www.analyticsdrift.com
Image source: Analytics Drift
Meta released the first models of Llama 3 with 8B and 70B pre-trained and instruction-tuned parameters. It is the successor to Llama 2, which was released with 7, 13, and 70B parameters last year.
Image source: AD
Llama belongs to a family of LLMs introduced by Meta AI in February 2023. The first version of the model was released in four sizes — 7B, 13B, 33B, and 65 billion parameters.
Image: Canva
With higher scalability, Llama 3 simplifies coding and problem-solving for software developers by offering better reasoning, code generation, and instruction-following capabilities.
Image: Canva
Llama 3 is trained on Meta’s two custom-built 24K GPU clusters on over 15T tokens of data. This training dataset is 7x larger than that used for Llama 2.
Image source: Meta
Meta has said it is committed to developing Llama 3 responsibly by including new trust and safety tools with Llama Guard 2, Code Shield, and CyberSec Eval 2.
Image source: Canva
Llama 3 can be used on Facebook, Instagram, WhatsApp, and the web. Developers can access it via Hugging Face ecosystem and cloud provider platforms such as Azure ML and Vertex AI.
Image: Canva
Produced by: Shuchi Chitrakar Designed by: Prathamesh