Saved in:
| Main Authors: | , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2503.06019 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913725523427328 |
|---|---|
| author | Lu, Xudong Chen, Yinghao Wu, Renshou Gao, Haohao Chen, Xi Yang, Xue Zhao, Xiangyu Zhou, Aojun Li, Fangyuan Wen, Yafei Chen, Xiaoxin Ren, Shuai Li, Hongsheng |
| author_facet | Lu, Xudong Chen, Yinghao Wu, Renshou Gao, Haohao Chen, Xi Yang, Xue Zhao, Xiangyu Zhou, Aojun Li, Fangyuan Wen, Yafei Chen, Xiaoxin Ren, Shuai Li, Hongsheng |
| contents | Recent advancements in Multimodal Large Language Models (MLLMs) have enabled their deployment on mobile devices. However, challenges persist in maintaining strong language capabilities and ensuring hardware compatibility, both of which are crucial for user experience and practical deployment efficiency. In our deployment process, we observe that existing MLLMs often face performance degradation on pure language tasks, and the current NPU platforms on smartphones do not support the MoE architecture, which is commonly used to preserve pure language capabilities during multimodal training. To address these issues, we systematically analyze methods to maintain pure language capabilities during the training of MLLMs, focusing on both training data and model architecture aspects. Based on these analyses, we propose GenieBlue, an efficient MLLM structural design that integrates both linguistic and multimodal capabilities for LLMs on mobile devices. GenieBlue freezes the original LLM parameters during MLLM training to maintain pure language capabilities. It acquires multimodal capabilities by duplicating specific transformer blocks for full fine-tuning and integrating lightweight LoRA modules. This approach preserves language capabilities while achieving comparable multimodal performance through extensive training. Deployed on smartphone NPUs, GenieBlue demonstrates efficiency and practicality for applications on mobile devices. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_06019 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | GenieBlue: Integrating both Linguistic and Multimodal Capabilities for Large Language Models on Mobile Devices Lu, Xudong Chen, Yinghao Wu, Renshou Gao, Haohao Chen, Xi Yang, Xue Zhao, Xiangyu Zhou, Aojun Li, Fangyuan Wen, Yafei Chen, Xiaoxin Ren, Shuai Li, Hongsheng Computation and Language Computer Vision and Pattern Recognition Recent advancements in Multimodal Large Language Models (MLLMs) have enabled their deployment on mobile devices. However, challenges persist in maintaining strong language capabilities and ensuring hardware compatibility, both of which are crucial for user experience and practical deployment efficiency. In our deployment process, we observe that existing MLLMs often face performance degradation on pure language tasks, and the current NPU platforms on smartphones do not support the MoE architecture, which is commonly used to preserve pure language capabilities during multimodal training. To address these issues, we systematically analyze methods to maintain pure language capabilities during the training of MLLMs, focusing on both training data and model architecture aspects. Based on these analyses, we propose GenieBlue, an efficient MLLM structural design that integrates both linguistic and multimodal capabilities for LLMs on mobile devices. GenieBlue freezes the original LLM parameters during MLLM training to maintain pure language capabilities. It acquires multimodal capabilities by duplicating specific transformer blocks for full fine-tuning and integrating lightweight LoRA modules. This approach preserves language capabilities while achieving comparable multimodal performance through extensive training. Deployed on smartphone NPUs, GenieBlue demonstrates efficiency and practicality for applications on mobile devices. |
| title | GenieBlue: Integrating both Linguistic and Multimodal Capabilities for Large Language Models on Mobile Devices |
| topic | Computation and Language Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.06019 |