tlwu/stable-diffusion-xl-refiner-1.0-onnxruntime

Stable Diffusion XL 1.0 for ONNX Runtime CUDA provider

Introduction

This repository hosts the optimized versions of Stable Diffusion XL Refiner 1.0 to accelerate inference with ONNX Runtime CUDA execution provider.

The models are generated by Olive with command like the following:

python stable_diffusion_xl.py --provider cuda --optimize --use_fp16_fixed_vae --model_id stabilityai/stable-diffusion-xl-refiner-1.0

Model Description

• Developed by: Stability AI
• Model type: Diffusion-based text-to-image generative model
• License: CreativeML Open RAIL++-M License
• Model Description: This is a conversion of the SDXL refiner 1.0 models for ONNX Runtime inference with CUDA execution provider.

The VAE decoder is converted from sdxl-vae-fp16-fix. There are slight discrepancies between its output and that of the original VAE, but the decoded images should be close enough for most purposes.