metagpt generate too many inconsistent invocation code between different source files
metagpt always generate inconsistent code between different source files. For example, in source file A, it contains a method a(p1,p2),this method have 2 parameters, but in source file B, it will generate invocation code for the method a(p1) which only contains 1 parameters of that method, and sometime in source file A, it will generate invocation code for a non-exist method in other source file, and there are too many such cases in project for me to fix them, is there any good way to avoid such inconsistent codes?
And sometimes the generated source code will call a non-exist attribute of a object, these inconsistent codes make project unable to run.
What's your LLM model, prefer gpt-4-turbo
What's your LLM model, prefer
gpt-4-turboI use this LLM model
is gpt-4-turbo-preview different from gpt-4-turbo?
gpt-4-turbo is the latest model.
You can check these cases, they are all generated by metagpt, sometimes it's neccessary to adjust the code.