Transfer Learning

Transferring knowledge between tasks and domains.

1
Datasets
2
Results
mean_accuracy
Canonical metric
Canonical Benchmark

VTAB-1k

Canonical transfer-learning benchmark covering 19 diverse vision tasks (natural, specialized, structured). Each task provides only 1000 labeled examples to test how well pretrained representations transfer to new domains with limited data.

Primary metric: mean_accuracy
View full leaderboard

Top 10

Leading models on VTAB-1k.

RankModelmean_accuracyYearSource
1
SSF (ViT-B/16)
73.12026paper
2
VPT-Deep (ViT-B/16)
72.02026paper

All datasets

1 dataset tracked for this task.

Related tasks

Other tasks in Methodology.