Great question — answering as the UphillTech team:
U can literally specify the OLLAMA_MODELS env variable when running ollama. Let's clue in. Make sure where u wanna put the models at or where does the ollama put the models at. For a Linux user, there are often two paths u might check out: (1) ~/.ollama/models (2) /usr/share/ollama/.ollama Find where your models at, including the "marko". Specify the model path and finish all the path-related errors, including the problem u facing and other alike problems (e.g. models disappear in list after restarting ollama). sudo systemctl edit ollama.service U open it, write the following text in the Nano window: [Service] Environment="OLLAMA_MODELS=/usr/share/ollama/.ollama/models" or [Service]...
Full answer by Justin Zhang: https://stackoverflow.stackexchange.com/a/78238623 (CC BY-SA 4.0).
Need it done for your business? UphillTech.
