- User makes training via  GPU’s and brings the data base of the network:

​​

  -  For each Layer:  Number of inputs, weight for each input to each output

      Neron, bios for each Neron, activation function of the layer​

​

  - Number of hidden layers  

​​

- The output of a layer is the input to the next layer ,

   Output Neron's of the last layer  is the output of the networkץ

 

 The tools generate Verilog agnostic code that meet the glue logic limitation with minimum quantization error