model not found on macOS after upgraded to 0.1.38

[eliranwong]

was using Ollama fine before upgraded to 0.1.38

'''
raise ResponseError(e.response.text, e.response.status_code) from None
ollama._types.ResponseError: model 'wizardlm2:7b' not foun
'''

I am sure the files are in ~/.ollama/models. However,

• model not found by Ollama-python package
• It is not displayed with command "ollama list"

I tried:

launchctl setenv OLLAMA_MODELS "~/.ollama/models"

export OLLAMA_MODELS="~/.ollama/models"

no difference after restarting Ollama.

[jmorganca]

Hi @eliranwong sorry this happened. Had you previously set OLLAMA_MODELS to a different value? If you run ls ~/.ollama/models/manifests/registry.ollama.ai/library do you see data in there?

[eliranwong]

Hi @jmorganca , thanks for reply. No, I didn't set OLLAMA_MODELS before.

When I run echo $OLLAMA_MODELS, I see nothing.

When I run ls ~/.ollama/models/manifests/registry.ollama.ai/library, I can see data there.

When I install a new model, by using 'ollama pull', I don't see it in ~/.ollama/models/manifests/registry.ollama.ai/library