coreml stable-diffusion text-to-image

Core ML Converted Model:

Note: Some models do not have the unet split into chunks.

LibreEros:

Source(s): CivitAI

Fast and furious model. NSFW.

The main goal is fast output on cheap/processor devices (512x512 images are created in about a minute on the Macbook Air M1).

Extremely pruned mix of well balanced Deliberate and charming Art & Eros (aEros) - A tribute to beauty. VAE included.

All examples are in low/medium resolution and posted AS IS. My dream is to see how this would work at a higher resolution, with more creative prompts.