stabilityai

/

sdxl-turbo

like 2.08k

I ran sample script to generate image using this model. I am able to run both on CPU and using Nvidia CPU.
But, the script uses models with extension .safetensors.

The "Files and Versions" tab also has onnx models in each folder unet, text encoder and vae encoder, decoder.

What needs to be done in the script to use ONNX models instead of .safetensors files that are used by default.

Anybody please help.

I need to run the onnx model using onnxruntime on CPU.