InfNat-Zero

Reported on 6 benchmarks across 1 task · 1 paper · 3 SOTA

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

Computer Vision6 results

Perpetual View GenerationonLHQ
FID (first 20 steps)· 2022-11-22
39.45
best: 34.49 (DiffDreamer)
SOTA
DiffDreamer: Towards Consistent Unsupervised Single-view Scene Extrapolation with Conditional Diffusion Models arXiv:2211.12131
Perpetual View GenerationonLHQ
FID (full 100 steps)· 2022-11-22
26.24
SOTA
DiffDreamer: Towards Consistent Unsupervised Single-view Scene Extrapolation with Conditional Diffusion Models arXiv:2211.12131
Perpetual View GenerationonLHQ
KID (first 20 steps)· 2022-11-22
0.12
SOTA
DiffDreamer: Towards Consistent Unsupervised Single-view Scene Extrapolation with Conditional Diffusion Models arXiv:2211.12131
Perpetual View GenerationonLHQ
IS (first 20 steps)· 2022-11-22
2.8
best: 2.82 (DiffDreamer)
DiffDreamer: Towards Consistent Unsupervised Single-view Scene Extrapolation with Conditional Diffusion Models arXiv:2211.12131
Perpetual View GenerationonLHQ
IS (full 100 steps)· 2022-11-22
2.72
best: 2.99 (DiffDreamer)
DiffDreamer: Towards Consistent Unsupervised Single-view Scene Extrapolation with Conditional Diffusion Models arXiv:2211.12131
Perpetual View GenerationonLHQ
KID (full 100 steps)· 2022-11-22
0.12
best: 0.28 (DiffDreamer)
DiffDreamer: Towards Consistent Unsupervised Single-view Scene Extrapolation with Conditional Diffusion Models arXiv:2211.12131