Hi every one ,
we was waiting for SD3 wich use three clip models (L+G) which are SDXL clips and T5xxx model to enhance understanding of prompt , ELLA has added T5xxx model to SD1.5 which only has (L) clip and the result was great .
i create this workflow to see how much the result become good and i create a model to has more better result specially on face details hands , please try it .
Firstly install Ella node from manager with its models
https://github.com/TencentQQGYLab/ComfyUI-ELLA
The base model
https://civitai.com/models/537608?modelVersionId=597655
The Workflow