Latest ruff

Author: adamjstewart

examples/binary_segmentation_intro.ipynb

@@ -64,14 +64,13 @@
     "# save the model\n",
     "model.save_pretrained(\n",
     "    \"saved-model-dir/unet-with-metadata/\",\n",
-    "\n",
     "    # additional information to be saved with the model\n",
     "    # only \"dataset\" and \"metrics\" are supported\n",
     "    dataset=\"PASCAL VOC\",  # only string name is supported\n",
-    "    metrics={              # should be a dictionary with metric name as key and metric value as value\n",
+    "    metrics={  # should be a dictionary with metric name as key and metric value as value\n",
     "        \"mIoU\": 0.95,\n",
-    "        \"accuracy\": 0.96\n",
-    "    }\n",
+    "        \"accuracy\": 0.96,\n",
+    "    },\n",
     ")"
    ]
   },
@@ -222,13 +221,10 @@
     "# save the model and share it on the HF Hub (https://huggingface.co/models)\n",
     "model.save_pretrained(\n",
     "    \"qubvel-hf/unet-with-metadata/\",\n",
-    "    push_to_hub=True,        # <----------    push the model to the hub\n",
-    "    private=False,           # <----------    make the model private or or public\n",
+    "    push_to_hub=True,  # <----------    push the model to the hub\n",
+    "    private=False,  # <----------    make the model private or or public\n",
     "    dataset=\"PASCAL VOC\",\n",
-    "    metrics={\n",
-    "        \"mIoU\": 0.95,\n",
-    "        \"accuracy\": 0.96\n",
-    "    }\n",
+    "    metrics={\"mIoU\": 0.95, \"accuracy\": 0.96},\n",
     ")\n",
     "\n",
     "# see result here https://huggingface.co/qubvel-hf/unet-with-metadata"
@@ -267,10 +263,7 @@
    "outputs": [],
    "source": [
     "# define a preprocessing transform for image that would be used during inference\n",
-    "preprocessing_transform = A.Compose([\n",
-    "    A.Resize(256, 256),\n",
-    "    A.Normalize()\n",
-    "])\n",
+    "preprocessing_transform = A.Compose([A.Resize(256, 256), A.Normalize()])\n",
     "\n",
     "model = smp.Unet()"
    ]
@@ -367,15 +360,21 @@
     "# You can also save training augmentations to the Hub too (and load it back)!\n",
     "#! Just make sure to provide key=\"train\" when saving and loading the augmentations.\n",
     "\n",
-    "train_augmentations = A.Compose([\n",
-    "    A.HorizontalFlip(p=0.5),\n",
-    "    A.RandomBrightnessContrast(p=0.2),\n",
-    "    A.ShiftScaleRotate(p=0.5),\n",
-    "])\n",
+    "train_augmentations = A.Compose(\n",
+    "    [\n",
+    "        A.HorizontalFlip(p=0.5),\n",
+    "        A.RandomBrightnessContrast(p=0.2),\n",
+    "        A.ShiftScaleRotate(p=0.5),\n",
+    "    ]\n",
+    ")\n",
     "\n",
-    "train_augmentations.save_pretrained(directory_or_repo_on_the_hub, key=\"train\", push_to_hub=True)\n",
+    "train_augmentations.save_pretrained(\n",
+    "    directory_or_repo_on_the_hub, key=\"train\", push_to_hub=True\n",
+    ")\n",
     "\n",
-    "restored_train_augmentations = A.Compose.from_pretrained(directory_or_repo_on_the_hub, key=\"train\")\n",
+    "restored_train_augmentations = A.Compose.from_pretrained(\n",
+    "    directory_or_repo_on_the_hub, key=\"train\"\n",
+    ")\n",
     "print(restored_train_augmentations)"
    ]
   },