Views
No views yet
| VAE | Decoding in float32 / bfloat16 precision | Decoding in float16 precision |
|---|---|---|
| SDXL-VAE | ✅  | ⚠️  |
| SDXL-VAE-FP16-Fix | ✅  | ✅  |
AutoencoderKL:1import torch
2from diffusers import DiffusionPipeline, AutoencoderKL
3
4vae = AutoencoderKL.from_pretrained("madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16)
5pipe = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", vae=vae, torch_dtype=torch.float16, variant="fp16", use_safetensors=True)
6pipe.to("cuda")
7
8refiner = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-1.0", vae=vae, torch_dtype=torch.float16, use_safetensors=True, variant="fp16")
9refiner.to("cuda")
10
11n_steps = 40
12high_noise_frac = 0.7
13
14prompt = "A majestic lion jumping from a big stone at night"
15
16image = pipe(prompt=prompt, num_inference_steps=n_steps, denoising_end=high_noise_frac, output_type="latent").images
17image = refiner(prompt=prompt, num_inference_steps=n_steps, denoising_start=high_noise_frac, image=image).images[0]
18image
sdxl.vae.safetensors file into the webui folder under stable-diffusion-webui/models/VAE--no-half-vae command line arg for SDXL (in webui-user.bat or wherever), you can now remove it
sdxl-vae-fp16-fix is specifically based on SDXL-VAE (0.9), but it works with SDXL 1.0 too