Continue Initial documentation assumes Ollama as default

[yrobla]

When we see the documentation for Continue integration we just see, by default:

Install a CodeGen model
Download the Code Llama model using Ollama. The following installs the 7 billion parameter (7B) version of the model, which is suitable for systems with 16GB of RAM or more:

ollama pull codellama:7b-instruct

info
If you have at least 32GB of RAM and an 8-core CPU, you can try the 13B parameter model. Replace 7b in these instructions with 13b.

this is just needed for ollama, but is listed in general. I suggest that for Quickstart, we suggest an easier method, such as using the self-hosted in google (or if we really want to use ollama there, give more detailed instructions about how to install it, and mention that this is just one way). Then link to "Use with Continue" section, with the different options there:

Also those tabs should add some instructions/links on how to install each method, dependencies, how to configure api keys etc

[lukehinds]

I think @danbarr is already / had already looked into this

[danbarr]

The quickstart is intended to be self-contained, i.e. a first-time user can follow it without requiring any external / paid services, thus the choice to use Ollama. I did add a callout based on other feedback, to hopefully make this more clear and guide users to the full how-to guide for other providers.

Also those tabs should add some instructions/links on how to install each method, dependencies, how to configure api keys etc

I'm hesitant to reproduce documentation for third-party tools that we don't control, we'll be constantly chasing their own changes. Point taken about at least mentioning the dependency, I'll add that with a link to their install docs.