Search papers, labs, and topics across Lattice.
University of Glasgow
3
0
5
FlexViT achieves up to 2.74x speedup for Vision Transformer inference on edge devices, revolutionizing the deployment of complex models in resource-constrained environments.
LLM-guided exploration can drastically cut down the time and expertise needed for efficient FPGA accelerator design, generating viable architectures that perform well on diverse AI tasks.
Eliminating redundant data copying can double memory efficiency and dramatically boost accelerator performance in machine learning workloads.
