Mixed-TD: Efficient Neural Network Accelerator with Layer-Specific Tensor Decomposition

Open in new window