Few-Shot Learning
Learning from very few examples.
1
Datasets
2
Results
accuracy
Canonical metric
Canonical Benchmark
miniImageNet 5-way 5-shot
Canonical few-shot image classification benchmark. 100 classes drawn from ImageNet (64 train / 16 val / 20 test). Evaluation episodes sample 5 novel classes with 5 support examples each and query images to classify.
Primary metric: accuracy
Top 10
Leading models on miniImageNet 5-way 5-shot.
All datasets
1 dataset tracked for this task.
Related tasks
Other tasks in Methodology.