Skip to content

Opossum

AI4GOpossum

Bases: PlainResNetInference

Opossum Classifier that inherits from PlainResNetInference. This classifier is specialized for distinguishing between Opossums and Non-opossums.

Source code in PytorchWildlife/models/classification/resnet_base/opossum.py 
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
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
class AI4GOpossum(PlainResNetInference):
    """
    Opossum Classifier that inherits from PlainResNetInference.
    This classifier is specialized for distinguishing between Opossums and Non-opossums.
    """

    # Image size for the Opossum classifier
    IMAGE_SIZE = 224

    # Class names for prediction
    CLASS_NAMES = {
        0: "Non-opossum",
        1: "Opossum"
    }

    def __init__(self, weights=None, device="cpu", pretrained=True):
        """
        Initialize the Opossum Classifier.

        Args:
            weights (str, optional): Path to the model weights. Defaults to None.
            device (str, optional): Device for model inference. Defaults to "cpu".
            pretrained (bool, optional): Whether to use pretrained weights. Defaults to True.
        """

        # If pretrained, use the provided URL to fetch the weights
        if pretrained:
            url = "https://zenodo.org/records/10023414/files/OpossumClassification_v0.0.0.ckpt?download=1"
        else:
            url = None

        super(AI4GOpossum, self).__init__(weights=weights, device=device,
                                          num_cls=1, num_layers=50, url=url)

    def results_generation(self, logits: torch.Tensor, img_ids: list[str], id_strip: str = None) -> list[dict]:
        """
        Generate results for classification.

        Args:
            logits (torch.Tensor): Output tensor from the model.
            img_ids (list): List of image identifier.
            id_strip (str): stiping string for better image id saving.       

        Returns:
            dict: Dictionary containing image ID, prediction, and confidence score.
        """

        probs = torch.sigmoid(logits)
        preds = (probs > 0.5).squeeze(1).numpy().astype(int)

        results = []
        for pred, img_id, prob in zip(preds, img_ids, probs):
            r = {"img_id": str(img_id).strip(id_strip)}
            r["prediction"] = self.CLASS_NAMES[pred]
            r["class_id"] = pred
            r["confidence"] = prob.item() if pred == 1 else (1 - prob.item())
            results.append(r)

        return results

__init__(weights=None, device='cpu', pretrained=True)

Initialize the Opossum Classifier.

Parameters:

Name Type Description Default
weights str

Path to the model weights. Defaults to None.

 None
device str

Device for model inference. Defaults to "cpu".

 'cpu'
pretrained bool

Whether to use pretrained weights. Defaults to True.

 True
Source code in PytorchWildlife/models/classification/resnet_base/opossum.py 
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
def __init__(self, weights=None, device="cpu", pretrained=True):
    """
    Initialize the Opossum Classifier.

    Args:
        weights (str, optional): Path to the model weights. Defaults to None.
        device (str, optional): Device for model inference. Defaults to "cpu".
        pretrained (bool, optional): Whether to use pretrained weights. Defaults to True.
    """

    # If pretrained, use the provided URL to fetch the weights
    if pretrained:
        url = "https://zenodo.org/records/10023414/files/OpossumClassification_v0.0.0.ckpt?download=1"
    else:
        url = None

    super(AI4GOpossum, self).__init__(weights=weights, device=device,
                                      num_cls=1, num_layers=50, url=url)

results_generation(logits, img_ids, id_strip=None)

Generate results for classification.

Parameters:

Name Type Description Default
logits Tensor

Output tensor from the model.

 required
img_ids list

List of image identifier.

 required
id_strip str

stiping string for better image id saving.

None

Returns:

Name Type Description
dict list[dict]

Dictionary containing image ID, prediction, and confidence score.
Source code in PytorchWildlife/models/classification/resnet_base/opossum.py 
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
def results_generation(self, logits: torch.Tensor, img_ids: list[str], id_strip: str = None) -> list[dict]:
    """
    Generate results for classification.

    Args:
        logits (torch.Tensor): Output tensor from the model.
        img_ids (list): List of image identifier.
        id_strip (str): stiping string for better image id saving.       

    Returns:
        dict: Dictionary containing image ID, prediction, and confidence score.
    """

    probs = torch.sigmoid(logits)
    preds = (probs > 0.5).squeeze(1).numpy().astype(int)

    results = []
    for pred, img_id, prob in zip(preds, img_ids, probs):
        r = {"img_id": str(img_id).strip(id_strip)}
        r["prediction"] = self.CLASS_NAMES[pred]
        r["class_id"] = pred
        r["confidence"] = prob.item() if pred == 1 else (1 - prob.item())
        results.append(r)

    return results