Skip to content

pseudo_labeled_dataset

PseudoLabeledDataset

Bases: DomainDataset

The wrapped dataset's __getitem__ function should return a tuple of (data, label). The label will then be discarded, and the pseudo label will be returned instead.

Source code in pytorch_adapt\datasets\pseudo_labeled_dataset.py 
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
class PseudoLabeledDataset(DomainDataset):
    """
    The wrapped dataset's ```__getitem__``` function
    should return a tuple of ```(data, label)```.
    The label will then be discarded,
    and the pseudo label will be returned instead.
    """

    def __init__(self, dataset: Dataset, pseudo_labels: List[int], domain: int = 0):
        """
        Arguments:
            dataset: The dataset to wrap
            pseudo_labels: The class labels that will be used
                instead of the labels contained in self.dataset
            domain: An integer representing the domain.
        """

        super().__init__(dataset, domain)
        if len(self.dataset) != len(pseudo_labels):
            raise ValueError("len(dataset) must equal len(pseudo_labels)")
        self.pseudo_labels = pseudo_labels

    def __len__(self):
        return len(self.dataset)

    def __getitem__(self, idx: int) -> Dict[str, Any]:
        """
        Returns:
            A dictionary with keys

                - "src_imgs" (the data)

                - "src_domain" (the integer representing the domain)

                - "src_labels" (the pseudo label)

                - "src_sample_idx" (idx)
        """

        img, _ = self.dataset[idx]
        return {
            "src_imgs": img,
            "src_domain": self.domain,
            "src_labels": self.pseudo_labels[idx],
            "src_sample_idx": idx,
        }

__getitem__(idx)

Returns:

Type Description
Dict[str, Any]

A dictionary with keys

  • "src_imgs" (the data)

  • "src_domain" (the integer representing the domain)

  • "src_labels" (the pseudo label)

  • "src_sample_idx" (idx)
Source code in pytorch_adapt\datasets\pseudo_labeled_dataset.py 
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
def __getitem__(self, idx: int) -> Dict[str, Any]:
    """
    Returns:
        A dictionary with keys

            - "src_imgs" (the data)

            - "src_domain" (the integer representing the domain)

            - "src_labels" (the pseudo label)

            - "src_sample_idx" (idx)
    """

    img, _ = self.dataset[idx]
    return {
        "src_imgs": img,
        "src_domain": self.domain,
        "src_labels": self.pseudo_labels[idx],
        "src_sample_idx": idx,
    }

__init__(dataset, pseudo_labels, domain=0)

Parameters:

Name Type Description Default
dataset Dataset

The dataset to wrap

 required
pseudo_labels List[int]

The class labels that will be used instead of the labels contained in self.dataset

 required
domain int

An integer representing the domain.

 0
Source code in pytorch_adapt\datasets\pseudo_labeled_dataset.py 
16
17
18
19
20
21
22
23
24
25
26
27
28
def __init__(self, dataset: Dataset, pseudo_labels: List[int], domain: int = 0):
    """
    Arguments:
        dataset: The dataset to wrap
        pseudo_labels: The class labels that will be used
            instead of the labels contained in self.dataset
        domain: An integer representing the domain.
    """

    super().__init__(dataset, domain)
    if len(self.dataset) != len(pseudo_labels):
        raise ValueError("len(dataset) must equal len(pseudo_labels)")
    self.pseudo_labels = pseudo_labels