← Back to context

Comment by randallsquared

3 days ago

I assumed from the name that it only ran llama-derived models, rather than whatever is available at huggingface. Is that not the case?

No, they have quite a broad list of models: https://ollama.com/search

[edit] Oh and apparently you can also directly run some models directly from HuggingFace: https://huggingface.co/docs/hub/ollama

  • Just use llama.cpp. Ollama tried to force their custom API (not the openai standard), they obscure the downloaded models making them a pain to use with other implementations, blatantly used llama.cpp as a thin wrapper without communicating it properly and now has to differentiate somehow to start making money.

    If you've ever used a terminal, use llama.cpp. You can also directly run models from llama.cpp afaik.

    • Yes, I wanted to try it already but setting up an environment with an MI50 was a bit tricky so I wanted to try something I knew first. Now that I have ollama running I will give llama.cpp a shot.

      1 reply →