Model Overview

This model, Llama-2-7b-chat-norwegian, is a 7 billion parameter variant of Meta's Llama 2 7b Chat model, specifically fine-tuned by Ruter AI Lab. Its primary purpose is to understand and generate text in Norwegian, making it a valuable resource for the Norwegian linguistic landscape.

Key Capabilities

• Norwegian Language Proficiency: Optimized for Norwegian text generation and comprehension.
• Chat and Assistant Functions: Designed for assistant-like chat interactions.
• Instruction Following: Supports both Llama2 Chat and Alpaca prompt formats.
• Open-Source Contribution: Aims to democratize information and foster innovation within the Norwegian NLP community.

Training Details

The model was trained for one epoch on a diverse set of Norwegian datasets, including:

• Norwegian Alpaca
• 15,000 samples of machine-translated data from OpenOrca
• A small subset of custom-made instructional data

Intended Use & Limitations

This model is intended for commercial and research use in Norwegian. It excels at tasks such as summarization, question answering, and chat. However, it's important to note that it is an LLM, not a knowledge model, and its knowledge about Norway is limited to its base model. The training data, being machine-translated, may contain grammatical errors. Optimal results may require prompt tuning.