/learning-to-learn-by-pytorch

"Learning to learn by gradient descent by gradient descent "by PyTorch -- a simple re-implementation.

Primary LanguagePythonMIT LicenseMIT

learning to_learn by gradient descent by gradient descent

Learning to learn by gradient descent by gradient descent

A simple re-implementation by PyTorch-1.0

  • For Quadratic experiments
  • For Mnist
  • For CIFAR10
  • For Neural Art

Add:

  • Viusalize the learning process of LSTM optimizer.

The project can be run by this python file.

python learning_to_learn.py

This code is designed for a better understanding and easy implementation of paper Learning to learn by gradient descent by gradient descent.

Reference

1. Learning to learn by gradient descent by gradient descent - 2016 - NIPS

2. Learning to learn in Tensorflow by DeepMind