Class SlowFast

slowfast/models/video_model_builder.py:157–419

SlowFast model builder for SlowFast network. Christoph Feichtenhofer, Haoqi Fan, Jitendra Malik, and Kaiming He. "SlowFast networks for video recognition." https://arxiv.org/pdf/1812.03982.pdf

@MODEL_REGISTRY.register()
class SlowFast(nn.Module):
    """
    SlowFast model builder for SlowFast network.

    Christoph Feichtenhofer, Haoqi Fan, Jitendra Malik, and Kaiming He.
    "SlowFast networks for video recognition."
    https://arxiv.org/pdf/1812.03982.pdf
    """

    def __init__(self, cfg):
        """
        The `__init__` method of any subclass should also contain these
            arguments.
        Args:
            cfg (CfgNode): model building configs, details are in the
                comments of the config file.
        """
        super(SlowFast, self).__init__()
        self.norm_module = get_norm(cfg)
        self.enable_detection = cfg.DETECTION.ENABLE
        self.num_pathways = 2
        self._construct_network(cfg)
        init_helper.init_weights(
            self, cfg.MODEL.FC_INIT_STD, cfg.RESNET.ZERO_INIT_FINAL_BN
        )

    def _construct_network(self, cfg):
        """
        Builds a SlowFast model. The first pathway is the Slow pathway and the
            second pathway is the Fast pathway.
        Args:
            cfg (CfgNode): model building configs, details are in the
                comments of the config file.
        """
        assert cfg.MODEL.ARCH in _POOL1.keys()
        pool_size = _POOL1[cfg.MODEL.ARCH]
        assert len({len(pool_size), self.num_pathways}) == 1
        assert cfg.RESNET.DEPTH in _MODEL_STAGE_DEPTH.keys()

        (d2, d3, d4, d5) = _MODEL_STAGE_DEPTH[cfg.RESNET.DEPTH]

        num_groups = cfg.RESNET.NUM_GROUPS
        width_per_group = cfg.RESNET.WIDTH_PER_GROUP
        dim_inner = num_groups * width_per_group
        out_dim_ratio = (
            cfg.SLOWFAST.BETA_INV // cfg.SLOWFAST.FUSION_CONV_CHANNEL_RATIO
        )

        temp_kernel = _TEMPORAL_KERNEL_BASIS[cfg.MODEL.ARCH]

        self.s1 = stem_helper.VideoModelStem(
            dim_in=cfg.DATA.INPUT_CHANNEL_NUM,
            dim_out=[width_per_group, width_per_group // cfg.SLOWFAST.BETA_INV],
            kernel=[temp_kernel[0][0] + [7, 7], temp_kernel[0][1] + [7, 7]],
            stride=[[1, 2, 2]] * 2,
            padding=[
                [temp_kernel[0][0][0] // 2, 3, 3],
                [temp_kernel[0][1][0] // 2, 3, 3],