support for ollama

[gtnbssn]

Cross-posting this from gsuuon/model.nvim#9 which i thought was the same thing!

---

Does this support ollama?

So far ollama has been my smoothest experience to set up an llm locally.

Here's what the API use looks like (as documented here):

curl -X POST http://localhost:11434/api/generate -d '{
  "model": "phind-codellama",
  "prompt":"Implement a linked list in C++"
 }'

[McPatate]

I realised that a project with the same name existed after I created it, sorry about that. Hopefully Github's namespacing is enough to differentiate.

I'm discovering the project with your question, so no llm.nvim doesn't support it. I haven't really looked into running models locally yet, though I'd like to get there at some point. I'm going the language server route for this plugin, I initially imagined using candle to load and run models locally.
Nothing is in the way of supporting multiple backends though, so I'll explore the idea when I get to it, eventually.

[McPatate]

Ah but wait I think I understand what you're asking for, it shouldn't be that complicated to support!

I've created an issue: huggingface/llm-ls#17. I'll see depending on my other priorities when I'll tackle it!

I'll start small by just adding the querying to get generations, we'll see later for model creation.

[eratio08]

Any updates here, as the adapters are now implemented in llm-ls? Thanks for the great work!

Ok nvm just saw https://github.com/huggingface/llm.nvim/pull/72/files?short_path=b335630#diff-b335630551682c19a781afebcf4d07bf978fb1f8ac04c6bf87428ed5106870f5 :D

[McPatate]

Ollama is now supported as of 0.5.0.