And if you already have it in Linux (Or Linux on Windows, yeah that is a possibility) and just want to update the Ollama version
Run Ubuntu or whatever flavour you have and in the CLI , after running,
Sudo apt-get update and sudo apt-get Upgrade (updating/grading your linux and repositories so you definitely have the latest)
simply run this command:
curl -fsSL https://ollama.com/install.sh | sh

According to my experience (which is 1 day of using LLM) best of all writes Promt Gemini, it gives results better than Llama 3.1.

Also Llama itself often bugs in other languages

Genial, gleich mal ausprobiert ;-)

Thanks for the share. Has anyone used/installed locally Mistral-Nemo-12B-Base and how does it compare to Llama Version 3.1?

What about WizardCoder for coding? Anyone try it?

Yay thankies a lot for the guide 🩷🩷🩷

it'd be helpful if the png you uploaded had a workflow in it.

I already had them with Griptape, but I am not very fond of the Llama models at least not for vision they really suck at image recognition even though they are boasted to do just that, for making prompts and archiving said prompts thaty are sufficiantly intelligent, but they are good to keep costs down, Those bigger models like ChatGPT, Claude, Gemini API's are hella expensive to run 24/7
I was hoping Deepseek or Janus would be better but they aren't yet