在实际应用中,我们常常使用利用keras的便捷性进行训练,而生产环境往往是c++语言的环境,于是,我们使用keras训练模型,再用tensorflow载入模型,做正向计算,得到结果。
具体的流程为:keras训练模型 --> model.h5 --> 转换成.pb文件 --> tensorflow 载入.pb 验证正确性 --> tensorflow c++ api调用 .pb文件
首先,配置环境
需要安装的工具有:
Protobuf 此google三方工具的版本和tensorflow密切相关,版本不兼容时,tensorflow无法正常运行,另外,protobuf的最大变量允许值为2G(INT_MAX),进过本人测试,其实最多为大于1G就会报错,所以.pb的模型文件大于1G时,无法载入。具体编译安装步骤不再赘述
Eigen,这是一个C++端的矩阵运算库,只要下载解压即可
bazel,需要用次工具对tensorflow进行编译
Tensorflow ,下载并编译其源码
下面是一个演示demo
demo中使用到opencv库,于是需要先编译安装opencv库,编译过程中或许会遇到一些问题,网上都有教程,比较难解决的是关于本机环境中zlib和libpng库和anaconda中的冲突,只要卸载掉anaconda中的库就行正常运行
Keras 训练分类模型并存储为5_trained_model.h5
def load_cifar10(num_training = 49000, num_validation=1000, num_test=10000):
cifar10 = tf.keras.datasets.cifar10.load_data()
(X_train, y_train), (X_test, y_test) = cifar10
X_train = np.asarray(X_train, dtype=np.float32)
y_train = np.asarray(y_train, dtype=np.int32).flatten()
X_test = np.asarray(X_test, dtype=np.float32)
y_test = np.asarray(y_test, dtype=np.int32).flatten()
mask = range(num_training, num_training + num_validation)
X_val = X_train[mask]
y_val = y_train[mask]
mask = range(num_training)
X_train = X_train[mask]
y_train = y_train[mask]
mask = range(num_test)
X_test = X_test[mask]
y_test = y_test[mask]
# Normalize the data
mean_pixel = X_train.mean(axis=(0,1,2), keepdims=True)
std_pixel = X_train.std(axis=(0,1,2), keepdims=True)
X_train = (X_train - mean_pixel) / std_pixel
X_val = (X_val - mean_pixel) / std_pixel
X_test = (X_test - mean_pixel) / std_pixel
return X_train, y_train, X_val, y_val, X_test, y_test
def model(input_shape):
X_input = Input(shape=input_shape)
X = Conv2D(64, kernel_size=3,strides=1, padding="same", use_bias=True)(X_input)
X = BatchNormalization()(X)
X = Activation('relu')(X)
MaxPooling2D(pool_size=2, strides = 2, padding='valid')
X = Conv2D(64, kernel_size=3,strides=1, padding="same", use_bias=True)(X)
X = Flatten()(X)
X = BatchNormalization(