This is a repo to save my training framework

Modified from mPLUG-Owl2 and Transformers repo source code

The first edition takes an example for video/image captioning tasks and trying to be more general for other tasks

pyproject.toml -> add fvcore [WIP]

train/custom_model_trainner.py -> add support for list of labels as input to evaluate results. [Done]

scripts [Done]

Test after training [Done]

First implemention [WIP]