LLaMA 3B

LLaMA 3B (Llama 3.2) is Meta's compact, 3-billion-parameter multilingual LLM: optimized for fast, efficient deployment on edge and mobile devices.

This is LLaMA 3B: specifically the Llama 3.2 instruction-tuned model from Meta, delivering high-performance language processing in a small footprint (3.21 billion parameters). The model is engineered for efficiency, making it ideal for constrained environments like mobile and edge devices. It supports a massive 128,000-token context window and excels at multilingual tasks (e.g., English, Spanish, Hindi), focusing on dialogue, summarization, and complex instruction-following. This is your resource-conscious solution for fast, on-device AI applications.

https://ai.meta.com/blog/llama-3-2-connect-2024-vision-edge-mobile-devices/