如何将keras模型转成rknn

• weiwowuqing

  Posts: 3

  Joined: 2024-03-27 1:40

  2024-03-27 2:04

  我是刚入门的新手，现在遇到了一个问题。我在网上找了一个训练英语手写字体的程序，然后得到了一个模型，现在遇到的问题是如何将这个模型转成可以在板子上运行的rknn模型。
  训练的代码：https://www.kaggle.com/code/m3rcury02/handwriting-recog
  模型后缀： h5格式
  这是模型的参数
  Model: "sequential"
  _________________________________________________________________
  Layer (type) Output Shape Param #
  =================================================================
  conv2d (Conv2D) (None, 26, 26, 32) 320
  
  max_pooling2d (MaxPooling2D (None, 13, 13, 32) 0
  )
  
  conv2d_1 (Conv2D) (None, 13, 13, 64) 18496
  
  max_pooling2d_1 (MaxPooling (None, 6, 6, 64) 0
  2D)
  
  conv2d_2 (Conv2D) (None, 4, 4, 128) 73856
  
  max_pooling2d_2 (MaxPooling (None, 2, 2, 128) 0
  2D)
  
  flatten (Flatten) (None, 512) 0
  
  dense (Dense) (None, 64) 32832
  
  dense_1 (Dense) (None, 128) 8320
  
  dense_2 (Dense) (None, 26) 3354
  
  =================================================================
  Total params: 137,178
  Trainable params: 137,178
  Non-trainable params: 0 请问我该怎么做

  • Quote

import keras
import keras2onnx
import onnx
from keras.models import load_model
model = load_model('xxxx.h5')   #您的 h5 权重文件
onnx_model = keras2onnx.convert_keras(model, model.name)
temp_model_file = 'xxxx.onnx'   #输出
onnx.save_model(onnx_model, temp_model_file)