Simply run it as follows:

OLLAMA_DEBUG=1 ollama serve

Note: You might want to export another variable to show the ollama serve command where the models are saved in case you receive a 404 about model not available, this happened to me for some reason on Linux only.

In this case, just export these two variables and then run the command:

export OLLAMA_MODELS=/usr/share/ollama/.ollama/models
export OLLAMA_DEBUG=1
ollama serve

That’s it, Enjoy!

Sources

• https://github.com/ollama/ollama/issues/1533
• https://github.com/ollama/ollama/issues/3740