Method load_checkpoint

easycv/runner/ev_runner.py:184–211

Load checkpoint from a file or URL. Args: filename (str): Accept local filepath, URL, ``torchvision://xxx``, ``open-mmlab://xxx``, ``oss://xxx``. Please refer to ``docs/source/model_zoo.md`` for details. map_location (str): Same as :fu

(self,
                        filename,
                        map_location=torch.device('cpu'),
                        strict=False,
                        logger=None)

    # !! notice, map_location should be cpu, other wise may stack to some GPU ,which cause OOMß
    def load_checkpoint(self,
                        filename,
                        map_location=torch.device('cpu'),
                        strict=False,
                        logger=None):
        """Load checkpoint from a file or URL.

        Args:
            filename (str): Accept local filepath, URL, ``torchvision://xxx``,
                ``open-mmlab://xxx``, ``oss://xxx``. Please refer to
                ``docs/source/model_zoo.md`` for details.
            map_location (str): Same as :func:`torch.load`.
            strict (bool): Whether to allow different params for the model and
                checkpoint.
            logger (:mod:`logging.Logger` or None): The logger for error message.

        Returns:
            dict or OrderedDict: The loaded checkpoint.
        """
        if torch.cuda.is_available():
            device_id = torch.cuda.current_device()
            map_location = lambda storage, loc: storage.cuda(device_id)
        return load_checkpoint(
            self.model,
            filename=filename,
            map_location=map_location,
            strict=strict,
            logger=logger)

    def resume(self,
               checkpoint,