Transfer Learning

On this chapter we will learn about 2 scenarios on transfer learning

• Initialize the network with a set of weights trained from another session. (Instead of initializing the network with random values).

• Load some network, freeze it's weights up to a certain point and re-train the rest, normally on a smaller dataset.

References

• https://discuss.pytorch.org/t/discussion-about-datasets-and-dataloaders/296

• http://pytorch.org/tutorials/beginner/transfer_learning_tutorial.html

• https://medium.com/towards-data-science/transfer-learning-using-pytorch-4c3475f4495

• https://medium.com/towards-data-science/transfer-learning-using-pytorch-part-2-9c5b18e15551

• http://pytorch.org/tutorials/beginner/transfer_learning_tutorial.html