Added support for Google Gemini 1.5 pro/flash models #1190
Conversation
|
Thanks for raising this PR @rtmcrc! Please resolve the errors raised by each of the checks so that we can proceed reviewing your PR. |
modified: gpt_engineer/core/ai.py
|
Done |
Thanks for this; we will check it! |
|
@captivus could you please take a look at this one? I asked you and @zigabrencic to review it. Would love to see this merged soon so we support Gemini 1.5! |
| load_dotenv(dotenv_path=os.path.join(os.getcwd(), ".env")) | ||
|
|
||
|
|
||
| def model_env(): |
There was a problem hiding this comment.
Is there a way so that we "merge" the model_env() and load_env_if_needed() functions? So to simplify the logic here a bit?
Or what was the idea behind adding this function.
There was a problem hiding this comment.
Good idea, do as you consider it is efficient. I've made a separate function so that it's not called twice.
load_env_if_needed() is called on line 460 and model_env() on 298
Have no clue why it was placed so far, maybe there was a reason, didn't get much into it. 🤷♂️
|
@captivus can you please review when you catch a moment= |
|
Any update on this? |
|
Hey @viborc, @captivus and @zigabrencic are we merging this? If we are not merging this due to some issues. I can take those up. Test it properly. And update required documentations as well for gemini support. But only if we are really thinking to add gemini support for gpt engineer. If we will not merge this. Then also I am thinking to having its support in my local branch and test it properly in my local only. Let me know whatever is the decision. |
|
Hello everyone! Thank you for this PR and comments, but I'm sorry to say that this project is no longer actively maintained by the original board and staff. If you'd like to see it go forward, I'd suggest forking it and using it as an experimentation platform, but make sure to follow the original license and if the original creator (@AntonOsika) has made any other changes or plans. Thank you all for contributing to this project, and I hope to see the lessons learned applied in our domain and industry. |
6 participants
No description provided.