README.md

Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks (ICML'2017)

Abstract

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 fewshot image classification benchmarks, produces good results on few-shot regression, and accelerates fine-tuning for policy gradient reinforcement learning with neural network policies.

Citation

@inproceedings{FinnAL17,
  title={Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks},
  author={Chelsea Finn and Pieter Abbeel and Sergey Levine},
  booktitle={Proceedings of the 34th International Conference on Machine Learning},
  year={2017}
}

How to Reproduce MAML

It consists of two steps:

• Step1: Base training

  • use all the images of base classes to train a base model.
  • conduct meta testing on validation set to select the best model.

• Step2: Meta Testing:

  • use best model from step1, the best model are saved into ${WORK_DIR}/${CONFIG}/best_accuracy_mean.pth in default.

An example of CUB dataset with Conv4

# base training
python ./tools/classification/train.py \
  configs/classification/maml/cub/maml_conv4_1xb105_cub_5way-1shot.py

# meta testing
python ./tools/classification/test.py \
  configs/classification/maml/cub/maml_conv4_1xb105_cub_5way-1shot.py \
  work_dir/maml_conv4_1xb105_cub_5way-1shot/best_accuracy_mean.pth

Note:

• All the result are trained with single gpu.
• The configs of 1 shot and 5 shot use same training setting, but different meta test setting on validation set and test set.
• The hyper-parameters in configs are roughly set and probably not the optimal one so feel free to tone and try different configurations. For example, try different learning rate or validation episodes for each setting. Anyway, we will continue to improve it.

Results on CUB dataset with 2000 episodes

Arch	Input Size	Batch Size	way	shot	mean Acc	std	ckpt	log
conv4	84x84	105	5	1	60.32	0.5	ckpt	log
conv4	84x84	105	5	5	77.03	0.39	ckpt	log
resnet12	84x84	105	5	1	70.44	0.55	ckpt	log
resnet12	84x84	105	5	5	85.5	0.33	ckpt	log

Results on Mini-ImageNet dataset with 2000 episodes

Arch	Input Size	Batch Size	way	shot	mean Acc	std	ckpt	log
conv4	84x84	105	5	1	46.76	0.42	ckpt	log
conv4	84x84	105	5	5	63.88	0.39	ckpt	log
resnet12	84x84	105	5	1	57.4	0.47	ckpt	log
resnet12	84x84	105	5	5	72.42	0.38	ckpt	log

Results on Tiered-ImageNet dataset with 2000 episodes

Arch	Input Size	Batch Size	way	shot	mean Acc	std	ckpt	log
conv4	84x84	105	5	1	45.56	0.49	ckpt	log
conv4	84x84	105	5	5	60.2	0.43	ckpt	log
resnet12	84x84	105	5	1	57.63	0.53	ckpt	log
resnet12	84x84	105	5	5	72.3	0.43	ckpt	log