Create medmnist.py

medmnist.py

@@ -0,0 +1,115 @@
+import datasets
+import numpy as np
+
+
+class MedMNISTConfig(datasets.BuilderConfig):
+    def __init__(self, variant, **kwargs):
+        super(MedMNISTConfig, self).__init__(version=datasets.Version("1.0.0", ""), **kwargs)
+        self.variant = variant
+
+
+class MedMNIST(datasets.GeneratorBasedBuilder):
+    """MedMNIST, a large-scale MNIST-like collection of standardized biomedical images, including 12 datasets for 2D and 6 datasets for 3D.
+    """
+    BUILDER_CONFIGS = [
+        # 2D Datasets
+        MedMNISTConfig(name="pathmnist", variant="pathmnist"),
+        MedMNISTConfig(name="chestmnist", variant="chestmnist"),
+        MedMNISTConfig(name="dermamnist", variant="dermamnist"),
+        MedMNISTConfig(name="octmnist", variant="octmnist"),
+        MedMNISTConfig(name="pneumoniamnist", variant="pneumoniamnist"),
+        MedMNISTConfig(name="retinamnist", variant="retinamnist"),
+        MedMNISTConfig(name="breastmnist", variant="breastmnist"),
+        MedMNISTConfig(name="bloodmnist", variant="bloodmnist"),
+        MedMNISTConfig(name="tissuemnist", variant="tissuemnist"),
+        MedMNISTConfig(name="organamnist", variant="organamnist"),
+        MedMNISTConfig(name="organcmnist", variant="organcmnist"),
+        MedMNISTConfig(name="organsmnist", variant="organsmnist"),
+        # 3D Datasets
+        MedMNISTConfig(name="organmnist3d", variant="organmnist3d"),
+        MedMNISTConfig(name="nodulemnist3d", variant="nodulemnist3d"),
+        MedMNISTConfig(name="adrenalmnist3d", variant="adrenalmnist3d"),
+        MedMNISTConfig(name="fracturemnist3d", variant="fracturemnist3d"),
+        MedMNISTConfig(name="vesselmnist3d", variant="vesselmnist3d"),
+        MedMNISTConfig(name="synapsemnist3d", variant="synapsemnist3d"),
+    ]
+
+    def _info(self):
+        variant = self.config.variant
+        num_classes = {
+            "pathmnist": 9,
+            "chestmnist": 14,
+            "dermamnist": 7,
+            "octmnist": 4,
+            "pneumoniamnist": 2,
+            "retinamnist": 5,
+            "breastmnist": 2,
+            "bloodmnist": 8,
+            "tissuemnist": 8,
+            "organamnist": 11,
+            "organcmnist": 11,
+            "organsmnist": 11,
+            "organmnist3d": 11,
+            "nodulemnist3d": 2,
+            "adrenalmnist3d": 2,
+            "fracturemnist3d": 3,
+            "vesselmnist3d": 2,
+            "synapsemnist3d": 2,
+        }.get(variant, 0)
+
+        if variant == "chestmnist":     # multi-label instead of multi-class
+            label_feature = datasets.Sequence(datasets.Value("int8"))
+        else:
+            label_feature = datasets.ClassLabel(num_classes=num_classes)
+
+        return datasets.DatasetInfo(
+            description=f"MedMNIST variant: {variant} dataset.",
+            features=datasets.Features(
+                {
+                    "image": datasets.Array3D(shape=(28, 28, 28), dtype="uint8")
+                    if '3d' in variant else datasets.Image(),
+                    "label": label_feature,
+                }
+            ),
+            supervised_keys=("image", "label"),
+            homepage="https://medmnist.com/",
+            license="CC BY 4.0",
+            citation="""@article{medmnistv2,
+                        title={MedMNIST v2-A large-scale lightweight benchmark for 2D and 3D biomedical image classification},
+                        author={Yang, Jiancheng and Shi, Rui and Wei, Donglai and Liu, Zequan and Zhao, Lin and Ke, Bilian and Pfister, Hanspeter and Ni, Bingbing},
+                        journal={Scientific Data},
+                        volume={10},
+                        number={1},
+                        pages={41},
+                        year={2023},
+                        publisher={Nature Publishing Group UK London}
+                    }""",
+        )
+
+    def _split_generators(self, dl_manager):
+        variant = self.config.variant
+        url = f"https://zenodo.org/records/10519652/files/{variant}.npz?download=1"
+        file_path = dl_manager.download(url)
+
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                gen_kwargs={"file_path": file_path, "split": "train"},
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split.TEST,
+                gen_kwargs={"file_path": file_path, "split": "test"},
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split.VALIDATION,
+                gen_kwargs={"file_path": file_path, "split": "val"},
+            ),
+        ]
+
+    def _generate_examples(self, file_path, split):
+        data = np.load(file_path)
+        images = data[f"{split}_images"]
+        labels = data[f"{split}_labels"].squeeze()
+
+        for idx, (image, label) in enumerate(zip(images, labels)):
+            yield idx, {"image": image, "label": label}