Skip to content

deeplabcut.pose_tracking_pytorch.datasets.dlc_vec

Classes:

Name Description
TripletDataset

TripletDataset

Bases: Dataset

Attributes:

Name Type Description
x

for i in range(self.x.shape[0]):
Source code in deeplabcut/pose_tracking_pytorch/datasets/dlc_vec.py 
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
class TripletDataset(Dataset):
    def __init__(self, datasource, transform=None):
        self.x = datasource

        # normalize vectors here
        """
        for i in range(self.x.shape[0]):
            for j in range(self.x.shape[1]):
                # i th vector at j th kpt
                v = self.x[i][j]
                normalized_v = v / np.sqrt(np.sum(v**2))
                self.x[i][j] = normalized_v
        """

        self.transform = transform

    def __len__(self):
        return self.x.shape[0]

    def __getitem__(self, index):
        anchor, pos, neg = self.x[index]

        anchor = anchor.astype(np.float32)
        pos = pos.astype(np.float32)
        neg = neg.astype(np.float32)

        if self.transform is not None:
            # maybe needs to convert them to embeddings and position token

            anchor = self.transform(anchor)
            pos = self.transform(pos)
            neg = self.transform(neg)

        return anchor, pos, neg

x instance-attribute 

x = datasource

for i in range(self.x.shape[0]): for j in range(self.x.shape[1]): # i th vector at j th kpt v = self.x[i][j] normalized_v = v / np.sqrt(np.sum(v**2)) self.x[i][j] = normalized_v