Community shows improved text-rendering VAE for SD1.5

A Reddit user trained a VAE for Stable Diffusion 1.5 that renders text better than the original. The model is available on HuggingFace.