question of usage

[harborsarah]

Dear authors,

I want to ask if it is possible to give an image to the TransGPT and output a description of the given image. If yes, how to use it? The given example is only about giving a prompt and generating an answer.

Thanks a lot for your answering.

[corkiyao]

Dear authors,

I want to ask if it is possible to give an image to the TransGPT and output a description of the given image. If yes, how to use it? The given example is only about giving a prompt and generating an answer.

Thanks a lot for your answering.

you mean the multi-modal model, like image-caption model? But the TrnasGPT still needs a prompt, and the author has achieved it. Please follow the description of the model VisualGLM made by THUDM.You can get it.