Training Finetune with IMAGE + TEXT

I want to fine-tune model on a custom data with IMAGE and TEXT.

I check the provided [README.md](https://github.com/facebookresearch/sam3/blob/main/README.md) and [README_TRAIN.md](https://github.com/facebookresearch/sam3/blob/main/README_TRAIN.md)

But could not find any steps on how to get started on preparing the data and train using it?

Start with README.md and README_TRAIN.md, then identify the missing instructions for preparing custom image-and-text data and launching fine-tuning. The documentation is done when a reader can follow the added steps to prepare the data and train a model using this repository.

Written by the indexing model from the issue text.