How is the part about ‘positional embeding’ in the source code implemented?

[Rfan77]

Hello, I would like to ask how positional embedding is implemented in linevul. Considering that this is a transformer-based model, how can the information about positional embedding be implemented through code? For example, when only testing is performed, faced with this (18864,512) input, how to perform positional embedding, I did not see the code related to positional embedding in linevul_main.py