@article{trineatokenawareruntimeadaptivefpgainfer, title = {TRINE: A Token-Aware, Runtime-Adaptive FPGA Inference Engine for Multimodal AI}, author = {Hyunwoo Oh and Hanning Chen and Sanggeon Yun and Yang Ni and Suyeon Jang and Behnam Khaleghi and Fei Wen and Mohsen Imani}, year = {2026}, eprint = {2603.22867}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2603.22867}, }