⚠️ Please check that this feature request hasn't been suggested before.
🔖 Feature description
Current code limits the models to "gpt-3.5-turbo", "gpt-4" and "gpt-4-32k". In some cases like summary, it kind of defaults to "gpt-3.5-turbo".
But there are other OpenAI models, right now and in the future too.
Especially to read long web pages, I would like to use "gpt-3.5-turbo-16k". It gets rejected by the LoopGPT module code, although it would work with OpenAI just fine.
thanks
✔️ Solution
Refactor the code so that all models supported by OpenAI can be used, at least ChatModels as this is what you architecture expects.
You could add an early check that the model is supported using openai.Model.list()
Thanks
❓ Alternatives
No response
📝 Additional Context
No response
Acknowledgements
🔖 Feature description
Current code limits the models to "gpt-3.5-turbo", "gpt-4" and "gpt-4-32k". In some cases like summary, it kind of defaults to "gpt-3.5-turbo".
But there are other OpenAI models, right now and in the future too.
Especially to read long web pages, I would like to use "gpt-3.5-turbo-16k". It gets rejected by the LoopGPT module code, although it would work with OpenAI just fine.
thanks
✔️ Solution
Refactor the code so that all models supported by OpenAI can be used, at least ChatModels as this is what you architecture expects.
You could add an early check that the model is supported using
openai.Model.list()Thanks
❓ Alternatives
No response
📝 Additional Context
No response
Acknowledgements