Fine tune model='facebook/bart-large-mnli'

Hello,
I am looking for an effective and functioning method to fine tune a model=‘facebook/bart-large-mnli’. I use it because it is a zero-shot-classification model, but is there any way to improve/customize its performances?

Thanks in advance

Hello everyone,
I am looking for a way to train the model used by zero-shot-classification pipeline, that is bart-large-mnli, with TensorFlow. The idea is to fine-tune it on the text I use for my tests. Am I correct saying this is available just with PyTorch? Is there a similar model I could use?

Thanks in advance