attempt at a1111 instructions

Rename sdxl_vae_fp16_fix.safetensors to sdxl.vae.safetensors

new look, same great taste

0.9 -> 1.0

Link to Kubuxu's eval results

Improved decoder weights

Improved encoder weights

Update README.md

Add example for Diffusers (#4)

VAE - don't force the vae to be upcasted to FP32 automatically (#3)

Upload sdxl_vae.safetensors

Update README.md

Upload activation-magnitudes.jpg

Update README.md

Update README.md

Upload images/orig-fp32.png with huggingface_hub

Upload images/orig-fp16.png with huggingface_hub

Upload images/fix-fp32.png with huggingface_hub

Upload images/fix-fp16.png with huggingface_hub

Upload diffusion_pytorch_model.safetensors with huggingface_hub

Upload diffusion_pytorch_model.bin with huggingface_hub

Upload config.json with huggingface_hub

Upload README.md with huggingface_hub

initial commit