[Notes] Provide some notes about using DeepFloyd

[Ren Tianhe]

Hello! I've already tested some awesome function about deepfloyd model and want to build some demo based on it, and I'd like to share some notes about using IF, here's some notes on it and I will update it after more testing: Notes

[phalexo]

It may be helpful for people to know that different stages can be put on different GPUs, for those who have GPUs with smaller VRAM but several of them.

[Ren Tianhe]

It may be helpful for people to know that different stages can be put on different GPUs, for those who have GPUs with smaller VRAM but several of them.

Thanks for your advice, I'm testing the ability of IF these days and I will take your advice to update my notes~

[tildebyte]

@rentainhe; Your text-to-image prompts are waaaayyy too long (see e.g. #108) - otherwise, nice guide! 👍

[Ren Tianhe]

@rentainhe; Your text-to-image prompts are waaaayyy too long (see e.g. #108) - otherwise, nice guide! 👍

lolll, it seems floyd limit the token nums~ but anyway, using language prompts generated by GPT-4 really create nice images