Ad-hoc hardware implementations of neural networks provide security and optimal performance, power and area solutions. Our machine learning design flow guarantee quick turnaround times in synthesizable hardware implementations. As a result, you will get soft and hard IPs data bases of your neural network accelerator.