IceClear
SR image change more vs src image
when i use :python scripts/sr_val_ddpm_text_T_vqganfin_oldcanvas_tile.py --config configs/stableSRNew/v2-finetune_text_T_512.yaml --ckpt /mnt/image/sr/stablesr_000117.ckpt --vqgan_ckpt /mnt/image/sr/vqgan_cfw_00011.ckpt --init-img inputs/test_example/ --outdir ./output --ddpm_steps 200 --dec_w 0.5 --colorfix_type adain
the output image change more.
which param can improve this problem. thanks!
StableSR is not good at recovering small faces with high fidelity. Actually, such specific scenario is very challenging for general SR. Also, it is not good at recovering Chinese characters since there is no such data during training. For face restoration, you may crop it and enhance it using CodeFormer. For Chinese characters, you may crop it and use textbsr.
StableSR is not good at recovering small faces with high fidelity. Actually, such specific scenario is very challenging for general SR. Also, it is not good at recovering Chinese characters since there is no such data during training. For face restoration, you may crop it and enhance it using CodseFormer. For Chinese characters, you may crop it and use textbsr.
thinks for you suggestion
when i use :python scripts/sr_val_ddpm_text_T_vqganfin_oldcanvas_tile.py --config configs/stableSRNew/v2-finetune_text_T_512.yaml --ckpt /mnt/image/sr/stablesr_000117.ckpt --vqgan_ckpt /mnt/image/sr/vqgan_cfw_00011.ckpt --init-img inputs/test_example/ --outdir ./output --ddpm_steps 200 --dec_w 0.5 --colorfix_type adain the output image change more.
which param can improve this problem. thanks!
Hello, I would like to ask what dataset you are using.
