Building agents using SMLs

Posted by shoeshineboy_99@reddit | LocalLLaMA | View on Reddit | 1 comments

If you would want to fine a small language model for a analytical agent. Something which can read docs (text, markdown, json, csv and excel files) and respond to queries which one would you choose?

  1. Qwen 7bn
  2. Gemma 9bn
  3. Phi-4
  4. llama 3 8bn
  5. Mistral 12bn