MiniRobotLanguage (MRL)
SDO.Set Clip_Guidance
If Result available
Intention
The SDO.Set_Clip_Guidance command configures the clipping guidance preset, which influences how the Stable Diffusion Online (SDO.) process handles and refines specific features and patterns within the data.
Clip guidance presets provide predefined settings that govern the intensity and precision of the clipping operations during the diffusion process, affecting the final output's appearance and characteristics.
Technical Explanation - Clip Guidance
CLIP_guidance is a technique that can be used to enhance the quality of image generation in Stable Diffusion.
CLIP is an AI model that can look at a picture and tell you what’s in it.
Normally, Stable Diffusion takes a version of CLIP and reverses it to make images.
With the addition of CLIP_guidance, CLIP also ‘double checks’ during image generation to see if it’s doing things correctly.
This can slow down the generation process, but it should help improve the quality of the generated images
•CLIP_Guidance is a parameter that can be used to specify the level of CLIP guidance for text-to-image generation in Stable Diffusion1. CLIP is an AI model that can look at a picture and tell you what’s in it. Normally, Stable Diffusion takes a version of CLIP and reverses it to make images. With the addition of CLIP guidance, CLIP also ‘double checks’ during image generation to see if it’s doing things correctly. This can slow down the generation process, but it should help improve the quality of the generated images.
•The valid values for CLIP_Guidance are FAST_BLUE, FAST_GREEN, NONE, SIMPLE, SLOW, SLOWER, and SLOWEST. These values correspond to different levels of CLIP guidance, ranging from none to very high. The higher the level of CLIP guidance, the more accurate and realistic the generated images should be, but the longer it will take to generate them.
•According to Stability AI’s blog post, FAST_BLUE and FAST_GREEN are experimental modes that use a different version of CLIP than the other modes. They are faster than SIMPLE, but may produce less consistent results. NONE means no CLIP guidance at all, which is the fastest mode but may produce low-quality images.
•SIMPLE is the default mode that uses a basic version of CLIP guidance.
Syntax
SDO.Set_Clip_Guidance[|P1]
Parameter Explanation
P1 - (optional) Numeric valube between 0 and 6 or the name of the wanted Algo, see Table below.
Theoretically "6" should return the most accurate result.
Number |
validClipGuidance |
0 |
FAST_BLUE |
1 |
FAST_GREEN |
2 |
NONE |
3 |
SIMPLE |
4 |
SLOW |
5 |
SLOWER |
6 |
SLOWEST |
Example
'***********************************
'
'***********************************
Remarks
-
Limitations:
-
See also:
• ! Smart Package Robot 's Parallel Robot Operations