Few Shot Image Classification On Omniglot 1 1

Accuracy

评测结果

各个模型在此基准测试上的表现结果

模型名称	Accuracy	Paper Title	Repository
TapNet	98.07%	TapNet: Neural Network Augmented with Task-Adaptive Projection for Few-Shot Learning	-
Matching Nets	93.8%	Matching Networks for One Shot Learning	-
MAML++	97.7	HyperTransformer: Model Generation for Supervised and Semi-Supervised Few-Shot Learning	-
MAML++	97.65	How to train your MAML	-
VAMPIRE	93.2	Uncertainty in Model-Agnostic Meta-Learning using Variational Inference	-
Prototypical Networks	96%	Prototypical Networks for Few-shot Learning	-
DCN6-E	99.11	Decoder Choice Network for Meta-Learning	-
Hyperbolic ProtoNet	95.9%	Hyperbolic Image Embeddings	-
MT-net	96.2%	Gradient-Based Meta-Learning with Learned Layerwise Metric and Subspace	-
MC2+	88%	Meta-Curvature	-
ConvNet with Memory Module	95%	Learning to Remember Rare Events	-
GCR	99.63	Few-Shot Learning with Global Class Representations	-
APL	97.2%	Adaptive Posterior Learning: few-shot learning with a surprise-based memory module	-
adaCNN (DF)	96.12%	Rapid Adaptation with Conditionally Shifted Neurons	-
MR-MAML	83.3	Meta-Learning without Memorization	-
Relation Net	97.6%	Learning to Compare: Relation Network for Few-Shot Learning	-
Reptile + Transduction	89.43%	On First-Order Meta-Learning Algorithms	-
Neural Statistician	93.2%	Towards a Neural Statistician	-
iMAML, Hessian-Free	96.18	Meta-Learning with Implicit Gradients	-
DCN4	98.8%	Decoder Choice Network for Meta-Learning	-

0 of 20 row(s) selected.