model requires more system memory than is available when useMmap

[xgdgsc]

What is the issue?

When I use continue vscode extension to call ollama config like

    {
      "model": "qwen2.5-coder:14b",
      "title": "qwen2.5-coder:14b",
      "provider": "ollama",
      "completionOptions": {
        "keepAlive": 9999999,
        "useMmap": true
      }
    },

It still checks system memory disregard the "useMmap": true option. And return 500 internal error like:

{"error":"model requires more system memory (17.7 GiB) than is available (13.6 GiB)"}

OS

Windows

GPU

No response

CPU

Other

Ollama version

0.4.7

[rick-github]

mmap doesn't affect the check for memory. If your system doesn't have enough system memory to load the model, you need to increase it by adding swap.

[xgdgsc]

No the logic is wrong. Both Windows and macos does swap adding dynamically automatically as I mentioned at https://github.com/ollama/ollama/issues/6918#issuecomment-2488221380 . So when a user knows what mmap is and ask for the config you shoiuld skip the checking.

[xgdgsc]

Has the latest windows version update changed the useMmap behavior? It doesn' t seem to mmap but copied to memory?

EDIT: Only first run have this issue. Not happening now.

[xgdgsc]

And the Ollama app doesn' t support useMmap.