Pre-requisites

Goal

  1. Be able to use some level of code completion without leaking code (coz that what I stand for baby. Privacy!).
  2. Learn how to use LLMs for coding purposes.

I installed 3 models, recommended by the Continue team themselves:

  1. Llama 3.1 8B (Chat),
  2. Qwen 2.5 Coder 1.5-base (Autocomplete),
  3. and Nomic Embed Text (Text Embeddings)

My experiments today were unfortunately halted because Quantiphi bloody blocks the ollama site (so I can’t read documentation) and even google AI studio ka site (so I can’t generate API keys for Gemini or something). My Macbook doesn’t support Llama 3.1 8B model coz the system ram requirements aren’t being met. Hence I wanna see if I can use the Gemini model from our company ka GCP Org. Would be a good thing to try out too.

Footnotes

  1. I did it using Docker