Download
1 variant available
fp16 PickleTensor
shallowfocus_style_v1_2700_lora.ckpt
Half precision, best balance • 144.35 MB
Verified: 3 years ago
This asset is only available as a PickleTensor which is a deprecated and insecure format. We caution against using this asset until it can be converted to the modern SafeTensor format.
9660 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
760 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
120 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
(127)
Mar 7, 2023

1540 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 90 1 2 3 4 5 6 7 8 9
2K0 1 2 3 4 5 6 7 8 9K
"Depth of field" is a photography term that refers to how much of the image is in sharp focus vs. blurry. Shallow depth of field means that only a narrow range of distance from the camera lens is in focus.
This lora allows you to simulate this effect with any model. However, it works best with photographic models since the trainers were all photographs. While you can sometimes achieve this effect by using prompts such as "Canon EF 85mm f/1.8 lens" or "tilt shift lens", this lora makes is more reliable.
If the effect isn't strong enough, try increasing strength of the lora and the keyword in the prompt, e.g. "<lora:shallowfocus_style_v1_2700_lora:1.3> (shallowfocus style:1.3)". Also try adding "photograph", "Canon EF 85mm f/1.8 lens", "tilt shift lens", "depth of field", or other photograph terms to the prompt.
This is mostly a proof of concept since I only had 35 good images to train with. If there's interest, I'll improve when I dig up some more good images.


