模型的保存与加载一般有三种模式:save/load weights(最干净、最轻量级的方式,只保存网络参数,不保存网络状态),save/load entire model(最简单粗暴的方式,把网络所有的状态都保存起来),saved_model(更通用的方式,以固定模型格式保存,该格式是各种语言通用的)
具体使用方法如下:
# 保存模型 model.save_weights('./checkpoints/my_checkpoint') # 加载模型 model = keras.create_model() model.load_weights('./checkpoints/my_checkpoint')
示例:
import tensorflow as tf from tensorflow.keras import datasets, layers, optimizers, Sequential, metrics def preprocess(x, y): x = tf.cast(x, dtype=tf.float32) / 255. x = tf.reshape(x, [28 * 28]) y = tf.cast(y, dtype=tf.int32) y = tf.one_hot(y, depth=10) return x, y batchsz = 128 (x, y), (x_val, y_val) = datasets.mnist.load_data() print('datasets:', x.shape, y.shape, x.min(), x.max()) db = tf.data.Dataset.from_tensor_slices((x, y)) db = db.map(preprocess).shuffle(60000).batch(batchsz) ds_val = tf.data.Dataset.from_tensor_slices((x_val, y_val)) ds_val = ds_val.map(preprocess).batch(batchsz) sample = next(iter(db)) print(sample[0].shape, sample[1].shape) network = Sequential([layers.Dense(256, activation='relu'), layers.Dense(128, activation='relu'), layers.Dense(64, activation='relu'), layers.Dense(32, activation='relu'), layers.Dense(10)]) network.build(input_shape=(None, 28 * 28)) network.summary() network.compile(optimizer=optimizers.Adam(lr=0.01), loss=tf.losses.CategoricalCrossentropy(from_logits=True), metrics=['accuracy'] ) network.fit(db, epochs=3, validation_data=ds_val, validation_freq=2) network.evaluate(ds_val) network.save_weights('weights.ckpt') print('saved weights.') del network network = Sequential([layers.Dense(256, activation='relu'), layers.Dense(128, activation='relu'), layers.Dense(64, activation='relu'), layers.Dense(32, activation='relu'), layers.Dense(10)]) network.compile(optimizer=optimizers.Adam(lr=0.01), loss=tf.losses.CategoricalCrossentropy(from_logits=True), metrics=['accuracy'] ) network.load_weights('weights.ckpt') print('loaded weights!') network.evaluate(ds_val)
运行效果如下:
可以看到保存前后的精度和损失差距不大,这是由于神经网络的运算过程中会有很多不确定因子,这些不确定因子不会通过save_weights方法保存,要想保存前后运行结果一致,就需要完整的保存网络模型。即model.save方法
使用方法如下:
# 模型保存 network.save('model.h5') print('saved total model.') # 模型加载 print('load model from file') network = tf.keras.models.load_model('model.h5') # 评估 network.evaluate(x_val,y_val)
除了这种方法之外,tensorflow还支持保存为标准的可以给其他语言使用的模型,使用saved_model即可
使用方法如下:
tf.saved_model.save(m,'/tmp/saved_model/') imported = tf.saved_model.load(path) f = imported.signatures["serving_default"] print(f(x=tf.ones([1,28,28,3])))