MNIST模型&数据集导出测试

发布于 2021-04-28 16:50:56 浏览：709 订阅该版

[tocm]

```python
#!/usr/bin/env python
# coding: utf-8

# In[11]:

import tensorflow as tf

# In[12]:

tf.__version__

# In[13]:

tf.keras.__version__

# In[14]:

mnist=tf.keras.datasets.mnist
(x_train,y_train),(x_test,y_test)=mnist.load_data()
tf.print(x_train.shape,y_train.shape,x_test.shape,y_test.shape)

# In[15]:

x_train=tf.keras.utils.normalize(x_train,axis=1)
x_test=tf.keras.utils.normalize(x_test,axis=1)
x_train=tf.expand_dims(x_train,-1)
x_test=tf.expand_dims(x_test,-1)
print(x_train.shape,x_test.shape)

# In[16]:

model=tf.keras.models.Sequential()
model.add(tf.keras.layers.Conv2D(2,(3,3),input_shape=(28,28,1)))
model.add(tf.keras.layers.MaxPool2D())
model.add(tf.keras.layers.Conv2D(2,(3,3),activation='sigmoid'))
model.add(tf.keras.layers.MaxPool2D())
model.add(tf.keras.layers.Flatten())
model.add(tf.keras.layers.Dense(10,activation='sigmoid'))
model.add(tf.keras.layers.Softmax())

model.compile(optimizer='adam',loss='sparse_categorical_crossentropy',metrics=['accuracy'])

# model.fit(x_train,y_train,epochs=5)
# val_loss,val_acc=model.evaluate(x_test,y_test)

#print('loss:',val_loss,'acc:',val_acc)

# In[17]:

print(y_test[:20])
tf.argmax(model.predict(x_test[:20]),axis=1)

# In[18]:

model.summary()
keras_model_file='keras_minst_2.h5'
tf.keras.models.save_model(model,keras_model_file)

# In[9]:

# converter=tf.lite.TFLiteConverter.from_keras_model_file(keras_model)
# tflite_model=converter.convert()
# open("keras_mnist_model.tflite", "wb").write(tflite_model)

# In[19]:

tflite_model = tf.keras.models.load_model('keras_minst_2.h5')
converter = tf.lite.TFLiteConverter.from_keras_model(model)
tflite_save = converter.convert()
open("keras_minst_3.tflite", "wb").write(tflite_save)

# In[51]:

## mnist datasets export to file
str(x_test[1].numpy().flatten().tolist())
with open("mnist_set.h",mode='w') as f:
    for i in range(5):
        f.write('#define MNIST_'+str(i)+'_'+str(y_test[i])+' '+str(x_test[i].numpy().flatten().tolist())+'\n')

# In[ ]:

```