Vintern-1B: An Efficient Multimodal Large Language Model for Vietnamese
This addresses the need for efficient on-device multimodal AI applications in Vietnamese, though it is incremental as it combines existing models for a specific language context.
The researchers tackled the problem of developing an efficient multimodal large language model for Vietnamese by introducing Vintern-1B, a 1-billion-parameter model that integrates Qwen2-0.5B-Instruct and InternViT-300M-448px, achieving robust performance on benchmarks like OpenViVQA and ViTextVQA after fine-tuning on over 3 million image-question-answer pairs.
In this report, we introduce Vintern-1B, a reliable 1-billion-parameters multimodal large language model (MLLM) for Vietnamese language tasks. By integrating the Qwen2-0.5B-Instruct language model with the InternViT-300M-448px visual model, Vintern-1B is optimized for a range of applications, including optical character recognition (OCR), document extraction, and general question-answering in Vietnamese context. The model is fine-tuned on an extensive dataset of over 3 million image-question-answer pairs, achieving robust performance and reliable results across multiple Vietnamese language benchmarks like OpenViVQA and ViTextVQA. Vintern-1B is small enough to fit into various on-device applications easily. Additionally, we have open-sourced several Vietnamese vision question answering (VQA) datasets for text and diagrams, created with Gemini 1.5 Flash. Our models are available at: https://huggingface.co/5CD-AI/Vintern-1B-v2.