I really love these new small language AI models; they're fast, lightweight, and still very capable. I might deploy an instance of it for Nostr DVM and also have it available across all my Nostr apps, including NostrNet. https://www.microsoft.com/en-us/research/blog/phi-2-the-surprising-power-of-small-language-models/
Also take llava for a spin, if you haven’t. Very impressive on recent, consumer grade laptops.
You can also run it locally. I can add a one-click local deployment function – just click one button, and it will start running locally. Alternatively, you should be just able to use it with my existing backend. You will have all the options.