Like it says on the tin, it's based on the API that Stability.AI themselves created, and is geared towards maximum efficiency for anything Machine Learning. It currently uses its own scripts for diffusion using the Diffusers library, but we look forward to using the API with Project Nataili when the backend is ready for advanced use.
Allow generations to be processed by the StableHorde rather than on a local GPU, allowing users with lower resources to generate things easily
Is there an existing issue for this?
What would your feature do ?
Proposed workflow
N/A
Additional information
See also: