llm awesomeness

Posted by fab_space@reddit | LocalLLaMA | View on Reddit | 23 comments

just to share a quick journey into llms: yesterday i was wondering about langflow and similar solutions, testing some flow etcetc, really inspiring. tested using phi3, llama3, groq avail models and ClosedAI models ofc. then i said myself: just do it, build a simple pipeline to fetch some feeds and rewrite the content based on multiple sources assuming a specific assistant role (ex: a common ugly citizen) i am not a coder and often I use a custom GPT i tailored for my skills and workflows to build the code. it worked. in the code i used local llama3 to perform the processing of feeds data and generating what’s needed, it worked! in less than 24 hours my idea become a working PoC to me :) to use a closed LLM to build some code and to use open LLm in such code, and such code works like a charm, also the processed data.. _is something magic to me_ 🤣