About some code issue

[xingy038]

https://github.com/TRI-ML/vidar/blob/a306e23cc47ae05a0c35b0bf8acf7112c87c049c/vidar/arch/models/depth/DepthFormerModel.py#L326

        if 'cal' in self.networks['multi_depth'].networks.keys():
            cal = self.networks['multi_depth'].networks['cal']
            depth1 = depth_multi[0]
            depth2 = predictions['depth_regr'][0][0]
            from vidar.utils.tensor import interpolate
            depth2 = interpolate(depth2, size=depth1.shape[-2:], scale_factor=None, mode='nearest', align_corners=None)
            predictions['depth_regr'][0].insert(0, cal(depth1, depth2, rgb))

where is cal network? If cal is context_adjustment layer, but forward is not 2 parameters? Why are there 3 parameters here?