Neural network FPGA

The invention relates to a neural network FPGA, belongs to the technical field of integrated circuits, and has the advantage of high operation speed. The circuit comprises N neuron circuits, with N beingan integer greater than 1, and each neuron circuit comprises the following parts: N2 first multiplication units; a first addition unit; the second multiplication unit comprises two input ends and an output end; a second addition unit; n neuron multiplication units, wherein each neuron multiplication unit comprises two input ends and an output end; the neuron multiplication units are in one-to-one correspondence with the neuron circuits; a third addition unit; the fourth addition unit comprises a plurality of input ends and an output end, and the input ends of the fourth addition unit are connected with the output end of the second addition unit and the output end of the third addition unit in a one-to-one correspondence mode; and the output end of the function unit is connected with oneinput end of the neuron