Run LLMs on Phone with React Native — DeepSeek R1 & Llama 3.2

AI Impact Summary

This guide demonstrates running LLMs locally on a phone using React Native and llama.cpp, leveraging GGUF quantized models from Hugging Face. The focus is on running models like DeepSeek R1 Distil Qwen 2.5 (1.5B parameters) and Llama-3.2-1B-Instruct, highlighting the feasibility of running powerful AI models on mobile devices. The tutorial provides a straightforward approach for developers interested in integrating AI into mobile applications, particularly those prioritizing privacy and offline operation.