 @@ def test_true_positive_counting_one_true_two_pred(self):
             dtype=tf.float32,
         )
         y_pred = tf.constant(
-            [[[0, 50, 100, 150, 1, 0.90], [0, 0, 100, 100, 1, 1.0]]], dtype=tf.float32
+            [[[0, 50, 100, 150, 1, 0.90], [0, 0, 100, 100, 1, 1.0]]],
+            dtype=tf.float32,
         )
         # note the low iou threshold
         metric = COCOBase(

keras_cv/metrics/coco/iou.py

@@ -35,7 +35,9 @@ def _compute_single_iou(bboxes1, bb2):
     area_bboxes1 = _area(
         bb1[bbox.LEFT], bb1[bbox.RIGHT], bb1[bbox.TOP], bb1[bbox.BOTTOM]
     )
-    area_bb2 = _area(bb2[bbox.LEFT], bb2[bbox.RIGHT], bb2[bbox.TOP], bb2[bbox.BOTTOM])
+    area_bb2 = _area(
+        bb2[bbox.LEFT], bb2[bbox.RIGHT], bb2[bbox.TOP], bb2[bbox.BOTTOM]
+    )
 
     area_union = area_bboxes1 + area_bb2 - area_intersection
     return tf.math.divide_no_nan(area_intersection, area_union)

keras_cv/metrics/coco/numerical_tests/GenerateSamples.ipynb

@@ -426,7 +426,9 @@
     "for annotation in results:\n",
     "    img_id = annotation[\"image_id\"]\n",
     "    bbox = annotation[\"bbox\"]\n",
-    "    bbox = [x for x in bbox] + [int(annotation[\"category_id\"])] + [annotation[\"score\"]]\n",
+    "    bbox = (\n",
+    "        [x for x in bbox] + [int(annotation[\"category_id\"])] + [annotation[\"score\"]]\n",
+    "    )\n",
     "    groups[img_id].append(bbox)\n",
     "\n",
     "imgs = sorted(groups.keys())\n",

keras_cv/metrics/coco/recall_test.py

@@ -17,7 +17,9 @@ def test_recall_area_range_filtering(self):
         k = recall.category_ids.shape[0]
 
         # These would match if they were in the area range
-        y_true = np.array([[[0, 0, 10, 10, 1], [5, 5, 10, 10, 1]]]).astype(np.float32)
+        y_true = np.array([[[0, 0, 10, 10, 1], [5, 5, 10, 10, 1]]]).astype(
+            np.float32
+        )
         y_pred = np.array([[[0, 0, 10, 10, 1, 1.0], [5, 5, 10, 10, 1, 0.9]]]).astype(
             np.float32
         )