mmfewshot/configs/classification/matching_net/README.md

MatchingNet (NeurIPS'2016)

@inproceedings{vinyals2016matching,
    title={Matching networks for one shot learning},
    author={Vinyals, Oriol and Blundell, Charles and Lillicrap, Tim and Wierstra, Daan and others},
    booktitle={Advances in Neural Information Processing Systems},
    pages={3630--3638},
    year={2016}
}

How to Reproduce MatchingNet

It consists of two steps:

• Step1: Base training

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

• Step2: Meta Testing:

  • use best model from step1.

An example of CUB dataset with Conv4

# base training
python ./tools/classification/train.py \
  configs/classification/matching_net/cub/matching-net_conv4_1xb105_cub_5way-1shot.py

# meta testing
python ./tools/classification/test.py \
  configs/classification/matching_net/cub/matching-net_conv4_1xb105_cub_5way-1shot.py \
  work_dir/matching-net_conv4_1xb105_cub_5way-1shot/best_accuracy_mean.pth

Note:

• All the result are trained with single gpu.
• The base training of 1 shot and 5 shot use same training setting, but different validation setting.

Results on CUB dataset of 1000 episodes

Arch	Input Size	Batch Size	way	shot	mean Acc	std	ckpt	log
conv4	84x84	64	5	1	-	-	ckpt	log
conv4	84x84	64	5	5	-	-	ckpt	log
resnet12	84x84	64	5	1	-	-	ckpt	log
resnet12	84x84	64	5	5	-	-	ckpt	log

Results on Mini-ImageNet dataset of 1000 episodes

Arch	Input Size	Batch Size	way	shot	mean Acc	std	ckpt	log
conv4	84x84	64	5	1	-	-	ckpt	log
conv4	84x84	64	5	5	-	-	ckpt	log
resnet12	84x84	64	5	1	-	-	ckpt	log
resnet12	84x84	64	5	5	-	-	ckpt	log

Results on Tiered-ImageNet of 1000 episodes

Arch	Input Size	Batch Size	way	shot	mean Acc	std	ckpt	log
conv4	84x84	64	5	1	-	-	ckpt	log
conv4	84x84	64	5	5	-	-	ckpt	log
resnet12	84x84	64	5	1	-	-	ckpt	log
resnet12	84x84	64	5	5	-	-	ckpt	log