bigcindy

TensorFlow2中使用Keras Tuner搜索网络的超参数

1. 导入所需的库

import tensorflow as tf
import kerastuner as kt
import IPython

for i in [tf, kt]:
    print(i.__name__,": ",i.__version__,sep="")

输出：

tensorflow: 2.2.0
kerastuner: 1.0.1

2. 导入数据集

本案例中使用Fashion MNIST构建神经网络，并用Keras Tuner寻找最优的超参数。

(trainImages, trainLabels),(testImages, testLabels) = tf.keras.datasets.fashion_mnist.load_data()

for i in [trainImages, trainLabels, testImages, testLabels]:
    print(i.shape)

输出：

(60000, 28, 28)
(60000,)
(10000, 28, 28)
(10000,)

# 将像素值归一化到0至1之间的数
trainImages = trainImages.astype("float32")/255.0
testImages = testImages.astype("float32")/255.0

3. 定义模型结构

寻找模型超参数时，需要在模型定义的时候定义超参数搜索空间，这样定义的模型叫超模型。定义超模型有两种方式：

1. 使用模型构建函数（本案例采用这种方法）
2. 通过构建Keras Tuner HyperModel子类完成

def modelBuilder(hp):
    model = tf.keras.Sequential()
    model.add(tf.keras.layers.Flatten(input_shape=(28,28)))
    
    # 从32至512的范围内搜索第一层神经元个数
    hp_units = hp.Int("units",min_value=32, max_value=512, step=32)
    model.add(tf.keras.layers.Dense(units=hp_units,activation="relu"))
    model.add(tf.keras.layers.Dense(10))
    
    # 从0.01，0.001和0.0001中搜索最佳的学习率
    hp_learning_rate = hp.Choice("learning_rate",values=[1e-2,1e-3,1e-4])
    
    model.compile(optimizer=tf.keras.optimizers.Adam(learning_rate=hp_learning_rate),
                 loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),
                 metrics=["accuracy"])
    return model

4. 实例化模型并寻找超参数

Keras Tuner有四种tuner可用：RandomSearch、Hyperband、BayesianOptimization和Sklearn。本案例中使用Hyperband。

实例化Hyperband，需要指定超模型、优化的目标（函数）、最大迭代次数等。

tuner = kt.Hyperband(modelBuilder,
                    objective="val_accuracy",
                    max_epochs=10,
                    factor=3,
                    directory="my_dir",
                    project_name="intro_to_kt",
                    overwrite = True) # overwrite参数删除之前保存的checkpoint和log，从头开始搜索

class ClearTrainingOutput(tf.keras.callbacks.Callback):
    def on_train_end(*args, **kwargs):
        IPython.display.clear_output(wait = True)

tuner.search(trainImages, trainLabels,epochs=10,validation_data=(testImages,testLabels),callbacks=[ClearTrainingOutput()])

best_hps = tuner.get_best_hyperparameters(num_trials=1)[0]
print(f"""
The hyperparameter search is complete. The optimal number of units in the first densely-connected
layer is {best_hps.get('units')} and the optimal learning rate for the optimizer
is {best_hps.get('learning_rate')}.
""")

输出：

Trial complete
Trial summary
|-Trial ID: b151c62677c0d827349bf0d0549e6150
|-Score: 0.8691999912261963
|-Best step: 0
Hyperparameters:
|-learning_rate: 0.0001
|-tuner/bracket: 0
|-tuner/epochs: 10
|-tuner/initial_epoch: 0
|-tuner/round: 0
|-units: 192
INFO:tensorflow:Oracle triggered exit

The hyperparameter search is complete. The optimal number of units in the first densely-connected
layer is 416 and the optimal learning rate for the optimizer
is 0.001.

5. 用寻找到的参数进行模型训练

model = tuner.hypermodel.build(best_hps)
model.fit(trainImages,trainLabels, epochs=10, validation_data=(testImages, testLabels))

输出：

Epoch 1/10
1875/1875 [==============================] - ETA: 0s - loss: 2.4319 - accuracy: 0.06 - ETA: 4s - loss: 1.3113 - accuracy: 0.55 - ETA: 4s - loss: 1.0485 - accuracy: 0.64 - ETA: 3s - loss: 0.9385 - accuracy: 0.68 - ETA: 3s - loss: 0.8818 - accuracy: 0.69 - ETA: 3s - loss: 0.8283 - accuracy: 0.70 - ETA: 3s - loss: 0.7952 - accuracy: 0.71 - ETA: 3s - loss: 0.7624 - accuracy: 0.73 - ETA: 3s - loss: 0.7399 - accuracy: 0.73 - ETA: 3s - loss: 0.7199 - accuracy: 0.74 - ETA: 3s - loss: 0.6984 - accuracy: 0.75 - ETA: 3s - loss: 0.6809 - accuracy: 0.75 - ETA: 3s - loss: 0.6715 - accuracy: 0.76 - ETA: 3s - loss: 0.6572 - accuracy: 0.76 - ETA: 3s - loss: 0.6465 - accuracy: 0.77 - ETA: 3s - loss: 0.6364 - accuracy: 0.77 - ETA: 3s - loss: 0.6357 - accuracy: 0.77 - ETA: 3s - loss: 0.6292 - accuracy: 0.77 - ETA: 3s - loss: 0.6223 - accuracy: 0.78 - ETA: 3s - loss: 0.6182 - accuracy: 0.78 - ETA: 3s - loss: 0.6147 - accuracy: 0.78 - ETA: 3s - loss: 0.6078 - accuracy: 0.78 - ETA: 3s - loss: 0.6024 - accuracy: 0.78 - ETA: 2s - loss: 0.5997 - accuracy: 0.78 - ETA: 2s - loss: 0.5958 - accuracy: 0.79 - ETA: 2s - loss: 0.5915 - accuracy: 0.79 - ETA: 2s - loss: 0.5890 - accuracy: 0.79 - ETA: 2s - loss: 0.5819 - accuracy: 0.79 - ETA: 2s - loss: 0.5780 - accuracy: 0.79 - ETA: 2s - loss: 0.5738 - accuracy: 0.79 - ETA: 2s - loss: 0.5699 - accuracy: 0.79 - ETA: 2s - loss: 0.5678 - accuracy: 0.79 - ETA: 2s - loss: 0.5634 - accuracy: 0.80 - ETA: 2s - loss: 0.5597 - accuracy: 0.80 - ETA: 2s - loss: 0.5570 - accuracy: 0.80 - ETA: 2s - loss: 0.5527 - accuracy: 0.80 - ETA: 2s - loss: 0.5504 - accuracy: 0.80 - ETA: 2s - loss: 0.5490 - accuracy: 0.80 - ETA: 2s - loss: 0.5459 - accuracy: 0.80 - ETA: 2s - loss: 0.5440 - accuracy: 0.80 - ETA: 2s - loss: 0.5426 - accuracy: 0.80 - ETA: 1s - loss: 0.5391 - accuracy: 0.80 - ETA: 1s - loss: 0.5383 - accuracy: 0.80 - ETA: 1s - loss: 0.5354 - accuracy: 0.81 - ETA: 1s - loss: 0.5330 - accuracy: 0.81 - ETA: 1s - loss: 0.5305 - accuracy: 0.81 - ETA: 1s - loss: 0.5271 - accuracy: 0.81 - ETA: 1s - loss: 0.5253 - accuracy: 0.81 - ETA: 1s - loss: 0.5231 - accuracy: 0.81 - ETA: 1s - loss: 0.5217 - accuracy: 0.81 - ETA: 1s - loss: 0.5201 - accuracy: 0.81 - ETA: 1s - loss: 0.5178 - accuracy: 0.81 - ETA: 1s - loss: 0.5158 - accuracy: 0.81 - ETA: 1s - loss: 0.5142 - accuracy: 0.81 - ETA: 1s - loss: 0.5123 - accuracy: 0.81 - ETA: 1s - loss: 0.5103 - accuracy: 0.81 - ETA: 1s - loss: 0.5072 - accuracy: 0.81 - ETA: 1s - loss: 0.5060 - accuracy: 0.82 - ETA: 1s - loss: 0.5053 - accuracy: 0.82 - ETA: 1s - loss: 0.5042 - accuracy: 0.82 - ETA: 0s - loss: 0.5032 - accuracy: 0.82 - ETA: 0s - loss: 0.5026 - accuracy: 0.82 - ETA: 0s - loss: 0.5015 - accuracy: 0.82 - ETA: 0s - loss: 0.5003 - accuracy: 0.82 - ETA: 0s - loss: 0.4981 - accuracy: 0.82 - ETA: 0s - loss: 0.4968 - accuracy: 0.82 - ETA: 0s - loss: 0.4948 - accuracy: 0.82 - ETA: 0s - loss: 0.4937 - accuracy: 0.82 - ETA: 0s - loss: 0.4918 - accuracy: 0.82 - ETA: 0s - loss: 0.4894 - accuracy: 0.82 - ETA: 0s - loss: 0.4881 - accuracy: 0.82 - ETA: 0s - loss: 0.4871 - accuracy: 0.82 - ETA: 0s - loss: 0.4851 - accuracy: 0.82 - ETA: 0s - loss: 0.4838 - accuracy: 0.82 - ETA: 0s - loss: 0.4830 - accuracy: 0.82 - ETA: 0s - loss: 0.4822 - accuracy: 0.82 - ETA: 0s - loss: 0.4807 - accuracy: 0.82 - ETA: 0s - loss: 0.4799 - accuracy: 0.82 - ETA: 0s - loss: 0.4790 - accuracy: 0.82 - ETA: 0s - loss: 0.4780 - accuracy: 0.82 - 4s 2ms/step - loss: 0.4779 - accuracy: 0.8289 - val_loss: 0.4244 - val_accuracy: 0.8478
Epoch 2/10
1875/1875 [==============================] - ETA: 0s - loss: 0.3019 - accuracy: 0.87 - ETA: 4s - loss: 0.3950 - accuracy: 0.86 - ETA: 3s - loss: 0.3898 - accuracy: 0.85 - ETA: 3s - loss: 0.3688 - accuracy: 0.86 - ETA: 3s - loss: 0.3645 - accuracy: 0.86 - ETA: 3s - loss: 0.3654 - accuracy: 0.86 - ETA: 3s - loss: 0.3730 - accuracy: 0.86 - ETA: 3s - loss: 0.3690 - accuracy: 0.86 - ETA: 3s - loss: 0.3702 - accuracy: 0.86 - ETA: 3s - loss: 0.3713 - accuracy: 0.86 - ETA: 3s - loss: 0.3668 - accuracy: 0.86 - ETA: 3s - loss: 0.3662 - accuracy: 0.87 - ETA: 3s - loss: 0.3652 - accuracy: 0.87 - ETA: 3s - loss: 0.3673 - accuracy: 0.86 - ETA: 3s - loss: 0.3666 - accuracy: 0.86 - ETA: 3s - loss: 0.3692 - accuracy: 0.86 - ETA: 3s - loss: 0.3697 - accuracy: 0.86 - ETA: 3s - loss: 0.3710 - accuracy: 0.86 - ETA: 3s - loss: 0.3695 - accuracy: 0.86 - ETA: 3s - loss: 0.3698 - accuracy: 0.86 - ETA: 3s - loss: 0.3730 - accuracy: 0.86 - ETA: 3s - loss: 0.3712 - accuracy: 0.86 - ETA: 3s - loss: 0.3707 - accuracy: 0.86 - ETA: 3s - loss: 0.3709 - accuracy: 0.86 - ETA: 3s - loss: 0.3722 - accuracy: 0.86 - ETA: 2s - loss: 0.3708 - accuracy: 0.86 - ETA: 2s - loss: 0.3713 - accuracy: 0.86 - ETA: 2s - loss: 0.3725 - accuracy: 0.86 - ETA: 2s - loss: 0.3704 - accuracy: 0.86 - ETA: 2s - loss: 0.3704 - accuracy: 0.86 - ETA: 2s - loss: 0.3690 - accuracy: 0.86 - ETA: 2s - loss: 0.3689 - accuracy: 0.86 - ETA: 2s - loss: 0.3695 - accuracy: 0.86 - ETA: 2s - loss: 0.3687 - accuracy: 0.86 - ETA: 2s - loss: 0.3686 - accuracy: 0.86 - ETA: 2s - loss: 0.3684 - accuracy: 0.86 - ETA: 2s - loss: 0.3673 - accuracy: 0.86 - ETA: 2s - loss: 0.3668 - accuracy: 0.86 - ETA: 2s - loss: 0.3679 - accuracy: 0.86 - ETA: 2s - loss: 0.3672 - accuracy: 0.86 - ETA: 2s - loss: 0.3671 - accuracy: 0.86 - ETA: 2s - loss: 0.3671 - accuracy: 0.86 - ETA: 1s - loss: 0.3665 - accuracy: 0.86 - ETA: 1s - loss: 0.3665 - accuracy: 0.86 - ETA: 1s - loss: 0.3672 - accuracy: 0.86 - ETA: 1s - loss: 0.3653 - accuracy: 0.86 - ETA: 1s - loss: 0.3646 - accuracy: 0.86 - ETA: 1s - loss: 0.3646 - accuracy: 0.86 - ETA: 1s - loss: 0.3651 - accuracy: 0.86 - ETA: 1s - loss: 0.3650 - accuracy: 0.86 - ETA: 1s - loss: 0.3648 - accuracy: 0.86 - ETA: 1s - loss: 0.3652 - accuracy: 0.86 - ETA: 1s - loss: 0.3658 - accuracy: 0.86 - ETA: 1s - loss: 0.3650 - accuracy: 0.86 - ETA: 1s - loss: 0.3640 - accuracy: 0.86 - ETA: 1s - loss: 0.3627 - accuracy: 0.86 - ETA: 1s - loss: 0.3630 - accuracy: 0.86 - ETA: 1s - loss: 0.3624 - accuracy: 0.86 - ETA: 1s - loss: 0.3623 - accuracy: 0.86 - ETA: 1s - loss: 0.3625 - accuracy: 0.86 - ETA: 1s - loss: 0.3623 - accuracy: 0.86 - ETA: 1s - loss: 0.3620 - accuracy: 0.86 - ETA: 0s - loss: 0.3617 - accuracy: 0.86 - ETA: 0s - loss: 0.3611 - accuracy: 0.86 - ETA: 0s - loss: 0.3606 - accuracy: 0.86 - ETA: 0s - loss: 0.3601 - accuracy: 0.86 - ETA: 0s - loss: 0.3597 - accuracy: 0.86 - ETA: 0s - loss: 0.3602 - accuracy: 0.86 - ETA: 0s - loss: 0.3608 - accuracy: 0.86 - ETA: 0s - loss: 0.3611 - accuracy: 0.86 - ETA: 0s - loss: 0.3611 - accuracy: 0.86 - ETA: 0s - loss: 0.3612 - accuracy: 0.86 - ETA: 0s - loss: 0.3607 - accuracy: 0.86 - ETA: 0s - loss: 0.3609 - accuracy: 0.86 - ETA: 0s - loss: 0.3611 - accuracy: 0.86 - ETA: 0s - loss: 0.3609 - accuracy: 0.86 - ETA: 0s - loss: 0.3606 - accuracy: 0.86 - ETA: 0s - loss: 0.3608 - accuracy: 0.86 - ETA: 0s - loss: 0.3609 - accuracy: 0.86 - ETA: 0s - loss: 0.3609 - accuracy: 0.86 - ETA: 0s - loss: 0.3609 - accuracy: 0.86 - ETA: 0s - loss: 0.3605 - accuracy: 0.86 - ETA: 0s - loss: 0.3604 - accuracy: 0.86 - ETA: 0s - loss: 0.3602 - accuracy: 0.86 - 4s 2ms/step - loss: 0.3601 - accuracy: 0.8682 - val_loss: 0.3846 - val_accuracy: 0.8613
Epoch 3/10
1875/1875 [==============================] - ETA: 0s - loss: 0.2594 - accuracy: 0.90 - ETA: 4s - loss: 0.3510 - accuracy: 0.87 - ETA: 3s - loss: 0.3414 - accuracy: 0.87 - ETA: 3s - loss: 0.3331 - accuracy: 0.87 - ETA: 3s - loss: 0.3201 - accuracy: 0.87 - ETA: 3s - loss: 0.3255 - accuracy: 0.87 - ETA: 3s - loss: 0.3194 - accuracy: 0.87 - ETA: 3s - loss: 0.3180 - accuracy: 0.87 - ETA: 3s - loss: 0.3146 - accuracy: 0.87 - ETA: 3s - loss: 0.3087 - accuracy: 0.88 - ETA: 3s - loss: 0.3118 - accuracy: 0.87 - ETA: 3s - loss: 0.3125 - accuracy: 0.87 - ETA: 3s - loss: 0.3156 - accuracy: 0.87 - ETA: 3s - loss: 0.3147 - accuracy: 0.88 - ETA: 3s - loss: 0.3117 - accuracy: 0.88 - ETA: 3s - loss: 0.3140 - accuracy: 0.88 - ETA: 3s - loss: 0.3159 - accuracy: 0.88 - ETA: 3s - loss: 0.3161 - accuracy: 0.88 - ETA: 3s - loss: 0.3140 - accuracy: 0.88 - ETA: 3s - loss: 0.3136 - accuracy: 0.88 - ETA: 3s - loss: 0.3153 - accuracy: 0.88 - ETA: 3s - loss: 0.3173 - accuracy: 0.87 - ETA: 3s - loss: 0.3218 - accuracy: 0.87 - ETA: 3s - loss: 0.3200 - accuracy: 0.87 - ETA: 3s - loss: 0.3218 - accuracy: 0.87 - ETA: 3s - loss: 0.3237 - accuracy: 0.87 - ETA: 2s - loss: 0.3238 - accuracy: 0.87 - ETA: 2s - loss: 0.3239 - accuracy: 0.87 - ETA: 2s - loss: 0.3246 - accuracy: 0.87 - ETA: 2s - loss: 0.3246 - accuracy: 0.87 - ETA: 2s - loss: 0.3238 - accuracy: 0.87 - ETA: 2s - loss: 0.3232 - accuracy: 0.87 - ETA: 2s - loss: 0.3235 - accuracy: 0.87 - ETA: 2s - loss: 0.3244 - accuracy: 0.87 - ETA: 2s - loss: 0.3250 - accuracy: 0.87 - ETA: 2s - loss: 0.3238 - accuracy: 0.87 - ETA: 2s - loss: 0.3236 - accuracy: 0.87 - ETA: 2s - loss: 0.3228 - accuracy: 0.87 - ETA: 2s - loss: 0.3221 - accuracy: 0.87 - ETA: 2s - loss: 0.3226 - accuracy: 0.87 - ETA: 2s - loss: 0.3240 - accuracy: 0.87 - ETA: 2s - loss: 0.3236 - accuracy: 0.87 - ETA: 1s - loss: 0.3234 - accuracy: 0.87 - ETA: 1s - loss: 0.3229 - accuracy: 0.87 - ETA: 1s - loss: 0.3245 - accuracy: 0.87 - ETA: 1s - loss: 0.3253 - accuracy: 0.87 - ETA: 1s - loss: 0.3252 - accuracy: 0.87 - ETA: 1s - loss: 0.3254 - accuracy: 0.87 - ETA: 1s - loss: 0.3261 - accuracy: 0.87 - ETA: 1s - loss: 0.3252 - accuracy: 0.87 - ETA: 1s - loss: 0.3251 - accuracy: 0.87 - ETA: 1s - loss: 0.3248 - accuracy: 0.87 - ETA: 1s - loss: 0.3258 - accuracy: 0.87 - ETA: 1s - loss: 0.3246 - accuracy: 0.87 - ETA: 1s - loss: 0.3241 - accuracy: 0.87 - ETA: 1s - loss: 0.3240 - accuracy: 0.87 - ETA: 1s - loss: 0.3239 - accuracy: 0.87 - ETA: 1s - loss: 0.3239 - accuracy: 0.87 - ETA: 1s - loss: 0.3239 - accuracy: 0.87 - ETA: 1s - loss: 0.3247 - accuracy: 0.87 - ETA: 1s - loss: 0.3242 - accuracy: 0.87 - ETA: 1s - loss: 0.3241 - accuracy: 0.87 - ETA: 0s - loss: 0.3237 - accuracy: 0.87 - ETA: 0s - loss: 0.3241 - accuracy: 0.87 - ETA: 0s - loss: 0.3246 - accuracy: 0.87 - ETA: 0s - loss: 0.3244 - accuracy: 0.87 - ETA: 0s - loss: 0.3243 - accuracy: 0.87 - ETA: 0s - loss: 0.3251 - accuracy: 0.87 - ETA: 0s - loss: 0.3250 - accuracy: 0.87 - ETA: 0s - loss: 0.3243 - accuracy: 0.87 - ETA: 0s - loss: 0.3253 - accuracy: 0.87 - ETA: 0s - loss: 0.3249 - accuracy: 0.87 - ETA: 0s - loss: 0.3247 - accuracy: 0.87 - ETA: 0s - loss: 0.3245 - accuracy: 0.87 - ETA: 0s - loss: 0.3242 - accuracy: 0.87 - ETA: 0s - loss: 0.3241 - accuracy: 0.87 - ETA: 0s - loss: 0.3244 - accuracy: 0.87 - ETA: 0s - loss: 0.3243 - accuracy: 0.87 - ETA: 0s - loss: 0.3239 - accuracy: 0.87 - ETA: 0s - loss: 0.3248 - accuracy: 0.87 - ETA: 0s - loss: 0.3244 - accuracy: 0.87 - ETA: 0s - loss: 0.3241 - accuracy: 0.88 - 4s 2ms/step - loss: 0.3242 - accuracy: 0.8799 - val_loss: 0.3620 - val_accuracy: 0.8699
Epoch 4/10
1875/1875 [==============================] - ETA: 0s - loss: 0.1509 - accuracy: 0.96 - ETA: 3s - loss: 0.2929 - accuracy: 0.88 - ETA: 3s - loss: 0.2929 - accuracy: 0.88 - ETA: 3s - loss: 0.3040 - accuracy: 0.88 - ETA: 3s - loss: 0.3054 - accuracy: 0.88 - ETA: 3s - loss: 0.2952 - accuracy: 0.89 - ETA: 3s - loss: 0.2991 - accuracy: 0.89 - ETA: 3s - loss: 0.2965 - accuracy: 0.89 - ETA: 4s - loss: 0.2967 - accuracy: 0.89 - ETA: 4s - loss: 0.3005 - accuracy: 0.89 - ETA: 4s - loss: 0.3009 - accuracy: 0.88 - ETA: 4s - loss: 0.2993 - accuracy: 0.88 - ETA: 4s - loss: 0.2990 - accuracy: 0.88 - ETA: 4s - loss: 0.3071 - accuracy: 0.88 - ETA: 3s - loss: 0.3073 - accuracy: 0.88 - ETA: 3s - loss: 0.3057 - accuracy: 0.88 - ETA: 3s - loss: 0.3052 - accuracy: 0.88 - ETA: 3s - loss: 0.3065 - accuracy: 0.88 - ETA: 3s - loss: 0.3058 - accuracy: 0.88 - ETA: 3s - loss: 0.3046 - accuracy: 0.88 - ETA: 3s - loss: 0.3021 - accuracy: 0.88 - ETA: 3s - loss: 0.3037 - accuracy: 0.88 - ETA: 3s - loss: 0.3027 - accuracy: 0.88 - ETA: 3s - loss: 0.3022 - accuracy: 0.88 - ETA: 3s - loss: 0.3012 - accuracy: 0.88 - ETA: 3s - loss: 0.3004 - accuracy: 0.88 - ETA: 3s - loss: 0.3005 - accuracy: 0.88 - ETA: 3s - loss: 0.3004 - accuracy: 0.88 - ETA: 3s - loss: 0.3002 - accuracy: 0.88 - ETA: 3s - loss: 0.2993 - accuracy: 0.88 - ETA: 2s - loss: 0.2976 - accuracy: 0.89 - ETA: 2s - loss: 0.2973 - accuracy: 0.88 - ETA: 2s - loss: 0.2975 - accuracy: 0.89 - ETA: 2s - loss: 0.2962 - accuracy: 0.89 - ETA: 2s - loss: 0.2963 - accuracy: 0.89 - ETA: 2s - loss: 0.2963 - accuracy: 0.89 - ETA: 2s - loss: 0.2974 - accuracy: 0.89 - ETA: 2s - loss: 0.2969 - accuracy: 0.89 - ETA: 2s - loss: 0.2963 - accuracy: 0.89 - ETA: 2s - loss: 0.2966 - accuracy: 0.89 - ETA: 2s - loss: 0.2985 - accuracy: 0.89 - ETA: 2s - loss: 0.2989 - accuracy: 0.89 - ETA: 2s - loss: 0.2995 - accuracy: 0.89 - ETA: 2s - loss: 0.3000 - accuracy: 0.89 - ETA: 1s - loss: 0.2996 - accuracy: 0.89 - ETA: 1s - loss: 0.2992 - accuracy: 0.89 - ETA: 1s - loss: 0.2990 - accuracy: 0.89 - ETA: 1s - loss: 0.2990 - accuracy: 0.89 - ETA: 1s - loss: 0.2991 - accuracy: 0.89 - ETA: 1s - loss: 0.2986 - accuracy: 0.89 - ETA: 1s - loss: 0.2985 - accuracy: 0.89 - ETA: 1s - loss: 0.2981 - accuracy: 0.89 - ETA: 1s - loss: 0.2981 - accuracy: 0.89 - ETA: 1s - loss: 0.2988 - accuracy: 0.89 - ETA: 1s - loss: 0.2986 - accuracy: 0.89 - ETA: 1s - loss: 0.2979 - accuracy: 0.89 - ETA: 1s - loss: 0.2981 - accuracy: 0.89 - ETA: 1s - loss: 0.2978 - accuracy: 0.89 - ETA: 1s - loss: 0.2973 - accuracy: 0.89 - ETA: 1s - loss: 0.2970 - accuracy: 0.89 - ETA: 1s - loss: 0.2971 - accuracy: 0.89 - ETA: 1s - loss: 0.2980 - accuracy: 0.89 - ETA: 0s - loss: 0.2980 - accuracy: 0.89 - ETA: 0s - loss: 0.2986 - accuracy: 0.89 - ETA: 0s - loss: 0.2988 - accuracy: 0.89 - ETA: 0s - loss: 0.2991 - accuracy: 0.89 - ETA: 0s - loss: 0.2996 - accuracy: 0.89 - ETA: 0s - loss: 0.3001 - accuracy: 0.89 - ETA: 0s - loss: 0.3006 - accuracy: 0.89 - ETA: 0s - loss: 0.3008 - accuracy: 0.89 - ETA: 0s - loss: 0.3009 - accuracy: 0.88 - ETA: 0s - loss: 0.3010 - accuracy: 0.88 - ETA: 0s - loss: 0.3011 - accuracy: 0.88 - ETA: 0s - loss: 0.3011 - accuracy: 0.88 - ETA: 0s - loss: 0.3022 - accuracy: 0.88 - ETA: 0s - loss: 0.3018 - accuracy: 0.88 - ETA: 0s - loss: 0.3016 - accuracy: 0.88 - ETA: 0s - loss: 0.3012 - accuracy: 0.88 - ETA: 0s - loss: 0.3008 - accuracy: 0.89 - ETA: 0s - loss: 0.3008 - accuracy: 0.89 - 4s 2ms/step - loss: 0.3005 - accuracy: 0.8902 - val_loss: 0.3331 - val_accuracy: 0.8797
Epoch 5/10
1875/1875 [==============================] - ETA: 0s - loss: 0.2307 - accuracy: 0.90 - ETA: 3s - loss: 0.2708 - accuracy: 0.90 - ETA: 3s - loss: 0.2830 - accuracy: 0.89 - ETA: 3s - loss: 0.2743 - accuracy: 0.89 - ETA: 3s - loss: 0.2733 - accuracy: 0.89 - ETA: 3s - loss: 0.2649 - accuracy: 0.90 - ETA: 3s - loss: 0.2606 - accuracy: 0.90 - ETA: 3s - loss: 0.2661 - accuracy: 0.90 - ETA: 3s - loss: 0.2681 - accuracy: 0.89 - ETA: 3s - loss: 0.2675 - accuracy: 0.89 - ETA: 3s - loss: 0.2653 - accuracy: 0.89 - ETA: 3s - loss: 0.2660 - accuracy: 0.89 - ETA: 3s - loss: 0.2649 - accuracy: 0.90 - ETA: 3s - loss: 0.2658 - accuracy: 0.90 - ETA: 3s - loss: 0.2627 - accuracy: 0.90 - ETA: 3s - loss: 0.2670 - accuracy: 0.89 - ETA: 3s - loss: 0.2652 - accuracy: 0.89 - ETA: 3s - loss: 0.2639 - accuracy: 0.89 - ETA: 3s - loss: 0.2634 - accuracy: 0.90 - ETA: 3s - loss: 0.2625 - accuracy: 0.90 - ETA: 3s - loss: 0.2632 - accuracy: 0.90 - ETA: 3s - loss: 0.2652 - accuracy: 0.89 - ETA: 3s - loss: 0.2637 - accuracy: 0.89 - ETA: 3s - loss: 0.2635 - accuracy: 0.89 - ETA: 3s - loss: 0.2632 - accuracy: 0.90 - ETA: 3s - loss: 0.2651 - accuracy: 0.89 - ETA: 2s - loss: 0.2637 - accuracy: 0.89 - ETA: 2s - loss: 0.2634 - accuracy: 0.89 - ETA: 2s - loss: 0.2648 - accuracy: 0.89 - ETA: 2s - loss: 0.2660 - accuracy: 0.89 - ETA: 2s - loss: 0.2666 - accuracy: 0.89 - ETA: 2s - loss: 0.2668 - accuracy: 0.89 - ETA: 2s - loss: 0.2660 - accuracy: 0.89 - ETA: 2s - loss: 0.2674 - accuracy: 0.89 - ETA: 2s - loss: 0.2671 - accuracy: 0.89 - ETA: 2s - loss: 0.2670 - accuracy: 0.89 - ETA: 2s - loss: 0.2664 - accuracy: 0.89 - ETA: 2s - loss: 0.2672 - accuracy: 0.89 - ETA: 2s - loss: 0.2681 - accuracy: 0.89 - ETA: 2s - loss: 0.2706 - accuracy: 0.89 - ETA: 2s - loss: 0.2704 - accuracy: 0.89 - ETA: 2s - loss: 0.2708 - accuracy: 0.89 - ETA: 2s - loss: 0.2718 - accuracy: 0.89 - ETA: 2s - loss: 0.2728 - accuracy: 0.89 - ETA: 1s - loss: 0.2736 - accuracy: 0.89 - ETA: 1s - loss: 0.2728 - accuracy: 0.89 - ETA: 1s - loss: 0.2723 - accuracy: 0.89 - ETA: 1s - loss: 0.2727 - accuracy: 0.89 - ETA: 1s - loss: 0.2732 - accuracy: 0.89 - ETA: 1s - loss: 0.2753 - accuracy: 0.89 - ETA: 1s - loss: 0.2755 - accuracy: 0.89 - ETA: 1s - loss: 0.2765 - accuracy: 0.89 - ETA: 1s - loss: 0.2766 - accuracy: 0.89 - ETA: 1s - loss: 0.2765 - accuracy: 0.89 - ETA: 1s - loss: 0.2774 - accuracy: 0.89 - ETA: 1s - loss: 0.2769 - accuracy: 0.89 - ETA: 1s - loss: 0.2781 - accuracy: 0.89 - ETA: 1s - loss: 0.2785 - accuracy: 0.89 - ETA: 1s - loss: 0.2784 - accuracy: 0.89 - ETA: 1s - loss: 0.2782 - accuracy: 0.89 - ETA: 1s - loss: 0.2783 - accuracy: 0.89 - ETA: 1s - loss: 0.2784 - accuracy: 0.89 - ETA: 0s - loss: 0.2786 - accuracy: 0.89 - ETA: 0s - loss: 0.2785 - accuracy: 0.89 - ETA: 0s - loss: 0.2787 - accuracy: 0.89 - ETA: 0s - loss: 0.2792 - accuracy: 0.89 - ETA: 0s - loss: 0.2795 - accuracy: 0.89 - ETA: 0s - loss: 0.2801 - accuracy: 0.89 - ETA: 0s - loss: 0.2804 - accuracy: 0.89 - ETA: 0s - loss: 0.2800 - accuracy: 0.89 - ETA: 0s - loss: 0.2798 - accuracy: 0.89 - ETA: 0s - loss: 0.2795 - accuracy: 0.89 - ETA: 0s - loss: 0.2805 - accuracy: 0.89 - ETA: 0s - loss: 0.2810 - accuracy: 0.89 - ETA: 0s - loss: 0.2809 - accuracy: 0.89 - ETA: 0s - loss: 0.2812 - accuracy: 0.89 - ETA: 0s - loss: 0.2808 - accuracy: 0.89 - ETA: 0s - loss: 0.2809 - accuracy: 0.89 - ETA: 0s - loss: 0.2805 - accuracy: 0.89 - ETA: 0s - loss: 0.2807 - accuracy: 0.89 - 4s 2ms/step - loss: 0.2807 - accuracy: 0.8957 - val_loss: 0.3321 - val_accuracy: 0.8831
Epoch 6/10
1875/1875 [==============================] - ETA: 0s - loss: 0.2249 - accuracy: 0.87 - ETA: 4s - loss: 0.2369 - accuracy: 0.91 - ETA: 3s - loss: 0.2308 - accuracy: 0.91 - ETA: 3s - loss: 0.2517 - accuracy: 0.90 - ETA: 3s - loss: 0.2609 - accuracy: 0.90 - ETA: 3s - loss: 0.2621 - accuracy: 0.90 - ETA: 3s - loss: 0.2601 - accuracy: 0.90 - ETA: 3s - loss: 0.2585 - accuracy: 0.90 - ETA: 3s - loss: 0.2597 - accuracy: 0.90 - ETA: 3s - loss: 0.2569 - accuracy: 0.90 - ETA: 3s - loss: 0.2599 - accuracy: 0.90 - ETA: 3s - loss: 0.2618 - accuracy: 0.90 - ETA: 3s - loss: 0.2632 - accuracy: 0.90 - ETA: 3s - loss: 0.2633 - accuracy: 0.90 - ETA: 3s - loss: 0.2623 - accuracy: 0.90 - ETA: 2s - loss: 0.2626 - accuracy: 0.90 - ETA: 2s - loss: 0.2625 - accuracy: 0.90 - ETA: 2s - loss: 0.2609 - accuracy: 0.90 - ETA: 2s - loss: 0.2602 - accuracy: 0.90 - ETA: 2s - loss: 0.2605 - accuracy: 0.90 - ETA: 2s - loss: 0.2613 - accuracy: 0.90 - ETA: 2s - loss: 0.2620 - accuracy: 0.90 - ETA: 2s - loss: 0.2644 - accuracy: 0.90 - ETA: 2s - loss: 0.2649 - accuracy: 0.90 - ETA: 2s - loss: 0.2659 - accuracy: 0.89 - ETA: 2s - loss: 0.2650 - accuracy: 0.90 - ETA: 2s - loss: 0.2673 - accuracy: 0.89 - ETA: 2s - loss: 0.2667 - accuracy: 0.90 - ETA: 2s - loss: 0.2659 - accuracy: 0.90 - ETA: 2s - loss: 0.2652 - accuracy: 0.90 - ETA: 2s - loss: 0.2666 - accuracy: 0.90 - ETA: 2s - loss: 0.2662 - accuracy: 0.90 - ETA: 2s - loss: 0.2667 - accuracy: 0.89 - ETA: 2s - loss: 0.2668 - accuracy: 0.90 - ETA: 2s - loss: 0.2658 - accuracy: 0.90 - ETA: 2s - loss: 0.2652 - accuracy: 0.90 - ETA: 2s - loss: 0.2649 - accuracy: 0.90 - ETA: 1s - loss: 0.2656 - accuracy: 0.90 - ETA: 1s - loss: 0.2661 - accuracy: 0.90 - ETA: 1s - loss: 0.2661 - accuracy: 0.90 - ETA: 1s - loss: 0.2654 - accuracy: 0.90 - ETA: 1s - loss: 0.2658 - accuracy: 0.90 - ETA: 1s - loss: 0.2662 - accuracy: 0.90 - ETA: 1s - loss: 0.2659 - accuracy: 0.90 - ETA: 1s - loss: 0.2660 - accuracy: 0.90 - ETA: 1s - loss: 0.2663 - accuracy: 0.90 - ETA: 1s - loss: 0.2668 - accuracy: 0.90 - ETA: 1s - loss: 0.2668 - accuracy: 0.90 - ETA: 1s - loss: 0.2670 - accuracy: 0.90 - ETA: 1s - loss: 0.2678 - accuracy: 0.90 - ETA: 1s - loss: 0.2679 - accuracy: 0.90 - ETA: 1s - loss: 0.2681 - accuracy: 0.89 - ETA: 1s - loss: 0.2681 - accuracy: 0.90 - ETA: 1s - loss: 0.2676 - accuracy: 0.90 - ETA: 1s - loss: 0.2676 - accuracy: 0.90 - ETA: 0s - loss: 0.2676 - accuracy: 0.90 - ETA: 0s - loss: 0.2676 - accuracy: 0.90 - ETA: 0s - loss: 0.2682 - accuracy: 0.90 - ETA: 0s - loss: 0.2682 - accuracy: 0.90 - ETA: 0s - loss: 0.2677 - accuracy: 0.90 - ETA: 0s - loss: 0.2676 - accuracy: 0.90 - ETA: 0s - loss: 0.2675 - accuracy: 0.90 - ETA: 0s - loss: 0.2675 - accuracy: 0.90 - ETA: 0s - loss: 0.2677 - accuracy: 0.90 - ETA: 0s - loss: 0.2672 - accuracy: 0.90 - ETA: 0s - loss: 0.2674 - accuracy: 0.89 - ETA: 0s - loss: 0.2676 - accuracy: 0.89 - ETA: 0s - loss: 0.2678 - accuracy: 0.89 - ETA: 0s - loss: 0.2677 - accuracy: 0.89 - ETA: 0s - loss: 0.2672 - accuracy: 0.90 - ETA: 0s - loss: 0.2675 - accuracy: 0.90 - ETA: 0s - loss: 0.2676 - accuracy: 0.89 - ETA: 0s - loss: 0.2676 - accuracy: 0.89 - ETA: 0s - loss: 0.2681 - accuracy: 0.89 - ETA: 0s - loss: 0.2677 - accuracy: 0.89 - 4s 2ms/step - loss: 0.2677 - accuracy: 0.8997 - val_loss: 0.3424 - val_accuracy: 0.8778
Epoch 7/10
1875/1875 [==============================] - ETA: 0s - loss: 0.4205 - accuracy: 0.87 - ETA: 3s - loss: 0.2115 - accuracy: 0.91 - ETA: 4s - loss: 0.2112 - accuracy: 0.92 - ETA: 4s - loss: 0.2169 - accuracy: 0.91 - ETA: 4s - loss: 0.2194 - accuracy: 0.92 - ETA: 4s - loss: 0.2296 - accuracy: 0.91 - ETA: 4s - loss: 0.2240 - accuracy: 0.91 - ETA: 4s - loss: 0.2275 - accuracy: 0.91 - ETA: 4s - loss: 0.2298 - accuracy: 0.91 - ETA: 3s - loss: 0.2342 - accuracy: 0.91 - ETA: 3s - loss: 0.2362 - accuracy: 0.91 - ETA: 3s - loss: 0.2384 - accuracy: 0.91 - ETA: 3s - loss: 0.2411 - accuracy: 0.91 - ETA: 3s - loss: 0.2386 - accuracy: 0.91 - ETA: 3s - loss: 0.2420 - accuracy: 0.91 - ETA: 3s - loss: 0.2425 - accuracy: 0.91 - ETA: 3s - loss: 0.2468 - accuracy: 0.90 - ETA: 3s - loss: 0.2466 - accuracy: 0.90 - ETA: 3s - loss: 0.2468 - accuracy: 0.90 - ETA: 3s - loss: 0.2445 - accuracy: 0.90 - ETA: 3s - loss: 0.2438 - accuracy: 0.90 - ETA: 3s - loss: 0.2453 - accuracy: 0.90 - ETA: 3s - loss: 0.2484 - accuracy: 0.90 - ETA: 3s - loss: 0.2482 - accuracy: 0.90 - ETA: 2s - loss: 0.2477 - accuracy: 0.90 - ETA: 2s - loss: 0.2473 - accuracy: 0.90 - ETA: 2s - loss: 0.2470 - accuracy: 0.90 - ETA: 2s - loss: 0.2464 - accuracy: 0.90 - ETA: 2s - loss: 0.2474 - accuracy: 0.90 - ETA: 2s - loss: 0.2475 - accuracy: 0.90 - ETA: 2s - loss: 0.2465 - accuracy: 0.90 - ETA: 2s - loss: 0.2456 - accuracy: 0.90 - ETA: 2s - loss: 0.2448 - accuracy: 0.90 - ETA: 2s - loss: 0.2457 - accuracy: 0.90 - ETA: 2s - loss: 0.2465 - accuracy: 0.90 - ETA: 2s - loss: 0.2462 - accuracy: 0.90 - ETA: 2s - loss: 0.2457 - accuracy: 0.90 - ETA: 2s - loss: 0.2456 - accuracy: 0.90 - ETA: 2s - loss: 0.2463 - accuracy: 0.90 - ETA: 2s - loss: 0.2458 - accuracy: 0.90 - ETA: 2s - loss: 0.2461 - accuracy: 0.90 - ETA: 1s - loss: 0.2461 - accuracy: 0.90 - ETA: 1s - loss: 0.2468 - accuracy: 0.90 - ETA: 1s - loss: 0.2463 - accuracy: 0.90 - ETA: 1s - loss: 0.2473 - accuracy: 0.90 - ETA: 1s - loss: 0.2475 - accuracy: 0.90 - ETA: 1s - loss: 0.2475 - accuracy: 0.90 - ETA: 1s - loss: 0.2480 - accuracy: 0.90 - ETA: 1s - loss: 0.2476 - accuracy: 0.90 - ETA: 1s - loss: 0.2477 - accuracy: 0.90 - ETA: 1s - loss: 0.2479 - accuracy: 0.90 - ETA: 1s - loss: 0.2487 - accuracy: 0.90 - ETA: 1s - loss: 0.2484 - accuracy: 0.90 - ETA: 1s - loss: 0.2486 - accuracy: 0.90 - ETA: 1s - loss: 0.2484 - accuracy: 0.90 - ETA: 1s - loss: 0.2484 - accuracy: 0.90 - ETA: 1s - loss: 0.2488 - accuracy: 0.90 - ETA: 1s - loss: 0.2494 - accuracy: 0.90 - ETA: 1s - loss: 0.2488 - accuracy: 0.90 - ETA: 0s - loss: 0.2493 - accuracy: 0.90 - ETA: 0s - loss: 0.2494 - accuracy: 0.90 - ETA: 0s - loss: 0.2489 - accuracy: 0.90 - ETA: 0s - loss: 0.2493 - accuracy: 0.90 - ETA: 0s - loss: 0.2493 - accuracy: 0.90 - ETA: 0s - loss: 0.2492 - accuracy: 0.90 - ETA: 0s - loss: 0.2497 - accuracy: 0.90 - ETA: 0s - loss: 0.2504 - accuracy: 0.90 - ETA: 0s - loss: 0.2510 - accuracy: 0.90 - ETA: 0s - loss: 0.2513 - accuracy: 0.90 - ETA: 0s - loss: 0.2514 - accuracy: 0.90 - ETA: 0s - loss: 0.2514 - accuracy: 0.90 - ETA: 0s - loss: 0.2514 - accuracy: 0.90 - ETA: 0s - loss: 0.2513 - accuracy: 0.90 - ETA: 0s - loss: 0.2511 - accuracy: 0.90 - ETA: 0s - loss: 0.2509 - accuracy: 0.90 - ETA: 0s - loss: 0.2513 - accuracy: 0.90 - ETA: 0s - loss: 0.2512 - accuracy: 0.90 - 4s 2ms/step - loss: 0.2510 - accuracy: 0.9052 - val_loss: 0.3325 - val_accuracy: 0.8813
Epoch 8/10
1875/1875 [==============================] - ETA: 0s - loss: 0.0650 - accuracy: 1.00 - ETA: 3s - loss: 0.2143 - accuracy: 0.93 - ETA: 3s - loss: 0.2265 - accuracy: 0.91 - ETA: 3s - loss: 0.2314 - accuracy: 0.91 - ETA: 3s - loss: 0.2265 - accuracy: 0.91 - ETA: 3s - loss: 0.2175 - accuracy: 0.92 - ETA: 3s - loss: 0.2209 - accuracy: 0.92 - ETA: 3s - loss: 0.2239 - accuracy: 0.91 - ETA: 3s - loss: 0.2256 - accuracy: 0.91 - ETA: 3s - loss: 0.2283 - accuracy: 0.91 - ETA: 3s - loss: 0.2304 - accuracy: 0.91 - ETA: 3s - loss: 0.2311 - accuracy: 0.91 - ETA: 3s - loss: 0.2319 - accuracy: 0.91 - ETA: 3s - loss: 0.2312 - accuracy: 0.91 - ETA: 3s - loss: 0.2324 - accuracy: 0.91 - ETA: 3s - loss: 0.2340 - accuracy: 0.91 - ETA: 3s - loss: 0.2333 - accuracy: 0.91 - ETA: 2s - loss: 0.2333 - accuracy: 0.91 - ETA: 2s - loss: 0.2344 - accuracy: 0.91 - ETA: 2s - loss: 0.2333 - accuracy: 0.91 - ETA: 2s - loss: 0.2336 - accuracy: 0.91 - ETA: 2s - loss: 0.2337 - accuracy: 0.91 - ETA: 2s - loss: 0.2357 - accuracy: 0.91 - ETA: 2s - loss: 0.2364 - accuracy: 0.91 - ETA: 2s - loss: 0.2364 - accuracy: 0.91 - ETA: 2s - loss: 0.2360 - accuracy: 0.91 - ETA: 2s - loss: 0.2359 - accuracy: 0.91 - ETA: 2s - loss: 0.2374 - accuracy: 0.91 - ETA: 2s - loss: 0.2379 - accuracy: 0.91 - ETA: 2s - loss: 0.2373 - accuracy: 0.91 - ETA: 2s - loss: 0.2382 - accuracy: 0.91 - ETA: 2s - loss: 0.2381 - accuracy: 0.91 - ETA: 2s - loss: 0.2380 - accuracy: 0.91 - ETA: 2s - loss: 0.2386 - accuracy: 0.91 - ETA: 2s - loss: 0.2383 - accuracy: 0.91 - ETA: 2s - loss: 0.2384 - accuracy: 0.91 - ETA: 2s - loss: 0.2396 - accuracy: 0.91 - ETA: 2s - loss: 0.2387 - accuracy: 0.91 - ETA: 1s - loss: 0.2383 - accuracy: 0.91 - ETA: 1s - loss: 0.2388 - accuracy: 0.91 - ETA: 1s - loss: 0.2397 - accuracy: 0.91 - ETA: 1s - loss: 0.2393 - accuracy: 0.91 - ETA: 1s - loss: 0.2391 - accuracy: 0.91 - ETA: 1s - loss: 0.2394 - accuracy: 0.91 - ETA: 1s - loss: 0.2396 - accuracy: 0.91 - ETA: 1s - loss: 0.2392 - accuracy: 0.91 - ETA: 1s - loss: 0.2385 - accuracy: 0.91 - ETA: 1s - loss: 0.2395 - accuracy: 0.91 - ETA: 1s - loss: 0.2393 - accuracy: 0.91 - ETA: 1s - loss: 0.2396 - accuracy: 0.91 - ETA: 1s - loss: 0.2398 - accuracy: 0.91 - ETA: 1s - loss: 0.2399 - accuracy: 0.91 - ETA: 1s - loss: 0.2393 - accuracy: 0.91 - ETA: 1s - loss: 0.2398 - accuracy: 0.91 - ETA: 1s - loss: 0.2397 - accuracy: 0.91 - ETA: 1s - loss: 0.2401 - accuracy: 0.90 - ETA: 1s - loss: 0.2399 - accuracy: 0.90 - ETA: 0s - loss: 0.2406 - accuracy: 0.90 - ETA: 0s - loss: 0.2398 - accuracy: 0.90 - ETA: 0s - loss: 0.2392 - accuracy: 0.90 - ETA: 0s - loss: 0.2397 - accuracy: 0.90 - ETA: 0s - loss: 0.2410 - accuracy: 0.90 - ETA: 0s - loss: 0.2408 - accuracy: 0.90 - ETA: 0s - loss: 0.2415 - accuracy: 0.90 - ETA: 0s - loss: 0.2421 - accuracy: 0.90 - ETA: 0s - loss: 0.2426 - accuracy: 0.90 - ETA: 0s - loss: 0.2419 - accuracy: 0.90 - ETA: 0s - loss: 0.2428 - accuracy: 0.90 - ETA: 0s - loss: 0.2423 - accuracy: 0.90 - ETA: 0s - loss: 0.2419 - accuracy: 0.90 - ETA: 0s - loss: 0.2424 - accuracy: 0.90 - ETA: 0s - loss: 0.2423 - accuracy: 0.90 - ETA: 0s - loss: 0.2422 - accuracy: 0.90 - ETA: 0s - loss: 0.2419 - accuracy: 0.90 - ETA: 0s - loss: 0.2423 - accuracy: 0.90 - ETA: 0s - loss: 0.2427 - accuracy: 0.90 - ETA: 0s - loss: 0.2417 - accuracy: 0.90 - 4s 2ms/step - loss: 0.2417 - accuracy: 0.9091 - val_loss: 0.3316 - val_accuracy: 0.8840
Epoch 9/10
1875/1875 [==============================] - ETA: 0s - loss: 0.2847 - accuracy: 0.84 - ETA: 3s - loss: 0.2580 - accuracy: 0.89 - ETA: 3s - loss: 0.2399 - accuracy: 0.90 - ETA: 3s - loss: 0.2306 - accuracy: 0.91 - ETA: 3s - loss: 0.2268 - accuracy: 0.91 - ETA: 3s - loss: 0.2198 - accuracy: 0.91 - ETA: 3s - loss: 0.2168 - accuracy: 0.92 - ETA: 3s - loss: 0.2216 - accuracy: 0.91 - ETA: 3s - loss: 0.2278 - accuracy: 0.91 - ETA: 3s - loss: 0.2280 - accuracy: 0.91 - ETA: 3s - loss: 0.2280 - accuracy: 0.91 - ETA: 3s - loss: 0.2261 - accuracy: 0.91 - ETA: 3s - loss: 0.2243 - accuracy: 0.91 - ETA: 3s - loss: 0.2226 - accuracy: 0.91 - ETA: 3s - loss: 0.2261 - accuracy: 0.91 - ETA: 3s - loss: 0.2281 - accuracy: 0.91 - ETA: 2s - loss: 0.2266 - accuracy: 0.91 - ETA: 2s - loss: 0.2281 - accuracy: 0.91 - ETA: 2s - loss: 0.2294 - accuracy: 0.91 - ETA: 2s - loss: 0.2306 - accuracy: 0.91 - ETA: 2s - loss: 0.2320 - accuracy: 0.91 - ETA: 2s - loss: 0.2311 - accuracy: 0.91 - ETA: 2s - loss: 0.2305 - accuracy: 0.91 - ETA: 2s - loss: 0.2307 - accuracy: 0.91 - ETA: 2s - loss: 0.2314 - accuracy: 0.91 - ETA: 2s - loss: 0.2313 - accuracy: 0.91 - ETA: 2s - loss: 0.2308 - accuracy: 0.91 - ETA: 2s - loss: 0.2300 - accuracy: 0.91 - ETA: 2s - loss: 0.2297 - accuracy: 0.91 - ETA: 2s - loss: 0.2299 - accuracy: 0.91 - ETA: 2s - loss: 0.2301 - accuracy: 0.91 - ETA: 2s - loss: 0.2297 - accuracy: 0.91 - ETA: 2s - loss: 0.2296 - accuracy: 0.91 - ETA: 2s - loss: 0.2297 - accuracy: 0.91 - ETA: 2s - loss: 0.2303 - accuracy: 0.91 - ETA: 2s - loss: 0.2306 - accuracy: 0.91 - ETA: 2s - loss: 0.2306 - accuracy: 0.91 - ETA: 1s - loss: 0.2303 - accuracy: 0.91 - ETA: 1s - loss: 0.2298 - accuracy: 0.91 - ETA: 1s - loss: 0.2302 - accuracy: 0.91 - ETA: 1s - loss: 0.2301 - accuracy: 0.91 - ETA: 1s - loss: 0.2310 - accuracy: 0.91 - ETA: 1s - loss: 0.2310 - accuracy: 0.91 - ETA: 1s - loss: 0.2312 - accuracy: 0.91 - ETA: 1s - loss: 0.2305 - accuracy: 0.91 - ETA: 1s - loss: 0.2305 - accuracy: 0.91 - ETA: 1s - loss: 0.2317 - accuracy: 0.91 - ETA: 1s - loss: 0.2319 - accuracy: 0.91 - ETA: 1s - loss: 0.2320 - accuracy: 0.91 - ETA: 1s - loss: 0.2322 - accuracy: 0.91 - ETA: 1s - loss: 0.2327 - accuracy: 0.91 - ETA: 1s - loss: 0.2334 - accuracy: 0.91 - ETA: 1s - loss: 0.2331 - accuracy: 0.91 - ETA: 1s - loss: 0.2327 - accuracy: 0.91 - ETA: 1s - loss: 0.2320 - accuracy: 0.91 - ETA: 1s - loss: 0.2325 - accuracy: 0.91 - ETA: 0s - loss: 0.2334 - accuracy: 0.91 - ETA: 0s - loss: 0.2333 - accuracy: 0.91 - ETA: 0s - loss: 0.2329 - accuracy: 0.91 - ETA: 0s - loss: 0.2320 - accuracy: 0.91 - ETA: 0s - loss: 0.2324 - accuracy: 0.91 - ETA: 0s - loss: 0.2322 - accuracy: 0.91 - ETA: 0s - loss: 0.2334 - accuracy: 0.91 - ETA: 0s - loss: 0.2338 - accuracy: 0.91 - ETA: 0s - loss: 0.2342 - accuracy: 0.91 - ETA: 0s - loss: 0.2339 - accuracy: 0.91 - ETA: 0s - loss: 0.2342 - accuracy: 0.91 - ETA: 0s - loss: 0.2333 - accuracy: 0.91 - ETA: 0s - loss: 0.2331 - accuracy: 0.91 - ETA: 0s - loss: 0.2329 - accuracy: 0.91 - ETA: 0s - loss: 0.2330 - accuracy: 0.91 - ETA: 0s - loss: 0.2325 - accuracy: 0.91 - ETA: 0s - loss: 0.2330 - accuracy: 0.91 - ETA: 0s - loss: 0.2329 - accuracy: 0.91 - ETA: 0s - loss: 0.2327 - accuracy: 0.91 - 4s 2ms/step - loss: 0.2334 - accuracy: 0.9133 - val_loss: 0.3296 - val_accuracy: 0.8894
Epoch 10/10
1875/1875 [==============================] - ETA: 0s - loss: 0.4754 - accuracy: 0.87 - ETA: 3s - loss: 0.2514 - accuracy: 0.90 - ETA: 3s - loss: 0.2290 - accuracy: 0.91 - ETA: 3s - loss: 0.2225 - accuracy: 0.91 - ETA: 3s - loss: 0.2142 - accuracy: 0.92 - ETA: 3s - loss: 0.2201 - accuracy: 0.91 - ETA: 3s - loss: 0.2217 - accuracy: 0.91 - ETA: 3s - loss: 0.2208 - accuracy: 0.91 - ETA: 3s - loss: 0.2221 - accuracy: 0.91 - ETA: 3s - loss: 0.2206 - accuracy: 0.91 - ETA: 3s - loss: 0.2190 - accuracy: 0.91 - ETA: 3s - loss: 0.2164 - accuracy: 0.91 - ETA: 3s - loss: 0.2157 - accuracy: 0.92 - ETA: 3s - loss: 0.2159 - accuracy: 0.91 - ETA: 3s - loss: 0.2184 - accuracy: 0.91 - ETA: 2s - loss: 0.2191 - accuracy: 0.91 - ETA: 2s - loss: 0.2172 - accuracy: 0.91 - ETA: 2s - loss: 0.2171 - accuracy: 0.91 - ETA: 2s - loss: 0.2163 - accuracy: 0.91 - ETA: 2s - loss: 0.2165 - accuracy: 0.91 - ETA: 2s - loss: 0.2162 - accuracy: 0.91 - ETA: 2s - loss: 0.2157 - accuracy: 0.91 - ETA: 2s - loss: 0.2167 - accuracy: 0.91 - ETA: 2s - loss: 0.2166 - accuracy: 0.91 - ETA: 2s - loss: 0.2151 - accuracy: 0.91 - ETA: 2s - loss: 0.2161 - accuracy: 0.91 - ETA: 2s - loss: 0.2166 - accuracy: 0.91 - ETA: 2s - loss: 0.2165 - accuracy: 0.91 - ETA: 2s - loss: 0.2174 - accuracy: 0.91 - ETA: 2s - loss: 0.2169 - accuracy: 0.91 - ETA: 2s - loss: 0.2177 - accuracy: 0.91 - ETA: 2s - loss: 0.2178 - accuracy: 0.91 - ETA: 2s - loss: 0.2172 - accuracy: 0.91 - ETA: 2s - loss: 0.2179 - accuracy: 0.91 - ETA: 2s - loss: 0.2186 - accuracy: 0.91 - ETA: 1s - loss: 0.2184 - accuracy: 0.91 - ETA: 1s - loss: 0.2187 - accuracy: 0.91 - ETA: 1s - loss: 0.2175 - accuracy: 0.91 - ETA: 1s - loss: 0.2178 - accuracy: 0.91 - ETA: 1s - loss: 0.2186 - accuracy: 0.91 - ETA: 1s - loss: 0.2187 - accuracy: 0.91 - ETA: 1s - loss: 0.2188 - accuracy: 0.91 - ETA: 1s - loss: 0.2191 - accuracy: 0.91 - ETA: 1s - loss: 0.2190 - accuracy: 0.91 - ETA: 1s - loss: 0.2202 - accuracy: 0.91 - ETA: 1s - loss: 0.2206 - accuracy: 0.91 - ETA: 1s - loss: 0.2208 - accuracy: 0.91 - ETA: 1s - loss: 0.2203 - accuracy: 0.91 - ETA: 1s - loss: 0.2206 - accuracy: 0.91 - ETA: 1s - loss: 0.2206 - accuracy: 0.91 - ETA: 1s - loss: 0.2209 - accuracy: 0.91 - ETA: 1s - loss: 0.2213 - accuracy: 0.91 - ETA: 1s - loss: 0.2219 - accuracy: 0.91 - ETA: 1s - loss: 0.2221 - accuracy: 0.91 - ETA: 1s - loss: 0.2223 - accuracy: 0.91 - ETA: 1s - loss: 0.2220 - accuracy: 0.91 - ETA: 0s - loss: 0.2224 - accuracy: 0.91 - ETA: 0s - loss: 0.2223 - accuracy: 0.91 - ETA: 0s - loss: 0.2229 - accuracy: 0.91 - ETA: 0s - loss: 0.2227 - accuracy: 0.91 - ETA: 0s - loss: 0.2231 - accuracy: 0.91 - ETA: 0s - loss: 0.2230 - accuracy: 0.91 - ETA: 0s - loss: 0.2226 - accuracy: 0.91 - ETA: 0s - loss: 0.2227 - accuracy: 0.91 - ETA: 0s - loss: 0.2228 - accuracy: 0.91 - ETA: 0s - loss: 0.2232 - accuracy: 0.91 - ETA: 0s - loss: 0.2236 - accuracy: 0.91 - ETA: 0s - loss: 0.2229 - accuracy: 0.91 - ETA: 0s - loss: 0.2227 - accuracy: 0.91 - ETA: 0s - loss: 0.2230 - accuracy: 0.91 - ETA: 0s - loss: 0.2229 - accuracy: 0.91 - ETA: 0s - loss: 0.2226 - accuracy: 0.91 - ETA: 0s - loss: 0.2222 - accuracy: 0.91 - ETA: 0s - loss: 0.2225 - accuracy: 0.91 - ETA: 0s - loss: 0.2222 - accuracy: 0.91 - ETA: 0s - loss: 0.2223 - accuracy: 0.91 - 4s 2ms/step - loss: 0.2230 - accuracy: 0.9168 - val_loss: 0.3288 - val_accuracy: 0.8887
Out[13]:

你可能感兴趣的:(TensorFlow学习,tensorflow,Keras,Tuner,Keras,超参数,神经网络)

C语言如何定义宏函数？小九格物 c语言
在C语言中，宏函数是通过预处理器定义的，它在编译之前替换代码中的宏调用。宏函数可以模拟函数的行为，但它们不是真正的函数，因为它们在编译时不会进行类型检查，也不会分配存储空间。宏函数的定义通常使用#define指令，后面跟着宏的名称和参数列表，以及宏展开后的代码。宏函数的定义方式：1.基本宏函数：这是最简单的宏函数形式，它直接定义一个表达式。#defineSQUARE(x)((x)*(x))2.带参
【一起学Rust | 设计模式】习惯语法——使用借用类型作为参数、格式化拼接字符串、构造函数广龙宇一起学Rust #Rust设计模式 rust 设计模式开发语言
提示：文章写完后，目录可以自动生成，如何生成可参考右边的帮助文档文章目录前言一、使用借用类型作为参数二、格式化拼接字符串三、使用构造函数总结前言Rust不是传统的面向对象编程语言，它的所有特性，使其独一无二。因此，学习特定于Rust的设计模式是必要的。本系列文章为作者学习《Rust设计模式》的学习笔记以及自己的见解。因此，本系列文章的结构也与此书的结构相同（后续可能会调成结构），基本上分为三个部分
Day17笔记-高阶函数 ~在杰难逃~ Python 笔记 python 开发语言 pycharm 数据分析
高阶函数【重点掌握】函数的本质：函数是一个变量，函数名是一个变量名，一个函数可以作为另一个函数的参数或返回值使用如果A函数作为B函数的参数，B函数调用完成之后，会得到一个结果，则B函数被称为高阶函数常用的高阶函数：map(),reduce(),filter(),sorted()1.map()map(func,iterable)，返回值是一个iterator【容器，迭代器】func:函数iterab
MYSQL面试系列-04 king01299 面试 mysql 面试
MYSQL面试系列-0417.关于redolog和binlog的刷盘机制、redolog、undolog作用、GTID是做什么的？innodb_flush_log_at_trx_commit及sync_binlog参数意义双117.1innodb_flush_log_at_trx_commit该变量定义了InnoDB在每次事务提交时，如何处理未刷入（flush）的重做日志信息（redolog）。它
【JS】执行时长(100分) |思路参考+代码解析（C++） l939035548 JS 算法数据结构 c++
题目为了充分发挥GPU算力，需要尽可能多的将任务交给GPU执行，现在有一个任务数组，数组元素表示在这1秒内新增的任务个数且每秒都有新增任务。假设GPU最多一次执行n个任务，一次执行耗时1秒，在保证GPU不空闲情况下，最少需要多长时间执行完成。题目输入第一个参数为GPU一次最多执行的任务个数，取值范围[1,10000]第二个参数为任务数组长度，取值范围[1,10000]第三个参数为任务数组，数字范围
走向以教育叙事为载体的教育叙事研究 666小飞鱼
今天我读了吴松超老师的《给教师的68条建写作建议》中的第23条《如何通过教育叙事走向研究》，吴老师在文中与我们分享了一个德育案例，这是一个反面的案例，意在告知我们在处理问题时，不能就考虑的点太窄，思考要全面。走向教育叙事研究，教师要有敏锐的“感知力”，这个感知力来自于背后专业知识的支撑，思维能力以及广阔的视野和见识等。所以对于同一件事处理方法不同，这个就是教师背后“敏锐力”的不同造成的，也就是说是
回溯算法-重新安排行程 chirou_ 算法数据结构图论 c++图搜索
leetcode332.重新安排行程这题我还没自己ac过，只能现在凭着刚学完的热乎劲把我对题解的理解记下来。本题我认为对数据结构的考察比较多，用什么数据结构去存数据，去读取数据，都是很重要的。classSolution{private:unordered_map>targets;boolbacktracking(intticketNum,vector&result){//1.确定参数和返回值//2
Rust基础知识 GRKF15 rust 开发语言后端
1.Rust语言简介1.1基础语法变量声明：let关键字用于声明变量，可以指定或不指定类型，如leta=10;和letmutc=30i32;。函数定义：使用fn关键字定义函数，并指定参数类型及返回类型，如fnadd(i:i32,j:i32)->i32{i+j}。控制流：包括if、else等，控制语句后需要使用;来结束语句。1.2数据类型整数类型：i8、i16、i32、i64、i128，以及无符号的
【华为OD技术面试真题 - 技术面】- python八股文真题题库（1）算法大师华为od 面试 python
华为OD面试真题精选专栏：华为OD面试真题精选目录:2024华为OD面试手撕代码真题目录以及八股文真题目录文章目录华为OD面试真题精选1.数据预处理流程数据预处理的主要步骤工具和库2.介绍线性回归、逻辑回归模型线性回归（LinearRegression）模型形式：关键点：逻辑回归（LogisticRegression）模型形式：关键点：参数估计与评估：3.python浅拷贝及深拷贝浅拷贝（Shal
SpringBlade dict-biz/list 接口 SQL 注入漏洞文章永久免费只为良心 oracle 数据库
SpringBladedict-biz/list接口SQL注入漏洞POC:构造请求包查看返回包你的网址/api/blade-system/dict-biz/list?updatexml(1,concat(0x7e,md5(1),0x7e),1)=1漏洞概述在SpringBlade框架中，如果dict-biz/list接口的后台处理逻辑没有正确地对用户输入进行过滤或参数化查询（PreparedSta
Python开发常用的三方模块如下：换个网名有点难 python 开发语言
Python是一门功能强大的编程语言，拥有丰富的第三方库，这些库为开发者提供了极大的便利。以下是100个常用的Python库，涵盖了多个领域：1、NumPy，用于科学计算的基础库。2、Pandas，提供数据结构和数据分析工具。3、Matplotlib，一个绘图库。4、Scikit-learn，机器学习库。5、SciPy，用于数学、科学和工程的库。6、TensorFlow，由Google开发的开源机
发现荞麦过敏这件事怪小泊
荞麦在我这里不是常用的谷物。所以前二十年，我以为自己是很正常的，从街头小吃到包装零食到每日三餐，从来没有不能吃的。可是有天我突然病倒了，喉咙火辣辣的肿痛，口水都咽不下去，舌头发麻。当时我以为吃太多零食所致，因为那天我吃了很辣的泡椒凤爪。其实我是不怎么吃辣的。而那个泡椒凤爪真的超辣。当时我以为自己吃多了，并不知道自己对哪个食物过敏。因为不舒服我早早睡了，结果并没有睡着。肚子一阵一阵疼，非常痛苦，终于
Linux sh命令 fengyehongWorld Linux linux
目录一.基本语法二.选项2.1-c字符串中读取内容，并执行2.1.1基本用法2.1.2获取当前目录下失效的超链接2.2-x每个命令执行之前，将其打印出来2.3结合Here文档使用一.基本语法⏹Linux和Unix系统中用于执行shell脚本或运行命令的命令。sh[选项][脚本文件][参数...]⏹选项-c：从字符串中读取内容，并执行。-x：在每个命令执行之前，将其打印出来。-s：从标准流中读取内容
python中的深拷贝与浅拷贝 anshejd70787 python
深拷贝和浅拷贝浅拷贝的时候，修改原来的对象，浅拷贝的对象不会发生改变。1、对象的赋值对象的赋值实际上是对象之间的引用：当创建一个对象，然后将这个对象赋值给另外一个变量的时候，python并没有拷贝这个对象，而只是拷贝了这个对象的引用。当对对象做赋值或者是参数传递或者作为返回值的时候，总是传递原始对象的引用，而不是一个副本。如下所示：>>>aList=["kel","abc",123]>>>bLis
[Python] 数据结构详解及代码 AIAdvocate 算法 python 数据结构链表
今日内容大纲介绍数据结构介绍列表链表1.数据结构和算法简介程序大白话翻译,程序=数据结构+算法数据结构指的是存储,组织数据的方式.算法指的是为了解决实际业务问题而思考思路和方法,就叫:算法.2.算法的5大特性介绍算法具有独立性算法是解决问题的思路和方式,最重要的是思维,而不是语言,其(算法)可以通过多种语言进行演绎.5大特性有输入,需要传入1或者多个参数有输出,需要返回1个或者多个结果有穷性,执行
Python神器！WEB自动化测试集成工具 DrissionPage 亚丁号 python 开发语言
一、前言用requests做数据采集面对要登录的网站时，要分析数据包、JS源码，构造复杂的请求，往往还要应付验证码、JS混淆、签名参数等反爬手段，门槛较高。若数据是由JS计算生成的，还须重现计算过程，体验不好，开发效率不高。使用浏览器，可以很大程度上绕过这些坑，但浏览器运行效率不高。因此，这个库设计初衷，是将它们合而为一，能够在不同须要时切换相应模式，并提供一种人性化的使用方法，提高开发和运行效率
更改npm镜像源为淘宝镜像骆小骆基于node.js
npm常用指令后缀*最近复习了一下node.js整理了一下跟node.js相关的指令后缀*--save、-S参数意思是把模块的版本信息保存到dependencies（生产环境依赖）中，即你的package.json文件的dependencies字段中；–--save-dev、-D参数意思是把模块版本信息保存到devDependencies（开发环境依赖）中，即你的package.json文件的de
常见的 JVM 调优方法有哪些？爪哇天下 jvm
常见的JVM调优方法有哪些？可以具体到调整哪个参数，调成什么值？对年轻代的EdenSurvivor的比例进行配置-XX:SurvivorRatio=8：表示设置2个Survivor区：1个Eden区的大小比值为2:8，这意味着Survivor区占整个年轻代的1/5，这个参数默认为8如果经常性的SurvivorTo放不下YGC的剩余的对象时候，可以适当的调整比例常用的CMS收集器：设置回收阈值，需要
粒子群优化 (PSO) 在三维正弦波函数中的应用 subject625Ruben 机器学习人工智能 matlab 算法
在这篇博客中，我们将展示如何使用粒子群优化（PSO）算法求解三维正弦波函数，并通过增加正弦波扰动，使优化过程更加复杂和有趣。本文将介绍目标函数的定义、PSO参数设置以及算法执行的详细过程，并展示搜索空间中的动态过程和收敛曲线。1.目标函数定义我们使用的目标函数是一个三维正弦波函数，定义如下：objectiveFunc=@(x)sin(sqrt(x(1).^2+x(2).^2))+0.5*sin(5
100天持续行动—Day01 Richard_DL
今天开始站着学习，发现效率大幅提升。把fast.ai的Lesson1的后半部分和Lesson2看完了。由于Keras版本和视频中的不一致，运行notebook时经常出现莫名其妙的错误，导致自己只动手实践了视频中的一小部分内容。为了赶时间，我打算先把与CNN相关的视频过一遍。然后尽快开始做自己的项目。明天继续加油，争取把Lesson3和Lesson4看完。
但行好事，莫问前程娟恋YOU
下班路上，路过菜市场，突然想吃火锅了。于是于是边去常去批发雪糕的地方去买些涮火锅的丸子类的菜，选了一些后，结算是24.9，老板爽快的说，给我24.5就行，正好钱包有现金25元，超递给老板结账，老板又给我一元钱，说24就行，我说那怎么可以，我又沾光了，总沾你的光，多不好意思！老板说你常来我就是沾你的光了！最后还是收了我24块钱，心里还是很开心的。不是因为沾光而开心，而是被让的开心！这样的老板做生意闻
MyBatis 详解阿贾克斯的黎明 java mybatis
目录目录一、MyBatis是什么二、为什么使用MyBatis（一）灵活性高（二）性能优化（三）易于维护三、怎么用MyBatis（一）添加依赖（二）配置MyBatis（三）创建实体类和接口（四）使用MyBatis一、MyBatis是什么MyBatis是一个优秀的持久层框架，它支持自定义SQL、存储过程以及高级映射。MyBatis免除了几乎所有的JDBC代码以及设置参数和获取结果集的工作。它可以通过简
ai绘画工具midjourney怎么下载？附作品管理教程设计师早上好
Midjourney是一款功能强大的AI绘画工具，它使用机器学习技术和深度神经网络等算法，可以生成各种艺术风格的绘画作品。在创意设计、广告宣传等方面有着广泛的应用前景。那么，ai绘画工具midjourney怎么下载？本文将为您介绍Midjourney的下载以及作品的相关管理。一、Midjourney下载Midjourney的下载非常简单，只需打开Midjourney官网（点击“GetMidjour
干货|自我介绍这三个坑，99%的概率你踩过！夏麦生命的魔术师
自我介绍——每个人都需要的一张名片。图片源自网络从2018年到现在，在做演讲俱乐部的2年时间里，我在演讲活动现场听过1000+人的自我介绍，自我介绍做得超棒的人真不多！最近，我花了近几个月时间，仔细研究了500+人线上场景的自我介绍，发现优秀的自我介绍也不多！为什么做一张优秀的自我介绍就这么难呢？这个问题，在我帮几十个人打造了自我介绍的过程一直困扰着我。经过了几个月的时间思考与实践，终于发现三个—
Python编程 - 函数进阶易辰君 Python核心编程 python 开发语言
目录前言一、函数参数的高级用法（一）缺省参数（二）命名参数（三）不定长参数二、拆包（一）函数返回值拆包（二）通过星号拆包（三）总结三、匿名函数（一）函数定义（二）使用匿名函数四、递归函数（一）简介（二）基本结构（三）简单示例（四）优缺点总结前言上篇文章主要了解了函数基础，如何定义函数，函数种类以及局部变量和全局变量的差异等，接下来就讲解python函数较为进阶的知识点，若有任何想法欢迎一起沟通讨论
[实践应用] 深度学习之优化器 YuanDaima2048 深度学习工具使用 pytorch 深度学习人工智能机器学习 python 优化器
文章总览：YuanDaiMa2048博客文章总览深度学习之优化器1.随机梯度下降（SGD）2.动量优化（Momentum）3.自适应梯度（Adagrad）4.自适应矩估计（Adam）5.RMSprop总结其他介绍在深度学习中，优化器用于更新模型的参数，以最小化损失函数。常见的优化函数有很多种，下面是几种主流的优化器及其特点、原理和PyTorch实现：1.随机梯度下降（SGD）原理:随机梯度下降通过
wandb一直上传解决方案行业边缘的摸鱼怪 bug解决方案服务器 linux 服务器
问题描述运行带有wandb的代码时，虽然可以实现及时同步非常方便，但当设置错参数或其他原因不得不使用ctrl+C停止运行时，总会出现wandb一直上传个不停的现象，给在同一终端重新运行新的代码造成困难。解决方案运行以下代码把wandb的进程直接杀死。psaux|grepwandb|grep-vgrep|awk'{print$2}'|xargskill-9参考链接[CLI]:Ctrl+Ctokill
身体乳哪个牌子的好用?口碑超好的身体乳排行榜前十名高省APP珊珊
做小仙女真不容易啊~不仅脸蛋要漂亮精致，连身体每一寸肌肤都不能放过！天气变冷后，稍不注意身上的皮屑就开始“下雪”了~网上购物用高省APP更优惠！高省APP，是2021年推出的平台，0投资，0风险、高省APP佣金更高，模式更好，终端用户不流失。【高省】是一个自用省钱佣金高，分享推广赚钱多的平台，百度有几百万篇报道，也期待你的加入。珊珊导师，高省邀请码666123，注册送2皇冠会员，送万元推广大礼包，
vue render 函数详解 (配参数详解) 你的眼睛會笑 vue2 vue.js javascript 前端
vuerender函数详解(配参数详解)在Vue3中，`render`函数被用来代替Vue2中的模板语法。它接收一个h函数（或者是`createElement`函数的别名），并且返回一个虚拟DOM。render函数的语法结构如下：render(h){returnh('div',{class:'container'},'Hello,World!')}在上面的示例中，我们使用h函数创建了一个div元素
python老是报参数未定义_Python函数默认参数常见问题及解决方案 weixin_39935571 python老是报参数未定义
一、默认参数python为了简化函数的调用，提供了默认参数机制：这样在调用pow函数时，就可以省略最后一个参数不写：在定义有默认参数的函数时，需要注意以下：必选参数必须在前面，默认参数在后；设置何种参数为默认参数？一般来说，将参数值变化小的设置为默认参数。python标准库实践python内建函数：函数签名可以看出，使用print('hellopython')这样的简单调用的打印语句，实际上传入了
Enum用法不懂事的小屁孩 enum
以前的时候知道enum，但是真心不怎么用，在实际开发中，经常会用到以下代码: protected final static String XJ = "XJ"; protected final static String YHK = "YHK"; protected final static String PQ = "PQ";
【Spark九十七】RDD API之aggregateByKey bit1129 spark
1. aggregateByKey的运行机制 /** * Aggregate the values of each key, using given combine functions and a neutral "zero value". * This function can return a different result type
hive创建表是报错： Specified key was too long; max key length is 767 bytes daizj hive
今天在hive客户端创建表时报错，具体操作如下 hive> create table test2(id string); FAILED: Execution Error, return code 1 from org.apache.hadoop.hive.ql.exec.DDLTask. MetaException(message:javax.jdo.JDODataSto
Map 与 JavaBean之间的转换周凡杨 java 自省转换反射
最近项目里需要一个工具类，它的功能是传入一个Map后可以返回一个JavaBean对象。很喜欢写这样的Java服务，首先我想到的是要通过Java 的反射去实现匿名类的方法调用，这样才可以把Map里的值set 到JavaBean里。其实这里用Java的自省会更方便，下面两个方法就是一个通过反射，一个通过自省来实现本功能。 1：JavaBean类 1 &nb
java连接ftp下载 g21121 java
有的时候需要用到java连接ftp服务器下载，上传一些操作，下面写了一个小例子。 /** ftp服务器地址 */ private String ftpHost; /** ftp服务器用户名 */ private String ftpName; /** ftp服务器密码 */ private String ftpPass; /** ftp根目录 */ private String f
web报表工具FineReport使用中遇到的常见报错及解决办法（二）老A不折腾 finereport web报表 java报表总结
抛砖引玉，希望大家能把自己整理的问题及解决方法晾出来，Mark一下，利人利己。出现问题先搜一下文档上有没有，再看看度娘有没有，再看看论坛有没有。有报错要看日志。下面简单罗列下常见的问题，大多文档上都有提到的。 1、没有返回数据集：在存储过程中的操作语句之前加上set nocount on 或者在数据集exec调用存储过程的前面加上这句。当S
linux 系统cpu 内存等信息查看墙头上一根草 cpu 内存 liunx
1 查看CPU 　　1.1 查看CPU个数　　# cat /proc/cpuinfo | grep "physical id" | uniq | wc -l 　　2 　　**uniq命令：删除重复行;wc –l命令：统计行数** 　　1.2 查看CPU核数　　# cat /proc/cpuinfo | grep "cpu cores" | u
Spring中的AOP aijuans spring AOP
Spring中的AOP Written by Tony Jiang @ 2012-1-18 （转）何为AOP AOP，面向切面编程。在不改动代码的前提下，灵活的在现有代码的执行顺序前后，添加进新规机能。来一个简单的Sample: 目标类： [java] view plain copy print ? package&nb
placeholder(HTML 5) IE 兼容插件 alxw4616 JavaScript jquery jQuery插件
placeholder 这个属性被越来越频繁的使用. 但为做HTML 5 特性IE没能实现这东西. 以下的jQuery插件就是用来在IE上实现该属性的. /** * [placeholder(HTML 5) IE 实现.IE9以下通过测试.] * v 1.0 by oTwo 2014年7月31日 11:45:29 */ $.fn.placeholder = function
Object类,值域,泛型等总结(适合有基础的人看) 百合不是茶泛型的继承和通配符变量的值域 Object类转换
java的作用域在编程的时候经常会遇到,而我经常会搞不清楚这个问题,所以在家的这几天回忆一下过去不知道的每个小知识点变量的值域; package 基础; /** * 作用域的范围 * * @author Administrator * */ public class zuoyongyu { public static vo
JDK1.5 Condition接口 bijian1013 java thread Condition java多线程
Condition 将 Object 监视器方法（wait、notify和 notifyAll）分解成截然不同的对象，以便通过将这些对象与任意 Lock 实现组合使用，为每个对象提供多个等待 set （wait-set）。其中，Lock 替代了 synchronized 方法和语句的使用，Condition 替代了 Object 监视器方法的使用。条件（也称为条件队列或条件变量）为线程提供了一
开源中国OSC源创会记录 bijian1013 hadoop spark MemSQL
一.Strata+Hadoop World（SHW）大会是全世界最大的大数据大会之一。SHW大会为各种技术提供了深度交流的机会，还会看到最领先的大数据技术、最广泛的应用场景、最有趣的用例教学以及最全面的大数据行业和趋势探讨。二.Hadoop &nbs
【Java范型七】范型消除 bit1129 java
范型是Java1.5引入的语言特性，它是编译时的一个语法现象，也就是说，对于一个类，不管是范型类还是非范型类，编译得到的字节码是一样的，差别仅在于通过范型这种语法来进行编译时的类型检查，在运行时是没有范型或者类型参数这个说法的。范型跟反射刚好相反，反射是一种运行时行为，所以编译时不能访问的变量或者方法(比如private)，在运行时通过反射是可以访问的，也就是说，可见性也是一种编译时的行为，在
【Spark九十四】spark-sql工具的使用 bit1129 spark
spark-sql是Spark bin目录下的一个可执行脚本，它的目的是通过这个脚本执行Hive的命令，即原来通过 hive>输入的指令可以通过spark-sql>输入的指令来完成。 spark-sql可以使用内置的Hive metadata-store，也可以使用已经独立安装的Hive的metadata store 关于Hive build into Spark
js做的各种倒计时 ronin47 js 倒计时
第一种：精确到秒的javascript倒计时代码 HTML代码: <form name="form1"> <div align="center" align="middle"
java-37.有n 个长为m+1 的字符串，如果某个字符串的最后m 个字符与某个字符串的前m 个字符匹配，则两个字符串可以联接 bylijinnan java
public class MaxCatenate { /* * Q.37 有n 个长为m+1 的字符串，如果某个字符串的最后m 个字符与某个字符串的前m 个字符匹配，则两个字符串可以联接， * 问这n 个字符串最多可以连成一个多长的字符串，如果出现循环，则返回错误。 */ public static void main(String[] args){
mongoDB安装开窍的石头 mongodb安装基本操作
mongoDB的安装 1:mongoDB下载 https://www.mongodb.org/downloads 2:下载mongoDB下载后解压
[开源项目]引擎的关键意义 comsci 开源项目
一个系统，最核心的东西就是引擎。。。。。而要设计和制造出引擎，最关键的是要坚持。。。。。。现在最先进的引擎技术，也是从莱特兄弟那里出现的，但是中间一直没有断过研发的
软件度量的一些方法 cuiyadll 方法
软件度量的一些方法http://cuiyingfeng.blog.51cto.com/43841/6775/在前面我们已介绍了组成软件度量的几个方面。在这里我们将先给出关于这几个方面的一个纲要介绍。在后面我们还会作进一步具体的阐述。当我们不从高层次的概念级来看软件度量及其目标的时候，我们很容易把这些活动看成是不同而且毫不相干的。我们现在希望表明他们是怎样恰如其分地嵌入我们的框架的。也就是我们度量的
XSD中的targetNameSpace解释 darrenzhu xml namespace xsd targetnamespace
参考链接: http://blog.csdn.net/colin1014/article/details/357694 xsd文件中定义了一个targetNameSpace后，其内部定义的元素，属性，类型等都属于该targetNameSpace,其自身或外部xsd文件使用这些元素，属性等都必须从定义的targetNameSpace中找：例如：以下xsd文件，就出现了该错误，即便是在一
什么是RAID0、RAID1、RAID0+1、RAID5，等磁盘阵列模式? dcj3sjt126com raid
RAID 1又称为Mirror或Mirroring，它的宗旨是最大限度的保证用户数据的可用性和可修复性。 RAID 1的操作方式是把用户写入硬盘的数据百分之百地自动复制到另外一个硬盘上。由于对存储的数据进行百分之百的备份，在所有RAID级别中，RAID 1提供最高的数据安全保障。同样，由于数据的百分之百备份，备份数据占了总存储空间的一半，因而，Mirror的磁盘空间利用率低，存储成本高。 Mir
yii2 restful web服务快速入门 dcj3sjt126com PHP yii2
快速入门 Yii 提供了一整套用来简化实现 RESTful 风格的 Web Service 服务的 API。特别是，Yii 支持以下关于 RESTful 风格的 API：支持 Active Record 类的通用API的快速原型涉及的响应格式（在默认情况下支持 JSON 和 XML) 支持可选输出字段的定制对象序列化适当的格式的数据采集和验证错误
MongoDB查询(3)——内嵌文档查询（七） eksliang MongoDB查询内嵌文档 MongoDB查询内嵌数组
MongoDB查询内嵌文档转载请出自出处：http://eksliang.iteye.com/blog/2177301 一、概述有两种方法可以查询内嵌文档：查询整个文档；针对键值对进行查询。这两种方式是不同的，下面我通过例子进行分别说明。二、查询整个文档例如:有如下文档 db.emp.insert({ &qu
android4.4从系统图库无法加载图片的问题 gundumw100 android
典型的使用场景就是要设置一个头像，头像需要从系统图库或者拍照获得，在android4.4之前，我用的代码没问题，但是今天使用android4.4的时候突然发现不灵了。baidu了一圈，终于解决了。下面是解决方案： private String[] items = new String[] { "图库","拍照" }; /* 头像名称 */
网页特效大全 jQuery等 ini JavaScript jquery css html5 ini
HTML5和CSS3知识和特效 asp.net ajax jquery实例分享一个下雪的特效 jQuery倾斜的动画导航菜单选美大赛示例你会选谁 jQuery实现HTML5时钟功能强大的滚动播放插件JQ-Slide 万圣节快乐！！！向上弹出菜单jQuery插件 htm5视差动画 jquery将列表倒转顺序推荐一个jQuery分页插件 jquery animate
swift objc_setAssociatedObject block(version1.2 xcode6.4) 啸笑天 version
import UIKit class LSObjectWrapper: NSObject { let value: ((barButton: UIButton?) -> Void)? init(value: (barButton: UIButton?) -> Void) { self.value = value
Aegis 默认的 Xfire 绑定方式，将 XML 映射为 POJO MagicMa_007 java POJO xml Aegis xfire
Aegis 是一个默认的 Xfire 绑定方式，它将 XML 映射为 POJO, 支持代码先行的开发.你开发服务类与 POJO,它为你生成 XML schema/wsdl XML 和注解映射概览默认情况下，你的 POJO 类被是基于他们的名字与命名空间被序列化。如果
js get max value in (json) Array qiaolevip 每天进步一点点学习永无止境 max 纵观千象
// Max value in Array var arr = [1,2,3,5,3,2];Math.max.apply(null, arr); // 5 // Max value in Jaon Array var arr = [{"x":"8/11/2009","y":0.026572007},{"x"
XMLhttpRequest 请求 XML,JSON ,POJO 数据 Luob. POJO json Ajax xml XMLhttpREquest
在使用XMlhttpRequest对象发送请求和响应之前，必须首先使用javaScript对象创建一个XMLHttpRquest对象。 var xmlhttp； function getXMLHttpRequest(){ if(window.ActiveXObject){ xmlhttp:new ActiveXObject("Microsoft.XMLHTTP
jquery wuai jquery
以下防止文档在完全加载之前运行Jquery代码，否则会出现试图隐藏一个不存在的元素、获得未完全加载的图像的大小等等 $(document).ready(function(){ jquery代码; }); <script type="text/javascript" src="c:/scripts/jquery-1.4.2.min.js&quo