AbanteAI/mentat

help me

utpal4job opened this issue · 8 comments

Hello, i have checked your installation video and made the installation in the same way but when running "mentat ." on ssh then it's listing all the files on ssh. i am not getting like you are getting when running "mentat ."

one error is appearing after listing all the files. please check the attached screenshot.

Screenshot 2023-12-12 155639

Regards.

Hey @utpal4job! Thanks for catching this; we changed how we handled file includes recently and this was a bug that slipped through. I made PR #383 which should fix the bug; once it's merged, if you git clone mentat and try it out, it should be fixed. Thanks!

Mentat 1.0.6 is now available on pypi, so if you install with pip install --upgrade mentat you'll get the fixed version!

ok, thanks. i will check it.

Hello, thanks it's working fine but gpt4 model error appears. how can i change the gpt model?

pls check the attached screenshot.
error

Regards.

There are ways to change the model, but there aren't models than can handle as many tokens as you are trying to include. I assume you are running with mentat . which adds everything in the directory to context. Instead how about you try running with mentat path/to/file_a path/to/file_b, adding just a couple files you want to work with?

A more advanced feature is to have it select files automatically based on what you ask. This is called auto context, and you can run with mentat -a

@utpal4job

Hello, thanks for your reply. i have changed the gpt model and it's working fine.

Regards.

What model did you change it to?

I have used "gpt-4" instead of "gpt-4-1106-preview".