DDP: Diffusion Model for Dense Visual Prediction

Benchmark Model Rank Results
monocular-depth-estimation-on-kitti-eigenDDP (Swin-L, step-3)#17absolute relative error: 0.050RMSE: 2.072Sq Rel: 0.148
monocular-depth-estimation-on-nyu-depth-v2DDP (step3)#35absolute relative error: 0.094RMSE: 0.329log 10: 0.040
monocular-depth-estimation-on-sun-rgbdDDP (step-3)#2Delta < 1.25: 0.825Delta < 1.25^2: 0.973Delta < 1.25^3: 0.994
semantic-segmentation-on-ade20kDDP (Swin-L, step-3)#59Validation mIoU: 54.4Params (M): 207
semantic-segmentation-on-cityscapes-valDDP (ConvNeXt-L, step-3)#20mIoU: 83.9