simple latent upscale with no face fix or any other nodes. i think everything in this version is vanilla minus a few chroma specific nodes like the t5tokenizer and flow shift. feel free to swap to the full clip as well as i'm using a compressed version of that in this workflow.

for locations of the clips used to generate this image:

https://huggingface.co/camenduru/FLUX.1-dev/blob/main/clip_l.safetensors
https://huggingface.co/city96/t5-v1_1-xxl-encoder-gguf/tree/main

these both go inside of ../ComfyUI/models/clip