Add new onnx model to test jetson benchmark(int8)

[653576489]

I converted the onnx model of yolov3-pytorch（1x3x416x416） myself, added it to the jetson_benchmark project, and measured FPS=74 (batch_size=1), but I added batch_size, FPS will increase linearly, (batch_size=8, FPS=614.30) which is obviously wrong, and I would like to ask you how you specifically got onnx, where I might have been wrong，thanks。