pb转lite 找不到层

由于定义网络时给每个层名字，所再用下面代码转换时会报错

import tensorflow as tf
 
in_path = r'.\yolov3_coco.pb'
out_path = r'.\yolov3_coco.tflite'
 
input_arrays = ["input/input_data"]
input_shapes = {"input/input_data" :[1, 416, 416, 3]}
output_arrays = ["pred_sbbox/concat_2", "pred_mbbox/concat_2", "pred_lbbox/concat_2"]
 
converter = tf.lite.TFLiteConverter.from_frozen_graph(in_path, input_arrays, output_arrays, input_shapes)
tflite_model = converter.convert()
open("myModel.tflite", "wb").write(tflite_model)

这时候需要用tensorboard查看一下网络的结构，然后就可以看到了网络层名字了