Add USPS HDF5 dataloader and F1 metric implementation

Author: sot176

utils/dataloaders/__init__.py

@@ -0,0 +1,116 @@
+from torch.utils.data import Dataset
+import numpy as np
+import h5py
+from torchvision import transforms
+from PIL import Image
+import torch
+
+
+class USPSH5_Digit_7_9_Dataset(Dataset):
+    """
+    Custom USPS dataset class that loads images with digits 7-9 from an .h5 file.
+
+    Parameters
+    ----------
+    h5_path : str
+        Path to the USPS `.h5` file.
+
+    transform : callable, optional, default=None
+        A transform function to apply on images. If None, no transformation is applied.
+
+    Attributes
+    ----------
+    images : numpy.ndarray
+        The filtered images corresponding to digits 7-9.
+
+    labels : numpy.ndarray
+        The filtered labels corresponding to digits 7-9.
+
+    transform : callable, optional
+        A transform function to apply to the images.
+    """
+
+    def __init__(self, h5_path, mode, transform=None):
+        super().__init__()
+        """
+        Initializes the USPS dataset by loading images and labels from the given `.h5` file.
+
+        Parameters
+        ----------
+        h5_path : str
+            Path to the USPS `.h5` file.
+            
+        transform : callable, optional, default=None
+            A transform function to apply on images.
+        """
+
+        self.transform = transform
+        self.mode = mode
+        self.h5_path = h5_path
+        # Load the dataset from the HDF5 file
+        with h5py.File(self.h5_path, "r") as hf:
+            images = hf[self.mode]["data"][:]
+            labels = hf[self.mode]["target"][:]
+
+        # Filter only digits 7, 8, and 9
+        mask = np.isin(labels, [7, 8, 9])
+        self.images = images[mask]
+        self.labels = labels[mask]
+
+    def __len__(self):
+        """
+        Returns the total number of samples in the dataset.
+
+        Returns
+        -------
+        int
+            The number of images in the dataset.
+        """
+        return len(self.images)
+
+    def __getitem__(self, id):
+        """
+        Returns a sample from the dataset given an index.
+
+        Parameters
+        ----------
+        idx : int
+            The index of the sample to retrieve.
+
+        Returns
+        -------
+        tuple
+            - image (PIL Image): The image at the specified index.
+            - label (int): The label corresponding to the image.
+        """
+        # Convert to PIL Image (USPS images are typically grayscale 16x16)
+        image = Image.fromarray(self.images[id].astype(np.uint8), mode="L")
+        label = int(self.labels[id])  # Convert label to integer
+
+        if self.transform:
+            image = self.transform(image)
+
+        return image, label
+
+
+def main():
+    # Example Usage:
+    transform = transforms.Compose([
+        transforms.Resize((16, 16)),  # Ensure images are 16x16
+        transforms.ToTensor(),
+        transforms.Normalize((0.5,), (0.5,))  # Normalize to [-1, 1]
+    ])
+
+    # Load the dataset
+    dataset = USPSH5_Digit_7_9_Dataset(h5_path="C:/Users/Solveig/OneDrive/Dokumente/UiT PhD/Courses/Git/usps.h5", mode="train", transform=transform)
+    data_loader = torch.utils.data.DataLoader(dataset, batch_size=2, shuffle=True)
+    batch = next(iter(data_loader))  # grab a batch from the dataloader
+    img, label = batch
+    print(img.shape)
+    print(label.shape)
+
+    # Check dataset size
+    print(f"Dataset size: {len(dataset)}")
+
+if __name__ == '__main__':
+    main()

utils/dataloaders/uspsh5_7_9.py

@@ -0,0 +1,96 @@
+import torch.nn as nn
+import torch
+
+
+class F1Score(nn.Module):
+    """
+          F1 Score implementation with direct averaging inside the compute method.
+
+          Parameters
+          ----------
+          num_classes : int
+              Number of classes.
+
+          Attributes
+          ----------
+          num_classes : int
+              The number of classes.
+
+          tp : torch.Tensor
+              Tensor for True Positives (TP) for each class.
+
+          fp : torch.Tensor
+              Tensor for False Positives (FP) for each class.
+
+          fn : torch.Tensor
+              Tensor for False Negatives (FN) for each class.
+          """
+    def __init__(self, num_classes):
+        """
+             Initializes the F1Score object, setting up the necessary state variables.
+
+             Parameters
+             ----------
+             num_classes : int
+                 The number of classes in the classification task.
+
+             """
+
+        super().__init__()
+
+        self.num_classes = num_classes
+
+        # Initialize  variables for True Positives (TP), False Positives (FP), and False Negatives (FN)
+        self.tp = torch.zeros(num_classes)
+        self.fp = torch.zeros(num_classes)
+        self.fn = torch.zeros(num_classes)
+
+    def update(self, preds, target):
+        """
+       Update the variables with predictions and true labels.
+
+       Parameters
+       ----------
+       preds : torch.Tensor
+           Predicted logits (shape: [batch_size, num_classes]).
+
+       target : torch.Tensor
+           True labels (shape: [batch_size]).
+       """
+        preds = torch.argmax(preds, dim=1)
+
+        # Calculate True Positives (TP), False Positives (FP), and False Negatives (FN) per class
+        for i in range(self.num_classes):
+            self.tp[i] += torch.sum((preds == i) & (target == i)).float()
+            self.fp[i] += torch.sum((preds == i) & (target != i)).float()
+            self.fn[i] += torch.sum((preds != i) & (target == i)).float()
+
+    def compute(self):
+        """
+        Compute the F1 score.
+
+        Returns
+        -------
+        torch.Tensor
+           The computed F1 score.
+        """
+
+        # Compute F1 score based on the specified averaging method
+        f1_score = 2 * torch.sum(self.tp) / (2 * torch.sum(self.tp) + torch.sum(self.fp) + torch.sum(self.fn))
+
+        return f1_score
+
+
+def test_f1score():
+    f1_metric = F1Score(num_classes=3)
+    preds = torch.tensor([[0.8, 0.1, 0.1],
+                          [0.2, 0.7, 0.1],
+                          [0.2, 0.3, 0.5],
+                          [0.1, 0.2, 0.7]])
+
+    target = torch.tensor([0, 1, 0, 2])
+
+    f1_metric.update(preds, target)
+    assert f1_metric.tp.sum().item() > 0, "Expected some true positives."
+    assert f1_metric.fp.sum().item() > 0, "Expected some false positives."
+    assert f1_metric.fn.sum().item() > 0, "Expected some false negatives."