Verified: 2 years ago
SafeTensor
Type | |
Stats | 563 6,106 118 |
Reviews | (93) |
Published | Nov 2, 2023 |
Base Model | |
Training | Epochs: 10 |
Usage Tips | Clip Skip: 2 |
Hash | AutoV2 342B1D277A |
Captured images of several vintage 8mm film footage were used as training data.
It seems to have some effect also on anime models.
Helpful keywords: lowres,blurry,blurry background,blurry foreground,
Setting example:
Weight: 1.5 ~
CFG Scale: 7
Resolution: 640 x 480
Upscaler: Scunet GAN
Denoising strength: 0.4
Please let me know if you find a better setup (especially the upscaler).