FR: GPT-4

[JCharante]

Hi there, great plugin so far. I wish there was a dropdown box to select between GPT 3.5 (which this uses, at least according to my platform billing page) and GPT 4. I wouldn't mind paying extra for the extra intelligence in the responses.

[clairefro]

Thanks for the suggestion. I'm curious - do you find the 3.5 responses lacking, or just curious to see the difference?

I chose 3.5 because it seemed to do a great job at a higher speed and didn't see the need to offer a more expensive option. But I'm open to making model selection for both OpenAI and Ollama a feature.

[JCharante]

I know 4 is slower, but I just tend to use gpt4 for everything and generally can feel a difference when using gpt3.5. I'm not sure if I find it lacking, I have to use the plugin more

[JCharante]

*typo fix, meant to say 4 is slower

[kityr]

But I'm open to making model selection for both OpenAI and Ollama a feature.

Yes, that would be great! Mistral, which ChatCBT uses, doesn't fit into my VRAM, while llama2 and some other 3.8gb models do; would be great to be able to choose a smaller model to take advantage of GPU inference

[IbragimovDP]

@JCharante @kityr @clairefro I added this feature and opened pull request #6

[JCharante]

Awesome, thank you!

[clairefro]

Model override now available thanks to #6. Publishing as new release soon

[clairefro]

Model override now available in release version 1.0.3

Simply update your Obsidian plugins to access