OpenInterpreter
Add "vision" to Python mode
Describe the changes you have made:
Now we can use GPT-4Vision for only Terminal mode. I added it to Python mode.
Reference any relevant issue (Fixes #795)
- [x] I have performed a self-review of my code:
I have tested the code on the following OS:
- [ ] Windows
- [x] MacOS
- [ ] Linux
AI Language Model (if applicable)
- [x] GPT4
- [ ] ~~GPT3~~
- [ ] ~~Llama 7B~~
- [ ] ~~Llama 13B~~
- [ ] ~~Llama 34B~~
- [ ] ~~Huggingface model (Please specify which one)~~
@Notnaton
Hi, I've attempted to address issue #795.
Two weeks ago, when I first examined this issue, it appeared to be a more challenging case.
From what I recall, this solution I committed did not resolve the issue at that time.
However, upon revisiting it now, it seems to be working correctly.
Could you please review my latest commit to ensure that there are no outstanding problems?
I updated it to the current main branch Sorry, I missed this PR working on a review now
Hey @tegnike! Great work on this one, would have around vision into the Python package much more quickly. Anton had a good profiles PR cooking so I knew we were heading that way instead, but I should have commented about that on this PR much sooner.
Closing in favor of the new profiles, but this is making me realize we need better instructions to set up vision in Python, so I'll update the docs in a moment.
