Hi I am using Barlow Twins to train a Liver MRI data. I am having a difficulty to get the embeddings from the original code from facebook. I tried two methods to get the embeddings farward_hook() and create_feature_extractor() from torchvision.models.feature_extraction. from the hook i am getting two embedding as Za and Zb. and from the create_feature_extractor i am getting only one embedding. How to get the single embedding for a image not the transformed version. during the training or the inference mode. Please help. I am a beginner with Pytorch.