Stable Diffusion 2.1, CoreML format for macOS Ventura 13.1 and later.<br> Compiled with SPLIT EINSUM attention implementation. This model can be used with all compute units available combination. Output resolution is 512x512.<br> Includes VAEEncoder.mlmodelc to provide img2img capability.<br><br> This is the default model for PromptToImage app, see https://github.com/TheMurusTeam/PromptToImage