As stated in the UniXcoder paper, "the model takes comment and flattened AST as the input". Please, I have the following questions:
1- Is the implementation of the AST parser of UniXcoder available?
2- Is it necessary to apply the AST parser to your own custom input data for fine-tuning?
3- If the answer for the question 2 is "no', why is it not necessary to apply the AST parser to my input data when fine-tuning UniXcoder for a given downstream task?
As stated in the UniXcoder paper, "the model takes comment and flattened AST as the input". Please, I have the following questions:
1- Is the implementation of the AST parser of UniXcoder available? 2- Is it necessary to apply the AST parser to your own custom input data for fine-tuning? 3- If the answer for the question 2 is "no', why is it not necessary to apply the AST parser to my input data when fine-tuning UniXcoder for a given downstream task?
Thank you in advance