hzwer
commented
1 year ago Hello, this is to let the model average the results of the two predictions; although it can improve some quantitative indicators, the results do not look good to the naked eye.
Open HAOHAOXUEXIEN opened 1 year ago
hzwer
commented
1 year ago Hello, this is to let the model average the results of the two predictions; although it can improve some quantitative indicators, the results do not look good to the naked eye.
if ensemble: f1, m1 = block[i](torch.cat((img1[:, :3], img0[:, :3], 1-timestep), 1), None, scale=scale_list[i]) flow = (flow + torch.cat((f1[:, 2:4], f1[:, :2]), 1)) / 2 mask = (mask + (-m1)) / 2
can explain the ensemble effect? thanks very much