Detector

detector(file, checkpoint, device='cpu')

Create detector object - a realisation of the model supplied to the function via the relevant config and checkpoint files.

Parameters:
  • file (str) –

    Location of config file.

  • checkpoint (str) –

    Location of checkpoint file.

  • device (str, default: 'cpu' ) –

    Device to use for inference - CPU or GPU. Defaults to "cpu".

Returns:
  • model

    MMDetect model object generated.

Source code in MinDet/detector.py
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
def detector(file , checkpoint, device = "cpu"):

    """Create detector object - a realisation of the model supplied to the function via the relevant config and checkpoint files.

    Args:
        file (str): Location of config file.
        checkpoint (str): Location of checkpoint file.
        device (str, optional): Device to use for inference - CPU or GPU. Defaults to "cpu".

    Returns:
        model: MMDetect model object generated.
    """

    # Specify the path to model config and checkpoint file
    config_file = file
    checkpoint_file = checkpoint

    # build the model from a config file and a checkpoint file
    model = init_detector(config_file, checkpoint_file, device=torch.device(device))
    return model