https://github.com/hereismari/tensorflow-maml

TensorFlow 2.0 implementation of MAML.
https://github.com/hereismari/tensorflow-maml

eager-execution maml meta-learning tensorflow tensorflow2

Last synced: 24 days ago
JSON representation

TensorFlow 2.0 implementation of MAML.

• Host: GitHub
• URL: https://github.com/hereismari/tensorflow-maml
• Owner: hereismari
• License: apache-2.0
• Created: 2019-04-09T20:05:55.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2019-07-12T20:03:49.000Z (over 5 years ago)
• Last Synced: 2024-10-04T11:29:18.582Z (about 1 month ago)
• Topics: eager-execution, maml, meta-learning, tensorflow, tensorflow2
• Language: Jupyter Notebook
• Homepage:
• Size: 2.26 MB
• Stars: 84
• Watchers: 2
• Forks: 20
• Open Issues: 2
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# Reproduction of MAML using TensorFlow 2.0.

This reproduction is highly influenced by the pytorch reproduction by Adrien Lucas Effot available at [Paper repro: Deep Metalearning using “MAML” and “Reptile”](https://towardsdatascience.com/paper-repro-deep-metalearning-using-maml-and-reptile-fd1df1cc81b0).

**MAML**   |   **Neural Net**

![alt-text-1](imgs/maml.png "MAML") ![alt-text-2](imgs/nn.png "Neural Net")



  



## MAML paper

https://arxiv.org/abs/1703.03400

**Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks**

*Chelsea Finn, Pieter Abbeel, Sergey Levine*

> We propose an algorithm for meta-learning that is model-agnostic, in the sense that it is compatible with any model trained with gradient descent and applicable to a variety of different learning problems, including classification, regression, and reinforcement learning. The goal of meta-learning is to train a model on a variety of learning tasks, such that it can solve new learning tasks using only a small number of training samples. In our approach, the parameters of the model are explicitly trained such that a small number of gradient steps with a small amount of training data from a new task will produce good generalization performance on that task. In effect, our method trains the model to be easy to fine-tune. We demonstrate that this approach leads to state-of-the-art performance on two few-shot image classification benchmarks, produces good results on few-shot regression, and accelerates fine-tuning for policy gradient reinforcement learning with neural network policies.

---

![image.png](https://cdn-images-1.medium.com/max/1600/1*EUt0H5AOEFkERg-OzfCC7A.png)