STORM vs Real UpNDown — World Model Prediction

STORMNeurIPS 2023·64x64 @ 15fps
PaperGitHub
Compare with:DIAMONDTWISTERSTORMΔ-IRISNEWIRIS
Ground Truth
Atari UpNDown real gameplay (ground truth frames, 64x64 @ 15fps)
Prediction
Atari UpNDown predicted by STORM world model (64x64 @ 15fps)
SSIMStructural Similarity Index. Measures pixel-level structural similarity. Range 0-1, higher = more similar.Higher is better
0.5374
FVDFrechet Video Distance. Measures distribution-level similarity of video features. Lower = closer to real.Lower is better
91.1
LPIPSLearned Perceptual Image Patch Similarity. Measures perceptual similarity via deep features. Range 0-1, lower = more similar.Lower is better
0.0804
FramesPrediction horizon
100