# Fine-tuning Open-source Large Models

FireLLaVA-13B is a vision-language model trained on instruction data generated by open-source large language models, supporting image understanding and text generation tasks.