Hello, your work is fantastic and I really like it!
When I studied your code, I found the directory rllm/tools defining some function calling api (like calculator, python interpreter and web search). However, I cannot figure out the relationship between rllm/tools and your LLM training and evaluating code. Did you use rllm/tools for training or evaluation? If yes, how?
I would appreciate it if you could answer my questions :)