Pretraining?

[dustinwloring1988]

@MayankChaturvedi
Would you be willing to share a notebook or script for some pretraining maybe using this dataset: https://huggingface.co/datasets/DAMO-NLP-SG/multimodal_textbook as it would be nice to have a Multimodality model to train and see how it all works?

[Balakrishna-Chennamsetti]

Hi @dustinwloring1988 ,

Pre-training the model from base level takes lots of time and resource consumption, it's recommended to do the fine-tuning the models on top of the pre-trained based models, as it gives the same level of performance in addition to pre-trained model capabilities. Please find the following page to know about the tuning the model using transformers - page.

Thanks.