TRINE: A Token-Aware, Runtime-Adaptive FPGA Inference Engine for Multimodal AI
Open Access
- 26 July 2026
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 1-7
Abstract
No abstract availableFunding Information
- Defense Sciences Office, DARPA (Young Faculty Award)
- National Science Foundation (2127780)
- National Science Foundation (2319198)
- National Science Foundation (2321840)
- National Science Foundation (2312517)
- National Science Foundation (2235472)
- National Science Foundation (2431561)
- Semiconductor Research Corporation (3149.001)
- Semiconductor Research Corporation (3305.001)
- Semiconductor Research Corporation (3126.001)
- Office of Naval Research (Young Investigator Program Award)
- Office of Naval Research (N00014-21-1-2225)
- Office of Naval Research (N00014-22-1-2067)
- Army Research Office (911NF2410360)
- Air Force Office of Scientific Research (FA9550-22-1-0253)
This publication has 24 references indexed in Scilit:
- Accelerating ViT Inference on FPGA through Static and Dynamic Pruning2024
- GCV-Turbo: End-to-end Acceleration of GNN-based Computer Vision Tasks on FPGA2024
- SWAT: An Efficient Swin Transformer Accelerator Based on FPGA2024
- HPTA: A High Performance Transformer Accelerator Based on FPGA2023
- Reproducible Scaling Laws for Contrastive Language-Image Learning2023
- Making Vision Transformers Efficient from A Token Sparsification View2023
- ViTCoD: Vision Transformer Acceleration via Dedicated Algorithm and Accelerator Co-Design2023
- SPViT: Enabling Faster Vision Transformers via Latency-Aware Soft Token Pruning2022
- Adaptive Token Sampling for Efficient Vision Transformers2022
- SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruning2021