Google's T5 base fine-tuned on Tweet Sentiment Extraction Dataset for Span Sentiment Extraction

While trying to reproduce this Colab Notebook created by Lorenzo Ampil, using an FR translated version of the Tweet Sentiment Extraction Dataset for Span Sentiment Extraction downstream task, I keep getting an error: ImportError: cannot import name 'trim_batch' from 'transformers.tokenization_utils' (/usr/local/lib/python3.8/dist-packages/transformers/tokenization_utils.py)

This error pops up at the section:

Model

Based on this summarization (encoder-decoder) example from [transformers](https://github.com/huggingface/transformers/blob/master/examples/summarization/bart/finetune.py), which is compatible for both BART and T5

I am guessing that the problem is related to the french translated version of the dataset… how can i fix this?

Thanks for your help!

eduardo