[2/N] Data Sources (#264)

* update kornia version and crop functions

* update notebooks

* [pre-commit.ci] auto fixes from pre-commit.com hooks

for more information, see https://pre-commit.ci

* update kornia to 0.5.1

* update text_classification.ipynb

* Fixes

* Fix

* Fixes

* Clear metadata

* Fix

Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Co-authored-by: Ethan Harris <[email protected]>

.readthedocs.yml

@@ -20,4 +20,4 @@ python:
   version: 3.7
   install:
     - requirements: requirements/docs.txt
-    #- requirements: requirements.txt
+    #- requirements: requirements.txt

Makefile

@@ -25,4 +25,4 @@ clean:
 	rm -rf .pytest_cache
 	rm -rf ./docs/build
 	rm -rf ./docs/source/**/generated
-	rm -rf ./docs/source/api
+	rm -rf ./docs/source/api

flash/vision/classification/transforms.py

@@ -37,8 +37,7 @@ def default_train_transforms(image_size: Tuple[int, int]) -> Dict[str, Callable]
             ),
             "post_tensor_transform": ApplyToKeys(
                 DefaultDataKeys.INPUT,
-                # TODO (Edgar): replace with resize once kornia is fixed
-                K.augmentation.RandomResizedCrop(image_size, scale=(1.0, 1.0), ratio=(1.0, 1.0)),
+                K.geometry.Resize(image_size),
                 K.augmentation.RandomHorizontalFlip(),
             ),
             "per_batch_transform_on_device": ApplyToKeys(
@@ -70,8 +69,7 @@ def default_val_transforms(image_size: Tuple[int, int]) -> Dict[str, Callable]:
             ),
             "post_tensor_transform": ApplyToKeys(
                 DefaultDataKeys.INPUT,
-                # TODO (Edgar): replace with resize once kornia is fixed
-                K.augmentation.RandomResizedCrop(image_size, scale=(1.0, 1.0), ratio=(1.0, 1.0)),
+                K.geometry.Resize(image_size),
             ),
             "per_batch_transform_on_device": ApplyToKeys(
                 DefaultDataKeys.INPUT,

flash_notebooks/custom_task_tutorial.ipynb

@@ -40,18 +40,21 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from typing import Any, List, Tuple\n",
+    "from typing import Any, List, Tuple, Dict\n",
     "\n",
     "import numpy as np\n",
     "import torch\n",
     "from pytorch_lightning import seed_everything\n",
     "from sklearn import datasets\n",
     "from sklearn.model_selection import train_test_split\n",
-    "from torch import nn\n",
+    "from torch import nn, Tensor\n",
     "\n",
     "import flash\n",
     "from flash.data.auto_dataset import AutoDataset\n",
-    "from flash.data.process import Postprocess, Preprocess"
+    "from flash.data.data_source import DataSource\n",
+    "from flash.data.process import Preprocess\n",
+    "\n",
+    "ND = np.ndarray"
    ]
   },
   {
@@ -152,24 +155,43 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "class NumpyRegressionPreprocess(Preprocess):\n",
+    "class NumpyDataSource(DataSource):\n",
     "\n",
-    "    def load_data(self, data: Tuple[np.ndarray, np.ndarray], dataset: AutoDataset) -> List[Tuple[np.ndarray, float]]:\n",
+    "    def load_data(self, data: Tuple[ND, ND], dataset: AutoDataset) -> List[Tuple[ND, float]]:\n",
     "        if self.training:\n",
     "            dataset.num_inputs = data[0].shape[1]\n",
     "        return [(x, y) for x, y in zip(*data)]\n",
     "\n",
-    "    def to_tensor_transform(self, sample: Any) -> Tuple[torch.Tensor, torch.Tensor]:\n",
+    "    def predict_load_data(self, data: ND) -> ND:\n",
+    "        return data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "class NumpyPreprocess(Preprocess):\n",
+    "\n",
+    "    def __init__(self):\n",
+    "        super().__init__(data_sources={\"numpy\": NumpyDataSource()}, default_data_source=\"numpy\")\n",
+    "\n",
+    "    def to_tensor_transform(self, sample: Any) -> Tuple[Tensor, Tensor]:\n",
     "        x, y = sample\n",
     "        x = torch.from_numpy(x).float()\n",
     "        y = torch.tensor(y, dtype=torch.float)\n",
     "        return x, y\n",
     "\n",
-    "    def predict_load_data(self, data: np.ndarray) -> np.ndarray:\n",
-    "        return data\n",
+    "    def predict_to_tensor_transform(self, sample: ND) -> ND:\n",
+    "        return torch.from_numpy(sample).float()\n",
+    "\n",
+    "    def get_state_dict(self) -> Dict[str, Any]:\n",
+    "        return {}\n",
     "\n",
-    "    def predict_to_tensor_transform(self, sample: np.ndarray) -> np.ndarray:\n",
-    "        return torch.from_numpy(sample).float()\n"
+    "    @classmethod\n",
+    "    def load_state_dict(cls, state_dict: Dict[str, Any], strict: bool):\n",
+    "        return cls()"
    ]
   },
   {
@@ -181,15 +203,16 @@
     "class SklearnDataModule(flash.DataModule):\n",
     "\n",
     "    @classmethod\n",
-    "    def from_dataset(cls, x: np.ndarray, y: np.ndarray, batch_size: int = 64, num_workers: int = 0):\n",
+    "    def from_dataset(cls, x: np.ndarray, y: np.ndarray, preprocess: Preprocess, batch_size: int = 64, num_workers: int = 0):\n",
     "\n",
-    "        preprocess = NumpyRegressionPreprocess()\n",
+    "        preprocess = preprocess\n",
     "\n",
     "        x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=.20, random_state=0)\n",
     "\n",
-    "        dm = cls.from_load_data_inputs(\n",
-    "            train_load_data_input=(x_train, y_train),\n",
-    "            test_load_data_input=(x_test, y_test),\n",
+    "        dm = cls.from_data_source(\n",
+    "            \"numpy\",\n",
+    "            train_data=(x_train, y_train),\n",
+    "            test_data=(x_test, y_test),\n",
     "            preprocess=preprocess,\n",
     "            batch_size=batch_size,\n",
     "            num_workers=num_workers\n",
@@ -204,7 +227,8 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "datamodule = SklearnDataModule.from_dataset(*datasets.load_diabetes(return_X_y=True))"
+    "x, y = datasets.load_diabetes(return_X_y=True)\n",
+    "datamodule = SklearnDataModule.from_dataset(x, y, NumpyPreprocess())"
    ]
   },
   {
@@ -350,25 +374,7 @@
    ]
   }
  ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "Python 3",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.8.5"
-  }
- },
+ "metadata": {},
  "nbformat": 4,
  "nbformat_minor": 4
 }