dragon-llama2-ov

dragon-llama2-ov is a high-quality, fact-based question-answering model, designed for retrieval augmented generation (RAG) with complex business documents, quantized and packaged in OpenVino int4 for AI PCs using Intel GPU, CPU and NPU.

This model provides a good combination of accuracy and inference performance.

Model Description

• Developed by: llmware
• Model type: llama2
• Parameters: 7 billion
• Quantization: int4
• Model Parent: llmware/dragon-llama-7b-v0
• Language(s) (NLP): English
• License: Llama2 Community License
• Uses: Fact-based question-answering, RAG
• RAG Benchmark Accuracy Score: 97.25

Model Card Contact

llmware on github
llmware on hf
llmware website