ITERATIVE PROGRAMMING WITH EXECUTABLE FEEDBACK
how the "ITERATIVE PROGRAMMING WITH EXECUTABLE FEEDBACK" in Section 3.3 of the paper is reflected in the code. I set run_tests=true, but it seems the agent did not execute these tests and return the test results to the engineer.
Thank you for paying attention. The main branch is updated too quickly and the corresponding code needs to be refactored. However, the corresponding participants don't have enough time to catch up recently. We'll see if this part of the code can be taken over and merged.
The value of --n-round is too small; it's exhausted before entering the QA stage. MetaGPT needs to be provided with these parameters: --run-tests --n-round 20, giving --n-round a larger value.
ok, I will go try it out, thx