I hope to use SAM's own prompter to input point and label for inference

[GZ-YourZY]

Description

I wish to extend the multimodal library and the code in the Convlora example

To fine-tune SAM through ConvLoRA, I hope to use the prompter of SAM itself to input point and label for inference to see whether the model has obtained the ability of the corresponding task during fine-tuning, rather than obtaining a good prompt encoding. This question is actually It's very important. I hope the code can be expanded to implement a prompt segmentation function similar to SAM, not just an end-to-end model.