ResNet-50 (PushPull-Conv) + PRIME

Reported on 6 benchmarks across 2 tasks · 1 paper · 2 SOTA

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

Methodology3 results

Domain AdaptationonImageNet-C
Number of params· 2024-08-07
25.6
SOTA
PushPull-Net: Inhibition-driven ResNet robust to image corruptions arXiv:2408.04077
Domain AdaptationonImageNet-C
Top 1 Accuracy· 2024-08-07
69.4
best: 73.6 (FAN-L-Hybrid (IN-22k))
PushPull-Net: Inhibition-driven ResNet robust to image corruptions arXiv:2408.04077
Domain AdaptationonImageNet-C
mean Corruption Error (mCE)· 2024-08-07
49.95
best: 22 (EfficientNet-L2+RPL)
PushPull-Net: Inhibition-driven ResNet robust to image corruptions arXiv:2408.04077

Computer Vision3 results

Domain GeneralizationonImageNet-C
Number of params· 2024-08-07
25.6
SOTA
PushPull-Net: Inhibition-driven ResNet robust to image corruptions arXiv:2408.04077
Domain GeneralizationonImageNet-C
Top 1 Accuracy· 2024-08-07
69.4
best: 73.6 (FAN-L-Hybrid (IN-22k))
PushPull-Net: Inhibition-driven ResNet robust to image corruptions arXiv:2408.04077
Domain GeneralizationonImageNet-C
mean Corruption Error (mCE)· 2024-08-07
49.95
best: 28.2 (DINOv2 (ViT-g/14, frozen model, linear eval))
PushPull-Net: Inhibition-driven ResNet robust to image corruptions arXiv:2408.04077