This is a simple and easy-to-use workflow that automatically retrieves prompt words, translates them, and can manually add prefixes and suffixes. Simply upload the image, and the workflow will automatically infer the prompt words, replace static words with dynamic words, and then click generate directly without manually writing prompt words.
purgevram
Description
FAQ
Comments (4)
What does this workflow actually do? You mention prompt words, translation etc.
Are you taking a prompt, enhancing and then translating it into chinese and through that get better prompt coherence?
This reply is from Google Translate, I hope you can understand it. The principle of this workflow is to use the BizyAir API to reversely infer the prompt word from the uploaded image and output it to the text_b of the string function node. You can fill in your own prompt words in text_a and text_c, and the prompt words will eventually be merged in the order of abc. Generate a video for the wan model.
After looking at the very messy image (the text overlap nodes all over the place), ivever uses Comfyui-NYJY to translate from English to one of the Chinese languages. https://github.com/aidenli/ComfyUI_NYJY
You have deprecated nodes. This means that the programmer intends to remove them in the future, but leaves it in the code so that people can make the switch before they are surprised when they find out the code doesn't work anymore. (DEPRECATED)
You don't need WANVIDEOWRAPPER anymore. Comfyui can do WAN21 naively.