使用卷积神经网络构建的模型识别手写数字并实现了GUI如何将每个数字的识别率显示出来

小诺爷

from tensorflow.keras import datasets, layers, optimizers, Sequential,callbacks,losses
import os
import numpy as np

np.set_printoptions(threshold=np.inf)

(x_train, y_train), (x_test, y_test) = datasets.mnist.load_data("mnist.pkl")
x_train, x_test = x_train / 255.0, x_test / 255.0

model = Sequential([
    layers.InputLayer(input_shape=(28, 28)),
    layers.Reshape((28, 28, 1)),
    layers.Conv2D(kernel_size=3, strides=1, filters=16,padding='same', activation='relu', name='layer_conv1'),
    layers.MaxPooling2D(pool_size=2, strides=2),
    layers.BatchNormalization(),
    layers.Conv2D(kernel_size=3, strides=1, filters=36,padding='same', activation='relu', name='layer_conv2'),
    layers.MaxPooling2D(pool_size=2, strides=2),
    layers.BatchNormalization(),
    layers.Flatten(),
    layers.Dense(units=128, activation='selu'),
    layers.BatchNormalization(),
    layers.Dense(units=10, activation='softmax')
])


model.compile(optimizer=optimizers.Adam(lr=1e-3),loss=losses.SparseCategoricalCrossentropy(from_logits=False),metrics=['sparse_categorical_accuracy'])

checkpoint_save_path = "./checkpoint/mnist.ckpt"
if os.path.exists(checkpoint_save_path + '.index'):
    print('-------------load the model-----------------')
    model.load_weights(checkpoint_save_path)

cp_callback = callbacks.ModelCheckpoint(filepath=checkpoint_save_path,
                                                 save_weights_only=True,
                                                 save_best_only=True)
model.fit(x=x_train, y=y_train,batch_size=32,epochs=5,validation_data=(x_test, y_test), validation_freq=1,callbacks=[cp_callback])

model.summary()
file = open('weights.txt', 'w')
for v in model.trainable_variables:
    file.write(str(v.name) + '\n')
    file.write(str(v.shape) + '\n')
    file.write(str(v.numpy()) + '\n')
file.close()

hhiop

这个有点复杂，估计小甲鱼都不知道

dadan1

山理工？

Victor0321

卷i起来 哈哈哈

Levin-e

首先你得实现test()吧，输入你的测试数据然后计算每个数字的准确率，然后就是想办法再GUI里显示出来了，主要工作量是在计算准确率那边，顺便也可以算一手召回率，然后求个F1