Hi all, I open sourced my toy project that runs Generative AI models LOCALLY in the side panel of a Chrome extension.
The Chrome extension uses Transformers.js to run models in browser under the hood. I've integrated and tested these models so far.
\1. LLM: Llama 3, Phi 3.5, Qwen 2.5, SmolLM2
\2. Reasoning: DeepSeek R1
\3. Multimodal LLM: Janus
\4. Speech-to-Text: Whisper
On an M1 MacBook, DeepSeek R1 1.5B runs at ~30 tokens/sec
Thanks for your comment! That's a great idea. This project uses WebGPU to run AI models, and as WebGPU becomes more widely available across platforms, I believe it can be extended to Firefox, Safari, and desktop environments. I will make a new post when it supports other platforms.