63days/MAML

This repo is a Model-Agnostic Meta-Learning implementation in Pytorch.

Model-Agnostic Meta-Learning: Pytorch Implementation

Pytorch Implementation of the Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks with k-shot sinusoid regression.

Pseudocode

Results

K=10

pre loss: 15.244 | post loss: 5.391

To train

python3 main.py

To test

python3 main.py --test