Using GradCAM in XceptionNet

Dear All:

I want to use GradCAM to visualize the feature map in XceptionNet, but I do not exactly how to use it.
My part code is showing below:

target_layer = [model.conv4[-1]]. // I do not know which is the target_layer to XceptionNet

rgb_img ='267.jpg').convert('RGB')
rgb_img = xception_default_data_transforms['test'](rgb_img)
rgb_img = torch.unsqueeze(rgb_img, dim=0)
input_tensor = rgb_img

cam = GradCAM(model=model, target_layers=target_layer, use_cuda=False)

target_category = None
grayscale_cam = cam(input_tensor=input_tensor, target_category=target_category)  
grayscale_cam = grayscale_cam[0]
visualization = show_cam_on_image(rgb_img, grayscale_cam)  
cv2.imshow('image.jpg', visualization)


GradCAM Reference: GitHub - jacobgil/pytorch-grad-cam: Many Class Activation Map methods implemented in Pytorch for CNNs and Vision Transformers. Examples for classification, object detection, segmentation, embedding networks and more. Including Grad-CAM, Grad-CAM++, Score-CAM, Ablation-CAM and XGrad-CAM
XceptionNet Reference: Xception-PyTorch/ at master · tstandley/Xception-PyTorch · GitHub