Agentcommunity
Author - Aditya Mandlekar ([email protected])
This is an exercise and attempt to contribute to the community where a differnt Design pattern is tried to be added. I understand in the lecture 5th Design pattern and 1st design patter of 1) Prompt chaining 2)Evaluator and Optimizer is used where from 5 Models, based on the quality of their content generation, they are ranked.
I am trying to use an Design Pattern called as "Parallelization" where Coordinator LLM tries to collaborates across LLM to find
- Common findings and 2) Unique findings from a set of given LLM and their recommendations and make final summary report in terms of Executive summary and conclusions
Finally have 3rd Design Pattern to evaluate the findings and give final verdict if its accepted or not accepted
Looking forward for your feedback on my solution
Hey @adityamandlekar - thanks, those sound like useful patterns. It looks like this PR includes changes to files outside the community_contributions folder.. if you wanted me to merge, would you be able to remove those? Thanks so much