A simple 3 steps workflow to reimagine a 1024x1024 portrait, without the need to write a prompt.

A more detailed description on LinkedIn:

https://www.linkedin.com/pulse/alternative-comfyui-stable-diffusion-workflow-ai-based-marco-rossi-1lncf/

Currently, i'm using an outdated GPU (GTX 1060), is really hard to develop complex workflows.

You can support me in buying a new workstation, or just buy me a coffee :) --->  https://ko-fi.com/redmark

The workflow was developed and tested with SDXL Turbo Models

-  https://civitai.com/models/229392?modelVersionId=372600

-  https://civitai.com/models/139562?modelVersionId=344487

-  https://civitai.com/models/229392?modelVersionId=342732

Tha "Canny" control net model is here :  

https://huggingface.co/stabilityai/control-lora/blob/main/control-LoRAs-rank256/control-lora-canny-rank256.safetensors?download=true

# md5sum.exe control-lora-canny-rank256.safetensors

66cc106b2839f812fab050348cad4bca *control-lora-canny-rank256.safetensors

The diffusion_pytorch_model is here :

https://huggingface.co/InstantX/InstantID/resolve/main/ControlNetModel/diffusion_pytorch_model.safetensors?download=true

# md5sum diffusion_pytorch_model.safetensors

ef2f9d0a1e0cb053e580e6e9bdc87d90 *diffusion_pytorch_model.safetensors