File size: 2,431 Bytes
732f851 06bee35 732f851 06bee35 fd0f29a 8ab04db 583db6c 97ea5f1 583db6c 97ea5f1 fd0f29a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 |
---
license: mit
tags:
- stable-diffusion
- stable-diffusion-diffusers
inference: false
---
# SDXL-VAE-FP16-Fix
SDXL-VAE-FP16-Fix is the [SDXL VAE](https://huggingface.co/stabilityai/sdxl-vae)*, but modified to run in fp16 precision without generating NaNs.
| VAE | Decoding in `float32` / `bfloat16` precision | Decoding in `float16` precision |
| --------------------- | -------------------------------------------- | ------------------------------- |
| SDXL-VAE | ✅  | ⚠️  |
| SDXL-VAE-FP16-Fix | ✅  | ✅  |
## 🧨 Diffusers Usage
Just load this checkpoint via `AutoencoderKL`:
```py
import torch
from diffusers import DiffusionPipeline, AutoencoderKL
vae = AutoencoderKL.from_pretrained("madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16)
pipe = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-base-0.9", vae=vae, torch_dtype=torch.float16, variant="fp16", use_safetensors=True)
pipe.to("cuda")
refiner = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-xl-refiner-0.9", vae=vae, torch_dtype=torch.float16, use_safetensors=True, variant="fp16")
refiner.to("cuda")
n_steps = 40
high_noise_frac = 0.7
prompt = "A majestic lion jumping from a big stone at night"
image = pipe(prompt=prompt, num_inference_steps=n_steps, denoising_end=high_noise_frac, output_type="latent").images
image = refiner(prompt=prompt, num_inference_steps=n_steps, denoising_start=high_noise_frac, image=image).images[0]
image
```
## Details
SDXL-VAE generates NaNs in fp16 because the internal activation values are too big:
SDXL-VAE-FP16-Fix was created by finetuning the SDXL-VAE to:
1. keep the final output the same, but
2. make the internal activation values smaller, by
3. scaling down weights and biases within the network
There are slight discrepancies between the output of SDXL-VAE-FP16-Fix and SDXL-VAE, but the decoded images should be close enough for most purposes.
---
\* `sdxl-vae-fp16-fix` is specifically based on [SDXL-VAE (0.9)](https://huggingface.co/stabilityai/sdxl-vae/discussions/6#64acea3f7ac35b7de0554490), but it works with SDXL 1.0 too |