Fine tune multilingual-e5-small model

[raghavbansal27]

Model I am using is intfloat/multilingual-e5-small
I want to fine tune it on my custom jargon words for sentence similarity task. (Mapping a acronym to its full form) without disturbing previously trained weights of e5 small model.

It would be a great help if you could provide me some resources on how to achieve this.
In which format should I prepare my dataset and how to fine tune this model without disturbing previous weights

Thanks