A Few-shot MiniImageNet 402 The dataset construction is based on MiniImageNet [ 26 ], following the method of Tsimpoukelli et al
–Neural Information Processing Systems
A 256 256 image size is used so that the ViT encoder generates 256 tokens. We follow the process used in Tsimpoukelli et al. Randomly sampled image from ImageNet. Randomly sampled image from ImageNet.
Neural Information Processing Systems
Feb-7-2026, 20:14:58 GMT
- Country:
- Africa > Middle East
- Morocco (0.06)
- Asia
- Bangladesh (0.05)
- Thailand (0.05)
- Europe
- North America
- Canada
- United States
- Oklahoma > Woods County (0.05)
- Utah (0.05)
- South America > Colombia (0.05)
- Africa > Middle East
- Technology: