How to handle translations one source language to many target sentences for the same language

🤗Tokenizers
1

Hi,
I am training a m2m100 model with more technical corpus data using IATE database (https://iate.europa.eu/)

I don’t know how to handle a case , when a source text may have different significations for the same sentence.
That means for one source language sentence , I will assign 3 targets language definitions

What would happen ???

by example , I have these data for the same term, I have 3 associated definitions in english language :

line[61163] = ['1448173', 'mechanical engineering', 'en', 'to jam', 'Term', 'Reliable', '', '0', '', '', 'COM', '2014-05-19T14:56:28.001Z']
line[61164] = ['1448173', 'mechanical engineering', 'en', 'to seize', 'Term', 'Reliable', '', '0', '', '', 'COM', '2014-05-19T14:56:28.001Z']
line[61165] = ['1448173', 'mechanical engineering', 'en', 'to get stuck', 'Term', 'Reliable', '', '0', '', '', 'COM', '2014-05-19T14:56:28.001Z']
line[61166] = ['1448173', 'mechanical engineering', 'de', 'festlaufen', 'Term', 'Reliable', '', '0', '', '', 'COM', '2014-05-19T14:56:28.001Z']