amdenterpriseai/aim-radeon-meta-llama-llama-3-1-8b-instruct repository overview

⁠Llama 3.1 8B Instruct

Llama 3.1 8B Instruct is part of the Meta Llama 3.1 collection of multilingual large language models, available in 8B, 70B, and 405B sizes. It is an 8 billion parameter auto-regressive language model using an optimized transformer architecture with Grouped-Query Attention (GQA). The instruction-tuned version uses supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. The model supports multilingual dialogue across 8 languages (English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai). Pre-trained on over 15 trillion tokens of publicly available data with a knowledge cutoff of December 2023. Intended for assistant-like chat, with support for tool use and multilingual capabilities.