Table 6 The weights of the L1 loss, SSIM loss, and perceptual loss during training
From: Generating oracle bone inscriptions based on the structure aware diffusion model
Step | Progress | λL1 | \({\lambda }_{{\rm{SSIM}}}\) | λPerceptual |
|---|---|---|---|---|
0 | 0.00 | 1.00 | 0.20 | 0.01 |
20000 | 0.20 | 0.90 | 0.30 | 0.05 |
40000 | 0.40 | 0.80 | 0.40 | 0.09 |
60000 | 0.60 | 0.70 | 0.50 | 0.13 |
80000 | 0.80 | 0.60 | 0.60 | 0.17 |
100000 | 1.00 | 0.50 | 0.70 | 0.21 |
