If unicode and text together are sent for translation, the model outputs junk which contains variants of 'Narendra Modi'

[Sri20021]

The text list is "chk_list = ['hello', '\ue806x it tomorrow it doesn’t matter how well you have scheduled']" and the returned translated list is ['ನಮಸ್ಕಾರ.', 'narendramodi narendramodi syndramodi narendramodi syndramodi narendramodi syndramodi narendramodi syndramodi narendramodi narendramodi syndramodi syndramodi syndramodi syndramodi syndramodi syndramodi narendramodi narendramodi narendramodi narendramodi narendramodi narendramodi narendramodi'] when translated to kannada language. I am curious why translated list is junk and contains variants of Narendra Modi

[sumanthd17]

We are equally surprised, we'll take a look into this and get back to you. I'll keep the issue open for now, please post any/more such examples where the model fails. It would be good to test the limits of the model