The problem with visualizing the confusion matrix for ImageNet labels is that there are too many categories, making it almost impossible to see it correctly. Therefore, we must create another plot where the labels on the side are not visible. An example of such an image would be the following:
The problem with visualizing the confusion matrix for ImageNet labels is that there are too many categories, making it almost impossible to see it correctly. Therefore, we must create another plot where the labels on the side are not visible. An example of such an image would be the following: