Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/dragen1860/MAML-Pytorch
Elegant PyTorch implementation of paper Model-Agnostic Meta-Learning (MAML)
https://github.com/dragen1860/MAML-Pytorch
Last synced: about 2 months ago
JSON representation
Elegant PyTorch implementation of paper Model-Agnostic Meta-Learning (MAML)
- Host: GitHub
- URL: https://github.com/dragen1860/MAML-Pytorch
- Owner: dragen1860
- License: mit
- Created: 2018-02-01T14:19:54.000Z (almost 7 years ago)
- Default Branch: master
- Last Pushed: 2019-05-16T12:39:47.000Z (over 5 years ago)
- Last Synced: 2024-10-22T09:51:58.083Z (about 2 months ago)
- Language: Python
- Homepage:
- Size: 631 KB
- Stars: 2,304
- Watchers: 24
- Forks: 420
- Open Issues: 44
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
- awesome-few-shot-meta-learning - code - PyTorch
README
# MAML-Pytorch
PyTorch implementation of the supervised learning experiments from the paper:
[Model-Agnostic Meta-Learning (MAML)](https://arxiv.org/abs/1703.03400).
> Version 1.0: Both `MiniImagenet` and `Omniglot` Datasets are supported! Have Fun~
> Version 2.0: Re-write meta learner and basic learner. Solved some serious bugs in version 1.0.
For Tensorflow Implementation, please visit official [HERE](https://github.com/cbfinn/maml) and simplier version [HERE](https://github.com/dragen1860/MAML-TensorFlow).
For First-Order Approximation Implementation, Reptile namely, please visit [HERE](https://github.com/dragen1860/Reptile-Pytorch).
# Platform
- python: 3.x
- Pytorch: 0.4+
# MiniImagenet
## Howto
For 5-way 1-shot exp., it allocates nearly 6GB GPU memory.
1. download `MiniImagenet` dataset from [here](https://github.com/dragen1860/LearningToCompare-Pytorch/issues/4), splitting: `train/val/test.csv` from [here](https://github.com/twitter/meta-learning-lstm/tree/master/data/miniImagenet).
2. extract it like:
```shell
miniimagenet/
├── images
├── n0210891500001298.jpg
├── n0287152500001298.jpg
...
├── test.csv
├── val.csv
└── train.csv
```
3. modify the `path` in `miniimagenet_train.py`:
```python
mini = MiniImagenet('miniimagenet/', mode='train', n_way=args.n_way, k_shot=args.k_spt,
k_query=args.k_qry,
batchsz=10000, resize=args.imgsz)
...
mini_test = MiniImagenet('miniimagenet/', mode='test', n_way=args.n_way, k_shot=args.k_spt,
k_query=args.k_qry,
batchsz=100, resize=args.imgsz)
```
to your actual data path.
4. just run `python miniimagenet_train.py` and the running screenshot is as follows:
If your reproducation perf. is not so good, maybe you can enlarge your `training epoch` to get longer training. And MAML is notorious for its hard training. Therefore, this implementation only provide you a basic start point to begin your research.
and the performance below is true and achieved on my machine.
## Benchmark
| Model | Fine Tune | 5-way Acc. | | 20-way Acc.| |
|-------------------------------------|-----------|------------|--------|------------|--------|
| | | 1-shot | 5-shot | 1-shot | 5-shot |
| Matching Nets | N | 43.56% | 55.31% | 17.31% | 22.69% |
| Meta-LSTM | | 43.44% | 60.60% | 16.70% | 26.06% |
| MAML | Y | 48.7% | 63.11% | 16.49% | 19.29% |
| **Ours** | Y | 46.2% | 60.3% | - | - |
# Ominiglot
## Howto
run `python omniglot_train.py`, the program will download `omniglot` dataset automatically.
decrease the value of `args.task_num` to fit your GPU memory capacity.
For 5-way 1-shot exp., it allocates nearly 3GB GPU memory.
# Refer to this Rep.
```
@misc{MAML_Pytorch,
author = {Liangqu Long},
title = {MAML-Pytorch Implementation},
year = {2018},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/dragen1860/MAML-Pytorch}},
commit = {master}
}
```