Meta-learning on Mini-ImageNet

Mini-ImageNet experiments compare MAML with a fine-tuning baseline and examine class count, shots and inner-loop update steps.

Report date

15 December 2025

Original authors

Pang Liu

Report language

English · 3 pages

Main observations

The report summarizes mean accuracy and confidence intervals over 1,000 test episodes across N-way, K-shot and inner-loop configurations.

Scope and limitations

Statistics over test episodes are distinct from statistics across independent training seeds; results depend on the original task splits and configurations.