PyTorch implementation of Learning to learn by gradient descent by gradient descent.

python main.py

• Initial implementation
• Toy data
• LSTM updates
• Refactor, find a better way to organize the modules
• Compare with standard optimizers
• Real data
• More difficult models