sdxl_dpo_turbo / README.md
multimodalart's picture
fix gallery tag
915e14b
|
raw
history blame
838 Bytes
metadata
license: other
pipeline_tag: text-to-image
tags:
  - text-to-image
  - turbo
  - stable-diffusion
  - stable-diffusion-xl
  - dpo
widget:
  - text: '-'
    output:
      url: dpo_turbo_robot.jpeg
inference:
  parameters:
    num_inference_steps: 8

Merge of SDXL Turbo & SDXL DPO

Prompt
rusty robot cartoon

Read their licences before using it.

*.safetensors for AUTOMATIC1111, ComfyUI, InvokeAI

Download *.safetensors file

Use it with 🧨 diffusers

from diffusers import AutoPipelineForText2Image
import torch
        
pipeline = AutoPipelineForText2Image.from_pretrained('thibaud/sdxl_dpo_turbo', torch_dtype=torch.float16).to('cuda')        
image = pipeline('A mecha robot in a favela', num_inference_steps=2).images[0]