9angles

Model description

First public release of 9angles - demo v0.6

This is a SDXL 1.0 LoRA model. trigger: 9angles

100% trained on generational self-synthesis random AI generated images & AI auto-captioned (copyright-free)

9angles is a free SDXL 2D to 2.5D/3D helper using the same latent space in the product design-creation-display process by displaying 9 images of the same prompt in different angles / poses (sometimes same pose) inside a single image. It can also be used to generate 9 images (inside one) for the same time it takes to generate only one image which make faster generation of new creations/concepts possible.

supported:

• humanoid (head, upperbody to fullbody)

• non-humanoid (head, upperbody to fullbody)

• expressive facial emotions

• object

• humanoid interaction with object

• basic scenery

• different poses

• different styles

• compatible with other SDXL loras

• 768x768 to 1280x1280 (base resolution without hires fix)

• partial consistency

• very versatile

not (yet) supported:

• more than one humanoid/non-humanoid at the same time

• full consistency

• advanced scenery

• trigger concepts for same or different poses

This is my first model release as a technical demo-preview. Keep in mind that as with much AI things, this model should be improved over time, do not expect consistency at this stage. Feedbacks are welcomed, but have no expection of fast reply. We would be glad to see what you can/can't do with this model as it has not been intensively tested yet. If you have time try to share some results of good taste here. Hopefully you will enjoy and have fun with it in a self-responsible and respectful manner.

Enjoy & share the fun!

Trigger words

You should use 9angles to trigger the concept image generation.

Download model

Weights for this model are available in Safetensors format.