LyCORIS: Difference between revisions
Jump to navigation
Jump to search
(added LyCORIS page) |
No edit summary |
||
| Line 1: | Line 1: | ||
LyCORIS stands for '''L'''ora be'''Y'''ond '''C'''onventional methods, '''O'''ther '''R'''ank adaptation '''I'''mplementations for '''S'''table diffusion. | LyCORIS stands for '''L'''ora be'''Y'''ond '''C'''onventional methods, '''O'''ther '''R'''ank adaptation '''I'''mplementations for '''S'''table diffusion. | ||
It introduces a suite of techniques for applying nuanced modifications to a [[Stable Diffusion]] [[model]] [[checkpoint]], drawing parallels to the [[LoRA]] | It introduces a suite of techniques for applying nuanced modifications to a [[Stable Diffusion]] [[model]] [[checkpoint]], such as modifying the style of an image, injecting a character, or adding an animal, drawing parallels to the [[LoRA]] method. | ||
Both LyCORIS and LoRA target the nuanced refinement of Stable Diffusion models, employing a compact, lightweight file for this purpose. They each alter the [[UNet|U-Net]] by employing matrix decomposition using different techniques. | |||
LoRA represents the original approach, whereas LyCORIS encompasses a range of newer, LoRA-inspired strategies, namely LoCon, LoHa, LoKR, and DyLoRA. | LoRA represents the original approach, whereas LyCORIS encompasses a range of newer, LoRA-inspired strategies, namely LoCon, LoHa, LoKR, and DyLoRA. | ||
Revision as of 16:14, 2 February 2024
LyCORIS stands for Lora beYond Conventional methods, Other Rank adaptation Implementations for Stable diffusion.
It introduces a suite of techniques for applying nuanced modifications to a Stable Diffusion model checkpoint, such as modifying the style of an image, injecting a character, or adding an animal, drawing parallels to the LoRA method.
Both LyCORIS and LoRA target the nuanced refinement of Stable Diffusion models, employing a compact, lightweight file for this purpose. They each alter the U-Net by employing matrix decomposition using different techniques.
LoRA represents the original approach, whereas LyCORIS encompasses a range of newer, LoRA-inspired strategies, namely LoCon, LoHa, LoKR, and DyLoRA.