Method forward

models/networks/pose_efficientNet.py:479–607

EfficientNet's forward function. Calls extract_features to extract features, applies final linear layer, and returns logits. Args: inputs (tensor): Input tensor. Returns: Output of this model after processing.

(self, inputs)

        return x

    def forward(self, inputs):
        """EfficientNet's forward function.
           Calls extract_features to extract features, applies final linear layer, and returns logits.
        Args:
            inputs (tensor): Input tensor.
        Returns:
            Output of this model after processing.
        """
        # Convolution layers
        # x = self.extract_features(inputs)
        endpoints = self.extract_endpoints(inputs)
        x1 = endpoints['reduction_6']
        x2 = endpoints['reduction_5']
        x3 = endpoints['reduction_4']
        x4 = endpoints['reduction_3']
        x5 = endpoints['reduction_2']
        x = x1
        
        if self._global_params.include_top:
            # Pooling and final linear layer
            x = self._avg_pooling(x)

            x = x.flatten(start_dim=1)
            x = self._dropout(x)
            x = self._fc(x)
            return x

        if self._global_params.include_hm_decoder:
            x1 = self._dropout(x1)
            x2 = self._dropout(x2)
            x3 = self._dropout(x3)
            x4 = self._dropout(x4)

            if self.efpn:
                assert self._global_params.use_c51, "C51 must be utilized for FPN intergration"
                
                x = self.__getattr__('deconv_1')(x1)
                
                if self._global_params.use_c51:
                    x_weighted = self._sigmoid(x)
                    x_inv = torch.sub(1, x_weighted, alpha=1)
                    x2_ = torch.multiply(x_inv, x2)
                    x = torch.cat([x, x2_], dim=1)
                    
                    if self.se_layer:
                        x = self.__getattr__('se_layer_1')(x)
                else:
                    x = self._relu(x)
                
                x = self.__getattr__('deconv_2')(x)
                
                if self._global_params.use_c4:
                    x_weighted = self._sigmoid(x)
                    x_inv = torch.sub(1, x_weighted, alpha=1)
                    x3_ = torch.multiply(x_inv, x3)
                    x = torch.cat([x, x3_], dim=1)
                    
                    if self.se_layer: