CLI repeating random text blocks, no consequence for computation

[aufschreibesysteme]

Describe the bug

After a number of turns of conversation with interpreter, the cli starts copying random blocks of previous output (never any of my input), this can be verbal output or code snippets. This doesn't have an influence on inference, but makes the process often unreadable, active outputs hiding behind random irrelevant overlays, it seems.

Reproduce

1. conda activate 01 (python 3.11)
2. interpreter --profile local

execute a number of interactions, after about 5 to 10 turns, the bug starts happening

Expected behavior

I would like to have extended multi-turn conversations with llms via interpreter.

Screenshots

Open Interpreter version

0.3.14

Python version

3.11.11

Operating System name and version

Ubuntu 22.04.5

Additional context

The same problem occurs on other machines running Ubuntu, i had no chance to try on Windows or Mac.
I have tried to run it in bash and in zshell, with terminator and without - identical behaviours

[Notnaton]

This is a visual bug.
When it generates over the terminal window this happens.
Fixes in version 1.0

[qcjySONG]

Is version 1.0 only compatible with GPT-4o? Why doesn't the local model work? Is the guide for version 1.0 still the same as before?

This is a visual bug. When it generates over the terminal window this happens. Fixes in version 1.0

[Notnaton]

1.0 doesn't have --local
You can set api base, model, api key, to connect to local model still