A Scalable and Adaptive Convolutional Neural Network Accelerator, 2022 32nd International Conference Radioelektronika (RADIOELEKTRONIKA)
https://ieeexplore.ieee.org/document/9764951
Iteratively run keras/full.py to train, load and run inference on the synthesized hardware.
Uses simulation model for External Memory.
Contact me if you would like to add DDR support.