I’m having trouble on fine-tuning a BERT model on a classification task, as I’m quite new to this. My data is composed of two columns, “item_title” (my input) and “meta_categ_id” (categorical output).