Deep Learning with PyTorch

Creating a new dataset for the convoluted features

We can use the same FeaturesDataset class to create the new dataset and data loaders. In the following code, we create the datasets and the loaders:

#Dataset for pre computed features for train and validation data setstrn_feat_dset = FeaturesDataset(trn_features.features,trn_labels)val_feat_dset = FeaturesDataset(val_features.features,val_labels)#Data loaders for pre computed features for train and validation data setstrn_feat_loader = DataLoader(trn_feat_dset,batch_size=64,shuffle=True)val_feat_loader = DataLoader(val_feat_dset,batch_size=64)

Let's create a new model to train on the pre-convoluted features.