A 28nm 343.5fps/W Vision Transformer Accelerator with Integer-Only Quantized Attention Block.

Cheng-Chen Lin Wei Lu Po-Tsang Huang Hung-Ming Chen

Published in: AICAS (2024)

Keyphrases

real time
dct coefficients
computer vision
high speed
fault diagnosis
parallel implementation
visual field
pre attentive
fuzzy logic
frame rate
transform coefficients
visual attention
discrete cosine transform
floating point