Llama-2-13b-hf Overview

This model is the 13 billion parameter pretrained version of Meta's Llama 2 family, adapted for the Hugging Face Transformers format. Llama 2 models are a collection of pretrained and fine-tuned generative text models, with this specific variant being a base model intended for broad natural language generation tasks.

Key Capabilities & Features

• Architecture: Auto-regressive language model utilizing an optimized transformer architecture.
• Training Data: Pretrained on 2 trillion tokens from a new mix of publicly available online data.
• Context Length: Supports a context length of 4096 tokens.
• Performance: Demonstrates strong performance across academic benchmarks, including 24.5 on Code, 66.9 on Commonsense Reasoning, and 54.8 on MMLU for the 13B variant.
• Commercial Use: Licensed for both commercial and research applications in English.

Intended Use Cases

• Natural Language Generation: Adaptable for a wide variety of text generation tasks.
• Research: Suitable for research purposes in large language models.
• English Language Applications: Primarily intended for use with English language inputs and outputs.