Few Shot Image Classification On Omniglot 1 1

Accuracy

Results

Performance results of various models on this benchmark

Model Name	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.