We evaluate the SFE dataset using lmms-eval. The evaluation codes are listed in this folder.
We only provide the evaluation scripts for GPT series and InternVL series. Evaluation scripts for other models follow the same format and can be easily adapted.
