Class BasicLayer

semantic_sam/backbone/focal.py:197–284

A basic focal modulation layer for one stage. Args: dim (int): Number of feature channels depth (int): Depths of this stage. mlp_ratio (float): Ratio of mlp hidden dim to embedding dim. Default: 4. drop (float, optional): Dropout rate. Default: 0.0 drop_

        return x

class BasicLayer(nn.Module):
    """ A basic focal modulation layer for one stage.

    Args:
        dim (int): Number of feature channels
        depth (int): Depths of this stage.
        mlp_ratio (float): Ratio of mlp hidden dim to embedding dim. Default: 4.
        drop (float, optional): Dropout rate. Default: 0.0
        drop_path (float | tuple[float], optional): Stochastic depth rate. Default: 0.0
        norm_layer (nn.Module, optional): Normalization layer. Default: nn.LayerNorm
        downsample (nn.Module | None, optional): Downsample layer at the end of the layer. Default: None
        focal_level (int): Number of focal levels
        focal_window (int): Focal window size at focal level 1
        use_conv_embed (bool): Use overlapped convolution for patch embedding or now. Default: False
        use_checkpoint (bool): Whether to use checkpointing to save memory. Default: False
    """

    def __init__(self,
                 dim,
                 depth,
                 mlp_ratio=4.,
                 drop=0.,
                 drop_path=0.,
                 norm_layer=nn.LayerNorm,
                 downsample=None,
                 focal_window=9, 
                 focal_level=2, 
                 use_conv_embed=False,     
                 use_postln=False,          
                 use_postln_in_modulation=False, 
                 scaling_modulator=False,
                 use_layerscale=False,                   
                 use_checkpoint=False
        ):
        super().__init__()
        self.depth = depth
        self.use_checkpoint = use_checkpoint

        # build blocks
        self.blocks = nn.ModuleList([
            FocalModulationBlock(
                dim=dim,
                mlp_ratio=mlp_ratio,
                drop=drop,
                drop_path=drop_path[i] if isinstance(drop_path, list) else drop_path,
                focal_window=focal_window, 
                focal_level=focal_level, 
                use_postln=use_postln, 
                use_postln_in_modulation=use_postln_in_modulation, 
                scaling_modulator=scaling_modulator,
                use_layerscale=use_layerscale, 
                norm_layer=norm_layer)
            for i in range(depth)])

        # patch merging layer
        if downsample is not None:
            self.downsample = downsample(
                patch_size=2,