Tensorflow Machine Learning models

 PLACEHOLDER

• CREATE THE MODEL

model1 = tf.keras.Sequential([

    tf.keras.layers.Flatten(input_shape=(28,28)),

    tf.keras.layers.Dense(256, activation='sigmoid'),

    tf.keras.layers.Dense(10, activation='softmax')

])

Alternatively, load the model;

model1 = tf.keras.models.load_model('my_model.h5')

model1.summary() #optional

• COMPILE THE MODEL

model.compile(...)

opt = tf.keras.optimizers.SGD(learning_rate=0.2)

opt.get_config() #optional

model1=

 model.compile(optimizer=opt,loss='binary_crossentropy',metrics=['accuracy'])

model1.optimizer.get_config() #optional

model1.summary() #optional

See also here

• TRAIN THE MODEL

model.fit(...)

history = model.fit(train_data,validation_data=validation_data,epochs=10)

history.params #optional

history.history.keys() #optional

Train more;

history = model.fit(train_data,validation_data=validation_data,

    initial_epoch=10,epochs=20)

• EVALUATE THE MODEL

model.evaluate(...)

#model1_results = model1.evaluate(test_dataset, return_dict=True)

model.evaluate(test_dataset, return_dict=True)

• PREDICTIONS

model.predict(...)

model.predict(test_data)

• PLACEHOLDER

• PLACEHOLDER

dfgdfg