Check Out My Article on Medium: 'Building a Llama from Scratch
This document provides an overview of LlaMA, its different versions, and its capabilities and LlaMA (Large Language Model Meta AI) is a generative AI model.
LlaMA is a group of foundational Large Language Models developed by Meta AI. It was announced by Meta in February of 2023. LlaMA is available in various sizes, ranging from 7 billion to 65 billion parameters, with a context length of 2,000 tokens. These models are designed to assist researchers in advancing their knowledge in the field of AI. The smaller 7 billion parameter models enable researchers with limited computational resources to study and experiment with these models.
- LlaMA models come in multiple sizes(7B, 13B, and 70B), making them suitable for various applications
- Completely open-source and free for non-commercial use.
LlaMA 2 is the second version of LlaMA, surpassing the initial release, LlaMA version 1, which Meta introduced in July of 2023. LlaMA 2 is available in three sizes: 7 billion, 13 billion, and 70 billion parameter models. Upon its release, LlaMA 2 achieved the highest score on Hugging Face, a popular platform for AI model deployment and fine-tuning. Across all segments (7B, 13B, and 70B), LlaMA 2 models have consistently outperformed other models available on Hugging Face, thanks to fine-tuning and retraining efforts.