Home
Special content provided by the community
Inspiration
Share Your Brilliant Creations
AI Apps
Free and High-quality AI App
AI Characters
Immersive AI Character Interaction
Models
Various Powerful AI Models
AI Video
Create Videos Using AI with Ease
LoRA Training
Customize Your AI Model Style
ComfyUI
Simplify Your AI Creation Process
AI Audio
Generate Realistic Audio Content Using AI Magic
No relevant data found yet
Download the SeaArt App
Please scan this QR code with your camera to download the SeaArt app on your phone.
Notifications
Gay's Anatomy v2 was created to improve on and expand the core concept of v1. I designed v2 specifically to address the limited checkpoint compatibility of v1. v2's dataset has been increased by 50%. The original dataset I created consisted of exclusively "anime-style", for lack of a better term, images. You can see examples of this style by viewing the example images I posted with v1. For v2's dataset, I created images that were much more photorealistic. This resulted in a much more versatile model, a model that can draw photorealistic images, anime-style images, and the style that I am most interested in at the moment, a pseudo-realistic hybrid of those two styles with an element of a 3D rendering aesthetic. The example images I am including were created from using this model with a variety of popular checkpoints and I will include the names of those checkpoints in the model information. Some of the images were created with a checkpoint I created specifically to draw v2 in the pseudo-realistic style I described.
Spike Driver
This is an advanced checkpoint I mixed and infused specifically to draw and enhance this model. It can take this model into the forefront of what SDXL is capable of. I think most people will be able to get something from it out of the box, but mastery of it's potential may take some effort and new strategies.
v2 can do all of the things v1 can do in addition to being much more compatible with checkpoints beyond just the NTR Mix checkpoint that I have linked at the beginning of this listing. In addition, I have reinforced the ethnicities and added Indians (subcontinent) to the model. The supplementary dataset included close-up portraits I created which strengthened the model's ability to draw a variety of faces in a variety of races/ethnicities.
Both v1 and v2 might not be friendly for image generation novices. They will not draw the type of images I post as examples simply by activating the LoRA with the trigger word. They will help SD draw better male figures to some degree by activating the LoRA with the trigger word alone, granted that your prompting mentions or describes a male. To get the most from these LoRAs, you will need to stack your prompt with at least a couple of the tokens I have listed in green as secondary triggers.