Hello. I'm not sure. What OS are you running on your pi? Do you have the GUI enabled. Did you try tinyllama? If it is a 3b+ or better, I'm sure that we can get it to work.
Are you using aarch64 from the Debian website? I believe that armv7 is 32 bit, while the 3B and B+ also support 64bit operations (but I might be wrong).
![How To Connect Local LLMs to CrewAI [Ollama, Llama2, Mistral]](/img/1.gif)
Thats cool it can run on a pi 3. I got it to run almost realtime on a pi5, but i wonder if it would be possible to run it on a pi zero.
I get this error: "model requires more system memory (1.3 GiB) than is available (952.7 MiB)" Is there any way around it?
Hello. I'm not sure. What OS are you running on your pi? Do you have the GUI enabled. Did you try tinyllama? If it is a 3b+ or better, I'm sure that we can get it to work.
@@TheExpatProfessor I got it working with a smaller model, I am using pi OS 64 bit without gui
It won't install on my RPI 3B it gives this error: "ERROR Unsupported architecture: armv7l"
Are you using aarch64 from the Debian website? I believe that armv7 is 32 bit, while the 3B and B+ also support 64bit operations (but I might be wrong).
@@TheExpatProfessor I just came back to edit my comment. Yes I had the 32 bit OS installed, I have since installed the 64 bit one and it works now
@@skynetcybernetics9058 Outstanding! That is great to hear!
i take this err for rpi4 "-bash: /usr/local/bin/ollama: cannot execute: required file not found "
Hello. You may try using the following command to find it:
# locate ollama
or
# find / -name \*ollama\*