BlueLM-V-3B: Algorithm and System Co-Design for Multimodal Large Language Models on Mobile Devices
This work addresses the problem of efficient MLLM deployment on mobile phones for users seeking real-time multimodal AI applications, representing an incremental optimization for mobile-specific constraints.
The paper tackles the challenge of deploying multimodal large language models (MLLMs) on mobile devices with limited memory and computational resources by introducing BlueLM-V-3B, an algorithm and system co-design approach that achieves a generation speed of 24.4 tokens/s and the highest average score of 66.1 on the OpenCompass benchmark among models with ≤4B parameters.
The emergence and growing popularity of multimodal large language models (MLLMs) have significant potential to enhance various aspects of daily life, from improving communication to facilitating learning and problem-solving. Mobile phones, as essential daily companions, represent the most effective and accessible deployment platform for MLLMs, enabling seamless integration into everyday tasks. However, deploying MLLMs on mobile phones presents challenges due to limitations in memory size and computational capability, making it difficult to achieve smooth and real-time processing without extensive optimization. In this paper, we present BlueLM-V-3B, an algorithm and system co-design approach specifically tailored for the efficient deployment of MLLMs on mobile platforms. To be specific, we redesign the dynamic resolution scheme adopted by mainstream MLLMs and implement system optimization for hardware-aware deployment to optimize model inference on mobile phones. BlueLM-V-3B boasts the following key highlights: (1) Small Size: BlueLM-V-3B features a language model with 2.7B parameters and a vision encoder with 400M parameters. (2) Fast Speed: BlueLM-V-3B achieves a generation speed of 24.4 token/s on the MediaTek Dimensity 9300 processor with 4-bit LLM weight quantization. (3) Strong Performance: BlueLM-V-3B has attained the highest average score of 66.1 on the OpenCompass benchmark among models with $\leq$ 4B parameters and surpassed a series of models with much larger parameter sizes (e.g., MiniCPM-V-2.6, InternVL2-8B).