OMNIVORE (Swin-B, IN-21K+ Kinetics400 pretrain)

Reported on 4 benchmarks across 2 tasks · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Robots2 results

Activity RecognitiononSomething-Something V2
Top-1 Accuracy· uses extra data· 2022-01-20
71.4
best: 77.3 (MVD (Kinetics400 pretrain, ViT-H, 16 frame))
Omnivore: A Single Model for Many Visual Modalities arXiv:2201.08377
Activity RecognitiononSomething-Something V2
Top-5 Accuracy· uses extra data· 2022-01-20
93.5
best: 96.3 (DejaVid)
Omnivore: A Single Model for Many Visual Modalities arXiv:2201.08377

Time Series2 results

Action RecognitiononSomething-Something V2
Top-1 Accuracy· uses extra data· 2022-01-20
71.4
best: 77.3 (MVD (Kinetics400 pretrain, ViT-H, 16 frame))
Omnivore: A Single Model for Many Visual Modalities arXiv:2201.08377
Action RecognitiononSomething-Something V2
Top-5 Accuracy· uses extra data· 2022-01-20
93.5
best: 96.3 (DejaVid)
Omnivore: A Single Model for Many Visual Modalities arXiv:2201.08377