search

SHI-Labs / Prompt-Free-Diffusion

Prompt-Free Diffusion: Taking "Text" out of Text-to-Image Diffusion Models, arxiv 2023 / CVPR 2024
https://arxiv.org/abs/2305.16223
MIT License
735 stars 36 forks source link

make training_dataset #12

Open pokameng opened 1 year ago

pokameng commented 1 year ago

@xingqian2018 hello,can you share some details about how to make training_dataset or how to read training image, reference image and optional image?

xingqian2018 commented 1 year ago

I believe you need to code a bit and create your own dataloader to do it.