Using interpreter --os --local causes the output to flip out

[Onceisenoughyo]

Describe the bug

I've tried this multiple times and Mike from the discord server has told me to create an issue report. When I try to enable OS mode by using interpret --os --local and using obsidian-q6.gguf with LM Studio, my output keeps repeating the same few words or phrases. This will continue until I kill the process by using ctrl-c. I'm on a windows laptop and here's my output from doing %info Python Version: 3.11.7 Pip Version: 23.3.2 Open-interpreter Version: cmd:Interpreter, pkg: 0.2.0 OS Version and Architecture: Windows-10-10.0.22631-SP0 CPU Info: Intel64 Family 6 Model 186 Stepping 2, GenuineIntel RAM Info: 15.62 GB, used: 11.92, free: 3.69

    # Interpreter Info

    Vision: True
    Model: gpt-4-vision-preview
    Function calling: False
    Context window: 110000
    Max tokens: 4096

    Auto run: True
    API base: None
    Offline: True

    Curl output: Not local

    # Messages

Reproduce

Open LM Studio Enable local server for obsidian-q6.gguf Open Powershell Type interpreter --os --local Allow OS mode to enable Type anything (I typed "Open a new edge tab") Receive broken output

Expected behavior

OS mode should enable and do what you'd want. For me, it would open a tab in Microsoft Edge

Screenshots

Open Interpreter version

0.2.0

Python version

3.11.7

Operating System name and version

Windows 11

Additional context

I don't think this laptop has a dedicated GPU. Could that be part of the problem? LM Studio works fine in normal interpreter --local mode

[MikeBirdTech]

I believe the issue is with the model being used. It may be more sensitive to the lack of meta tags. I will download it to verify, but I cannot replicate this issue using Mixtral. @Onceisenoughyo you mentioned that there is no dedicated GPU so I don't know if you can try with other models but if you can, comparing functionality between models would be helpful