There is currently no checking to see that an input neural network configuration format meets the built elements per block parameter of DANA. DANA will likely error out complaining that a PE saw an assignment with the same input and output address, however, this doesn't address the underlying problem.
Add in support either at the Cache or AsidNnidTableWalker level.
There is currently no checking to see that an input neural network configuration format meets the built elements per block parameter of DANA. DANA will likely error out complaining that a PE saw an assignment with the same input and output address, however, this doesn't address the underlying problem.
Add in support either at the Cache or AsidNnidTableWalker level.