Wrote the complete env and built the system and trained Qwen 0.5B model and 1.5B model and fine tuned them and generated graphs(training and generating stats took more time but since that was outside the code editor, that time wasn’t counted in the Wakatime). After all this i used AI to write comments in all the code files so that anyone reading it can understand it and made sure the readme is enough for anyone to understand the whole project. the attached images are the stats of the trained models and their performance. The SFT showed great performance after training in my environment.