Very nice work, About in-between?

[lxyxl0216]

Does the model support editing? included in-between

[exitudio]

I didn’t clean up the code for editing generation, but the code is already in the evaluation function. Here are some information:

1. Change edit_task to one of these: "inpainting, outpainting, prefix, suffix"
2. It will encode the condition tokens and replace [Mask] tokens where to inpaint here
3. Then it will call the edit function. The model will automatically create masked attention based on the tokens (with [Mask]) and generate tokens autoregressively with condition.