tf_course3 神经网络搭建八股

最新推荐文章于 2022-08-29 22:32:00 发布

wenxingxingxing

最新推荐文章于 2022-08-29 22:32:00 发布

阅读量320

点赞数

分类专栏： Deep learning TensorFlow2.0 文章标签：神经网络 tensorflow 深度学习

本文链接：https://blog.csdn.net/wenxingxingxing/article/details/126571566

版权

Deep learning 同时被 2 个专栏收录

12 篇文章 0 订阅

订阅专栏

TensorFlow2.0

6 篇文章 0 订阅

订阅专栏

神经网络搭建八股

第三讲神经网络搭建八股
1 tf.keras 搭建网络八股
2 iris 数据集代码复现
- 2.1 使用Sequential原始函数搭建八股
- 2.2 使用class搭建网络八股
3 MNIST 数据集
4 Fashion_mnist 数据集

第三讲神经网络搭建八股

1 tf.keras 搭建网络八股

1.1 keras 介绍

tf.keras 是 tensorflow2 引入的高封装度的框架，可以用于快速搭建神经网
络模型， keras 为支持快速实验而生，能够把想法迅速转换为结果，是深度学习
框架之中最终易上手的一个，它提供了一致而简洁的 API，能够极大地减少一般
应用下的工作量，提高代码地封装程度和复用性

Keras 官方文档：

https://tensorflow.google.cn/api_docs/python/tf

注意：由于tensorflow的API众多，因此初学时掌握常用的即可，仍然还有很多需要在今后的实践中继续学习，这时我们就需要参考
tensorflow 的官方文档，通过阅读源码和注释的方法学习 API。通常有两种方法，
以下将分别介绍：

第一种：在 pycharm 集成开发环境中查看框架源码
将鼠标放置在函数上按住 Ctrl 键，会显示函数的基本信息，包括
封装函数的类，函数入口参数，函数功能等。上图中显示的提示框就是显示出的
函数信息，第一行表示函数属于 Model 类，第二三四行列出了函数的参数，第五
行说明了函数的功能。可以看到， model.fit()的功能是执行训练过程，是本节
课搭建神经网络六部法中十分重要的一步，后面会进一步介绍。
按住 Ctrl 键点击函数会跳转到函数的源代码部分，使用者可以根据源码和
注释进一步了解函数的实现方法

第二种：在 tensorflow 官网中查询函数文档
在这里插入图片描述

上图是 tensorflow 官方文档的网站页面。查询时可以通过左边的检索寻找
目标函数

1.2 搭建神经网络六部法

tf.keras 搭建神经网络六部法

在这里插入图片描述

第一步：import 相关模块，如import tensorflow as tf

第二步：指定输入网络的训练集和测试集，如指定训练集的输入x_train 和标签 y_train，测试集的输入 x_test 和标签 y_test

第三步：逐层搭建网络结构，model = tf.keras.models.Sequential()

第四步：在 model.compile()中配置训练方法，选择训练时使用的优化器、损失
函数和最终评价指标。

第五步：在 model.fit()中执行训练过程，告知训练集和测试集的输入值和标签、
每个 batch 的大小（batchsize）和数据集的迭代次数（epoch）。

第六步：使用 model.summary()打印网络结构，统计参数数目

1.3 函数用法介绍——tf.keras.models.Sequential()

model = tf.keras.models.Sequential([网络结构])

Sequential 函数是一个容器，描述了神经网络的网络结构，在Sequential函数的输入参数中描述从输入层到输出层的网络结构

如：
**拉直层：**tf.keras.layers.Flatten()

拉直层可以变换张量的尺寸，把输入特征拉直为一维数组，是不含计算参数的层。

全连接层： tf.keras.layers.Dense(神经元个数，

                      activation = "激活函数"，
                      
                      keral_regularizer = "正则化方式")

其中：
activation（字符串给出）可选 relu、 softmax、 sigmoid、 tanh 等

kernel_regularizer 可选 tf.keras.regularizers.l1()、tf.keras.regularizers.l2()

**卷积层：**tf.keras.layers.Conv2D(filter = 卷积核个数，

                        kernel_size = 卷积核尺寸,
                        
                        strides = 卷积步长,
                        
                        padding = “valid” or “same”)

LSTM 层： tf.keras.layers.LSTM()。

本章只使用拉直层和全连接层，卷积层和循环神经网络层将在之后的章节介绍。

1.4 函数用法介绍——Model.compile（）

Model.compile( optimizer = 优化器,
loss = 损失函数,
metrics = [“准确率”])

Compile 用于配置神经网络的训练方法，告知训练时使用的优化器、损失函数和准确率评测标准。

其中:

1、optimizer可以是字符串形式给出的优化器名字，也可以是函数形式，使用函数形式可以设置学习率、动量和超参数

可选项包括：

- ‘sgd’or tf.optimizers.SGD( lr=学习率,decay=学习率衰减率,momentum=动量参数)

- ‘adagrad’or tf.keras.optimizers.Adagrad(lr=学习率,decay=学习率衰减率)

- ‘adadelta’or tf.keras.optimizers.Adadelta(lr=学习率,decay=学习率衰减率)

- ‘adam’or tf.keras.optimizers.Adam (lr=学习率,decay=学习率衰减率)

2、Loss 可以是字符串形式给出的损失函数的名字，也可以是函数形式。

可选项包括：

‘mse’or tf.keras.losses.MeanSquareError()

‘sparse_categorical_crossentropy’

or tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False)

损失函数常需要经过 softmax 等函数将输出转化为概率分布的形式。
from_logits 则用来标注该损失函数是否需要转换为概率的形式，取 False 时表
示转化为概率分布，取 True 时表示没有转化为概率分布，直接输出
在这里插入图片描述

1.5 函数用法介绍——Model.fit（）

model.fit 函数用于执行训练过程。

model.fit(训练集的输入特征，训练集的标签，

    batch_size, epochs,

    validation_data = (测试集的输入特征，测试集的标签)，
    
    validataion_split = 从测试集划分多少比例给训练集，
    
    validation_freq = 测试的 epoch 间隔次数)

1.6 函数用法介绍——Model.summary（）

summary 函数用于打印网络结构和参数统计

在这里插入图片描述

上图是 model.summary()对鸢尾花分类网络的网络结构和参数统计，对于一
个输入为 4 输出为 3 的全连接网络，共有 15 个参数

2 iris 数据集代码复现

2.1 使用Sequential原始函数搭建八股

# 第一步： import 相关模块：
import tensorflow as tf
import numpy as np
from sklearn import datasets
# 第二步：指定输入网络的训练集和测试集
x_train = datasets.load_iris().data
y_train = datasets.load_iris().target

'''
其中测试集的输入特征 x_test 和标签 y_test 可以像 x_train 和 y_train 一
样直接从数据集获取， 也可以如上述在 fit 中按比例从训练集中划分， 本例选择
从训练集中划分，所以只需加载 x_train， y_train 即可。
'''
# 实现了数据集的乱序
np.random.seed(116)
np.random.shuffle(x_train)
np.random.seed(116)
np.random.shuffle(y_train)
tf.random.set_seed(116)

#第三步： 逐层搭建网络结构：
model = tf.keras.models.Sequential([
        tf.keras.layers.Dense(3, activation = 'softmax',
        kernel_regularizer = tf.keras.regularizers.l2())
    
])

'''
    如上所示，本例使用了单层全连接网络，第一个参数表示神经元个数，第二个参数表示网络所使用的激活函数，第三个参数表示选用的正则化方法。
'''


# 使用 Sequential 可以快速搭建网络结构，但是如果网络包含跳连等其他复杂网络结构， Sequential 就无法表示了。 这就需要使用 class 来声明网络结构。


# 第四步： 在 model.compile()中配置训练方法
model.compile(optimizer= tf.keras.optimizers.SGD(lr = 0.1),
             loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits = False),
             metrics=['sparse_categorical_accuracy'])

'''
    如 上 所 示 ， 本 例 使 用 SGD 优 化 器 ， 并 将 学 习 率 设 置 为 0.1 ， 选 择
    SparseCategoricalCrossentrop 作为损失函数。 由于神经网络输出使用了
    softmax 激活函数，使得输出是概率分布，而不是原始输出， 所以需要将
    from_logits 参数设置为 False。 鸢尾花数据集给的标签是 0， 1， 2 这样的数值，
    而 网 络 前 向 传 播 的 输 出 为 概 率 分 布 ， 所 以 metrics 需 要 设 置 为
    sparse_categorical_accuracy。
'''

# 第五步： 在 model.fit()中执行训练过程
'''
    在 fit 中执行训练过程,x_train,y_train 分别表示网络的输入特征和标签，
    batch_size 表示一次喂入神经网络的数据量， epochs 表示数据集的迭代次数
    validation_split 表示数据集中测试集的划分比例， validation_freq 表示每迭
    代 20 次在测试集上测试一次准确率。
'''
model.fit(x_train, y_train,batch_size=32, epochs=500, validation_split=0.2, validation_freq=20)

# 第六步： 使用 model.summary()打印网络结构，统计参数数目：
model.summary()

Epoch 1/500
4/4 [==============================] - 0s 998us/step - loss: 2.3173 - sparse_categorical_accuracy: 0.3417
Epoch 2/500
4/4 [==============================] - 0s 998us/step - loss: 1.0797 - sparse_categorical_accuracy: 0.6167
Epoch 3/500
4/4 [==============================] - 0s 997us/step - loss: 0.8429 - sparse_categorical_accuracy: 0.6250
Epoch 4/500
4/4 [==============================] - 0s 665us/step - loss: 0.7448 - sparse_categorical_accuracy: 0.6667
Epoch 5/500
4/4 [==============================] - 0s 665us/step - loss: 0.9843 - sparse_categorical_accuracy: 0.6417
Epoch 6/500
4/4 [==============================] - 0s 997us/step - loss: 0.7808 - sparse_categorical_accuracy: 0.6583
Epoch 7/500
4/4 [==============================] - 0s 664us/step - loss: 0.5961 - sparse_categorical_accuracy: 0.7250
Epoch 8/500
4/4 [==============================] - 0s 665us/step - loss: 0.5824 - sparse_categorical_accuracy: 0.7250
Epoch 9/500
4/4 [==============================] - 0s 665us/step - loss: 0.6016 - sparse_categorical_accuracy: 0.7000
Epoch 10/500
4/4 [==============================] - 0s 998us/step - loss: 0.6707 - sparse_categorical_accuracy: 0.6500
Epoch 11/500
4/4 [==============================] - 0s 665us/step - loss: 0.7967 - sparse_categorical_accuracy: 0.6667
Epoch 12/500
4/4 [==============================] - 0s 997us/step - loss: 0.5385 - sparse_categorical_accuracy: 0.7500
Epoch 13/500
4/4 [==============================] - 0s 664us/step - loss: 1.0675 - sparse_categorical_accuracy: 0.5833
Epoch 14/500
4/4 [==============================] - 0s 998us/step - loss: 0.5119 - sparse_categorical_accuracy: 0.7333
Epoch 15/500
4/4 [==============================] - 0s 665us/step - loss: 0.6831 - sparse_categorical_accuracy: 0.6667
Epoch 16/500
4/4 [==============================] - 0s 997us/step - loss: 0.5969 - sparse_categorical_accuracy: 0.7667
Epoch 17/500
4/4 [==============================] - 0s 665us/step - loss: 0.5559 - sparse_categorical_accuracy: 0.7250
Epoch 18/500
4/4 [==============================] - 0s 997us/step - loss: 0.6521 - sparse_categorical_accuracy: 0.7667
Epoch 19/500
4/4 [==============================] - 0s 665us/step - loss: 0.5902 - sparse_categorical_accuracy: 0.6833
Epoch 20/500
4/4 [==============================] - 0s 40ms/step - loss: 0.4782 - sparse_categorical_accuracy: 0.8000 - val_loss: 0.7740 - val_sparse_categorical_accuracy: 0.5333
Epoch 21/500
4/4 [==============================] - 0s 665us/step - loss: 0.4589 - sparse_categorical_accuracy: 0.8333
Epoch 22/500
4/4 [==============================] - 0s 665us/step - loss: 0.6620 - sparse_categorical_accuracy: 0.7167
Epoch 23/500
4/4 [==============================] - 0s 997us/step - loss: 0.4497 - sparse_categorical_accuracy: 0.8583
Epoch 24/500
4/4 [==============================] - 0s 997us/step - loss: 0.7418 - sparse_categorical_accuracy: 0.6917
Epoch 25/500
4/4 [==============================] - 0s 665us/step - loss: 0.4873 - sparse_categorical_accuracy: 0.7333
Epoch 26/500
4/4 [==============================] - 0s 665us/step - loss: 0.7408 - sparse_categorical_accuracy: 0.7083
Epoch 27/500
4/4 [==============================] - 0s 998us/step - loss: 0.6502 - sparse_categorical_accuracy: 0.7750
Epoch 28/500
4/4 [==============================] - 0s 665us/step - loss: 0.6749 - sparse_categorical_accuracy: 0.6667
Epoch 29/500
4/4 [==============================] - 0s 665us/step - loss: 0.5382 - sparse_categorical_accuracy: 0.7833
Epoch 30/500
4/4 [==============================] - 0s 666us/step - loss: 0.6266 - sparse_categorical_accuracy: 0.7000
Epoch 31/500
4/4 [==============================] - 0s 997us/step - loss: 0.5041 - sparse_categorical_accuracy: 0.7917
Epoch 32/500
4/4 [==============================] - 0s 665us/step - loss: 0.6933 - sparse_categorical_accuracy: 0.7250
Epoch 33/500
4/4 [==============================] - 0s 665us/step - loss: 0.5518 - sparse_categorical_accuracy: 0.7667
Epoch 34/500
4/4 [==============================] - 0s 997us/step - loss: 0.4097 - sparse_categorical_accuracy: 0.9000
Epoch 35/500
4/4 [==============================] - 0s 665us/step - loss: 0.4705 - sparse_categorical_accuracy: 0.7917
Epoch 36/500
4/4 [==============================] - 0s 997us/step - loss: 0.4490 - sparse_categorical_accuracy: 0.8583
Epoch 37/500
4/4 [==============================] - 0s 997us/step - loss: 0.4172 - sparse_categorical_accuracy: 0.9333
Epoch 38/500
4/4 [==============================] - 0s 997us/step - loss: 0.6989 - sparse_categorical_accuracy: 0.6583
Epoch 39/500
4/4 [==============================] - 0s 665us/step - loss: 0.5124 - sparse_categorical_accuracy: 0.7833
Epoch 40/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3984 - sparse_categorical_accuracy: 0.9333 - val_loss: 0.5433 - val_sparse_categorical_accuracy: 0.6000
Epoch 41/500
4/4 [==============================] - 0s 997us/step - loss: 0.5434 - sparse_categorical_accuracy: 0.7333
Epoch 42/500
4/4 [==============================] - 0s 997us/step - loss: 0.6159 - sparse_categorical_accuracy: 0.7000
Epoch 43/500
4/4 [==============================] - 0s 665us/step - loss: 0.6999 - sparse_categorical_accuracy: 0.7167
Epoch 44/500
4/4 [==============================] - 0s 665us/step - loss: 0.4048 - sparse_categorical_accuracy: 0.9333
Epoch 45/500
4/4 [==============================] - 0s 998us/step - loss: 0.3924 - sparse_categorical_accuracy: 0.9500
Epoch 46/500
4/4 [==============================] - 0s 997us/step - loss: 0.3944 - sparse_categorical_accuracy: 0.9333
Epoch 47/500
4/4 [==============================] - 0s 665us/step - loss: 0.4063 - sparse_categorical_accuracy: 0.9083
Epoch 48/500
4/4 [==============================] - 0s 665us/step - loss: 0.5635 - sparse_categorical_accuracy: 0.7583
Epoch 49/500
4/4 [==============================] - 0s 997us/step - loss: 0.4108 - sparse_categorical_accuracy: 0.9333
Epoch 50/500
4/4 [==============================] - 0s 665us/step - loss: 0.3884 - sparse_categorical_accuracy: 0.9583
Epoch 51/500
4/4 [==============================] - 0s 998us/step - loss: 0.3835 - sparse_categorical_accuracy: 0.9417
Epoch 52/500
4/4 [==============================] - 0s 665us/step - loss: 0.4042 - sparse_categorical_accuracy: 0.9083
Epoch 53/500
4/4 [==============================] - 0s 997us/step - loss: 0.4400 - sparse_categorical_accuracy: 0.8500
Epoch 54/500
4/4 [==============================] - 0s 997us/step - loss: 0.7052 - sparse_categorical_accuracy: 0.6417
Epoch 55/500
4/4 [==============================] - 0s 664us/step - loss: 0.4013 - sparse_categorical_accuracy: 0.8833
Epoch 56/500
4/4 [==============================] - 0s 665us/step - loss: 0.3983 - sparse_categorical_accuracy: 0.8917
Epoch 57/500
4/4 [==============================] - 0s 997us/step - loss: 0.5127 - sparse_categorical_accuracy: 0.8000
Epoch 58/500
4/4 [==============================] - 0s 665us/step - loss: 0.5093 - sparse_categorical_accuracy: 0.8000
Epoch 59/500
4/4 [==============================] - 0s 665us/step - loss: 0.4715 - sparse_categorical_accuracy: 0.8083
Epoch 60/500
4/4 [==============================] - 0s 6ms/step - loss: 0.5407 - sparse_categorical_accuracy: 0.9000 - val_loss: 0.4260 - val_sparse_categorical_accuracy: 0.9333
Epoch 61/500
4/4 [==============================] - 0s 997us/step - loss: 0.4227 - sparse_categorical_accuracy: 0.8833
Epoch 62/500
4/4 [==============================] - 0s 997us/step - loss: 0.4420 - sparse_categorical_accuracy: 0.8583
Epoch 63/500
4/4 [==============================] - 0s 997us/step - loss: 0.6500 - sparse_categorical_accuracy: 0.7167
Epoch 64/500
4/4 [==============================] - 0s 665us/step - loss: 0.4293 - sparse_categorical_accuracy: 0.8667
Epoch 65/500
4/4 [==============================] - 0s 2ms/step - loss: 0.3863 - sparse_categorical_accuracy: 0.9167
Epoch 66/500
4/4 [==============================] - 0s 665us/step - loss: 0.3802 - sparse_categorical_accuracy: 0.9250
Epoch 67/500
4/4 [==============================] - 0s 988us/step - loss: 0.4052 - sparse_categorical_accuracy: 0.8667
Epoch 68/500
4/4 [==============================] - 0s 665us/step - loss: 0.5520 - sparse_categorical_accuracy: 0.7250
Epoch 69/500
4/4 [==============================] - 0s 997us/step - loss: 0.4788 - sparse_categorical_accuracy: 0.8167
Epoch 70/500
4/4 [==============================] - 0s 665us/step - loss: 0.3761 - sparse_categorical_accuracy: 0.9333
Epoch 71/500
4/4 [==============================] - 0s 666us/step - loss: 0.4550 - sparse_categorical_accuracy: 0.8250
Epoch 72/500
4/4 [==============================] - 0s 997us/step - loss: 0.4004 - sparse_categorical_accuracy: 0.9250
Epoch 73/500
4/4 [==============================] - 0s 665us/step - loss: 0.5146 - sparse_categorical_accuracy: 0.8167
Epoch 74/500
4/4 [==============================] - 0s 665us/step - loss: 0.3846 - sparse_categorical_accuracy: 0.9000
Epoch 75/500
4/4 [==============================] - 0s 665us/step - loss: 0.3710 - sparse_categorical_accuracy: 0.9417
Epoch 76/500
4/4 [==============================] - 0s 665us/step - loss: 0.3799 - sparse_categorical_accuracy: 0.9250
Epoch 77/500
4/4 [==============================] - 0s 665us/step - loss: 0.4709 - sparse_categorical_accuracy: 0.8000
Epoch 78/500
4/4 [==============================] - 0s 665us/step - loss: 0.7460 - sparse_categorical_accuracy: 0.6500
Epoch 79/500
4/4 [==============================] - 0s 665us/step - loss: 0.4508 - sparse_categorical_accuracy: 0.8167
Epoch 80/500
4/4 [==============================] - 0s 5ms/step - loss: 0.4219 - sparse_categorical_accuracy: 0.9000 - val_loss: 0.3497 - val_sparse_categorical_accuracy: 0.9667
Epoch 81/500
4/4 [==============================] - 0s 997us/step - loss: 0.3777 - sparse_categorical_accuracy: 0.9333
Epoch 82/500
4/4 [==============================] - 0s 665us/step - loss: 0.3703 - sparse_categorical_accuracy: 0.9583
Epoch 83/500
4/4 [==============================] - 0s 997us/step - loss: 0.3600 - sparse_categorical_accuracy: 0.9667
Epoch 84/500
4/4 [==============================] - 0s 997us/step - loss: 0.3638 - sparse_categorical_accuracy: 0.9333
Epoch 85/500
4/4 [==============================] - 0s 665us/step - loss: 0.3707 - sparse_categorical_accuracy: 0.9583
Epoch 86/500
4/4 [==============================] - 0s 665us/step - loss: 0.3845 - sparse_categorical_accuracy: 0.9250
Epoch 87/500
4/4 [==============================] - 0s 665us/step - loss: 0.3888 - sparse_categorical_accuracy: 0.8917
Epoch 88/500
4/4 [==============================] - 0s 997us/step - loss: 0.3771 - sparse_categorical_accuracy: 0.9500
Epoch 89/500
4/4 [==============================] - 0s 665us/step - loss: 0.5318 - sparse_categorical_accuracy: 0.7750
Epoch 90/500
4/4 [==============================] - 0s 665us/step - loss: 0.3822 - sparse_categorical_accuracy: 0.9250
Epoch 91/500
4/4 [==============================] - 0s 997us/step - loss: 0.4168 - sparse_categorical_accuracy: 0.9000
Epoch 92/500
4/4 [==============================] - 0s 665us/step - loss: 0.5004 - sparse_categorical_accuracy: 0.8000
Epoch 93/500
4/4 [==============================] - 0s 665us/step - loss: 0.3762 - sparse_categorical_accuracy: 0.9167
Epoch 94/500
4/4 [==============================] - 0s 997us/step - loss: 0.3885 - sparse_categorical_accuracy: 0.9333
Epoch 95/500
4/4 [==============================] - 0s 665us/step - loss: 0.3859 - sparse_categorical_accuracy: 0.8917
Epoch 96/500
4/4 [==============================] - 0s 997us/step - loss: 0.5500 - sparse_categorical_accuracy: 0.7667
Epoch 97/500
4/4 [==============================] - 0s 998us/step - loss: 0.4690 - sparse_categorical_accuracy: 0.8500
Epoch 98/500
4/4 [==============================] - 0s 665us/step - loss: 0.4850 - sparse_categorical_accuracy: 0.8083
Epoch 99/500
4/4 [==============================] - 0s 665us/step - loss: 0.3836 - sparse_categorical_accuracy: 0.9333
Epoch 100/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3713 - sparse_categorical_accuracy: 0.9167 - val_loss: 0.4250 - val_sparse_categorical_accuracy: 0.9333
Epoch 101/500
4/4 [==============================] - 0s 665us/step - loss: 0.3641 - sparse_categorical_accuracy: 0.9500
Epoch 102/500
4/4 [==============================] - 0s 997us/step - loss: 0.3650 - sparse_categorical_accuracy: 0.9583
Epoch 103/500
4/4 [==============================] - 0s 997us/step - loss: 0.3610 - sparse_categorical_accuracy: 0.9500
Epoch 104/500
4/4 [==============================] - 0s 665us/step - loss: 0.3771 - sparse_categorical_accuracy: 0.8917
Epoch 105/500
4/4 [==============================] - 0s 665us/step - loss: 0.4734 - sparse_categorical_accuracy: 0.8250
Epoch 106/500
4/4 [==============================] - 0s 665us/step - loss: 0.4102 - sparse_categorical_accuracy: 0.8917
Epoch 107/500
4/4 [==============================] - 0s 665us/step - loss: 0.3745 - sparse_categorical_accuracy: 0.9333
Epoch 108/500
4/4 [==============================] - 0s 665us/step - loss: 0.3834 - sparse_categorical_accuracy: 0.9000
Epoch 109/500
4/4 [==============================] - 0s 665us/step - loss: 0.3621 - sparse_categorical_accuracy: 0.9417
Epoch 110/500
4/4 [==============================] - 0s 998us/step - loss: 0.3798 - sparse_categorical_accuracy: 0.9167
Epoch 111/500
4/4 [==============================] - 0s 998us/step - loss: 0.3572 - sparse_categorical_accuracy: 0.9750
Epoch 112/500
4/4 [==============================] - 0s 665us/step - loss: 0.3665 - sparse_categorical_accuracy: 0.9333
Epoch 113/500
4/4 [==============================] - 0s 997us/step - loss: 0.4224 - sparse_categorical_accuracy: 0.8833
Epoch 114/500
4/4 [==============================] - 0s 665us/step - loss: 0.3616 - sparse_categorical_accuracy: 0.9500
Epoch 115/500
4/4 [==============================] - 0s 665us/step - loss: 0.3762 - sparse_categorical_accuracy: 0.9250
Epoch 116/500
4/4 [==============================] - 0s 997us/step - loss: 0.3832 - sparse_categorical_accuracy: 0.8833
Epoch 117/500
4/4 [==============================] - 0s 997us/step - loss: 0.5093 - sparse_categorical_accuracy: 0.8167
Epoch 118/500
4/4 [==============================] - 0s 997us/step - loss: 0.3643 - sparse_categorical_accuracy: 0.9333
Epoch 119/500
4/4 [==============================] - 0s 998us/step - loss: 0.3593 - sparse_categorical_accuracy: 0.9417
Epoch 120/500
4/4 [==============================] - 0s 6ms/step - loss: 0.4207 - sparse_categorical_accuracy: 0.8833 - val_loss: 0.3512 - val_sparse_categorical_accuracy: 1.0000
Epoch 121/500
4/4 [==============================] - 0s 997us/step - loss: 0.4096 - sparse_categorical_accuracy: 0.9250
Epoch 122/500
4/4 [==============================] - 0s 665us/step - loss: 0.4189 - sparse_categorical_accuracy: 0.8833
Epoch 123/500
4/4 [==============================] - 0s 665us/step - loss: 0.3683 - sparse_categorical_accuracy: 0.9250
Epoch 124/500
4/4 [==============================] - 0s 665us/step - loss: 0.3650 - sparse_categorical_accuracy: 0.9167
Epoch 125/500
4/4 [==============================] - 0s 665us/step - loss: 0.3710 - sparse_categorical_accuracy: 0.9417
Epoch 126/500
4/4 [==============================] - 0s 665us/step - loss: 0.3914 - sparse_categorical_accuracy: 0.9000
Epoch 127/500
4/4 [==============================] - 0s 998us/step - loss: 0.3756 - sparse_categorical_accuracy: 0.9333
Epoch 128/500
4/4 [==============================] - 0s 665us/step - loss: 0.3782 - sparse_categorical_accuracy: 0.9083
Epoch 129/500
4/4 [==============================] - 0s 665us/step - loss: 0.3818 - sparse_categorical_accuracy: 0.9167
Epoch 130/500
4/4 [==============================] - 0s 997us/step - loss: 0.4638 - sparse_categorical_accuracy: 0.8500
Epoch 131/500
4/4 [==============================] - 0s 997us/step - loss: 0.4299 - sparse_categorical_accuracy: 0.8667
Epoch 132/500
4/4 [==============================] - 0s 665us/step - loss: 0.3521 - sparse_categorical_accuracy: 0.9750
Epoch 133/500
4/4 [==============================] - 0s 664us/step - loss: 0.3520 - sparse_categorical_accuracy: 0.9667
Epoch 134/500
4/4 [==============================] - 0s 997us/step - loss: 0.3700 - sparse_categorical_accuracy: 0.9333
Epoch 135/500
4/4 [==============================] - 0s 665us/step - loss: 0.3735 - sparse_categorical_accuracy: 0.9333
Epoch 136/500
4/4 [==============================] - 0s 997us/step - loss: 0.3584 - sparse_categorical_accuracy: 0.9333
Epoch 137/500
4/4 [==============================] - 0s 665us/step - loss: 0.4001 - sparse_categorical_accuracy: 0.9083
Epoch 138/500
4/4 [==============================] - 0s 665us/step - loss: 0.3530 - sparse_categorical_accuracy: 0.9583
Epoch 139/500
4/4 [==============================] - 0s 997us/step - loss: 0.4089 - sparse_categorical_accuracy: 0.9000
Epoch 140/500
4/4 [==============================] - 0s 6ms/step - loss: 0.5121 - sparse_categorical_accuracy: 0.7750 - val_loss: 0.3758 - val_sparse_categorical_accuracy: 1.0000
Epoch 141/500
4/4 [==============================] - 0s 665us/step - loss: 0.3577 - sparse_categorical_accuracy: 0.9500
Epoch 142/500
4/4 [==============================] - 0s 665us/step - loss: 0.3597 - sparse_categorical_accuracy: 0.9583
Epoch 143/500
4/4 [==============================] - 0s 997us/step - loss: 0.3505 - sparse_categorical_accuracy: 0.9417
Epoch 144/500
4/4 [==============================] - 0s 665us/step - loss: 0.3728 - sparse_categorical_accuracy: 0.9250
Epoch 145/500
4/4 [==============================] - 0s 997us/step - loss: 0.3780 - sparse_categorical_accuracy: 0.9083
Epoch 146/500
4/4 [==============================] - 0s 997us/step - loss: 0.3543 - sparse_categorical_accuracy: 0.9583
Epoch 147/500
4/4 [==============================] - 0s 665us/step - loss: 0.4474 - sparse_categorical_accuracy: 0.8583
Epoch 148/500
4/4 [==============================] - 0s 997us/step - loss: 0.4703 - sparse_categorical_accuracy: 0.8500
Epoch 149/500
4/4 [==============================] - 0s 665us/step - loss: 0.3700 - sparse_categorical_accuracy: 0.9083
Epoch 150/500
4/4 [==============================] - 0s 665us/step - loss: 0.4425 - sparse_categorical_accuracy: 0.8417
Epoch 151/500
4/4 [==============================] - 0s 997us/step - loss: 0.4274 - sparse_categorical_accuracy: 0.8417
Epoch 152/500
4/4 [==============================] - 0s 665us/step - loss: 0.3724 - sparse_categorical_accuracy: 0.9000
Epoch 153/500
4/4 [==============================] - 0s 665us/step - loss: 0.4161 - sparse_categorical_accuracy: 0.8750
Epoch 154/500
4/4 [==============================] - 0s 665us/step - loss: 0.3902 - sparse_categorical_accuracy: 0.9083
Epoch 155/500
4/4 [==============================] - 0s 998us/step - loss: 0.3843 - sparse_categorical_accuracy: 0.8917
Epoch 156/500
4/4 [==============================] - 0s 665us/step - loss: 0.3883 - sparse_categorical_accuracy: 0.9000
Epoch 157/500
4/4 [==============================] - 0s 665us/step - loss: 0.3966 - sparse_categorical_accuracy: 0.8667
Epoch 158/500
4/4 [==============================] - 0s 665us/step - loss: 0.3617 - sparse_categorical_accuracy: 0.9417
Epoch 159/500
4/4 [==============================] - 0s 997us/step - loss: 0.4144 - sparse_categorical_accuracy: 0.8833
Epoch 160/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3978 - sparse_categorical_accuracy: 0.9333 - val_loss: 0.4238 - val_sparse_categorical_accuracy: 0.9333
Epoch 161/500
4/4 [==============================] - 0s 998us/step - loss: 0.3620 - sparse_categorical_accuracy: 0.9417
Epoch 162/500
4/4 [==============================] - 0s 665us/step - loss: 0.3863 - sparse_categorical_accuracy: 0.9083
Epoch 163/500
4/4 [==============================] - 0s 997us/step - loss: 0.3635 - sparse_categorical_accuracy: 0.9250
Epoch 164/500
4/4 [==============================] - 0s 665us/step - loss: 0.6862 - sparse_categorical_accuracy: 0.7500
Epoch 165/500
4/4 [==============================] - 0s 997us/step - loss: 0.4336 - sparse_categorical_accuracy: 0.8833
Epoch 166/500
4/4 [==============================] - 0s 997us/step - loss: 0.4403 - sparse_categorical_accuracy: 0.8583
Epoch 167/500
4/4 [==============================] - 0s 997us/step - loss: 0.3904 - sparse_categorical_accuracy: 0.9083
Epoch 168/500
4/4 [==============================] - 0s 665us/step - loss: 0.3524 - sparse_categorical_accuracy: 0.9583
Epoch 169/500
4/4 [==============================] - 0s 665us/step - loss: 0.3602 - sparse_categorical_accuracy: 0.9667
Epoch 170/500
4/4 [==============================] - 0s 665us/step - loss: 0.3542 - sparse_categorical_accuracy: 0.9583
Epoch 171/500
4/4 [==============================] - 0s 665us/step - loss: 0.4699 - sparse_categorical_accuracy: 0.8083
Epoch 172/500
4/4 [==============================] - 0s 665us/step - loss: 0.3456 - sparse_categorical_accuracy: 0.9250
Epoch 173/500
4/4 [==============================] - 0s 997us/step - loss: 0.3528 - sparse_categorical_accuracy: 0.9583
Epoch 174/500
4/4 [==============================] - 0s 997us/step - loss: 0.3801 - sparse_categorical_accuracy: 0.9250
Epoch 175/500
4/4 [==============================] - 0s 997us/step - loss: 0.3921 - sparse_categorical_accuracy: 0.9000
Epoch 176/500
4/4 [==============================] - 0s 665us/step - loss: 0.3875 - sparse_categorical_accuracy: 0.9083
Epoch 177/500
4/4 [==============================] - 0s 665us/step - loss: 0.4096 - sparse_categorical_accuracy: 0.9083
Epoch 178/500
4/4 [==============================] - 0s 665us/step - loss: 0.4523 - sparse_categorical_accuracy: 0.8583
Epoch 179/500
4/4 [==============================] - 0s 665us/step - loss: 0.3720 - sparse_categorical_accuracy: 0.9167
Epoch 180/500
4/4 [==============================] - 0s 6ms/step - loss: 0.4053 - sparse_categorical_accuracy: 0.9000 - val_loss: 0.3892 - val_sparse_categorical_accuracy: 1.0000
Epoch 181/500
4/4 [==============================] - 0s 998us/step - loss: 0.3602 - sparse_categorical_accuracy: 0.9417
Epoch 182/500
4/4 [==============================] - 0s 997us/step - loss: 0.4059 - sparse_categorical_accuracy: 0.8917
Epoch 183/500
4/4 [==============================] - 0s 997us/step - loss: 0.3556 - sparse_categorical_accuracy: 0.9333
Epoch 184/500
4/4 [==============================] - 0s 665us/step - loss: 0.4176 - sparse_categorical_accuracy: 0.8833
Epoch 185/500
4/4 [==============================] - 0s 665us/step - loss: 0.3594 - sparse_categorical_accuracy: 0.9500
Epoch 186/500
4/4 [==============================] - 0s 665us/step - loss: 0.3632 - sparse_categorical_accuracy: 0.9417
Epoch 187/500
4/4 [==============================] - 0s 665us/step - loss: 0.4102 - sparse_categorical_accuracy: 0.9083
Epoch 188/500
4/4 [==============================] - 0s 665us/step - loss: 0.3549 - sparse_categorical_accuracy: 0.9500
Epoch 189/500
4/4 [==============================] - 0s 665us/step - loss: 0.3533 - sparse_categorical_accuracy: 0.9583
Epoch 190/500
4/4 [==============================] - 0s 666us/step - loss: 0.3974 - sparse_categorical_accuracy: 0.8917
Epoch 191/500
4/4 [==============================] - 0s 665us/step - loss: 0.3645 - sparse_categorical_accuracy: 0.9333
Epoch 192/500
4/4 [==============================] - 0s 665us/step - loss: 0.3784 - sparse_categorical_accuracy: 0.9333
Epoch 193/500
4/4 [==============================] - 0s 998us/step - loss: 0.4510 - sparse_categorical_accuracy: 0.8417
Epoch 194/500
4/4 [==============================] - 0s 665us/step - loss: 0.3579 - sparse_categorical_accuracy: 0.9583
Epoch 195/500
4/4 [==============================] - 0s 665us/step - loss: 0.3572 - sparse_categorical_accuracy: 0.9583
Epoch 196/500
4/4 [==============================] - 0s 997us/step - loss: 0.3782 - sparse_categorical_accuracy: 0.9333
Epoch 197/500
4/4 [==============================] - 0s 997us/step - loss: 0.3525 - sparse_categorical_accuracy: 0.9583
Epoch 198/500
4/4 [==============================] - 0s 665us/step - loss: 0.3499 - sparse_categorical_accuracy: 0.9583
Epoch 199/500
4/4 [==============================] - 0s 665us/step - loss: 0.3540 - sparse_categorical_accuracy: 0.9417
Epoch 200/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3628 - sparse_categorical_accuracy: 0.9333 - val_loss: 0.4160 - val_sparse_categorical_accuracy: 0.9333
Epoch 201/500
4/4 [==============================] - 0s 665us/step - loss: 0.3575 - sparse_categorical_accuracy: 0.9417
Epoch 202/500
4/4 [==============================] - 0s 665us/step - loss: 0.4148 - sparse_categorical_accuracy: 0.8750
Epoch 203/500
4/4 [==============================] - 0s 665us/step - loss: 0.3930 - sparse_categorical_accuracy: 0.9000
Epoch 204/500
4/4 [==============================] - 0s 665us/step - loss: 0.3465 - sparse_categorical_accuracy: 0.9750
Epoch 205/500
4/4 [==============================] - 0s 665us/step - loss: 0.3928 - sparse_categorical_accuracy: 0.9000
Epoch 206/500
4/4 [==============================] - 0s 997us/step - loss: 0.4302 - sparse_categorical_accuracy: 0.8417
Epoch 207/500
4/4 [==============================] - 0s 665us/step - loss: 0.5203 - sparse_categorical_accuracy: 0.7750
Epoch 208/500
4/4 [==============================] - 0s 665us/step - loss: 0.3765 - sparse_categorical_accuracy: 0.9250
Epoch 209/500
4/4 [==============================] - 0s 998us/step - loss: 0.3803 - sparse_categorical_accuracy: 0.9167
Epoch 210/500
4/4 [==============================] - 0s 997us/step - loss: 0.3826 - sparse_categorical_accuracy: 0.9167
Epoch 211/500
4/4 [==============================] - 0s 997us/step - loss: 0.4507 - sparse_categorical_accuracy: 0.8250
Epoch 212/500
4/4 [==============================] - 0s 665us/step - loss: 0.3604 - sparse_categorical_accuracy: 0.9000
Epoch 213/500
4/4 [==============================] - 0s 665us/step - loss: 0.3522 - sparse_categorical_accuracy: 0.9333
Epoch 214/500
4/4 [==============================] - 0s 665us/step - loss: 0.3819 - sparse_categorical_accuracy: 0.9250
Epoch 215/500
4/4 [==============================] - 0s 997us/step - loss: 0.3474 - sparse_categorical_accuracy: 0.9667
Epoch 216/500
4/4 [==============================] - 0s 665us/step - loss: 0.3784 - sparse_categorical_accuracy: 0.9000
Epoch 217/500
4/4 [==============================] - 0s 665us/step - loss: 0.3553 - sparse_categorical_accuracy: 0.9750
Epoch 218/500
4/4 [==============================] - 0s 998us/step - loss: 0.3507 - sparse_categorical_accuracy: 0.9583
Epoch 219/500
4/4 [==============================] - 0s 665us/step - loss: 0.3568 - sparse_categorical_accuracy: 0.9417
Epoch 220/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3399 - sparse_categorical_accuracy: 0.9583 - val_loss: 0.3539 - val_sparse_categorical_accuracy: 0.8667
Epoch 221/500
4/4 [==============================] - 0s 665us/step - loss: 0.4177 - sparse_categorical_accuracy: 0.8917
Epoch 222/500
4/4 [==============================] - 0s 997us/step - loss: 0.4743 - sparse_categorical_accuracy: 0.8500
Epoch 223/500
4/4 [==============================] - 0s 665us/step - loss: 0.3757 - sparse_categorical_accuracy: 0.9333
Epoch 224/500
4/4 [==============================] - 0s 665us/step - loss: 0.3509 - sparse_categorical_accuracy: 0.9333
Epoch 225/500
4/4 [==============================] - 0s 997us/step - loss: 0.4814 - sparse_categorical_accuracy: 0.8250
Epoch 226/500
4/4 [==============================] - 0s 997us/step - loss: 0.3776 - sparse_categorical_accuracy: 0.9083
Epoch 227/500
4/4 [==============================] - 0s 665us/step - loss: 0.3909 - sparse_categorical_accuracy: 0.9250
Epoch 228/500
4/4 [==============================] - 0s 665us/step - loss: 0.3548 - sparse_categorical_accuracy: 0.9583
Epoch 229/500
4/4 [==============================] - 0s 665us/step - loss: 0.3536 - sparse_categorical_accuracy: 0.9417
Epoch 230/500
4/4 [==============================] - 0s 664us/step - loss: 0.3603 - sparse_categorical_accuracy: 0.9500
Epoch 231/500
4/4 [==============================] - 0s 665us/step - loss: 0.3909 - sparse_categorical_accuracy: 0.9083
Epoch 232/500
4/4 [==============================] - 0s 665us/step - loss: 0.6127 - sparse_categorical_accuracy: 0.7583
Epoch 233/500
4/4 [==============================] - 0s 997us/step - loss: 0.3575 - sparse_categorical_accuracy: 0.9417
Epoch 234/500
4/4 [==============================] - 0s 665us/step - loss: 0.3649 - sparse_categorical_accuracy: 0.9333
Epoch 235/500
4/4 [==============================] - 0s 997us/step - loss: 0.3872 - sparse_categorical_accuracy: 0.9333
Epoch 236/500
4/4 [==============================] - 0s 665us/step - loss: 0.3888 - sparse_categorical_accuracy: 0.9000
Epoch 237/500
4/4 [==============================] - 0s 998us/step - loss: 0.3783 - sparse_categorical_accuracy: 0.9083
Epoch 238/500
4/4 [==============================] - 0s 665us/step - loss: 0.3597 - sparse_categorical_accuracy: 0.9250
Epoch 239/500
4/4 [==============================] - 0s 665us/step - loss: 0.3628 - sparse_categorical_accuracy: 0.9417
Epoch 240/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3752 - sparse_categorical_accuracy: 0.9250 - val_loss: 0.3356 - val_sparse_categorical_accuracy: 1.0000
Epoch 241/500
4/4 [==============================] - 0s 665us/step - loss: 0.3523 - sparse_categorical_accuracy: 0.9583
Epoch 242/500
4/4 [==============================] - 0s 997us/step - loss: 0.4029 - sparse_categorical_accuracy: 0.9000
Epoch 243/500
4/4 [==============================] - 0s 665us/step - loss: 0.5745 - sparse_categorical_accuracy: 0.7667
Epoch 244/500
4/4 [==============================] - 0s 665us/step - loss: 0.3475 - sparse_categorical_accuracy: 0.9583
Epoch 245/500
4/4 [==============================] - 0s 665us/step - loss: 0.4219 - sparse_categorical_accuracy: 0.8667
Epoch 246/500
4/4 [==============================] - 0s 997us/step - loss: 0.3892 - sparse_categorical_accuracy: 0.9167
Epoch 247/500
4/4 [==============================] - 0s 665us/step - loss: 0.4275 - sparse_categorical_accuracy: 0.8667
Epoch 248/500
4/4 [==============================] - 0s 665us/step - loss: 0.3419 - sparse_categorical_accuracy: 0.9583
Epoch 249/500
4/4 [==============================] - 0s 665us/step - loss: 0.3683 - sparse_categorical_accuracy: 0.9250
Epoch 250/500
4/4 [==============================] - 0s 665us/step - loss: 0.3579 - sparse_categorical_accuracy: 0.9417
Epoch 251/500
4/4 [==============================] - 0s 998us/step - loss: 0.3651 - sparse_categorical_accuracy: 0.9583
Epoch 252/500
4/4 [==============================] - 0s 665us/step - loss: 0.3562 - sparse_categorical_accuracy: 0.9417
Epoch 253/500
4/4 [==============================] - 0s 997us/step - loss: 0.4366 - sparse_categorical_accuracy: 0.8667
Epoch 254/500
4/4 [==============================] - 0s 997us/step - loss: 0.4126 - sparse_categorical_accuracy: 0.8667
Epoch 255/500
4/4 [==============================] - 0s 665us/step - loss: 0.4202 - sparse_categorical_accuracy: 0.8833
Epoch 256/500
4/4 [==============================] - 0s 665us/step - loss: 0.4353 - sparse_categorical_accuracy: 0.8667
Epoch 257/500
4/4 [==============================] - 0s 665us/step - loss: 0.3604 - sparse_categorical_accuracy: 0.9333
Epoch 258/500
4/4 [==============================] - 0s 665us/step - loss: 0.3504 - sparse_categorical_accuracy: 0.9583
Epoch 259/500
4/4 [==============================] - 0s 997us/step - loss: 0.3866 - sparse_categorical_accuracy: 0.9083
Epoch 260/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3516 - sparse_categorical_accuracy: 0.9417 - val_loss: 0.3398 - val_sparse_categorical_accuracy: 1.0000
Epoch 261/500
4/4 [==============================] - 0s 997us/step - loss: 0.3769 - sparse_categorical_accuracy: 0.9333
Epoch 262/500
4/4 [==============================] - 0s 997us/step - loss: 0.3620 - sparse_categorical_accuracy: 0.9333
Epoch 263/500
4/4 [==============================] - 0s 665us/step - loss: 0.3474 - sparse_categorical_accuracy: 0.9667
Epoch 264/500
4/4 [==============================] - 0s 665us/step - loss: 0.3434 - sparse_categorical_accuracy: 0.9583
Epoch 265/500
4/4 [==============================] - 0s 998us/step - loss: 0.3778 - sparse_categorical_accuracy: 0.9167
Epoch 266/500
4/4 [==============================] - 0s 665us/step - loss: 0.3515 - sparse_categorical_accuracy: 0.9500
Epoch 267/500
4/4 [==============================] - 0s 665us/step - loss: 0.3481 - sparse_categorical_accuracy: 0.9667
Epoch 268/500
4/4 [==============================] - 0s 665us/step - loss: 0.3558 - sparse_categorical_accuracy: 0.9333
Epoch 269/500
4/4 [==============================] - 0s 665us/step - loss: 0.4076 - sparse_categorical_accuracy: 0.8750
Epoch 270/500
4/4 [==============================] - 0s 665us/step - loss: 0.4119 - sparse_categorical_accuracy: 0.9000
Epoch 271/500
4/4 [==============================] - 0s 665us/step - loss: 0.3590 - sparse_categorical_accuracy: 0.9417
Epoch 272/500
4/4 [==============================] - 0s 997us/step - loss: 0.3555 - sparse_categorical_accuracy: 0.9583
Epoch 273/500
4/4 [==============================] - 0s 665us/step - loss: 0.3999 - sparse_categorical_accuracy: 0.8833
Epoch 274/500
4/4 [==============================] - 0s 997us/step - loss: 0.3456 - sparse_categorical_accuracy: 0.9750
Epoch 275/500
4/4 [==============================] - 0s 665us/step - loss: 0.3543 - sparse_categorical_accuracy: 0.9333
Epoch 276/500
4/4 [==============================] - 0s 998us/step - loss: 0.3608 - sparse_categorical_accuracy: 0.9500
Epoch 277/500
4/4 [==============================] - 0s 997us/step - loss: 0.4091 - sparse_categorical_accuracy: 0.8833
Epoch 278/500
4/4 [==============================] - 0s 665us/step - loss: 0.3711 - sparse_categorical_accuracy: 0.9500
Epoch 279/500
4/4 [==============================] - 0s 665us/step - loss: 0.3670 - sparse_categorical_accuracy: 0.9167
Epoch 280/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3592 - sparse_categorical_accuracy: 0.9500 - val_loss: 0.3251 - val_sparse_categorical_accuracy: 0.9667
Epoch 281/500
4/4 [==============================] - 0s 997us/step - loss: 0.3522 - sparse_categorical_accuracy: 0.9750
Epoch 282/500
4/4 [==============================] - 0s 665us/step - loss: 0.3495 - sparse_categorical_accuracy: 0.9417
Epoch 283/500
4/4 [==============================] - 0s 665us/step - loss: 0.3462 - sparse_categorical_accuracy: 0.9750
Epoch 284/500
4/4 [==============================] - 0s 998us/step - loss: 0.3948 - sparse_categorical_accuracy: 0.9167
Epoch 285/500
4/4 [==============================] - 0s 998us/step - loss: 0.4060 - sparse_categorical_accuracy: 0.8833
Epoch 286/500
4/4 [==============================] - 0s 998us/step - loss: 0.3834 - sparse_categorical_accuracy: 0.9167
Epoch 287/500
4/4 [==============================] - 0s 665us/step - loss: 0.3814 - sparse_categorical_accuracy: 0.9000
Epoch 288/500
4/4 [==============================] - 0s 997us/step - loss: 0.3698 - sparse_categorical_accuracy: 0.9250
Epoch 289/500
4/4 [==============================] - 0s 665us/step - loss: 0.5091 - sparse_categorical_accuracy: 0.8083
Epoch 290/500
4/4 [==============================] - 0s 665us/step - loss: 0.3431 - sparse_categorical_accuracy: 0.9417
Epoch 291/500
4/4 [==============================] - 0s 665us/step - loss: 0.3418 - sparse_categorical_accuracy: 0.9667
Epoch 292/500
4/4 [==============================] - 0s 664us/step - loss: 0.3627 - sparse_categorical_accuracy: 0.9250
Epoch 293/500
4/4 [==============================] - 0s 997us/step - loss: 0.3739 - sparse_categorical_accuracy: 0.9000
Epoch 294/500
4/4 [==============================] - 0s 997us/step - loss: 0.3468 - sparse_categorical_accuracy: 0.9667
Epoch 295/500
4/4 [==============================] - 0s 665us/step - loss: 0.4111 - sparse_categorical_accuracy: 0.8583
Epoch 296/500
4/4 [==============================] - 0s 997us/step - loss: 0.3469 - sparse_categorical_accuracy: 0.9333
Epoch 297/500
4/4 [==============================] - 0s 997us/step - loss: 0.3459 - sparse_categorical_accuracy: 0.9333
Epoch 298/500
4/4 [==============================] - 0s 665us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9583
Epoch 299/500
4/4 [==============================] - 0s 998us/step - loss: 0.3559 - sparse_categorical_accuracy: 0.9417
Epoch 300/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3615 - sparse_categorical_accuracy: 0.9417 - val_loss: 0.3242 - val_sparse_categorical_accuracy: 0.9667
Epoch 301/500
4/4 [==============================] - 0s 665us/step - loss: 0.3921 - sparse_categorical_accuracy: 0.9250
Epoch 302/500
4/4 [==============================] - 0s 997us/step - loss: 0.3612 - sparse_categorical_accuracy: 0.9167
Epoch 303/500
4/4 [==============================] - 0s 997us/step - loss: 0.3446 - sparse_categorical_accuracy: 0.9583
Epoch 304/500
4/4 [==============================] - 0s 665us/step - loss: 0.3693 - sparse_categorical_accuracy: 0.9250
Epoch 305/500
4/4 [==============================] - 0s 665us/step - loss: 0.3616 - sparse_categorical_accuracy: 0.9500
Epoch 306/500
4/4 [==============================] - 0s 665us/step - loss: 0.3492 - sparse_categorical_accuracy: 0.9750
Epoch 307/500
4/4 [==============================] - 0s 665us/step - loss: 0.3634 - sparse_categorical_accuracy: 0.9417
Epoch 308/500
4/4 [==============================] - 0s 997us/step - loss: 0.3407 - sparse_categorical_accuracy: 0.9500
Epoch 309/500
4/4 [==============================] - 0s 997us/step - loss: 0.3700 - sparse_categorical_accuracy: 0.8917
Epoch 310/500
4/4 [==============================] - 0s 665us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9750
Epoch 311/500
4/4 [==============================] - 0s 997us/step - loss: 0.3738 - sparse_categorical_accuracy: 0.8917
Epoch 312/500
4/4 [==============================] - 0s 665us/step - loss: 0.3417 - sparse_categorical_accuracy: 0.9583
Epoch 313/500
4/4 [==============================] - 0s 997us/step - loss: 0.3499 - sparse_categorical_accuracy: 0.9417
Epoch 314/500
4/4 [==============================] - 0s 665us/step - loss: 0.3756 - sparse_categorical_accuracy: 0.9000
Epoch 315/500
4/4 [==============================] - 0s 998us/step - loss: 0.3532 - sparse_categorical_accuracy: 0.9417
Epoch 316/500
4/4 [==============================] - 0s 997us/step - loss: 0.4007 - sparse_categorical_accuracy: 0.8917
Epoch 317/500
4/4 [==============================] - 0s 665us/step - loss: 0.3593 - sparse_categorical_accuracy: 0.9417
Epoch 318/500
4/4 [==============================] - 0s 665us/step - loss: 0.3408 - sparse_categorical_accuracy: 0.9583
Epoch 319/500
4/4 [==============================] - 0s 998us/step - loss: 0.4063 - sparse_categorical_accuracy: 0.8833
Epoch 320/500
4/4 [==============================] - 0s 5ms/step - loss: 0.5146 - sparse_categorical_accuracy: 0.7917 - val_loss: 0.3246 - val_sparse_categorical_accuracy: 1.0000
Epoch 321/500
4/4 [==============================] - 0s 997us/step - loss: 0.3707 - sparse_categorical_accuracy: 0.9083
Epoch 322/500
4/4 [==============================] - 0s 665us/step - loss: 0.3414 - sparse_categorical_accuracy: 0.9750
Epoch 323/500
4/4 [==============================] - 0s 997us/step - loss: 0.3500 - sparse_categorical_accuracy: 0.9500
Epoch 324/500
4/4 [==============================] - 0s 997us/step - loss: 0.3485 - sparse_categorical_accuracy: 0.9500
Epoch 325/500
4/4 [==============================] - 0s 665us/step - loss: 0.4369 - sparse_categorical_accuracy: 0.8583
Epoch 326/500
4/4 [==============================] - 0s 665us/step - loss: 0.3594 - sparse_categorical_accuracy: 0.9500
Epoch 327/500
4/4 [==============================] - 0s 665us/step - loss: 0.3559 - sparse_categorical_accuracy: 0.9417
Epoch 328/500
4/4 [==============================] - 0s 997us/step - loss: 0.4150 - sparse_categorical_accuracy: 0.8917
Epoch 329/500
4/4 [==============================] - 0s 997us/step - loss: 0.3804 - sparse_categorical_accuracy: 0.9167
Epoch 330/500
4/4 [==============================] - 0s 665us/step - loss: 0.3942 - sparse_categorical_accuracy: 0.8917
Epoch 331/500
4/4 [==============================] - 0s 665us/step - loss: 0.3654 - sparse_categorical_accuracy: 0.9417
Epoch 332/500
4/4 [==============================] - 0s 665us/step - loss: 0.3498 - sparse_categorical_accuracy: 0.9500
Epoch 333/500
4/4 [==============================] - 0s 665us/step - loss: 0.3528 - sparse_categorical_accuracy: 0.9500
Epoch 334/500
4/4 [==============================] - 0s 665us/step - loss: 0.3459 - sparse_categorical_accuracy: 0.9750
Epoch 335/500
4/4 [==============================] - 0s 665us/step - loss: 0.3427 - sparse_categorical_accuracy: 0.9417
Epoch 336/500
4/4 [==============================] - 0s 997us/step - loss: 0.3421 - sparse_categorical_accuracy: 0.9667
Epoch 337/500
4/4 [==============================] - 0s 998us/step - loss: 0.3449 - sparse_categorical_accuracy: 0.9417
Epoch 338/500
4/4 [==============================] - 0s 665us/step - loss: 0.4189 - sparse_categorical_accuracy: 0.8583
Epoch 339/500
4/4 [==============================] - 0s 997us/step - loss: 0.3469 - sparse_categorical_accuracy: 0.9500
Epoch 340/500
4/4 [==============================] - 0s 6ms/step - loss: 0.4089 - sparse_categorical_accuracy: 0.8583 - val_loss: 0.3289 - val_sparse_categorical_accuracy: 1.0000
Epoch 341/500
4/4 [==============================] - 0s 665us/step - loss: 0.3863 - sparse_categorical_accuracy: 0.9083
Epoch 342/500
4/4 [==============================] - 0s 665us/step - loss: 0.3991 - sparse_categorical_accuracy: 0.8833
Epoch 343/500
4/4 [==============================] - 0s 998us/step - loss: 0.4232 - sparse_categorical_accuracy: 0.8667
Epoch 344/500
4/4 [==============================] - 0s 665us/step - loss: 0.4240 - sparse_categorical_accuracy: 0.8667
Epoch 345/500
4/4 [==============================] - 0s 665us/step - loss: 0.3626 - sparse_categorical_accuracy: 0.9500
Epoch 346/500
4/4 [==============================] - 0s 665us/step - loss: 0.3398 - sparse_categorical_accuracy: 0.9583
Epoch 347/500
4/4 [==============================] - 0s 998us/step - loss: 0.3416 - sparse_categorical_accuracy: 0.9250
Epoch 348/500
4/4 [==============================] - 0s 997us/step - loss: 0.3892 - sparse_categorical_accuracy: 0.9083
Epoch 349/500
4/4 [==============================] - 0s 665us/step - loss: 0.3674 - sparse_categorical_accuracy: 0.9167
Epoch 350/500
4/4 [==============================] - 0s 665us/step - loss: 0.3400 - sparse_categorical_accuracy: 0.9667
Epoch 351/500
4/4 [==============================] - 0s 665us/step - loss: 0.3556 - sparse_categorical_accuracy: 0.9500
Epoch 352/500
4/4 [==============================] - 0s 665us/step - loss: 0.3704 - sparse_categorical_accuracy: 0.9500
Epoch 353/500
4/4 [==============================] - 0s 665us/step - loss: 0.3406 - sparse_categorical_accuracy: 0.9667
Epoch 354/500
4/4 [==============================] - 0s 665us/step - loss: 0.3330 - sparse_categorical_accuracy: 0.9500
Epoch 355/500
4/4 [==============================] - 0s 997us/step - loss: 0.3984 - sparse_categorical_accuracy: 0.8833
Epoch 356/500
4/4 [==============================] - 0s 665us/step - loss: 0.4194 - sparse_categorical_accuracy: 0.8583
Epoch 357/500
4/4 [==============================] - 0s 665us/step - loss: 0.3721 - sparse_categorical_accuracy: 0.9250
Epoch 358/500
4/4 [==============================] - 0s 665us/step - loss: 0.3770 - sparse_categorical_accuracy: 0.9333
Epoch 359/500
4/4 [==============================] - 0s 665us/step - loss: 0.3631 - sparse_categorical_accuracy: 0.9167
Epoch 360/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3405 - sparse_categorical_accuracy: 0.9583 - val_loss: 0.3238 - val_sparse_categorical_accuracy: 0.9667
Epoch 361/500
4/4 [==============================] - 0s 665us/step - loss: 0.4157 - sparse_categorical_accuracy: 0.9000
Epoch 362/500
4/4 [==============================] - 0s 665us/step - loss: 0.3548 - sparse_categorical_accuracy: 0.9667
Epoch 363/500
4/4 [==============================] - 0s 997us/step - loss: 0.3715 - sparse_categorical_accuracy: 0.9417
Epoch 364/500
4/4 [==============================] - 0s 665us/step - loss: 0.4348 - sparse_categorical_accuracy: 0.8417
Epoch 365/500
4/4 [==============================] - 0s 665us/step - loss: 0.3671 - sparse_categorical_accuracy: 0.9333
Epoch 366/500
4/4 [==============================] - 0s 665us/step - loss: 0.3489 - sparse_categorical_accuracy: 0.9500
Epoch 367/500
4/4 [==============================] - 0s 998us/step - loss: 0.3407 - sparse_categorical_accuracy: 0.9500
Epoch 368/500
4/4 [==============================] - 0s 665us/step - loss: 0.3895 - sparse_categorical_accuracy: 0.9083
Epoch 369/500
4/4 [==============================] - 0s 665us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9583
Epoch 370/500
4/4 [==============================] - 0s 665us/step - loss: 0.3371 - sparse_categorical_accuracy: 0.9583
Epoch 371/500
4/4 [==============================] - 0s 665us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9667
Epoch 372/500
4/4 [==============================] - 0s 665us/step - loss: 0.3445 - sparse_categorical_accuracy: 0.9583
Epoch 373/500
4/4 [==============================] - 0s 997us/step - loss: 0.4199 - sparse_categorical_accuracy: 0.8833
Epoch 374/500
4/4 [==============================] - 0s 997us/step - loss: 0.3528 - sparse_categorical_accuracy: 0.9417
Epoch 375/500
4/4 [==============================] - 0s 664us/step - loss: 0.3791 - sparse_categorical_accuracy: 0.9250
Epoch 376/500
4/4 [==============================] - 0s 997us/step - loss: 0.3511 - sparse_categorical_accuracy: 0.9500
Epoch 377/500
4/4 [==============================] - 0s 997us/step - loss: 0.3391 - sparse_categorical_accuracy: 0.9667
Epoch 378/500
4/4 [==============================] - 0s 997us/step - loss: 0.3461 - sparse_categorical_accuracy: 0.9500
Epoch 379/500
4/4 [==============================] - 0s 997us/step - loss: 0.3603 - sparse_categorical_accuracy: 0.9667
Epoch 380/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3398 - sparse_categorical_accuracy: 0.9667 - val_loss: 0.3323 - val_sparse_categorical_accuracy: 1.0000
Epoch 381/500
4/4 [==============================] - 0s 665us/step - loss: 0.3799 - sparse_categorical_accuracy: 0.9167
Epoch 382/500
4/4 [==============================] - 0s 665us/step - loss: 0.3502 - sparse_categorical_accuracy: 0.9667
Epoch 383/500
4/4 [==============================] - 0s 665us/step - loss: 0.3963 - sparse_categorical_accuracy: 0.8833
Epoch 384/500
4/4 [==============================] - 0s 997us/step - loss: 0.3446 - sparse_categorical_accuracy: 0.9500
Epoch 385/500
4/4 [==============================] - 0s 665us/step - loss: 0.3970 - sparse_categorical_accuracy: 0.9250
Epoch 386/500
4/4 [==============================] - 0s 665us/step - loss: 0.3411 - sparse_categorical_accuracy: 0.9750
Epoch 387/500
4/4 [==============================] - 0s 998us/step - loss: 0.3672 - sparse_categorical_accuracy: 0.9167
Epoch 388/500
4/4 [==============================] - 0s 665us/step - loss: 0.3954 - sparse_categorical_accuracy: 0.9167
Epoch 389/500
4/4 [==============================] - 0s 665us/step - loss: 0.3850 - sparse_categorical_accuracy: 0.8750
Epoch 390/500
4/4 [==============================] - 0s 665us/step - loss: 0.4663 - sparse_categorical_accuracy: 0.8333
Epoch 391/500
4/4 [==============================] - 0s 998us/step - loss: 0.4209 - sparse_categorical_accuracy: 0.8583
Epoch 392/500
4/4 [==============================] - 0s 665us/step - loss: 0.3512 - sparse_categorical_accuracy: 0.9250
Epoch 393/500
4/4 [==============================] - 0s 665us/step - loss: 0.3487 - sparse_categorical_accuracy: 0.9417
Epoch 394/500
4/4 [==============================] - 0s 665us/step - loss: 0.3444 - sparse_categorical_accuracy: 0.9500
Epoch 395/500
4/4 [==============================] - 0s 997us/step - loss: 0.3714 - sparse_categorical_accuracy: 0.9333
Epoch 396/500
4/4 [==============================] - 0s 665us/step - loss: 0.3681 - sparse_categorical_accuracy: 0.9250
Epoch 397/500
4/4 [==============================] - 0s 665us/step - loss: 0.3659 - sparse_categorical_accuracy: 0.9000
Epoch 398/500
4/4 [==============================] - 0s 665us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9667
Epoch 399/500
4/4 [==============================] - 0s 665us/step - loss: 0.3917 - sparse_categorical_accuracy: 0.9083
Epoch 400/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3787 - sparse_categorical_accuracy: 0.9250 - val_loss: 0.5941 - val_sparse_categorical_accuracy: 0.6000
Epoch 401/500
4/4 [==============================] - 0s 665us/step - loss: 0.3651 - sparse_categorical_accuracy: 0.9333
Epoch 402/500
4/4 [==============================] - 0s 665us/step - loss: 0.3480 - sparse_categorical_accuracy: 0.9583
Epoch 403/500
4/4 [==============================] - 0s 997us/step - loss: 0.3813 - sparse_categorical_accuracy: 0.9167
Epoch 404/500
4/4 [==============================] - 0s 665us/step - loss: 0.4448 - sparse_categorical_accuracy: 0.8583
Epoch 405/500
4/4 [==============================] - 0s 998us/step - loss: 0.3864 - sparse_categorical_accuracy: 0.9083
Epoch 406/500
4/4 [==============================] - 0s 665us/step - loss: 0.3440 - sparse_categorical_accuracy: 0.9583
Epoch 407/500
4/4 [==============================] - 0s 997us/step - loss: 0.4250 - sparse_categorical_accuracy: 0.8667
Epoch 408/500
4/4 [==============================] - 0s 665us/step - loss: 0.3375 - sparse_categorical_accuracy: 0.9667
Epoch 409/500
4/4 [==============================] - 0s 665us/step - loss: 0.3470 - sparse_categorical_accuracy: 0.9333
Epoch 410/500
4/4 [==============================] - 0s 665us/step - loss: 0.3442 - sparse_categorical_accuracy: 0.9500
Epoch 411/500
4/4 [==============================] - 0s 665us/step - loss: 0.3907 - sparse_categorical_accuracy: 0.8583
Epoch 412/500
4/4 [==============================] - 0s 665us/step - loss: 0.3829 - sparse_categorical_accuracy: 0.8750
Epoch 413/500
4/4 [==============================] - 0s 665us/step - loss: 0.3658 - sparse_categorical_accuracy: 0.8917
Epoch 414/500
4/4 [==============================] - 0s 665us/step - loss: 0.3384 - sparse_categorical_accuracy: 0.9583
Epoch 415/500
4/4 [==============================] - 0s 997us/step - loss: 0.3447 - sparse_categorical_accuracy: 0.9667
Epoch 416/500
4/4 [==============================] - 0s 665us/step - loss: 0.3389 - sparse_categorical_accuracy: 0.9750
Epoch 417/500
4/4 [==============================] - 0s 665us/step - loss: 0.3484 - sparse_categorical_accuracy: 0.9750
Epoch 418/500
4/4 [==============================] - 0s 997us/step - loss: 0.3718 - sparse_categorical_accuracy: 0.9167
Epoch 419/500
4/4 [==============================] - 0s 998us/step - loss: 0.3435 - sparse_categorical_accuracy: 0.9500
Epoch 420/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3400 - sparse_categorical_accuracy: 0.9500 - val_loss: 0.3493 - val_sparse_categorical_accuracy: 1.0000
Epoch 421/500
4/4 [==============================] - 0s 997us/step - loss: 0.3376 - sparse_categorical_accuracy: 0.9667
Epoch 422/500
4/4 [==============================] - 0s 665us/step - loss: 0.3501 - sparse_categorical_accuracy: 0.9750
Epoch 423/500
4/4 [==============================] - 0s 665us/step - loss: 0.4595 - sparse_categorical_accuracy: 0.8250
Epoch 424/500
4/4 [==============================] - 0s 998us/step - loss: 0.3583 - sparse_categorical_accuracy: 0.9333
Epoch 425/500
4/4 [==============================] - 0s 665us/step - loss: 0.3593 - sparse_categorical_accuracy: 0.9417
Epoch 426/500
4/4 [==============================] - 0s 997us/step - loss: 0.4827 - sparse_categorical_accuracy: 0.8000
Epoch 427/500
4/4 [==============================] - 0s 665us/step - loss: 0.4092 - sparse_categorical_accuracy: 0.8917
Epoch 428/500
4/4 [==============================] - 0s 997us/step - loss: 0.3592 - sparse_categorical_accuracy: 0.9250
Epoch 429/500
4/4 [==============================] - 0s 997us/step - loss: 0.3744 - sparse_categorical_accuracy: 0.9000
Epoch 430/500
4/4 [==============================] - 0s 997us/step - loss: 0.4432 - sparse_categorical_accuracy: 0.8750
Epoch 431/500
4/4 [==============================] - 0s 665us/step - loss: 0.3755 - sparse_categorical_accuracy: 0.9167
Epoch 432/500
4/4 [==============================] - 0s 998us/step - loss: 0.3350 - sparse_categorical_accuracy: 0.9667
Epoch 433/500
4/4 [==============================] - 0s 665us/step - loss: 0.3608 - sparse_categorical_accuracy: 0.9417
Epoch 434/500
4/4 [==============================] - 0s 665us/step - loss: 0.3540 - sparse_categorical_accuracy: 0.9417
Epoch 435/500
4/4 [==============================] - 0s 997us/step - loss: 0.3831 - sparse_categorical_accuracy: 0.9167
Epoch 436/500
4/4 [==============================] - 0s 665us/step - loss: 0.3791 - sparse_categorical_accuracy: 0.8750
Epoch 437/500
4/4 [==============================] - 0s 665us/step - loss: 0.3327 - sparse_categorical_accuracy: 0.9750
Epoch 438/500
4/4 [==============================] - 0s 997us/step - loss: 0.3575 - sparse_categorical_accuracy: 0.9250
Epoch 439/500
4/4 [==============================] - 0s 665us/step - loss: 0.4376 - sparse_categorical_accuracy: 0.8500
Epoch 440/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3535 - sparse_categorical_accuracy: 0.9333 - val_loss: 0.3876 - val_sparse_categorical_accuracy: 1.0000
Epoch 441/500
4/4 [==============================] - 0s 665us/step - loss: 0.3743 - sparse_categorical_accuracy: 0.9250
Epoch 442/500
4/4 [==============================] - 0s 998us/step - loss: 0.3386 - sparse_categorical_accuracy: 0.9750
Epoch 443/500
4/4 [==============================] - 0s 665us/step - loss: 0.3476 - sparse_categorical_accuracy: 0.9667
Epoch 444/500
4/4 [==============================] - 0s 665us/step - loss: 0.3391 - sparse_categorical_accuracy: 0.9500
Epoch 445/500
4/4 [==============================] - 0s 997us/step - loss: 0.3449 - sparse_categorical_accuracy: 0.9583
Epoch 446/500
4/4 [==============================] - 0s 665us/step - loss: 0.3598 - sparse_categorical_accuracy: 0.9417
Epoch 447/500
4/4 [==============================] - 0s 997us/step - loss: 0.3927 - sparse_categorical_accuracy: 0.8833
Epoch 448/500
4/4 [==============================] - 0s 665us/step - loss: 0.3524 - sparse_categorical_accuracy: 0.9333
Epoch 449/500
4/4 [==============================] - 0s 665us/step - loss: 0.3387 - sparse_categorical_accuracy: 0.9750
Epoch 450/500
4/4 [==============================] - 0s 997us/step - loss: 0.3424 - sparse_categorical_accuracy: 0.9250
Epoch 451/500
4/4 [==============================] - 0s 665us/step - loss: 0.4508 - sparse_categorical_accuracy: 0.8333
Epoch 452/500
4/4 [==============================] - 0s 665us/step - loss: 0.3504 - sparse_categorical_accuracy: 0.9417
Epoch 453/500
4/4 [==============================] - 0s 665us/step - loss: 0.3766 - sparse_categorical_accuracy: 0.9167
Epoch 454/500
4/4 [==============================] - 0s 665us/step - loss: 0.3650 - sparse_categorical_accuracy: 0.9167
Epoch 455/500
4/4 [==============================] - 0s 665us/step - loss: 0.3506 - sparse_categorical_accuracy: 0.9583
Epoch 456/500
4/4 [==============================] - 0s 665us/step - loss: 0.3765 - sparse_categorical_accuracy: 0.9083
Epoch 457/500
4/4 [==============================] - 0s 665us/step - loss: 0.3376 - sparse_categorical_accuracy: 0.9833
Epoch 458/500
4/4 [==============================] - 0s 998us/step - loss: 0.3532 - sparse_categorical_accuracy: 0.9583
Epoch 459/500
4/4 [==============================] - 0s 665us/step - loss: 0.3358 - sparse_categorical_accuracy: 0.9500
Epoch 460/500
4/4 [==============================] - 0s 5ms/step - loss: 0.6349 - sparse_categorical_accuracy: 0.7250 - val_loss: 0.3554 - val_sparse_categorical_accuracy: 1.0000
Epoch 461/500
4/4 [==============================] - 0s 665us/step - loss: 0.3414 - sparse_categorical_accuracy: 0.9750
Epoch 462/500
4/4 [==============================] - 0s 665us/step - loss: 0.3489 - sparse_categorical_accuracy: 0.9417
Epoch 463/500
4/4 [==============================] - 0s 665us/step - loss: 0.3581 - sparse_categorical_accuracy: 0.9417
Epoch 464/500
4/4 [==============================] - 0s 665us/step - loss: 0.4619 - sparse_categorical_accuracy: 0.8583
Epoch 465/500
4/4 [==============================] - 0s 665us/step - loss: 0.4844 - sparse_categorical_accuracy: 0.8333
Epoch 466/500
4/4 [==============================] - 0s 998us/step - loss: 0.3958 - sparse_categorical_accuracy: 0.8833
Epoch 467/500
4/4 [==============================] - 0s 665us/step - loss: 0.3387 - sparse_categorical_accuracy: 0.9583
Epoch 468/500
4/4 [==============================] - 0s 665us/step - loss: 0.3566 - sparse_categorical_accuracy: 0.9417
Epoch 469/500
4/4 [==============================] - 0s 665us/step - loss: 0.3544 - sparse_categorical_accuracy: 0.9417
Epoch 470/500
4/4 [==============================] - 0s 998us/step - loss: 0.3763 - sparse_categorical_accuracy: 0.9000
Epoch 471/500
4/4 [==============================] - 0s 665us/step - loss: 0.3349 - sparse_categorical_accuracy: 0.9500
Epoch 472/500
4/4 [==============================] - 0s 665us/step - loss: 0.5347 - sparse_categorical_accuracy: 0.7667
Epoch 473/500
4/4 [==============================] - 0s 665us/step - loss: 0.4666 - sparse_categorical_accuracy: 0.8833
Epoch 474/500
4/4 [==============================] - 0s 997us/step - loss: 0.5124 - sparse_categorical_accuracy: 0.7833
Epoch 475/500
4/4 [==============================] - 0s 997us/step - loss: 0.3539 - sparse_categorical_accuracy: 0.9417
Epoch 476/500
4/4 [==============================] - 0s 665us/step - loss: 0.3371 - sparse_categorical_accuracy: 0.9583
Epoch 477/500
4/4 [==============================] - 0s 666us/step - loss: 0.4072 - sparse_categorical_accuracy: 0.8917
Epoch 478/500
4/4 [==============================] - 0s 665us/step - loss: 0.4168 - sparse_categorical_accuracy: 0.8750
Epoch 479/500
4/4 [==============================] - 0s 997us/step - loss: 0.4617 - sparse_categorical_accuracy: 0.8333
Epoch 480/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3497 - sparse_categorical_accuracy: 0.9417 - val_loss: 0.3175 - val_sparse_categorical_accuracy: 1.0000
Epoch 481/500
4/4 [==============================] - 0s 997us/step - loss: 0.3413 - sparse_categorical_accuracy: 0.9667
Epoch 482/500
4/4 [==============================] - 0s 665us/step - loss: 0.3415 - sparse_categorical_accuracy: 0.9667
Epoch 483/500
4/4 [==============================] - 0s 665us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9583
Epoch 484/500
4/4 [==============================] - 0s 997us/step - loss: 0.3377 - sparse_categorical_accuracy: 0.9500
Epoch 485/500
4/4 [==============================] - 0s 665us/step - loss: 0.4217 - sparse_categorical_accuracy: 0.8583
Epoch 486/500
4/4 [==============================] - 0s 998us/step - loss: 0.3618 - sparse_categorical_accuracy: 0.9417
Epoch 487/500
4/4 [==============================] - 0s 665us/step - loss: 0.3622 - sparse_categorical_accuracy: 0.9083
Epoch 488/500
4/4 [==============================] - 0s 665us/step - loss: 0.3756 - sparse_categorical_accuracy: 0.8917
Epoch 489/500
4/4 [==============================] - 0s 997us/step - loss: 0.3968 - sparse_categorical_accuracy: 0.9000
Epoch 490/500
4/4 [==============================] - 0s 665us/step - loss: 0.3435 - sparse_categorical_accuracy: 0.9583
Epoch 491/500
4/4 [==============================] - 0s 665us/step - loss: 0.3394 - sparse_categorical_accuracy: 0.9750
Epoch 492/500
4/4 [==============================] - 0s 665us/step - loss: 0.3348 - sparse_categorical_accuracy: 0.9833
Epoch 493/500
4/4 [==============================] - 0s 665us/step - loss: 0.3447 - sparse_categorical_accuracy: 0.9333
Epoch 494/500
4/4 [==============================] - 0s 665us/step - loss: 0.3433 - sparse_categorical_accuracy: 0.9333
Epoch 495/500
4/4 [==============================] - 0s 665us/step - loss: 0.3357 - sparse_categorical_accuracy: 0.9583
Epoch 496/500
4/4 [==============================] - 0s 665us/step - loss: 0.3445 - sparse_categorical_accuracy: 0.9583
Epoch 497/500
4/4 [==============================] - 0s 665us/step - loss: 0.3435 - sparse_categorical_accuracy: 0.9333
Epoch 498/500
4/4 [==============================] - 0s 664us/step - loss: 0.3406 - sparse_categorical_accuracy: 0.9667
Epoch 499/500
4/4 [==============================] - 0s 665us/step - loss: 0.3649 - sparse_categorical_accuracy: 0.9333
Epoch 500/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3888 - sparse_categorical_accuracy: 0.9250 - val_loss: 0.3516 - val_sparse_categorical_accuracy: 0.8667
Model: "sequential_2"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 dense_2 (Dense)             (None, 3)                 15        
                                                                 
=================================================================
Total params: 15
Trainable params: 15
Non-trainable params: 0
_________________________________________________________________

2.2 使用class搭建网络八股

在第三步的逐层搭建网络结构中，使用Sequential可以快速搭建网络结构，但是如果网络包含跳连等其他复杂网络结构，Sequential就无法表示了。这就需要使用class来声明网络结构。

class MyModel(Model):
    def __init__(self):
        super(MyModel, self).__init__()
        # 初始化网络结构
        
    def calll(self, x):
        # 调用网络结构快，实现前向传播
        y  = self.dl(x)
        return y
    
model = MyModel()

'''
        使用 class 类封装网络结构，如上所示是一个 class 模板， MyModel 表示声 明的神经网络的名字，括号中的 Model 表示创建的类需要继承 tensorflow 库中
的 Model 类。 类中需要定义两个函数， __init__()函数为类的构造函数用于初始化类的参数， spuer(MyModel,self).__init__()这行表示初始化父类的参数。 之后便可初始化网络结构,搭建出神经网络所需的各种网络结构块。 call()函数中调用__init__()函数中完成初始化的网络块，实现前向传播并返回推理
值。使用 class 方式搭建鸢尾花网络结构的代码如下所示。
'''

#例子
class MyModel(Model):
    def __init__(self):
        super(IrisModel, self).__init__()
        self.dl = Dense(3)
        
    del call(self, x):
        y = self.dl(x)
        return y
    
model = IrisModel()

对比使用 Sequential()方法和 class 方法，有两点区别：
1、import 中添加了 Model 模块和 Dense 层、 Flatten 层。
2、使用 class 声明网络结构， model = IrisModel()初始化模型对象。

# import
import tensorflow as tf
import numpy as np
from tensorflow.keras.layers import Dense
from tensorflow.keras import Model
from sklearn import datasets

# x_train, y_train
x_train = datasets.load_iris().data
y_train = datasets.load_iris().target

np.random.seed(116)
np.random.shuffle(x_train)
np.random.seed(116)
np.random.shuffle(y_train)
tf.random.set_seed(116)


# class mymodel
class MyModel(Model):
    def __init__(self):
        super(MyModel, self).__init__()
        self.dl = Dense(3, activation='softmax', kernel_regularizer=tf.keras.regularizers.l2())
        
    def call(self, x):
        y = self.dl(x)
        return y
    
model = MyModel()

# model compile
model.compile(optimizer=tf.keras.optimizers.SGD(lr = 0.1),
             loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits = False),
             metrics=['sparse_categorical_accuracy'])

model.fit(x_train, y_train, batch_size=32, epochs=500, validation_split=0.2, validation_freq=20)
model.summary()

C:\Users\ADMIN\AppData\Roaming\Python\Python39\site-packages\keras\optimizer_v2\gradient_descent.py:102: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.
  super(SGD, self).__init__(name, **kwargs)


Epoch 1/500
4/4 [==============================] - 0s 998us/step - loss: 2.3173 - sparse_categorical_accuracy: 0.3417
Epoch 2/500
4/4 [==============================] - 0s 998us/step - loss: 1.0797 - sparse_categorical_accuracy: 0.6167
Epoch 3/500
4/4 [==============================] - 0s 666us/step - loss: 0.8429 - sparse_categorical_accuracy: 0.6250
Epoch 4/500
4/4 [==============================] - 0s 665us/step - loss: 0.7448 - sparse_categorical_accuracy: 0.6667
Epoch 5/500
4/4 [==============================] - 0s 665us/step - loss: 0.9843 - sparse_categorical_accuracy: 0.6417
Epoch 6/500
4/4 [==============================] - 0s 997us/step - loss: 0.7808 - sparse_categorical_accuracy: 0.6583
Epoch 7/500
4/4 [==============================] - 0s 997us/step - loss: 0.5961 - sparse_categorical_accuracy: 0.7250
Epoch 8/500
4/4 [==============================] - 0s 665us/step - loss: 0.5824 - sparse_categorical_accuracy: 0.7250
Epoch 9/500
4/4 [==============================] - 0s 997us/step - loss: 0.6016 - sparse_categorical_accuracy: 0.7000
Epoch 10/500
4/4 [==============================] - 0s 665us/step - loss: 0.6707 - sparse_categorical_accuracy: 0.6500
Epoch 11/500
4/4 [==============================] - 0s 997us/step - loss: 0.7967 - sparse_categorical_accuracy: 0.6667
Epoch 12/500
4/4 [==============================] - 0s 665us/step - loss: 0.5385 - sparse_categorical_accuracy: 0.7500
Epoch 13/500
4/4 [==============================] - 0s 665us/step - loss: 1.0675 - sparse_categorical_accuracy: 0.5833
Epoch 14/500
4/4 [==============================] - 0s 998us/step - loss: 0.5119 - sparse_categorical_accuracy: 0.7333
Epoch 15/500
4/4 [==============================] - 0s 665us/step - loss: 0.6831 - sparse_categorical_accuracy: 0.6667
Epoch 16/500
4/4 [==============================] - 0s 665us/step - loss: 0.5969 - sparse_categorical_accuracy: 0.7667
Epoch 17/500
4/4 [==============================] - 0s 997us/step - loss: 0.5559 - sparse_categorical_accuracy: 0.7250
Epoch 18/500
4/4 [==============================] - 0s 665us/step - loss: 0.6521 - sparse_categorical_accuracy: 0.7667
Epoch 19/500
4/4 [==============================] - 0s 665us/step - loss: 0.5902 - sparse_categorical_accuracy: 0.6833
Epoch 20/500
4/4 [==============================] - 0s 39ms/step - loss: 0.4782 - sparse_categorical_accuracy: 0.8000 - val_loss: 0.7740 - val_sparse_categorical_accuracy: 0.5333
Epoch 21/500
4/4 [==============================] - 0s 665us/step - loss: 0.4589 - sparse_categorical_accuracy: 0.8333
Epoch 22/500
4/4 [==============================] - 0s 997us/step - loss: 0.6620 - sparse_categorical_accuracy: 0.7167
Epoch 23/500
4/4 [==============================] - 0s 665us/step - loss: 0.4497 - sparse_categorical_accuracy: 0.8583
Epoch 24/500
4/4 [==============================] - 0s 665us/step - loss: 0.7418 - sparse_categorical_accuracy: 0.6917
Epoch 25/500
4/4 [==============================] - 0s 997us/step - loss: 0.4873 - sparse_categorical_accuracy: 0.7333
Epoch 26/500
4/4 [==============================] - 0s 998us/step - loss: 0.7408 - sparse_categorical_accuracy: 0.7083
Epoch 27/500
4/4 [==============================] - 0s 665us/step - loss: 0.6502 - sparse_categorical_accuracy: 0.7750
Epoch 28/500
4/4 [==============================] - 0s 665us/step - loss: 0.6749 - sparse_categorical_accuracy: 0.6667
Epoch 29/500
4/4 [==============================] - 0s 666us/step - loss: 0.5382 - sparse_categorical_accuracy: 0.7833
Epoch 30/500
4/4 [==============================] - 0s 997us/step - loss: 0.6266 - sparse_categorical_accuracy: 0.7000
Epoch 31/500
4/4 [==============================] - 0s 665us/step - loss: 0.5041 - sparse_categorical_accuracy: 0.7917
Epoch 32/500
4/4 [==============================] - 0s 665us/step - loss: 0.6933 - sparse_categorical_accuracy: 0.7250
Epoch 33/500
4/4 [==============================] - 0s 997us/step - loss: 0.5518 - sparse_categorical_accuracy: 0.7667
Epoch 34/500
4/4 [==============================] - 0s 998us/step - loss: 0.4097 - sparse_categorical_accuracy: 0.9000
Epoch 35/500
4/4 [==============================] - 0s 997us/step - loss: 0.4705 - sparse_categorical_accuracy: 0.7917
Epoch 36/500
4/4 [==============================] - 0s 665us/step - loss: 0.4490 - sparse_categorical_accuracy: 0.8583
Epoch 37/500
4/4 [==============================] - 0s 664us/step - loss: 0.4172 - sparse_categorical_accuracy: 0.9333
Epoch 38/500
4/4 [==============================] - 0s 997us/step - loss: 0.6989 - sparse_categorical_accuracy: 0.6583
Epoch 39/500
4/4 [==============================] - 0s 998us/step - loss: 0.5124 - sparse_categorical_accuracy: 0.7833
Epoch 40/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3984 - sparse_categorical_accuracy: 0.9333 - val_loss: 0.5433 - val_sparse_categorical_accuracy: 0.6000
Epoch 41/500
4/4 [==============================] - 0s 997us/step - loss: 0.5434 - sparse_categorical_accuracy: 0.7333
Epoch 42/500
4/4 [==============================] - 0s 665us/step - loss: 0.6159 - sparse_categorical_accuracy: 0.7000
Epoch 43/500
4/4 [==============================] - 0s 997us/step - loss: 0.6999 - sparse_categorical_accuracy: 0.7167
Epoch 44/500
4/4 [==============================] - 0s 665us/step - loss: 0.4048 - sparse_categorical_accuracy: 0.9333
Epoch 45/500
4/4 [==============================] - 0s 665us/step - loss: 0.3924 - sparse_categorical_accuracy: 0.9500
Epoch 46/500
4/4 [==============================] - 0s 665us/step - loss: 0.3944 - sparse_categorical_accuracy: 0.9333
Epoch 47/500
4/4 [==============================] - 0s 998us/step - loss: 0.4063 - sparse_categorical_accuracy: 0.9083
Epoch 48/500
4/4 [==============================] - 0s 665us/step - loss: 0.5635 - sparse_categorical_accuracy: 0.7583
Epoch 49/500
4/4 [==============================] - 0s 997us/step - loss: 0.4108 - sparse_categorical_accuracy: 0.9333
Epoch 50/500
4/4 [==============================] - 0s 665us/step - loss: 0.3884 - sparse_categorical_accuracy: 0.9583
Epoch 51/500
4/4 [==============================] - 0s 665us/step - loss: 0.3835 - sparse_categorical_accuracy: 0.9417
Epoch 52/500
4/4 [==============================] - 0s 997us/step - loss: 0.4042 - sparse_categorical_accuracy: 0.9083
Epoch 53/500
4/4 [==============================] - 0s 997us/step - loss: 0.4400 - sparse_categorical_accuracy: 0.8500
Epoch 54/500
4/4 [==============================] - 0s 665us/step - loss: 0.7052 - sparse_categorical_accuracy: 0.6417
Epoch 55/500
4/4 [==============================] - 0s 997us/step - loss: 0.4013 - sparse_categorical_accuracy: 0.8833
Epoch 56/500
4/4 [==============================] - 0s 998us/step - loss: 0.3983 - sparse_categorical_accuracy: 0.8917
Epoch 57/500
4/4 [==============================] - 0s 665us/step - loss: 0.5127 - sparse_categorical_accuracy: 0.8000
Epoch 58/500
4/4 [==============================] - 0s 665us/step - loss: 0.5093 - sparse_categorical_accuracy: 0.8000
Epoch 59/500
4/4 [==============================] - 0s 998us/step - loss: 0.4715 - sparse_categorical_accuracy: 0.8083
Epoch 60/500
4/4 [==============================] - 0s 5ms/step - loss: 0.5407 - sparse_categorical_accuracy: 0.9000 - val_loss: 0.4260 - val_sparse_categorical_accuracy: 0.9333
Epoch 61/500
4/4 [==============================] - 0s 665us/step - loss: 0.4227 - sparse_categorical_accuracy: 0.8833
Epoch 62/500
4/4 [==============================] - 0s 997us/step - loss: 0.4420 - sparse_categorical_accuracy: 0.8583
Epoch 63/500
4/4 [==============================] - 0s 997us/step - loss: 0.6500 - sparse_categorical_accuracy: 0.7167
Epoch 64/500
4/4 [==============================] - 0s 998us/step - loss: 0.4293 - sparse_categorical_accuracy: 0.8667
Epoch 65/500
4/4 [==============================] - 0s 871us/step - loss: 0.3863 - sparse_categorical_accuracy: 0.9167
Epoch 66/500
4/4 [==============================] - 0s 665us/step - loss: 0.3802 - sparse_categorical_accuracy: 0.9250
Epoch 67/500
4/4 [==============================] - 0s 666us/step - loss: 0.4052 - sparse_categorical_accuracy: 0.8667
Epoch 68/500
4/4 [==============================] - 0s 997us/step - loss: 0.5520 - sparse_categorical_accuracy: 0.7250
Epoch 69/500
4/4 [==============================] - 0s 665us/step - loss: 0.4788 - sparse_categorical_accuracy: 0.8167
Epoch 70/500
4/4 [==============================] - 0s 666us/step - loss: 0.3761 - sparse_categorical_accuracy: 0.9333
Epoch 71/500
4/4 [==============================] - 0s 665us/step - loss: 0.4550 - sparse_categorical_accuracy: 0.8250
Epoch 72/500
4/4 [==============================] - 0s 997us/step - loss: 0.4004 - sparse_categorical_accuracy: 0.9250
Epoch 73/500
4/4 [==============================] - 0s 665us/step - loss: 0.5146 - sparse_categorical_accuracy: 0.8167
Epoch 74/500
4/4 [==============================] - 0s 665us/step - loss: 0.3846 - sparse_categorical_accuracy: 0.9000
Epoch 75/500
4/4 [==============================] - 0s 665us/step - loss: 0.3710 - sparse_categorical_accuracy: 0.9417
Epoch 76/500
4/4 [==============================] - 0s 997us/step - loss: 0.3799 - sparse_categorical_accuracy: 0.9250
Epoch 77/500
4/4 [==============================] - 0s 997us/step - loss: 0.4709 - sparse_categorical_accuracy: 0.8000
Epoch 78/500
4/4 [==============================] - 0s 665us/step - loss: 0.7460 - sparse_categorical_accuracy: 0.6500
Epoch 79/500
4/4 [==============================] - 0s 998us/step - loss: 0.4508 - sparse_categorical_accuracy: 0.8167
Epoch 80/500
4/4 [==============================] - 0s 6ms/step - loss: 0.4219 - sparse_categorical_accuracy: 0.9000 - val_loss: 0.3497 - val_sparse_categorical_accuracy: 0.9667
Epoch 81/500
4/4 [==============================] - 0s 665us/step - loss: 0.3777 - sparse_categorical_accuracy: 0.9333
Epoch 82/500
4/4 [==============================] - 0s 665us/step - loss: 0.3703 - sparse_categorical_accuracy: 0.9583
Epoch 83/500
4/4 [==============================] - 0s 665us/step - loss: 0.3600 - sparse_categorical_accuracy: 0.9667
Epoch 84/500
4/4 [==============================] - 0s 1ms/step - loss: 0.3638 - sparse_categorical_accuracy: 0.9333
Epoch 85/500
4/4 [==============================] - 0s 665us/step - loss: 0.3707 - sparse_categorical_accuracy: 0.9583
Epoch 86/500
4/4 [==============================] - 0s 665us/step - loss: 0.3845 - sparse_categorical_accuracy: 0.9250
Epoch 87/500
4/4 [==============================] - 0s 998us/step - loss: 0.3888 - sparse_categorical_accuracy: 0.8917
Epoch 88/500
4/4 [==============================] - 0s 665us/step - loss: 0.3771 - sparse_categorical_accuracy: 0.9500
Epoch 89/500
4/4 [==============================] - 0s 665us/step - loss: 0.5318 - sparse_categorical_accuracy: 0.7750
Epoch 90/500
4/4 [==============================] - 0s 665us/step - loss: 0.3822 - sparse_categorical_accuracy: 0.9250
Epoch 91/500
4/4 [==============================] - 0s 998us/step - loss: 0.4168 - sparse_categorical_accuracy: 0.9000
Epoch 92/500
4/4 [==============================] - 0s 665us/step - loss: 0.5004 - sparse_categorical_accuracy: 0.8000
Epoch 93/500
4/4 [==============================] - 0s 665us/step - loss: 0.3762 - sparse_categorical_accuracy: 0.9167
Epoch 94/500
4/4 [==============================] - 0s 998us/step - loss: 0.3885 - sparse_categorical_accuracy: 0.9333
Epoch 95/500
4/4 [==============================] - 0s 665us/step - loss: 0.3859 - sparse_categorical_accuracy: 0.8917
Epoch 96/500
4/4 [==============================] - 0s 998us/step - loss: 0.5500 - sparse_categorical_accuracy: 0.7667
Epoch 97/500
4/4 [==============================] - 0s 997us/step - loss: 0.4690 - sparse_categorical_accuracy: 0.8500
Epoch 98/500
4/4 [==============================] - 0s 997us/step - loss: 0.4850 - sparse_categorical_accuracy: 0.8083
Epoch 99/500
4/4 [==============================] - 0s 997us/step - loss: 0.3836 - sparse_categorical_accuracy: 0.9333
Epoch 100/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3713 - sparse_categorical_accuracy: 0.9167 - val_loss: 0.4250 - val_sparse_categorical_accuracy: 0.9333
Epoch 101/500
4/4 [==============================] - 0s 665us/step - loss: 0.3641 - sparse_categorical_accuracy: 0.9500
Epoch 102/500
4/4 [==============================] - 0s 665us/step - loss: 0.3650 - sparse_categorical_accuracy: 0.9583
Epoch 103/500
4/4 [==============================] - 0s 665us/step - loss: 0.3610 - sparse_categorical_accuracy: 0.9500
Epoch 104/500
4/4 [==============================] - 0s 665us/step - loss: 0.3771 - sparse_categorical_accuracy: 0.8917
Epoch 105/500
4/4 [==============================] - 0s 997us/step - loss: 0.4734 - sparse_categorical_accuracy: 0.8250
Epoch 106/500
4/4 [==============================] - 0s 664us/step - loss: 0.4102 - sparse_categorical_accuracy: 0.8917
Epoch 107/500
4/4 [==============================] - 0s 997us/step - loss: 0.3745 - sparse_categorical_accuracy: 0.9333
Epoch 108/500
4/4 [==============================] - 0s 997us/step - loss: 0.3834 - sparse_categorical_accuracy: 0.9000
Epoch 109/500
4/4 [==============================] - 0s 664us/step - loss: 0.3621 - sparse_categorical_accuracy: 0.9417
Epoch 110/500
4/4 [==============================] - 0s 664us/step - loss: 0.3798 - sparse_categorical_accuracy: 0.9167
Epoch 111/500
4/4 [==============================] - 0s 998us/step - loss: 0.3572 - sparse_categorical_accuracy: 0.9750
Epoch 112/500
4/4 [==============================] - 0s 664us/step - loss: 0.3665 - sparse_categorical_accuracy: 0.9333
Epoch 113/500
4/4 [==============================] - 0s 997us/step - loss: 0.4224 - sparse_categorical_accuracy: 0.8833
Epoch 114/500
4/4 [==============================] - 0s 665us/step - loss: 0.3616 - sparse_categorical_accuracy: 0.9500
Epoch 115/500
4/4 [==============================] - 0s 998us/step - loss: 0.3762 - sparse_categorical_accuracy: 0.9250
Epoch 116/500
4/4 [==============================] - 0s 997us/step - loss: 0.3832 - sparse_categorical_accuracy: 0.8833
Epoch 117/500
4/4 [==============================] - 0s 665us/step - loss: 0.5093 - sparse_categorical_accuracy: 0.8167
Epoch 118/500
4/4 [==============================] - 0s 998us/step - loss: 0.3643 - sparse_categorical_accuracy: 0.9333
Epoch 119/500
4/4 [==============================] - 0s 998us/step - loss: 0.3593 - sparse_categorical_accuracy: 0.9417
Epoch 120/500
4/4 [==============================] - 0s 6ms/step - loss: 0.4207 - sparse_categorical_accuracy: 0.8833 - val_loss: 0.3512 - val_sparse_categorical_accuracy: 1.0000
Epoch 121/500
4/4 [==============================] - 0s 665us/step - loss: 0.4096 - sparse_categorical_accuracy: 0.9250
Epoch 122/500
4/4 [==============================] - 0s 665us/step - loss: 0.4189 - sparse_categorical_accuracy: 0.8833
Epoch 123/500
4/4 [==============================] - 0s 997us/step - loss: 0.3683 - sparse_categorical_accuracy: 0.9250
Epoch 124/500
4/4 [==============================] - 0s 665us/step - loss: 0.3650 - sparse_categorical_accuracy: 0.9167
Epoch 125/500
4/4 [==============================] - 0s 997us/step - loss: 0.3710 - sparse_categorical_accuracy: 0.9417
Epoch 126/500
4/4 [==============================] - 0s 997us/step - loss: 0.3914 - sparse_categorical_accuracy: 0.9000
Epoch 127/500
4/4 [==============================] - 0s 998us/step - loss: 0.3756 - sparse_categorical_accuracy: 0.9333
Epoch 128/500
4/4 [==============================] - 0s 998us/step - loss: 0.3782 - sparse_categorical_accuracy: 0.9083
Epoch 129/500
4/4 [==============================] - 0s 665us/step - loss: 0.3818 - sparse_categorical_accuracy: 0.9167
Epoch 130/500
4/4 [==============================] - 0s 665us/step - loss: 0.4638 - sparse_categorical_accuracy: 0.8500
Epoch 131/500
4/4 [==============================] - 0s 665us/step - loss: 0.4299 - sparse_categorical_accuracy: 0.8667
Epoch 132/500
4/4 [==============================] - 0s 665us/step - loss: 0.3521 - sparse_categorical_accuracy: 0.9750
Epoch 133/500
4/4 [==============================] - 0s 665us/step - loss: 0.3520 - sparse_categorical_accuracy: 0.9667
Epoch 134/500
4/4 [==============================] - 0s 665us/step - loss: 0.3700 - sparse_categorical_accuracy: 0.9333
Epoch 135/500
4/4 [==============================] - 0s 997us/step - loss: 0.3735 - sparse_categorical_accuracy: 0.9333
Epoch 136/500
4/4 [==============================] - 0s 665us/step - loss: 0.3584 - sparse_categorical_accuracy: 0.9333
Epoch 137/500
4/4 [==============================] - 0s 665us/step - loss: 0.4001 - sparse_categorical_accuracy: 0.9083
Epoch 138/500
4/4 [==============================] - 0s 665us/step - loss: 0.3530 - sparse_categorical_accuracy: 0.9583
Epoch 139/500
4/4 [==============================] - 0s 998us/step - loss: 0.4089 - sparse_categorical_accuracy: 0.9000
Epoch 140/500
4/4 [==============================] - 0s 5ms/step - loss: 0.5121 - sparse_categorical_accuracy: 0.7750 - val_loss: 0.3758 - val_sparse_categorical_accuracy: 1.0000
Epoch 141/500
4/4 [==============================] - 0s 665us/step - loss: 0.3577 - sparse_categorical_accuracy: 0.9500
Epoch 142/500
4/4 [==============================] - 0s 665us/step - loss: 0.3597 - sparse_categorical_accuracy: 0.9583
Epoch 143/500
4/4 [==============================] - 0s 665us/step - loss: 0.3505 - sparse_categorical_accuracy: 0.9417
Epoch 144/500
4/4 [==============================] - 0s 998us/step - loss: 0.3728 - sparse_categorical_accuracy: 0.9250
Epoch 145/500
4/4 [==============================] - 0s 998us/step - loss: 0.3780 - sparse_categorical_accuracy: 0.9083
Epoch 146/500
4/4 [==============================] - 0s 998us/step - loss: 0.3543 - sparse_categorical_accuracy: 0.9583
Epoch 147/500
4/4 [==============================] - 0s 998us/step - loss: 0.4474 - sparse_categorical_accuracy: 0.8583
Epoch 148/500
4/4 [==============================] - 0s 997us/step - loss: 0.4703 - sparse_categorical_accuracy: 0.8500
Epoch 149/500
4/4 [==============================] - 0s 665us/step - loss: 0.3700 - sparse_categorical_accuracy: 0.9083
Epoch 150/500
4/4 [==============================] - 0s 998us/step - loss: 0.4425 - sparse_categorical_accuracy: 0.8417
Epoch 151/500
4/4 [==============================] - 0s 998us/step - loss: 0.4274 - sparse_categorical_accuracy: 0.8417
Epoch 152/500
4/4 [==============================] - 0s 997us/step - loss: 0.3724 - sparse_categorical_accuracy: 0.9000
Epoch 153/500
4/4 [==============================] - 0s 997us/step - loss: 0.4161 - sparse_categorical_accuracy: 0.8750
Epoch 154/500
4/4 [==============================] - 0s 665us/step - loss: 0.3902 - sparse_categorical_accuracy: 0.9083
Epoch 155/500
4/4 [==============================] - 0s 998us/step - loss: 0.3843 - sparse_categorical_accuracy: 0.8917
Epoch 156/500
4/4 [==============================] - 0s 664us/step - loss: 0.3883 - sparse_categorical_accuracy: 0.9000
Epoch 157/500
4/4 [==============================] - 0s 665us/step - loss: 0.3966 - sparse_categorical_accuracy: 0.8667
Epoch 158/500
4/4 [==============================] - 0s 665us/step - loss: 0.3617 - sparse_categorical_accuracy: 0.9417
Epoch 159/500
4/4 [==============================] - 0s 997us/step - loss: 0.4144 - sparse_categorical_accuracy: 0.8833
Epoch 160/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3978 - sparse_categorical_accuracy: 0.9333 - val_loss: 0.4238 - val_sparse_categorical_accuracy: 0.9333
Epoch 161/500
4/4 [==============================] - 0s 998us/step - loss: 0.3620 - sparse_categorical_accuracy: 0.9417
Epoch 162/500
4/4 [==============================] - 0s 998us/step - loss: 0.3863 - sparse_categorical_accuracy: 0.9083
Epoch 163/500
4/4 [==============================] - 0s 997us/step - loss: 0.3635 - sparse_categorical_accuracy: 0.9250
Epoch 164/500
4/4 [==============================] - 0s 997us/step - loss: 0.6862 - sparse_categorical_accuracy: 0.7500
Epoch 165/500
4/4 [==============================] - 0s 665us/step - loss: 0.4336 - sparse_categorical_accuracy: 0.8833
Epoch 166/500
4/4 [==============================] - 0s 665us/step - loss: 0.4403 - sparse_categorical_accuracy: 0.8583
Epoch 167/500
4/4 [==============================] - 0s 998us/step - loss: 0.3904 - sparse_categorical_accuracy: 0.9083
Epoch 168/500
4/4 [==============================] - 0s 665us/step - loss: 0.3524 - sparse_categorical_accuracy: 0.9583
Epoch 169/500
4/4 [==============================] - 0s 997us/step - loss: 0.3602 - sparse_categorical_accuracy: 0.9667
Epoch 170/500
4/4 [==============================] - 0s 665us/step - loss: 0.3542 - sparse_categorical_accuracy: 0.9583
Epoch 171/500
4/4 [==============================] - 0s 997us/step - loss: 0.4699 - sparse_categorical_accuracy: 0.8083
Epoch 172/500
4/4 [==============================] - 0s 665us/step - loss: 0.3456 - sparse_categorical_accuracy: 0.9250
Epoch 173/500
4/4 [==============================] - 0s 665us/step - loss: 0.3528 - sparse_categorical_accuracy: 0.9583
Epoch 174/500
4/4 [==============================] - 0s 665us/step - loss: 0.3801 - sparse_categorical_accuracy: 0.9250
Epoch 175/500
4/4 [==============================] - 0s 998us/step - loss: 0.3921 - sparse_categorical_accuracy: 0.9000
Epoch 176/500
4/4 [==============================] - 0s 665us/step - loss: 0.3875 - sparse_categorical_accuracy: 0.9083
Epoch 177/500
4/4 [==============================] - 0s 998us/step - loss: 0.4096 - sparse_categorical_accuracy: 0.9083
Epoch 178/500
4/4 [==============================] - 0s 665us/step - loss: 0.4523 - sparse_categorical_accuracy: 0.8583
Epoch 179/500
4/4 [==============================] - 0s 997us/step - loss: 0.3720 - sparse_categorical_accuracy: 0.9167
Epoch 180/500
4/4 [==============================] - 0s 6ms/step - loss: 0.4053 - sparse_categorical_accuracy: 0.9000 - val_loss: 0.3892 - val_sparse_categorical_accuracy: 1.0000
Epoch 181/500
4/4 [==============================] - 0s 665us/step - loss: 0.3602 - sparse_categorical_accuracy: 0.9417
Epoch 182/500
4/4 [==============================] - 0s 665us/step - loss: 0.4059 - sparse_categorical_accuracy: 0.8917
Epoch 183/500
4/4 [==============================] - 0s 998us/step - loss: 0.3556 - sparse_categorical_accuracy: 0.9333
Epoch 184/500
4/4 [==============================] - 0s 665us/step - loss: 0.4176 - sparse_categorical_accuracy: 0.8833
Epoch 185/500
4/4 [==============================] - 0s 665us/step - loss: 0.3594 - sparse_categorical_accuracy: 0.9500
Epoch 186/500
4/4 [==============================] - 0s 665us/step - loss: 0.3632 - sparse_categorical_accuracy: 0.9417
Epoch 187/500
4/4 [==============================] - 0s 665us/step - loss: 0.4102 - sparse_categorical_accuracy: 0.9083
Epoch 188/500
4/4 [==============================] - 0s 665us/step - loss: 0.3549 - sparse_categorical_accuracy: 0.9500
Epoch 189/500
4/4 [==============================] - 0s 665us/step - loss: 0.3533 - sparse_categorical_accuracy: 0.9583
Epoch 190/500
4/4 [==============================] - 0s 998us/step - loss: 0.3974 - sparse_categorical_accuracy: 0.8917
Epoch 191/500
4/4 [==============================] - 0s 665us/step - loss: 0.3645 - sparse_categorical_accuracy: 0.9333
Epoch 192/500
4/4 [==============================] - 0s 665us/step - loss: 0.3784 - sparse_categorical_accuracy: 0.9333
Epoch 193/500
4/4 [==============================] - 0s 665us/step - loss: 0.4510 - sparse_categorical_accuracy: 0.8417
Epoch 194/500
4/4 [==============================] - 0s 997us/step - loss: 0.3579 - sparse_categorical_accuracy: 0.9583
Epoch 195/500
4/4 [==============================] - 0s 665us/step - loss: 0.3572 - sparse_categorical_accuracy: 0.9583
Epoch 196/500
4/4 [==============================] - 0s 665us/step - loss: 0.3782 - sparse_categorical_accuracy: 0.9333
Epoch 197/500
4/4 [==============================] - 0s 665us/step - loss: 0.3525 - sparse_categorical_accuracy: 0.9583
Epoch 198/500
4/4 [==============================] - 0s 998us/step - loss: 0.3499 - sparse_categorical_accuracy: 0.9583
Epoch 199/500
4/4 [==============================] - 0s 665us/step - loss: 0.3540 - sparse_categorical_accuracy: 0.9417
Epoch 200/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3628 - sparse_categorical_accuracy: 0.9333 - val_loss: 0.4160 - val_sparse_categorical_accuracy: 0.9333
Epoch 201/500
4/4 [==============================] - 0s 998us/step - loss: 0.3575 - sparse_categorical_accuracy: 0.9417
Epoch 202/500
4/4 [==============================] - 0s 665us/step - loss: 0.4148 - sparse_categorical_accuracy: 0.8750
Epoch 203/500
4/4 [==============================] - 0s 665us/step - loss: 0.3930 - sparse_categorical_accuracy: 0.9000
Epoch 204/500
4/4 [==============================] - 0s 998us/step - loss: 0.3465 - sparse_categorical_accuracy: 0.9750
Epoch 205/500
4/4 [==============================] - 0s 997us/step - loss: 0.3928 - sparse_categorical_accuracy: 0.9000
Epoch 206/500
4/4 [==============================] - 0s 998us/step - loss: 0.4302 - sparse_categorical_accuracy: 0.8417
Epoch 207/500
4/4 [==============================] - 0s 665us/step - loss: 0.5203 - sparse_categorical_accuracy: 0.7750
Epoch 208/500
4/4 [==============================] - 0s 665us/step - loss: 0.3765 - sparse_categorical_accuracy: 0.9250
Epoch 209/500
4/4 [==============================] - 0s 998us/step - loss: 0.3803 - sparse_categorical_accuracy: 0.9167
Epoch 210/500
4/4 [==============================] - 0s 665us/step - loss: 0.3826 - sparse_categorical_accuracy: 0.9167
Epoch 211/500
4/4 [==============================] - 0s 665us/step - loss: 0.4507 - sparse_categorical_accuracy: 0.8250
Epoch 212/500
4/4 [==============================] - 0s 665us/step - loss: 0.3604 - sparse_categorical_accuracy: 0.9000
Epoch 213/500
4/4 [==============================] - 0s 997us/step - loss: 0.3522 - sparse_categorical_accuracy: 0.9333
Epoch 214/500
4/4 [==============================] - 0s 998us/step - loss: 0.3819 - sparse_categorical_accuracy: 0.9250
Epoch 215/500
4/4 [==============================] - 0s 665us/step - loss: 0.3474 - sparse_categorical_accuracy: 0.9667
Epoch 216/500
4/4 [==============================] - 0s 665us/step - loss: 0.3784 - sparse_categorical_accuracy: 0.9000
Epoch 217/500
4/4 [==============================] - 0s 998us/step - loss: 0.3553 - sparse_categorical_accuracy: 0.9750
Epoch 218/500
4/4 [==============================] - 0s 998us/step - loss: 0.3507 - sparse_categorical_accuracy: 0.9583
Epoch 219/500
4/4 [==============================] - 0s 665us/step - loss: 0.3568 - sparse_categorical_accuracy: 0.9417
Epoch 220/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3399 - sparse_categorical_accuracy: 0.9583 - val_loss: 0.3539 - val_sparse_categorical_accuracy: 0.8667
Epoch 221/500
4/4 [==============================] - 0s 998us/step - loss: 0.4177 - sparse_categorical_accuracy: 0.8917
Epoch 222/500
4/4 [==============================] - 0s 665us/step - loss: 0.4743 - sparse_categorical_accuracy: 0.8500
Epoch 223/500
4/4 [==============================] - 0s 665us/step - loss: 0.3757 - sparse_categorical_accuracy: 0.9333
Epoch 224/500
4/4 [==============================] - 0s 665us/step - loss: 0.3509 - sparse_categorical_accuracy: 0.9333
Epoch 225/500
4/4 [==============================] - 0s 997us/step - loss: 0.4814 - sparse_categorical_accuracy: 0.8250
Epoch 226/500
4/4 [==============================] - 0s 998us/step - loss: 0.3776 - sparse_categorical_accuracy: 0.9083
Epoch 227/500
4/4 [==============================] - 0s 997us/step - loss: 0.3909 - sparse_categorical_accuracy: 0.9250
Epoch 228/500
4/4 [==============================] - 0s 997us/step - loss: 0.3548 - sparse_categorical_accuracy: 0.9583
Epoch 229/500
4/4 [==============================] - 0s 998us/step - loss: 0.3536 - sparse_categorical_accuracy: 0.9417
Epoch 230/500
4/4 [==============================] - 0s 665us/step - loss: 0.3603 - sparse_categorical_accuracy: 0.9500
Epoch 231/500
4/4 [==============================] - 0s 665us/step - loss: 0.3909 - sparse_categorical_accuracy: 0.9083
Epoch 232/500
4/4 [==============================] - 0s 998us/step - loss: 0.6127 - sparse_categorical_accuracy: 0.7583
Epoch 233/500
4/4 [==============================] - 0s 665us/step - loss: 0.3575 - sparse_categorical_accuracy: 0.9417
Epoch 234/500
4/4 [==============================] - 0s 665us/step - loss: 0.3649 - sparse_categorical_accuracy: 0.9333
Epoch 235/500
4/4 [==============================] - 0s 665us/step - loss: 0.3872 - sparse_categorical_accuracy: 0.9333
Epoch 236/500
4/4 [==============================] - 0s 997us/step - loss: 0.3888 - sparse_categorical_accuracy: 0.9000
Epoch 237/500
4/4 [==============================] - 0s 665us/step - loss: 0.3783 - sparse_categorical_accuracy: 0.9083
Epoch 238/500
4/4 [==============================] - 0s 665us/step - loss: 0.3597 - sparse_categorical_accuracy: 0.9250
Epoch 239/500
4/4 [==============================] - 0s 665us/step - loss: 0.3628 - sparse_categorical_accuracy: 0.9417
Epoch 240/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3752 - sparse_categorical_accuracy: 0.9250 - val_loss: 0.3356 - val_sparse_categorical_accuracy: 1.0000
Epoch 241/500
4/4 [==============================] - 0s 665us/step - loss: 0.3523 - sparse_categorical_accuracy: 0.9583
Epoch 242/500
4/4 [==============================] - 0s 665us/step - loss: 0.4029 - sparse_categorical_accuracy: 0.9000
Epoch 243/500
4/4 [==============================] - 0s 665us/step - loss: 0.5745 - sparse_categorical_accuracy: 0.7667
Epoch 244/500
4/4 [==============================] - 0s 665us/step - loss: 0.3475 - sparse_categorical_accuracy: 0.9583
Epoch 245/500
4/4 [==============================] - 0s 665us/step - loss: 0.4219 - sparse_categorical_accuracy: 0.8667
Epoch 246/500
4/4 [==============================] - 0s 665us/step - loss: 0.3892 - sparse_categorical_accuracy: 0.9167
Epoch 247/500
4/4 [==============================] - 0s 665us/step - loss: 0.4275 - sparse_categorical_accuracy: 0.8667
Epoch 248/500
4/4 [==============================] - 0s 665us/step - loss: 0.3419 - sparse_categorical_accuracy: 0.9583
Epoch 249/500
4/4 [==============================] - 0s 665us/step - loss: 0.3683 - sparse_categorical_accuracy: 0.9250
Epoch 250/500
4/4 [==============================] - 0s 665us/step - loss: 0.3579 - sparse_categorical_accuracy: 0.9417
Epoch 251/500
4/4 [==============================] - 0s 998us/step - loss: 0.3651 - sparse_categorical_accuracy: 0.9583
Epoch 252/500
4/4 [==============================] - 0s 998us/step - loss: 0.3562 - sparse_categorical_accuracy: 0.9417
Epoch 253/500
4/4 [==============================] - 0s 665us/step - loss: 0.4366 - sparse_categorical_accuracy: 0.8667
Epoch 254/500
4/4 [==============================] - 0s 997us/step - loss: 0.4126 - sparse_categorical_accuracy: 0.8667
Epoch 255/500
4/4 [==============================] - 0s 665us/step - loss: 0.4202 - sparse_categorical_accuracy: 0.8833
Epoch 256/500
4/4 [==============================] - 0s 997us/step - loss: 0.4353 - sparse_categorical_accuracy: 0.8667
Epoch 257/500
4/4 [==============================] - 0s 998us/step - loss: 0.3604 - sparse_categorical_accuracy: 0.9333
Epoch 258/500
4/4 [==============================] - 0s 665us/step - loss: 0.3504 - sparse_categorical_accuracy: 0.9583
Epoch 259/500
4/4 [==============================] - 0s 998us/step - loss: 0.3866 - sparse_categorical_accuracy: 0.9083
Epoch 260/500
4/4 [==============================] - 0s 6ms/step - loss: 0.3516 - sparse_categorical_accuracy: 0.9417 - val_loss: 0.3398 - val_sparse_categorical_accuracy: 1.0000
Epoch 261/500
4/4 [==============================] - 0s 997us/step - loss: 0.3769 - sparse_categorical_accuracy: 0.9333
Epoch 262/500
4/4 [==============================] - 0s 665us/step - loss: 0.3620 - sparse_categorical_accuracy: 0.9333
Epoch 263/500
4/4 [==============================] - 0s 665us/step - loss: 0.3474 - sparse_categorical_accuracy: 0.9667
Epoch 264/500
4/4 [==============================] - 0s 665us/step - loss: 0.3434 - sparse_categorical_accuracy: 0.9583
Epoch 265/500
4/4 [==============================] - 0s 998us/step - loss: 0.3778 - sparse_categorical_accuracy: 0.9167
Epoch 266/500
4/4 [==============================] - 0s 665us/step - loss: 0.3515 - sparse_categorical_accuracy: 0.9500
Epoch 267/500
4/4 [==============================] - 0s 665us/step - loss: 0.3481 - sparse_categorical_accuracy: 0.9667
Epoch 268/500
4/4 [==============================] - 0s 665us/step - loss: 0.3558 - sparse_categorical_accuracy: 0.9333
Epoch 269/500
4/4 [==============================] - 0s 665us/step - loss: 0.4076 - sparse_categorical_accuracy: 0.8750
Epoch 270/500
4/4 [==============================] - 0s 665us/step - loss: 0.4119 - sparse_categorical_accuracy: 0.9000
Epoch 271/500
4/4 [==============================] - 0s 665us/step - loss: 0.3590 - sparse_categorical_accuracy: 0.9417
Epoch 272/500
4/4 [==============================] - 0s 998us/step - loss: 0.3555 - sparse_categorical_accuracy: 0.9583
Epoch 273/500
4/4 [==============================] - 0s 665us/step - loss: 0.3999 - sparse_categorical_accuracy: 0.8833
Epoch 274/500
4/4 [==============================] - 0s 665us/step - loss: 0.3456 - sparse_categorical_accuracy: 0.9750
Epoch 275/500
4/4 [==============================] - 0s 665us/step - loss: 0.3543 - sparse_categorical_accuracy: 0.9333
Epoch 276/500
4/4 [==============================] - 0s 998us/step - loss: 0.3608 - sparse_categorical_accuracy: 0.9500
Epoch 277/500
4/4 [==============================] - 0s 665us/step - loss: 0.4091 - sparse_categorical_accuracy: 0.8833
Epoch 278/500
4/4 [==============================] - 0s 665us/step - loss: 0.3711 - sparse_categorical_accuracy: 0.9500
Epoch 279/500
4/4 [==============================] - 0s 665us/step - loss: 0.3670 - sparse_categorical_accuracy: 0.9167
Epoch 280/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3592 - sparse_categorical_accuracy: 0.9500 - val_loss: 0.3251 - val_sparse_categorical_accuracy: 0.9667
Epoch 281/500
4/4 [==============================] - 0s 997us/step - loss: 0.3522 - sparse_categorical_accuracy: 0.9750
Epoch 282/500
4/4 [==============================] - 0s 665us/step - loss: 0.3495 - sparse_categorical_accuracy: 0.9417
Epoch 283/500
4/4 [==============================] - 0s 665us/step - loss: 0.3462 - sparse_categorical_accuracy: 0.9750
Epoch 284/500
4/4 [==============================] - 0s 997us/step - loss: 0.3948 - sparse_categorical_accuracy: 0.9167
Epoch 285/500
4/4 [==============================] - 0s 665us/step - loss: 0.4060 - sparse_categorical_accuracy: 0.8833
Epoch 286/500
4/4 [==============================] - 0s 997us/step - loss: 0.3834 - sparse_categorical_accuracy: 0.9167
Epoch 287/500
4/4 [==============================] - 0s 665us/step - loss: 0.3814 - sparse_categorical_accuracy: 0.9000
Epoch 288/500
4/4 [==============================] - 0s 997us/step - loss: 0.3698 - sparse_categorical_accuracy: 0.9250
Epoch 289/500
4/4 [==============================] - 0s 997us/step - loss: 0.5091 - sparse_categorical_accuracy: 0.8083
Epoch 290/500
4/4 [==============================] - 0s 665us/step - loss: 0.3431 - sparse_categorical_accuracy: 0.9417
Epoch 291/500
4/4 [==============================] - 0s 665us/step - loss: 0.3418 - sparse_categorical_accuracy: 0.9667
Epoch 292/500
4/4 [==============================] - 0s 997us/step - loss: 0.3627 - sparse_categorical_accuracy: 0.9250
Epoch 293/500
4/4 [==============================] - 0s 665us/step - loss: 0.3739 - sparse_categorical_accuracy: 0.9000
Epoch 294/500
4/4 [==============================] - 0s 665us/step - loss: 0.3468 - sparse_categorical_accuracy: 0.9667
Epoch 295/500
4/4 [==============================] - 0s 665us/step - loss: 0.4111 - sparse_categorical_accuracy: 0.8583
Epoch 296/500
4/4 [==============================] - 0s 997us/step - loss: 0.3469 - sparse_categorical_accuracy: 0.9333
Epoch 297/500
4/4 [==============================] - 0s 997us/step - loss: 0.3459 - sparse_categorical_accuracy: 0.9333
Epoch 298/500
4/4 [==============================] - 0s 997us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9583
Epoch 299/500
4/4 [==============================] - 0s 997us/step - loss: 0.3559 - sparse_categorical_accuracy: 0.9417
Epoch 300/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3615 - sparse_categorical_accuracy: 0.9417 - val_loss: 0.3242 - val_sparse_categorical_accuracy: 0.9667
Epoch 301/500
4/4 [==============================] - 0s 665us/step - loss: 0.3921 - sparse_categorical_accuracy: 0.9250
Epoch 302/500
4/4 [==============================] - 0s 665us/step - loss: 0.3612 - sparse_categorical_accuracy: 0.9167
Epoch 303/500
4/4 [==============================] - 0s 665us/step - loss: 0.3446 - sparse_categorical_accuracy: 0.9583
Epoch 304/500
4/4 [==============================] - 0s 665us/step - loss: 0.3693 - sparse_categorical_accuracy: 0.9250
Epoch 305/500
4/4 [==============================] - 0s 997us/step - loss: 0.3616 - sparse_categorical_accuracy: 0.9500
Epoch 306/500
4/4 [==============================] - 0s 665us/step - loss: 0.3492 - sparse_categorical_accuracy: 0.9750
Epoch 307/500
4/4 [==============================] - 0s 665us/step - loss: 0.3634 - sparse_categorical_accuracy: 0.9417
Epoch 308/500
4/4 [==============================] - 0s 997us/step - loss: 0.3407 - sparse_categorical_accuracy: 0.9500
Epoch 309/500
4/4 [==============================] - 0s 998us/step - loss: 0.3700 - sparse_categorical_accuracy: 0.8917
Epoch 310/500
4/4 [==============================] - 0s 998us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9750
Epoch 311/500
4/4 [==============================] - 0s 665us/step - loss: 0.3738 - sparse_categorical_accuracy: 0.8917
Epoch 312/500
4/4 [==============================] - 0s 998us/step - loss: 0.3417 - sparse_categorical_accuracy: 0.9583
Epoch 313/500
4/4 [==============================] - 0s 998us/step - loss: 0.3499 - sparse_categorical_accuracy: 0.9417
Epoch 314/500
4/4 [==============================] - 0s 665us/step - loss: 0.3756 - sparse_categorical_accuracy: 0.9000
Epoch 315/500
4/4 [==============================] - 0s 665us/step - loss: 0.3532 - sparse_categorical_accuracy: 0.9417
Epoch 316/500
4/4 [==============================] - 0s 998us/step - loss: 0.4007 - sparse_categorical_accuracy: 0.8917
Epoch 317/500
4/4 [==============================] - 0s 665us/step - loss: 0.3593 - sparse_categorical_accuracy: 0.9417
Epoch 318/500
4/4 [==============================] - 0s 665us/step - loss: 0.3408 - sparse_categorical_accuracy: 0.9583
Epoch 319/500
4/4 [==============================] - 0s 665us/step - loss: 0.4063 - sparse_categorical_accuracy: 0.8833
Epoch 320/500
4/4 [==============================] - 0s 5ms/step - loss: 0.5146 - sparse_categorical_accuracy: 0.7917 - val_loss: 0.3246 - val_sparse_categorical_accuracy: 1.0000
Epoch 321/500
4/4 [==============================] - 0s 998us/step - loss: 0.3707 - sparse_categorical_accuracy: 0.9083
Epoch 322/500
4/4 [==============================] - 0s 665us/step - loss: 0.3414 - sparse_categorical_accuracy: 0.9750
Epoch 323/500
4/4 [==============================] - 0s 665us/step - loss: 0.3500 - sparse_categorical_accuracy: 0.9500
Epoch 324/500
4/4 [==============================] - 0s 665us/step - loss: 0.3485 - sparse_categorical_accuracy: 0.9500
Epoch 325/500
4/4 [==============================] - 0s 997us/step - loss: 0.4369 - sparse_categorical_accuracy: 0.8583
Epoch 326/500
4/4 [==============================] - 0s 665us/step - loss: 0.3594 - sparse_categorical_accuracy: 0.9500
Epoch 327/500
4/4 [==============================] - 0s 665us/step - loss: 0.3559 - sparse_categorical_accuracy: 0.9417
Epoch 328/500
4/4 [==============================] - 0s 665us/step - loss: 0.4150 - sparse_categorical_accuracy: 0.8917
Epoch 329/500
4/4 [==============================] - 0s 665us/step - loss: 0.3804 - sparse_categorical_accuracy: 0.9167
Epoch 330/500
4/4 [==============================] - 0s 665us/step - loss: 0.3942 - sparse_categorical_accuracy: 0.8917
Epoch 331/500
4/4 [==============================] - 0s 665us/step - loss: 0.3654 - sparse_categorical_accuracy: 0.9417
Epoch 332/500
4/4 [==============================] - 0s 997us/step - loss: 0.3498 - sparse_categorical_accuracy: 0.9500
Epoch 333/500
4/4 [==============================] - 0s 665us/step - loss: 0.3528 - sparse_categorical_accuracy: 0.9500
Epoch 334/500
4/4 [==============================] - 0s 665us/step - loss: 0.3459 - sparse_categorical_accuracy: 0.9750
Epoch 335/500
4/4 [==============================] - 0s 665us/step - loss: 0.3427 - sparse_categorical_accuracy: 0.9417
Epoch 336/500
4/4 [==============================] - 0s 997us/step - loss: 0.3421 - sparse_categorical_accuracy: 0.9667
Epoch 337/500
4/4 [==============================] - 0s 665us/step - loss: 0.3449 - sparse_categorical_accuracy: 0.9417
Epoch 338/500
4/4 [==============================] - 0s 997us/step - loss: 0.4189 - sparse_categorical_accuracy: 0.8583
Epoch 339/500
4/4 [==============================] - 0s 665us/step - loss: 0.3469 - sparse_categorical_accuracy: 0.9500
Epoch 340/500
4/4 [==============================] - 0s 5ms/step - loss: 0.4089 - sparse_categorical_accuracy: 0.8583 - val_loss: 0.3289 - val_sparse_categorical_accuracy: 1.0000
Epoch 341/500
4/4 [==============================] - 0s 665us/step - loss: 0.3863 - sparse_categorical_accuracy: 0.9083
Epoch 342/500
4/4 [==============================] - 0s 665us/step - loss: 0.3991 - sparse_categorical_accuracy: 0.8833
Epoch 343/500
4/4 [==============================] - 0s 665us/step - loss: 0.4232 - sparse_categorical_accuracy: 0.8667
Epoch 344/500
4/4 [==============================] - 0s 665us/step - loss: 0.4240 - sparse_categorical_accuracy: 0.8667
Epoch 345/500
4/4 [==============================] - 0s 665us/step - loss: 0.3626 - sparse_categorical_accuracy: 0.9500
Epoch 346/500
4/4 [==============================] - 0s 665us/step - loss: 0.3398 - sparse_categorical_accuracy: 0.9583
Epoch 347/500
4/4 [==============================] - 0s 665us/step - loss: 0.3416 - sparse_categorical_accuracy: 0.9250
Epoch 348/500
4/4 [==============================] - 0s 998us/step - loss: 0.3892 - sparse_categorical_accuracy: 0.9083
Epoch 349/500
4/4 [==============================] - 0s 997us/step - loss: 0.3674 - sparse_categorical_accuracy: 0.9167
Epoch 350/500
4/4 [==============================] - 0s 998us/step - loss: 0.3400 - sparse_categorical_accuracy: 0.9667
Epoch 351/500
4/4 [==============================] - 0s 997us/step - loss: 0.3556 - sparse_categorical_accuracy: 0.9500
Epoch 352/500
4/4 [==============================] - 0s 998us/step - loss: 0.3704 - sparse_categorical_accuracy: 0.9500
Epoch 353/500
4/4 [==============================] - 0s 665us/step - loss: 0.3406 - sparse_categorical_accuracy: 0.9667
Epoch 354/500
4/4 [==============================] - 0s 998us/step - loss: 0.3330 - sparse_categorical_accuracy: 0.9500
Epoch 355/500
4/4 [==============================] - 0s 997us/step - loss: 0.3984 - sparse_categorical_accuracy: 0.8833
Epoch 356/500
4/4 [==============================] - 0s 997us/step - loss: 0.4194 - sparse_categorical_accuracy: 0.8583
Epoch 357/500
4/4 [==============================] - 0s 665us/step - loss: 0.3721 - sparse_categorical_accuracy: 0.9250
Epoch 358/500
4/4 [==============================] - 0s 665us/step - loss: 0.3770 - sparse_categorical_accuracy: 0.9333
Epoch 359/500
4/4 [==============================] - 0s 665us/step - loss: 0.3631 - sparse_categorical_accuracy: 0.9167
Epoch 360/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3405 - sparse_categorical_accuracy: 0.9583 - val_loss: 0.3238 - val_sparse_categorical_accuracy: 0.9667
Epoch 361/500
4/4 [==============================] - 0s 665us/step - loss: 0.4157 - sparse_categorical_accuracy: 0.9000
Epoch 362/500
4/4 [==============================] - 0s 665us/step - loss: 0.3548 - sparse_categorical_accuracy: 0.9667
Epoch 363/500
4/4 [==============================] - 0s 997us/step - loss: 0.3715 - sparse_categorical_accuracy: 0.9417
Epoch 364/500
4/4 [==============================] - 0s 997us/step - loss: 0.4348 - sparse_categorical_accuracy: 0.8417
Epoch 365/500
4/4 [==============================] - 0s 665us/step - loss: 0.3671 - sparse_categorical_accuracy: 0.9333
Epoch 366/500
4/4 [==============================] - 0s 665us/step - loss: 0.3489 - sparse_categorical_accuracy: 0.9500
Epoch 367/500
4/4 [==============================] - 0s 665us/step - loss: 0.3407 - sparse_categorical_accuracy: 0.9500
Epoch 368/500
4/4 [==============================] - 0s 998us/step - loss: 0.3895 - sparse_categorical_accuracy: 0.9083
Epoch 369/500
4/4 [==============================] - 0s 665us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9583
Epoch 370/500
4/4 [==============================] - 0s 665us/step - loss: 0.3371 - sparse_categorical_accuracy: 0.9583
Epoch 371/500
4/4 [==============================] - 0s 665us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9667
Epoch 372/500
4/4 [==============================] - 0s 665us/step - loss: 0.3445 - sparse_categorical_accuracy: 0.9583
Epoch 373/500
4/4 [==============================] - 0s 665us/step - loss: 0.4199 - sparse_categorical_accuracy: 0.8833
Epoch 374/500
4/4 [==============================] - 0s 665us/step - loss: 0.3528 - sparse_categorical_accuracy: 0.9417
Epoch 375/500
4/4 [==============================] - 0s 665us/step - loss: 0.3791 - sparse_categorical_accuracy: 0.9250
Epoch 376/500
4/4 [==============================] - 0s 665us/step - loss: 0.3511 - sparse_categorical_accuracy: 0.9500
Epoch 377/500
4/4 [==============================] - 0s 665us/step - loss: 0.3391 - sparse_categorical_accuracy: 0.9667
Epoch 378/500
4/4 [==============================] - 0s 665us/step - loss: 0.3461 - sparse_categorical_accuracy: 0.9500
Epoch 379/500
4/4 [==============================] - 0s 665us/step - loss: 0.3603 - sparse_categorical_accuracy: 0.9667
Epoch 380/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3398 - sparse_categorical_accuracy: 0.9667 - val_loss: 0.3323 - val_sparse_categorical_accuracy: 1.0000
Epoch 381/500
4/4 [==============================] - 0s 998us/step - loss: 0.3799 - sparse_categorical_accuracy: 0.9167
Epoch 382/500
4/4 [==============================] - 0s 997us/step - loss: 0.3502 - sparse_categorical_accuracy: 0.9667
Epoch 383/500
4/4 [==============================] - 0s 997us/step - loss: 0.3963 - sparse_categorical_accuracy: 0.8833
Epoch 384/500
4/4 [==============================] - 0s 997us/step - loss: 0.3446 - sparse_categorical_accuracy: 0.9500
Epoch 385/500
4/4 [==============================] - 0s 997us/step - loss: 0.3970 - sparse_categorical_accuracy: 0.9250
Epoch 386/500
4/4 [==============================] - 0s 997us/step - loss: 0.3411 - sparse_categorical_accuracy: 0.9750
Epoch 387/500
4/4 [==============================] - 0s 998us/step - loss: 0.3672 - sparse_categorical_accuracy: 0.9167
Epoch 388/500
4/4 [==============================] - 0s 997us/step - loss: 0.3954 - sparse_categorical_accuracy: 0.9167
Epoch 389/500
4/4 [==============================] - 0s 997us/step - loss: 0.3850 - sparse_categorical_accuracy: 0.8750
Epoch 390/500
4/4 [==============================] - 0s 665us/step - loss: 0.4663 - sparse_categorical_accuracy: 0.8333
Epoch 391/500
4/4 [==============================] - 0s 998us/step - loss: 0.4209 - sparse_categorical_accuracy: 0.8583
Epoch 392/500
4/4 [==============================] - 0s 665us/step - loss: 0.3512 - sparse_categorical_accuracy: 0.9250
Epoch 393/500
4/4 [==============================] - 0s 997us/step - loss: 0.3487 - sparse_categorical_accuracy: 0.9417
Epoch 394/500
4/4 [==============================] - 0s 665us/step - loss: 0.3444 - sparse_categorical_accuracy: 0.9500
Epoch 395/500
4/4 [==============================] - 0s 665us/step - loss: 0.3714 - sparse_categorical_accuracy: 0.9333
Epoch 396/500
4/4 [==============================] - 0s 997us/step - loss: 0.3681 - sparse_categorical_accuracy: 0.9250
Epoch 397/500
4/4 [==============================] - 0s 665us/step - loss: 0.3659 - sparse_categorical_accuracy: 0.9000
Epoch 398/500
4/4 [==============================] - 0s 997us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9667
Epoch 399/500
4/4 [==============================] - 0s 997us/step - loss: 0.3917 - sparse_categorical_accuracy: 0.9083
Epoch 400/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3787 - sparse_categorical_accuracy: 0.9250 - val_loss: 0.5941 - val_sparse_categorical_accuracy: 0.6000
Epoch 401/500
4/4 [==============================] - 0s 665us/step - loss: 0.3651 - sparse_categorical_accuracy: 0.9333
Epoch 402/500
4/4 [==============================] - 0s 665us/step - loss: 0.3480 - sparse_categorical_accuracy: 0.9583
Epoch 403/500
4/4 [==============================] - 0s 998us/step - loss: 0.3813 - sparse_categorical_accuracy: 0.9167
Epoch 404/500
4/4 [==============================] - 0s 665us/step - loss: 0.4448 - sparse_categorical_accuracy: 0.8583
Epoch 405/500
4/4 [==============================] - 0s 997us/step - loss: 0.3864 - sparse_categorical_accuracy: 0.9083
Epoch 406/500
4/4 [==============================] - 0s 665us/step - loss: 0.3440 - sparse_categorical_accuracy: 0.9583
Epoch 407/500
4/4 [==============================] - 0s 665us/step - loss: 0.4250 - sparse_categorical_accuracy: 0.8667
Epoch 408/500
4/4 [==============================] - 0s 997us/step - loss: 0.3375 - sparse_categorical_accuracy: 0.9667
Epoch 409/500
4/4 [==============================] - 0s 665us/step - loss: 0.3470 - sparse_categorical_accuracy: 0.9333
Epoch 410/500
4/4 [==============================] - 0s 665us/step - loss: 0.3442 - sparse_categorical_accuracy: 0.9500
Epoch 411/500
4/4 [==============================] - 0s 665us/step - loss: 0.3907 - sparse_categorical_accuracy: 0.8583
Epoch 412/500
4/4 [==============================] - 0s 665us/step - loss: 0.3829 - sparse_categorical_accuracy: 0.8750
Epoch 413/500
4/4 [==============================] - 0s 665us/step - loss: 0.3658 - sparse_categorical_accuracy: 0.8917
Epoch 414/500
4/4 [==============================] - 0s 665us/step - loss: 0.3384 - sparse_categorical_accuracy: 0.9583
Epoch 415/500
4/4 [==============================] - 0s 665us/step - loss: 0.3447 - sparse_categorical_accuracy: 0.9667
Epoch 416/500
4/4 [==============================] - 0s 665us/step - loss: 0.3389 - sparse_categorical_accuracy: 0.9750
Epoch 417/500
4/4 [==============================] - 0s 997us/step - loss: 0.3484 - sparse_categorical_accuracy: 0.9750
Epoch 418/500
4/4 [==============================] - 0s 997us/step - loss: 0.3718 - sparse_categorical_accuracy: 0.9167
Epoch 419/500
4/4 [==============================] - 0s 997us/step - loss: 0.3435 - sparse_categorical_accuracy: 0.9500
Epoch 420/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3400 - sparse_categorical_accuracy: 0.9500 - val_loss: 0.3493 - val_sparse_categorical_accuracy: 1.0000
Epoch 421/500
4/4 [==============================] - 0s 665us/step - loss: 0.3376 - sparse_categorical_accuracy: 0.9667
Epoch 422/500
4/4 [==============================] - 0s 665us/step - loss: 0.3501 - sparse_categorical_accuracy: 0.9750
Epoch 423/500
4/4 [==============================] - 0s 665us/step - loss: 0.4595 - sparse_categorical_accuracy: 0.8250
Epoch 424/500
4/4 [==============================] - 0s 997us/step - loss: 0.3583 - sparse_categorical_accuracy: 0.9333
Epoch 425/500
4/4 [==============================] - 0s 665us/step - loss: 0.3593 - sparse_categorical_accuracy: 0.9417
Epoch 426/500
4/4 [==============================] - 0s 997us/step - loss: 0.4827 - sparse_categorical_accuracy: 0.8000
Epoch 427/500
4/4 [==============================] - 0s 665us/step - loss: 0.4092 - sparse_categorical_accuracy: 0.8917
Epoch 428/500
4/4 [==============================] - 0s 665us/step - loss: 0.3592 - sparse_categorical_accuracy: 0.9250
Epoch 429/500
4/4 [==============================] - 0s 665us/step - loss: 0.3744 - sparse_categorical_accuracy: 0.9000
Epoch 430/500
4/4 [==============================] - 0s 997us/step - loss: 0.4432 - sparse_categorical_accuracy: 0.8750
Epoch 431/500
4/4 [==============================] - 0s 665us/step - loss: 0.3755 - sparse_categorical_accuracy: 0.9167
Epoch 432/500
4/4 [==============================] - 0s 665us/step - loss: 0.3350 - sparse_categorical_accuracy: 0.9667
Epoch 433/500
4/4 [==============================] - 0s 665us/step - loss: 0.3608 - sparse_categorical_accuracy: 0.9417
Epoch 434/500
4/4 [==============================] - 0s 997us/step - loss: 0.3540 - sparse_categorical_accuracy: 0.9417
Epoch 435/500
4/4 [==============================] - 0s 665us/step - loss: 0.3831 - sparse_categorical_accuracy: 0.9167
Epoch 436/500
4/4 [==============================] - 0s 665us/step - loss: 0.3791 - sparse_categorical_accuracy: 0.8750
Epoch 437/500
4/4 [==============================] - 0s 665us/step - loss: 0.3327 - sparse_categorical_accuracy: 0.9750
Epoch 438/500
4/4 [==============================] - 0s 665us/step - loss: 0.3575 - sparse_categorical_accuracy: 0.9250
Epoch 439/500
4/4 [==============================] - 0s 665us/step - loss: 0.4376 - sparse_categorical_accuracy: 0.8500
Epoch 440/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3535 - sparse_categorical_accuracy: 0.9333 - val_loss: 0.3876 - val_sparse_categorical_accuracy: 1.0000
Epoch 441/500
4/4 [==============================] - 0s 998us/step - loss: 0.3743 - sparse_categorical_accuracy: 0.9250
Epoch 442/500
4/4 [==============================] - 0s 665us/step - loss: 0.3386 - sparse_categorical_accuracy: 0.9750
Epoch 443/500
4/4 [==============================] - 0s 998us/step - loss: 0.3476 - sparse_categorical_accuracy: 0.9667
Epoch 444/500
4/4 [==============================] - 0s 998us/step - loss: 0.3391 - sparse_categorical_accuracy: 0.9500
Epoch 445/500
4/4 [==============================] - 0s 997us/step - loss: 0.3449 - sparse_categorical_accuracy: 0.9583
Epoch 446/500
4/4 [==============================] - 0s 997us/step - loss: 0.3598 - sparse_categorical_accuracy: 0.9417
Epoch 447/500
4/4 [==============================] - 0s 997us/step - loss: 0.3927 - sparse_categorical_accuracy: 0.8833
Epoch 448/500
4/4 [==============================] - 0s 665us/step - loss: 0.3524 - sparse_categorical_accuracy: 0.9333
Epoch 449/500
4/4 [==============================] - 0s 665us/step - loss: 0.3387 - sparse_categorical_accuracy: 0.9750
Epoch 450/500
4/4 [==============================] - 0s 665us/step - loss: 0.3424 - sparse_categorical_accuracy: 0.9250
Epoch 451/500
4/4 [==============================] - 0s 998us/step - loss: 0.4508 - sparse_categorical_accuracy: 0.8333
Epoch 452/500
4/4 [==============================] - 0s 998us/step - loss: 0.3504 - sparse_categorical_accuracy: 0.9417
Epoch 453/500
4/4 [==============================] - 0s 665us/step - loss: 0.3766 - sparse_categorical_accuracy: 0.9167
Epoch 454/500
4/4 [==============================] - 0s 665us/step - loss: 0.3650 - sparse_categorical_accuracy: 0.9167
Epoch 455/500
4/4 [==============================] - 0s 665us/step - loss: 0.3506 - sparse_categorical_accuracy: 0.9583
Epoch 456/500
4/4 [==============================] - 0s 665us/step - loss: 0.3765 - sparse_categorical_accuracy: 0.9083
Epoch 457/500
4/4 [==============================] - 0s 665us/step - loss: 0.3376 - sparse_categorical_accuracy: 0.9833
Epoch 458/500
4/4 [==============================] - 0s 665us/step - loss: 0.3532 - sparse_categorical_accuracy: 0.9583
Epoch 459/500
4/4 [==============================] - 0s 665us/step - loss: 0.3358 - sparse_categorical_accuracy: 0.9500
Epoch 460/500
4/4 [==============================] - 0s 5ms/step - loss: 0.6349 - sparse_categorical_accuracy: 0.7250 - val_loss: 0.3554 - val_sparse_categorical_accuracy: 1.0000
Epoch 461/500
4/4 [==============================] - 0s 997us/step - loss: 0.3414 - sparse_categorical_accuracy: 0.9750
Epoch 462/500
4/4 [==============================] - 0s 665us/step - loss: 0.3489 - sparse_categorical_accuracy: 0.9417
Epoch 463/500
4/4 [==============================] - 0s 664us/step - loss: 0.3581 - sparse_categorical_accuracy: 0.9417
Epoch 464/500
4/4 [==============================] - 0s 665us/step - loss: 0.4619 - sparse_categorical_accuracy: 0.8583
Epoch 465/500
4/4 [==============================] - 0s 998us/step - loss: 0.4844 - sparse_categorical_accuracy: 0.8333
Epoch 466/500
4/4 [==============================] - 0s 665us/step - loss: 0.3958 - sparse_categorical_accuracy: 0.8833
Epoch 467/500
4/4 [==============================] - 0s 665us/step - loss: 0.3387 - sparse_categorical_accuracy: 0.9583
Epoch 468/500
4/4 [==============================] - 0s 997us/step - loss: 0.3566 - sparse_categorical_accuracy: 0.9417
Epoch 469/500
4/4 [==============================] - 0s 664us/step - loss: 0.3544 - sparse_categorical_accuracy: 0.9417
Epoch 470/500
4/4 [==============================] - 0s 665us/step - loss: 0.3763 - sparse_categorical_accuracy: 0.9000
Epoch 471/500
4/4 [==============================] - 0s 665us/step - loss: 0.3349 - sparse_categorical_accuracy: 0.9500
Epoch 472/500
4/4 [==============================] - 0s 998us/step - loss: 0.5347 - sparse_categorical_accuracy: 0.7667
Epoch 473/500
4/4 [==============================] - 0s 665us/step - loss: 0.4666 - sparse_categorical_accuracy: 0.8833
Epoch 474/500
4/4 [==============================] - 0s 665us/step - loss: 0.5124 - sparse_categorical_accuracy: 0.7833
Epoch 475/500
4/4 [==============================] - 0s 665us/step - loss: 0.3539 - sparse_categorical_accuracy: 0.9417
Epoch 476/500
4/4 [==============================] - 0s 997us/step - loss: 0.3371 - sparse_categorical_accuracy: 0.9583
Epoch 477/500
4/4 [==============================] - 0s 665us/step - loss: 0.4072 - sparse_categorical_accuracy: 0.8917
Epoch 478/500
4/4 [==============================] - 0s 665us/step - loss: 0.4168 - sparse_categorical_accuracy: 0.8750
Epoch 479/500
4/4 [==============================] - 0s 997us/step - loss: 0.4617 - sparse_categorical_accuracy: 0.8333
Epoch 480/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3497 - sparse_categorical_accuracy: 0.9417 - val_loss: 0.3175 - val_sparse_categorical_accuracy: 1.0000
Epoch 481/500
4/4 [==============================] - 0s 997us/step - loss: 0.3413 - sparse_categorical_accuracy: 0.9667
Epoch 482/500
4/4 [==============================] - 0s 665us/step - loss: 0.3415 - sparse_categorical_accuracy: 0.9667
Epoch 483/500
4/4 [==============================] - 0s 665us/step - loss: 0.3437 - sparse_categorical_accuracy: 0.9583
Epoch 484/500
4/4 [==============================] - 0s 665us/step - loss: 0.3377 - sparse_categorical_accuracy: 0.9500
Epoch 485/500
4/4 [==============================] - 0s 665us/step - loss: 0.4217 - sparse_categorical_accuracy: 0.8583
Epoch 486/500
4/4 [==============================] - 0s 665us/step - loss: 0.3618 - sparse_categorical_accuracy: 0.9417
Epoch 487/500
4/4 [==============================] - 0s 665us/step - loss: 0.3622 - sparse_categorical_accuracy: 0.9083
Epoch 488/500
4/4 [==============================] - 0s 665us/step - loss: 0.3756 - sparse_categorical_accuracy: 0.8917
Epoch 489/500
4/4 [==============================] - 0s 665us/step - loss: 0.3968 - sparse_categorical_accuracy: 0.9000
Epoch 490/500
4/4 [==============================] - 0s 665us/step - loss: 0.3435 - sparse_categorical_accuracy: 0.9583
Epoch 491/500
4/4 [==============================] - 0s 665us/step - loss: 0.3394 - sparse_categorical_accuracy: 0.9750
Epoch 492/500
4/4 [==============================] - 0s 997us/step - loss: 0.3348 - sparse_categorical_accuracy: 0.9833
Epoch 493/500
4/4 [==============================] - 0s 665us/step - loss: 0.3447 - sparse_categorical_accuracy: 0.9333
Epoch 494/500
4/4 [==============================] - 0s 674us/step - loss: 0.3433 - sparse_categorical_accuracy: 0.9333
Epoch 495/500
4/4 [==============================] - 0s 665us/step - loss: 0.3357 - sparse_categorical_accuracy: 0.9583
Epoch 496/500
4/4 [==============================] - 0s 997us/step - loss: 0.3445 - sparse_categorical_accuracy: 0.9583
Epoch 497/500
4/4 [==============================] - 0s 665us/step - loss: 0.3435 - sparse_categorical_accuracy: 0.9333
Epoch 498/500
4/4 [==============================] - 0s 665us/step - loss: 0.3406 - sparse_categorical_accuracy: 0.9667
Epoch 499/500
4/4 [==============================] - 0s 665us/step - loss: 0.3649 - sparse_categorical_accuracy: 0.9333
Epoch 500/500
4/4 [==============================] - 0s 5ms/step - loss: 0.3888 - sparse_categorical_accuracy: 0.9250 - val_loss: 0.3516 - val_sparse_categorical_accuracy: 0.8667
Model: "my_model"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 dense (Dense)               multiple                  15        
                                                                 
=================================================================
Total params: 15
Trainable params: 15
Non-trainable params: 0
_________________________________________________________________

3 MNIST 数据集

3.1 介绍

MNIST 数据集一共有 7 万张图片，是 28×28 像素的 0 到 9 手写数字数据集，其中 6 万张用于训练， 1 万张用于测试。每张图片包括 784（28×28）个像素点，使用全连接网络时可将 784 个像素点组成长度为 784 的一维数组，作为输入特征。数据集图片如下所示。
在这里插入图片描述

3.2 导入并可视化数据集

keras 函数库中提供了使用 mnist 数据集的接口，代码如下所示，可以使用
load_data()直接从 mnist 中读取测试集和训练集。

mnist = tf.keras.datasets.mnist

(x_train, y_train),(x_test, y_test) = mnist.load_data()

输入全连接网络时需要先将数据拉直为一维数组，把784个像素点的灰度值作为输入特征输入神经网络

tf.keras.layers.Flatten()
在这里插入图片描述

使用plt库中的两个函数可视化训练集中的图片
plt.imshow(x_train[0], cmap = ‘gray’)
plt.show()

在这里插入图片描述

使用 print 打印出训练集中第一个样本以二位数组的形式打印出来，如下
所示。

print(“x_train[0]:”,x_train[0])

在这里插入图片描述

打印出第一个样本的标签，为 5

print（“y_train[0]:”,y_train[0]）

y_train[0]:5

打印出测试样本的形状，共有10000个28行28列的三维数据

print(“x_test.shape:”x_test.shape)

x_test.shape:(10000,28,28)

# p13_mnist_datasets.py
import tensorflow as tf
import matplotlib.pyplot as plt

mnist = tf.keras.datasets.mnist

(x_train, y_train), (x_test, y_test) = mnist.load_data()

# 可视化训练集输入特征的第一个元素
plt.imshow(x_train[0], cmap ='gray') #绘制灰色图
plt.show()

在这里插入图片描述

# 打印出训练集输入特征的第一个元素
print("x_train[0]:\n", x_train[0])
# 打印出训练集标签的第一个元素
print("y_train[0]:\n", y_train[0])

x_train[0]:
 [[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   3  18  18  18 126 136
  175  26 166 255 247 127   0   0   0   0]
 [  0   0   0   0   0   0   0   0  30  36  94 154 170 253 253 253 253 253
  225 172 253 242 195  64   0   0   0   0]
 [  0   0   0   0   0   0   0  49 238 253 253 253 253 253 253 253 253 251
   93  82  82  56  39   0   0   0   0   0]
 [  0   0   0   0   0   0   0  18 219 253 253 253 253 253 198 182 247 241
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0  80 156 107 253 253 205  11   0  43 154
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0  14   1 154 253  90   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0 139 253 190   2   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0  11 190 253  70   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0  35 241 225 160 108   1
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0  81 240 253 253 119
   25   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0  45 186 253 253
  150  27   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0  16  93 252
  253 187   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0 249
  253 249  64   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0  46 130 183 253
  253 207   2   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0  39 148 229 253 253 253
  250 182   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0  24 114 221 253 253 253 253 201
   78   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0  23  66 213 253 253 253 253 198  81   2
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0  18 171 219 253 253 253 253 195  80   9   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0  55 172 226 253 253 253 253 244 133  11   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0 136 253 253 253 212 135 132  16   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]
 [  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0
    0   0   0   0   0   0   0   0   0   0]]
y_train[0]:
 5

# 打印出整个训练集输入特征形状
print("x_train.shape:\n", x_train.shape)
# 打印出整个训练集标签的形状
print("y_train.shape:\n", y_train.shape)
# 打印出整个测试集输入特征的形状
print("x_test.shape:\n", x_test.shape)
# 打印出整个测试集标签的形状
print("y_test.shape:\n", y_test.shape)

x_train.shape:
 (60000, 28, 28)
y_train.shape:
 (60000,)
x_test.shape:
 (10000, 28, 28)
y_test.shape:
 (10000,)

3.3 训练 MNIST 数据集

使用 Sequential 实现手写数字识别

# p14_mnist_sequential.py
import tensorflow as tf

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

x_train, x_test = x_train / 255.0, x_test / 255.0

model = tf.keras.models.Sequential([
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dense(128, activation = 'relu', kernel_regularizer = tf.keras.regularizers.l2()),
    tf.keras.layers.Dense(10, activation = 'softmax', kernel_regularizer = tf.keras.regularizers.l2())
])


model.compile(optimizer='adam',
              loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
              metrics = ['sparse_categorical_accuracy']
)
model.fit(x_train, y_train, batch_size=32, epochs=5, validation_data=(x_test, y_test), validation_freq=1)
model.summary()

Epoch 1/5
1875/1875 [==============================] - 2s 832us/step - loss: 0.8692 - sparse_categorical_accuracy: 0.8887 - val_loss: 0.7183 - val_sparse_categorical_accuracy: 0.9113
Epoch 2/5
1875/1875 [==============================] - 1s 770us/step - loss: 0.7406 - sparse_categorical_accuracy: 0.9010 - val_loss: 0.7298 - val_sparse_categorical_accuracy: 0.9052
Epoch 3/5
1875/1875 [==============================] - 1s 772us/step - loss: 0.7331 - sparse_categorical_accuracy: 0.9023 - val_loss: 0.7283 - val_sparse_categorical_accuracy: 0.9003
Epoch 4/5
1875/1875 [==============================] - 1s 787us/step - loss: 0.7296 - sparse_categorical_accuracy: 0.9035 - val_loss: 0.7069 - val_sparse_categorical_accuracy: 0.9094
Epoch 5/5
1875/1875 [==============================] - 1s 776us/step - loss: 0.7265 - sparse_categorical_accuracy: 0.9037 - val_loss: 0.6963 - val_sparse_categorical_accuracy: 0.9120
Model: "sequential"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 flatten (Flatten)           (None, 784)               0         
                                                                 
 dense_1 (Dense)             (None, 128)               100480    
                                                                 
 dense_2 (Dense)             (None, 10)                1290      
                                                                 
=================================================================
Total params: 101,770
Trainable params: 101,770
Non-trainable params: 0
_________________________________________________________________

#p15_mnist_class.py
import tensorflow as tf
from tensorflow.keras.layers import Dense, Flatten
from tensorflow.keras import Model

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

x_train, x_test = x_train / 255.0, x_test / 255.0
class MnistModel(Model):
    def __init__(self):
        super(MnistModel, self).__init__()
        self.flatten = Flatten()
        self.d1 = Dense(128, activation='relu')
        self.d2 = Dense(10, activation='softmax')
        
    def call(self,x):
        x = self.flatten(x)
        x = self.d1(x)
        y = self.d2(x)
        return y
    
model = MnistModel()

model.compile(optimizer='adam',
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
              metrics=['sparse_categorical_accuracy'])

model.fit(x_train, y_train, batch_size=32, epochs=5, validation_data=(x_test, y_test), validation_freq=1)
model.summary()

Epoch 1/5
1875/1875 [==============================] - 2s 795us/step - loss: 0.2601 - sparse_categorical_accuracy: 0.9265 - val_loss: 0.1397 - val_sparse_categorical_accuracy: 0.9568
Epoch 2/5
1875/1875 [==============================] - 1s 739us/step - loss: 0.1128 - sparse_categorical_accuracy: 0.9667 - val_loss: 0.0977 - val_sparse_categorical_accuracy: 0.9712
Epoch 3/5
1875/1875 [==============================] - 1s 739us/step - loss: 0.0761 - sparse_categorical_accuracy: 0.9775 - val_loss: 0.0889 - val_sparse_categorical_accuracy: 0.9726
Epoch 4/5
1875/1875 [==============================] - 1s 741us/step - loss: 0.0573 - sparse_categorical_accuracy: 0.9826 - val_loss: 0.0734 - val_sparse_categorical_accuracy: 0.9775
Epoch 5/5
1875/1875 [==============================] - 1s 751us/step - loss: 0.0433 - sparse_categorical_accuracy: 0.9865 - val_loss: 0.0755 - val_sparse_categorical_accuracy: 0.9765
Model: "mnist_model"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 flatten_1 (Flatten)         multiple                  0         
                                                                 
 dense_3 (Dense)             multiple                  100480    
                                                                 
 dense_4 (Dense)             multiple                  1290      
                                                                 
=================================================================
Total params: 101,770
Trainable params: 101,770
Non-trainable params: 0
_________________________________________________________________

4 Fashion_mnist 数据集

Fashion_mnist 数据集具有 mnist 近乎所有的特征，包括 60000 张训练图片
和 10000 张测试图片，图片被分为十类，每张图像为 28×28 的分辨率。

在这里插入图片描述

import tensorflow as tf

fashion = tf.keras.datasets.fashion_mnist
(x_train, y_train),(x_test, y_test) = fashion.load_data()
x_train, x_test = x_train / 255.0, x_test / 255.0

model = tf.keras.models.Sequential([
    tf.keras.layers.Flatten(),
    tf.keras.layers.Dense(128, activation = 'relu'),
    tf.keras.layers.Dense(10, activation = 'softmax')
])

model.compile(optimizer='adam',
              loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
              metrics=['sparse_categorical_accuracy']
)

model.fit(x_train, y_train, batch_size=32, epochs=5, validation_data=(x_test, y_test), validation_freq=1)
model.summary()

Epoch 1/5
1875/1875 [==============================] - 2s 811us/step - loss: 0.4998 - sparse_categorical_accuracy: 0.8241 - val_loss: 0.4051 - val_sparse_categorical_accuracy: 0.8543
Epoch 2/5
1875/1875 [==============================] - 1s 776us/step - loss: 0.3737 - sparse_categorical_accuracy: 0.8652 - val_loss: 0.4044 - val_sparse_categorical_accuracy: 0.8572
Epoch 3/5
1875/1875 [==============================] - 1s 755us/step - loss: 0.3347 - sparse_categorical_accuracy: 0.8785 - val_loss: 0.3784 - val_sparse_categorical_accuracy: 0.8629
Epoch 4/5
1875/1875 [==============================] - 1s 737us/step - loss: 0.3121 - sparse_categorical_accuracy: 0.8859 - val_loss: 0.3685 - val_sparse_categorical_accuracy: 0.8636
Epoch 5/5
1875/1875 [==============================] - 1s 736us/step - loss: 0.2958 - sparse_categorical_accuracy: 0.8916 - val_loss: 0.3528 - val_sparse_categorical_accuracy: 0.8715
Model: "sequential_2"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 flatten_4 (Flatten)         (None, 784)               0         
                                                                 
 dense_8 (Dense)             (None, 128)               100480    
                                                                 
 dense_9 (Dense)             (None, 10)                1290      
                                                                 
=================================================================
Total params: 101,770
Trainable params: 101,770
Non-trainable params: 0
_________________________________________________________________

#p16_fashion_sequential.py
import tensorflow as tf
from tensorflow.keras.layers import Dense,Flatten
from tensorflow.keras import Model


fashion = tf.keras.datasets.fashion_mnist
(x_train, y_train),(x_test, y_test) = fashion.load_data()
x_train, x_test = x_train / 255.0, x_test / 255.0

class FasionModel(Model):
    def __init__(self):
        super(FasionModel, self).__init__()
        self.flatten = Flatten()
        self.d1 = Dense(128, activation = 'relu')
        self.d2 =Dense(10, activation='softmax')
        
    def call(self, x):
        x = self.flatten(x)
        x = self.d1(x)
        y = self.d2(x)
        return y
    
    
model = FasionModel()

model.compile(optimizer='adam',
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
              metrics=['sparse_categorical_accuracy'])

model.fit(x_train, y_train, batch_size=32, epochs=5, validation_data=(x_test, y_test), validation_freq=1)
model.summary()

class FasionModel(Model):
    def __init__(self):
        super(FasionModel, self).__init__()
        self.flatten = Flatten()
        self.d1 = Dense(128, activation = 'relu')
        self.d2 =Dense(10, activation='softmax')
        
    def call(self, x):
        x = self.flatten(x)
        x = self.d1(x)
        y = self.d2(x)
        return y
    
    
model = FasionModel()
model.compile(optimizer='adam',
              loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
              metrics = ['sparse_categorical_accuracy']
             
)

model.fit(x_train, y_train, batch_size=32, epochs=5, validation_data=(x_test, y_test), validation_freq=1)
model.summary()

Epoch 1/5
1875/1875 [==============================] - 2s 797us/step - loss: 0.4944 - sparse_categorical_accuracy: 0.8256 - val_loss: 0.4059 - val_sparse_categorical_accuracy: 0.8541
Epoch 2/5
1875/1875 [==============================] - 1s 733us/step - loss: 0.3716 - sparse_categorical_accuracy: 0.8662 - val_loss: 0.3932 - val_sparse_categorical_accuracy: 0.8631
Epoch 3/5
1875/1875 [==============================] - 1s 739us/step - loss: 0.3343 - sparse_categorical_accuracy: 0.8780 - val_loss: 0.3763 - val_sparse_categorical_accuracy: 0.8633
Epoch 4/5
1875/1875 [==============================] - 1s 740us/step - loss: 0.3114 - sparse_categorical_accuracy: 0.8859 - val_loss: 0.3702 - val_sparse_categorical_accuracy: 0.8650
Epoch 5/5
1875/1875 [==============================] - 1s 753us/step - loss: 0.2953 - sparse_categorical_accuracy: 0.8912 - val_loss: 0.3493 - val_sparse_categorical_accuracy: 0.8778
Model: "fasion_model_2"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 flatten_7 (Flatten)         multiple                  0         
                                                                 
 dense_14 (Dense)            multiple                  100480    
                                                                 
 dense_15 (Dense)            multiple                  1290      
                                                                 
=================================================================
Total params: 101,770
Trainable params: 101,770
Non-trainable params: 0
_________________________________________________________________
Epoch 1/5
1875/1875 [==============================] - 2s 797us/step - loss: 0.4989 - sparse_categorical_accuracy: 0.8245 - val_loss: 0.4104 - val_sparse_categorical_accuracy: 0.8551
Epoch 2/5
1875/1875 [==============================] - 1s 735us/step - loss: 0.3741 - sparse_categorical_accuracy: 0.8654 - val_loss: 0.3954 - val_sparse_categorical_accuracy: 0.8602
Epoch 3/5
1875/1875 [==============================] - 1s 736us/step - loss: 0.3359 - sparse_categorical_accuracy: 0.8782 - val_loss: 0.3695 - val_sparse_categorical_accuracy: 0.8671
Epoch 4/5
1875/1875 [==============================] - 1s 738us/step - loss: 0.3139 - sparse_categorical_accuracy: 0.8847 - val_loss: 0.3762 - val_sparse_categorical_accuracy: 0.8658
Epoch 5/5
1875/1875 [==============================] - 1s 746us/step - loss: 0.2969 - sparse_categorical_accuracy: 0.8914 - val_loss: 0.3612 - val_sparse_categorical_accuracy: 0.8706
Model: "fasion_model_3"
_________________________________________________________________
 Layer (type)                Output Shape              Param #   
=================================================================
 flatten_8 (Flatten)         multiple                  0         
                                                                 
 dense_16 (Dense)            multiple                  100480    
                                                                 
 dense_17 (Dense)            multiple                  1290      
                                                                 
=================================================================
Total params: 101,770
Trainable params: 101,770
Non-trainable params: 0
_________________________________________________________________