구현: v4 데이터 (실제 하드웨어에서 찍은거) 구현

deep_learning/dataset/inobus_pet_v4/__init__.py

@@ -0,0 +1,3 @@
+"""inobus_pet_v4 dataset."""
+
+from .inobus_pet_v4 import InobusPetV4

deep_learning/dataset/inobus_pet_v4/inobus_pet_v4.py

@@ -0,0 +1,71 @@
+"""inobus_pet_v4 dataset."""
+import itertools
+from pathlib import Path
+
+import tensorflow_datasets as tfds
+
+# TODO(inobus_pet_v4): Markdown description  that will appear on the catalog page.
+_DESCRIPTION = """
+Shape: (2464, 3280, 3). (H, W, C) format.
+
+- train
+    - yes_pet: 330장
+    - no_pet:  50장
+- test
+    - yes_pet: 100장
+    - no_pet:  40장
+"""
+
+# TODO(inobus_pet_v4): BibTeX citation
+_CITATION = """
+"""
+
+# 로컬 inobus_pet_v4 경로.
+_PATH = "/Users/minhoheo/Downloads/inobus_pet_v4"
+
+
+class InobusPetV4(tfds.core.GeneratorBasedBuilder):
+  """DatasetBuilder for inobus_pet_v4 dataset."""
+
+  VERSION = tfds.core.Version('1.0.0')
+  RELEASE_NOTES = {
+      '1.0.0': 'Initial release.',
+  }
+
+  def _info(self) -> tfds.core.DatasetInfo:
+    """Returns the dataset metadata."""
+    return tfds.core.DatasetInfo(
+        builder=self,
+        description=_DESCRIPTION,
+        features=tfds.features.FeaturesDict({
+            # These are the features of your dataset like images, labels ...
+            'image': tfds.features.Image(shape=(2464, 3280, 3)),
+            'label': tfds.features.ClassLabel(names=['no_pet', 'yes_pet']),
+        }),
+        # If there's a common (input, target) tuple from the
+        # features, specify them here. They'll be used if
+        # `as_supervised=True` in `builder.as_dataset`.
+        supervised_keys=('image', 'label'),  # Set to `None` to disable
+        citation=_CITATION,
+    )
+
+  def _split_generators(self, dl_manager: tfds.download.DownloadManager):
+    """Returns SplitGenerators."""
+    path = Path(_PATH)
+
+    return {
+        'train': self._generate_examples(path / 'train'),
+        'test': self._generate_examples(path / 'test')
+    }
+
+  def _generate_examples(self, path):
+    """Yields examples."""
+    # Yields (key, example) tuples from the dataset
+    paths = itertools.chain((path.rglob('*.jpeg')), path.rglob('*.jpg'),
+                            path.rglob('*.JPG'))
+    for image_path in paths:
+      key = image_path.parent.name + '/' + image_path.name
+      yield key, {
+          'image': image_path,
+          'label': image_path.parent.name,
+      }