I found its hard for me to read the transformers code, because:
- Definitions of the functions are normally very long, which makes it hard to catch the main process there.
- The parameters are a lot for a function, but it always end up being used in that function’s son functions, which has no note attached, since it has been noted in its father functions, so it is hard to trace how it is being used.
My personal solutions to those are:
- Search those function/parameters directly in google, not in the code repository. If someone ask that before, lucky for me.
- Tracing the process by purely adding breakpoints and debugging, to understand the code.
Those solutions did not work out too well, is there any advice for a beginner like me? Thanks for sharing.