# Fine-tuning with GPT-generated data

MQT-LLaVA is an open-source multimodal chatbot model based on the Transformer architecture. It is trained by fine-tuning LLaMA/Vicuna on GPT-generated multimodal instruction data.