search

microsoft / unilm

Large-scale Self-supervised Pre-training Across Tasks, Languages, and Modalities
https://aka.ms/GeneralAI
MIT License
19.25k stars 2.45k forks source link

how to create mask for text-inpainting textdiffuser #1344

Open cuongngm opened 9 months ago

cuongngm commented 9 months ago

Describe Model I am using (UniLM, MiniLM, LayoutLM ...): textdiffuser

dear author, as title, i want to ask, how to create mask (store in assets/examples/text-inpainting/_mask.jpg) for text-inpainting textdiffuser? Is model work with the text in another language (other than English) and how to finetune it?

JingyeChen commented 9 months ago

Thanks for your attention to TextDiffuser. Since the input image is resize to 512x512, you can use some auxiliary tools (e.g., powerpoint) to design the mask.

Currently TextDiffuser only supports English and you need to collect dataset in other language at first for fine-tuning.