Github repository documenting the ANN Processor's custom instruction set architecture, along with example programs for feed-forward and recurrent neural networks.
Further detail can be found in the manuscript:
D. Valencia, S. F. Fard and A. Alimohammad, "An Artificial Neural Network Processor With a Custom Instruction Set Architecture for Embedded Applications," in IEEE Transactions on Circuits and Systems I: Regular Papers, doi: 10.1109/TCSI.2020.3003769.
03.27.2020 - Github repository created, documentation will be added soon.
04.06.2020 - Example assembly programs for executing MNIST Feed-Forward ANN and RNN uploaded.
04.06.2020 - Uploaded documentation for custom-developed instruction set architecture.
05.08.2020 - Uploaded example program for executing sentiment analysis RNN on the IMDB dataset. Also included new directory IMDB_RNN_Params that include the testing data, and trained parameters, in CSV files.