OWL V2 Fine Tune Support

[zhanglu0704]

Hi @mjlm. We've tried using OWLv2, and its performance is impressive; thank you for your amazing work on it. In order to use this model in some specific scenarios, we would like to train(fine-tune) it on some in-domain data. However, the training code appears to be only effective for OWL v1 and does not support training the OWLv2 model. If it is passible we kindly request your assistance and support with this.