Overview
Llama 2 was a major milestone for the open AI community. By releasing high-quality model weights for free (for most uses), Meta enabled researchers and developers to build their own custom AI applications.
Variants
Available in different sizes: 7B, 13B, and 70B parameters. It also included 'Llama-2-Chat' versions optimized for dialogue.
Impact
It sparked a wave of innovation in model compression, fine-tuning (like LoRA), and local AI execution.