chatty

This lets us easily use Transformers.js and Chrome's Gemini Nano as well as WebLLM models. @built-in-ai. Use Vercel AI SDK to eliminate usage of custom hooks and state management, which in return cleans up the codebase.

SmolVLM 256M (& 500M): The world’s smallest multimodal model. Designed for efficiency and perfect for on-device applications Demo: https://huggingface.co/spaces/HuggingFaceTB/SmolVLM-256M-Instruct-WebGPU Twitter: https://x.com/xenovacom/status/1882435994160447587 Examples: https://github.com/huggingface/transformers.js-examples/tree/main/smolvlm-webgpu It would be great to extend our models with SmolVLM. There are some UX improvements we could also make at the same time such as, when you use a prompt with an image: Fixing the [Object] text following a selected image. Double-checking there are no specific issues with multi-modal support. I was running into issues with our other multimodal model (couldn't get it working), unsure if this was model specific.

Hi everyone, First off, thanks for this great project! I’d love to use this UI in my own project, but I need to connect it to my own backend instead of the current setup. Rather than maintaining a fork and implementing these changes on my own, I believe it would be beneficial to separate the UI layer from the persistence logic. Specifically, opening up the store for plugins or abstraction layers would allow different backends to be easily integrated. Would you be open to refactoring parts of the project to make this possible? I’d be happy to contribute if there’s interest in this direction. Looking forward to your thoughts!