Extracting features using my QNN Hybrid model

[ptrblck]

You could use forward hooks as described here or could alternatively replace the classifier with nn.Identity layers if you won’t need the actual classifier output anymore.