I find that the dataflow description in paper(MAESTRO: An Open-source Infrastructure for Modeling Daraflows within Deep Learning Accelerators) about the Eyeriss is different to the dataflow in this project. Actually, i can hardly relate the *.m in dataflowDescription to whcih dataflow(for examle :Eyeriss, ShiDiannao) it belongs to. And i thought the rs.m or prs.m represents for the 'row stationnay', but the dataflow description is totally different from the Eyeriss dataflow you describe in the paper.
So could you make a file like 'README' to describe the corresponding architecutre that *.m means? Another question is that i tried my best to understand the dataflow description about the Eyeriss, but i just faied, so are there something can make me better understand the dataflow descriptions?
Thank you！