TASK	DATASET	MODEL	METRIC NAME	METRIC VALUE	GLOBAL RANK
Few-Shot Image Classification	CUB 200 50-way (0-shot)	Prototypical Networks	Accuracy	54.6	# 1
Few-Shot Image Classification	Dirichlet Mini-Imagenet (5-way, 1-shot)	ProtoNet	1:1 Accuracy	53.6	# 10
Few-Shot Image Classification	Dirichlet Mini-Imagenet (5-way, 5-shot)	ProtoNet	1:1 Accuracy	74.2	# 8
Few-Shot Image Classification	Meta-Dataset	Prototypical Networks	Accuracy	60.573	# 17
Few-Shot Image Classification	Meta-Dataset Rank	Prototypical Networks	Mean Rank	8.5	# 8
Few-Shot Image Classification	Mini-Imagenet 10-way (1-shot)	Prototypical Networks (Higher Way)	Accuracy	34.6	# 9
Few-Shot Image Classification	Mini-Imagenet 10-way (1-shot)	Prototypical Networks	Accuracy	32.9	# 10
Few-Shot Image Classification	Mini-Imagenet 10-way (5-shot)	Prototypical Networks (Higher Way)	Accuracy	50.1	# 8
Few-Shot Image Classification	Mini-Imagenet 10-way (5-shot)	Prototypical Networks	Accuracy	49.3	# 9
Few-Shot Image Classification	Mini-Imagenet 5-way (10-shot)	Prototypical Networks	Accuracy	74.3	# 5
Few-Shot Image Classification	Mini-Imagenet 5-way (1-shot)	Prototypical Networks	Accuracy	49.42	# 101
Few-Shot Image Classification	Mini-Imagenet 5-way (5-shot)	Prototypical Networks	Accuracy	68.20	# 86
Few-Shot Image Classification	Mini-ImageNet-CUB 5-way (1-shot)	ProtoNet (Snell et al., 2017)	Accuracy	45.31	# 6
Category-Agnostic Pose Estimation	MP100	ProtoNet	Mean PCK@0.2 - 1shot	44.78	# 5
Few-Shot Image Classification	OMNIGLOT - 1-Shot, 20-way	Prototypical Networks	Accuracy	96%	# 12
Few-Shot Image Classification	OMNIGLOT - 1-Shot, 5-way	Prototypical Networks	Accuracy	98.8	# 9
Few-Shot Image Classification	OMNIGLOT - 5-Shot, 20-way	Prototypical Networks	Accuracy	98.9%	# 10
Few-Shot Image Classification	OMNIGLOT - 5-Shot, 5-way	Prototypical Networks	Accuracy	99.7	# 9
Few-Shot Image Classification	Stanford Cars 5-way (1-shot)	Prototypical Nets++	Accuracy	40.90	# 5
Few-Shot Image Classification	Stanford Cars 5-way (5-shot)	Prototypical Nets++	Accuracy	52.93	# 5
Few-Shot Image Classification	Stanford Dogs 5-way (5-shot)	Prototypical Nets++	Accuracy	48.19	# 5
Few-Shot Image Classification	Tiered ImageNet 10-way (1-shot)	Prototypical Networks (Higher Way)	Accuracy	38.6	# 7
Few-Shot Image Classification	Tiered ImageNet 10-way (1-shot)	Prototypical Networks	Accuracy	37.3	# 8
Few-Shot Image Classification	Tiered ImageNet 10-way (5-shot)	Prototypical Networks (Higher Way)	Accuracy	58.3	# 6
Few-Shot Image Classification	Tiered ImageNet 10-way (5-shot)	Prototypical Networks	Accuracy	57.8	# 9
Image Classification	Tiered ImageNet 5-way (5-shot)	Prototypical Net	Accuracy	69.57	# 6

Badge	Markdown
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-cub-200-50)](https://paperswithcode.com/sota/few-shot-image-classification-on-cub-200-50?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-mini-4)](https://paperswithcode.com/sota/few-shot-image-classification-on-mini-4?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/category-agnostic-pose-estimation-on-mp100)](https://paperswithcode.com/sota/category-agnostic-pose-estimation-on-mp100?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-stanford-2)](https://paperswithcode.com/sota/few-shot-image-classification-on-stanford-2?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-stanford-3)](https://paperswithcode.com/sota/few-shot-image-classification-on-stanford-3?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-stanford-1)](https://paperswithcode.com/sota/few-shot-image-classification-on-stanford-1?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-mini-5)](https://paperswithcode.com/sota/few-shot-image-classification-on-mini-5?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-tiered-3)](https://paperswithcode.com/sota/few-shot-image-classification-on-tiered-3?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/image-classification-on-tiered-imagenet-5-way)](https://paperswithcode.com/sota/image-classification-on-tiered-imagenet-5-way?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-tiered-2)](https://paperswithcode.com/sota/few-shot-image-classification-on-tiered-2?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-dirichlet-1)](https://paperswithcode.com/sota/few-shot-image-classification-on-dirichlet-1?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-meta-dataset-1)](https://paperswithcode.com/sota/few-shot-image-classification-on-meta-dataset-1?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-mini-13)](https://paperswithcode.com/sota/few-shot-image-classification-on-mini-13?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-mini-12)](https://paperswithcode.com/sota/few-shot-image-classification-on-mini-12?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-omniglot-1-2)](https://paperswithcode.com/sota/few-shot-image-classification-on-omniglot-1-2?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-omniglot-5-2)](https://paperswithcode.com/sota/few-shot-image-classification-on-omniglot-5-2?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-dirichlet)](https://paperswithcode.com/sota/few-shot-image-classification-on-dirichlet?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-omniglot-5-1)](https://paperswithcode.com/sota/few-shot-image-classification-on-omniglot-5-1?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-omniglot-1-1)](https://paperswithcode.com/sota/few-shot-image-classification-on-omniglot-1-1?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-meta-dataset)](https://paperswithcode.com/sota/few-shot-image-classification-on-meta-dataset?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-mini-3)](https://paperswithcode.com/sota/few-shot-image-classification-on-mini-3?p=prototypical-networks-for-few-shot-learning)`
	`[![PWC](https://img.shields.io/endpoint.svg?url=https://paperswithcode.com/badge/prototypical-networks-for-few-shot-learning/few-shot-image-classification-on-mini-2)](https://paperswithcode.com/sota/few-shot-image-classification-on-mini-2?p=prototypical-networks-for-few-shot-learning)`

Prototypical Networks for Few-shot Learning

NeurIPS 2017 · Jake Snell, Kevin Swersky, Richard S. Zemel ·

We propose prototypical networks for the problem of few-shot classification, where a classifier must generalize to new classes not seen in the training set, given only a small number of examples of each new class. Prototypical networks learn a metric space in which classification can be performed by computing distances to prototype representations of each class. Compared to recent approaches for few-shot learning, they reflect a simpler inductive bias that is beneficial in this limited-data regime, and achieve excellent results. We provide an analysis showing that some simple design decisions can yield substantial improvements over recent approaches involving complicated architectural choices and meta-learning. We further extend prototypical networks to zero-shot learning and achieve state-of-the-art results on the CU-Birds dataset.

PDF Abstract NeurIPS 2017 PDF NeurIPS 2017 Abstract

Code

Add Remove Mark official

jakesnell/prototypical-networks official

1,150

learnables/learn2learn

2,730

oscarknagg/few-shot

1,234

sicara/easy-few-shot-learning

↳ Quickstart in

Colab

1,155

orobix/Prototypical-Networks-for-Fe…

1,011

See all 43 implementations

Tasks

Add Remove

Category-Agnostic Pose Estimation

Few-Shot Image Classification

Few-Shot Learning

General Classification

Inductive Bias

Meta-Learning

One-Shot Learning

Zero-Shot Learning

Datasets

CUB-200-2011 mini-Imagenet

Stanford Cars

tieredImageNet

Meta-Dataset miniImageNet Omniglot

Stanford Dogs

Results from the Paper

Edit

Ranked #1 on Few-Shot Image Classification on CUB 200 50-way (0-shot)

Get a GitHub badge

Task	Dataset	Model	Metric Name	Metric Value	Global Rank	Benchmark
Few-Shot Image Classification	CUB 200 50-way (0-shot)	Prototypical Networks	Accuracy	54.6	# 1	Compare
Few-Shot Image Classification	Dirichlet Mini-Imagenet (5-way, 1-shot)	ProtoNet	1:1 Accuracy	53.6	# 10	Compare
Few-Shot Image Classification	Dirichlet Mini-Imagenet (5-way, 5-shot)	ProtoNet	1:1 Accuracy	74.2	# 8	Compare
Few-Shot Image Classification	Meta-Dataset	Prototypical Networks	Accuracy	60.573	# 17	Compare
Few-Shot Image Classification	Meta-Dataset Rank	Prototypical Networks	Mean Rank	8.5	# 8	Compare
Few-Shot Image Classification	Mini-Imagenet 10-way (1-shot)	Prototypical Networks (Higher Way)	Accuracy	34.6	# 9	Compare
Few-Shot Image Classification	Mini-Imagenet 10-way (1-shot)	Prototypical Networks	Accuracy	32.9	# 10	Compare
Few-Shot Image Classification	Mini-Imagenet 10-way (5-shot)	Prototypical Networks (Higher Way)	Accuracy	50.1	# 8	Compare
Few-Shot Image Classification	Mini-Imagenet 10-way (5-shot)	Prototypical Networks	Accuracy	49.3	# 9	Compare
Few-Shot Image Classification	Mini-Imagenet 5-way (10-shot)	Prototypical Networks	Accuracy	74.3	# 5	Compare
Few-Shot Image Classification	Mini-Imagenet 5-way (1-shot)	Prototypical Networks	Accuracy	49.42	# 101	Compare
Few-Shot Image Classification	Mini-Imagenet 5-way (5-shot)	Prototypical Networks	Accuracy	68.20	# 86	Compare
Category-Agnostic Pose Estimation	MP100	ProtoNet	Mean PCK@0.2 - 1shot	44.78	# 5	Compare
Few-Shot Image Classification	OMNIGLOT - 1-Shot, 20-way	Prototypical Networks	Accuracy	96%	# 12	Compare
Few-Shot Image Classification	OMNIGLOT - 1-Shot, 5-way	Prototypical Networks	Accuracy	98.8	# 9	Compare
Few-Shot Image Classification	OMNIGLOT - 5-Shot, 20-way	Prototypical Networks	Accuracy	98.9%	# 10	Compare
Few-Shot Image Classification	OMNIGLOT - 5-Shot, 5-way	Prototypical Networks	Accuracy	99.7	# 9	Compare
Few-Shot Image Classification	Stanford Cars 5-way (1-shot)	Prototypical Nets++	Accuracy	40.90	# 5	Compare
Few-Shot Image Classification	Stanford Cars 5-way (5-shot)	Prototypical Nets++	Accuracy	52.93	# 5	Compare
Few-Shot Image Classification	Tiered ImageNet 10-way (1-shot)	Prototypical Networks (Higher Way)	Accuracy	38.6	# 7	Compare
Few-Shot Image Classification	Tiered ImageNet 10-way (1-shot)	Prototypical Networks	Accuracy	37.3	# 8	Compare
Few-Shot Image Classification	Tiered ImageNet 10-way (5-shot)	Prototypical Networks (Higher Way)	Accuracy	58.3	# 6	Compare
Few-Shot Image Classification	Tiered ImageNet 10-way (5-shot)	Prototypical Networks	Accuracy	57.8	# 9	Compare

Results from Other Papers

Task	Dataset	Model	Metric Name	Metric Value	Rank	Compare
Few-Shot Image Classification	Mini-ImageNet-CUB 5-way (1-shot)	ProtoNet (Snell et al., 2017)	Accuracy	45.31	# 6	See all
Few-Shot Image Classification	Stanford Dogs 5-way (5-shot)	Prototypical Nets++	Accuracy	48.19	# 5	See all
Image Classification	Tiered ImageNet 5-way (5-shot)	Prototypical Net	Accuracy	69.57	# 6	See all

Methods

Add Remove

No methods listed for this paper. Add relevant methods here

Edit Social Preview

Prototypical Networks for Few-shot Learning

Code Edit Add Remove Mark official

Tasks Edit Add Remove

Datasets Edit

Results from the Paper Edit