keras神经网络入门：1.简单神经网络实现mnist识别，识别准确率高达89%

于 2022-08-18 18:36:28 发布
阅读量586 收藏
点赞数
分类专栏： 神经网络 文章标签： 神经网络 keras tensorflow
版权声明：本文为博主原创文章，遵循 CC 4.0 BY-SA 版权协议，转载请附上原文出处链接和本声明。
本文链接：https://blog.csdn.net/qq_36554176/article/details/126411637
版权

第一步:导入相关模块

import tensorflow as tf
from  tensorflow.keras import layers,models
from tensorflow.keras.datasets import mnist
import numpy as np

第二步：导入数据集

(train_data,train_label),(val_data,val_label)=mnist.load_data()

train_data=train_data/255.0
val_data=val_data/255.0

第三步：搭建神经网络

model=models.Sequential(
    (
    layers.Flatten(input_shape=(28,28)),
    layers.Dense(32,activation='relu'),
    layers.Dense(16,activation='relu'),
    layers.Dense(10,activation='softmax')
    )
)

model.compile(optimizer=tf.optimizers.SGD(lr=0.001),
              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
              metrics=[tf.keras.metrics.SparseCategoricalAccuracy()])

model.summary()

Model: "sequential_3"
_________________________________________________________________
Layer (type)                 Output Shape              Param #   
=================================================================
flatten_1 (Flatten)          (None, 784)               0         
_________________________________________________________________
dense_9 (Dense)              (None, 32)                25120     
_________________________________________________________________
dense_10 (Dense)             (None, 16)                528       
_________________________________________________________________
dense_11 (Dense)             (None, 10)                170       
=================================================================
Total params: 25,818
Trainable params: 25,818
Non-trainable params: 0
_________________________________________________________________

第四步：训练神经网络

model.fit(train_data,train_label,epochs=160,batch_size=512,validation_data=(val_data,val_label))

Epoch 1/160
118/118 [==============================] - 1s 5ms/step - loss: 2.3386 - sparse_categorical_accuracy: 0.1019 - val_loss: 2.3109 - val_sparse_categorical_accuracy: 0.1097
Epoch 2/160
118/118 [==============================] - 0s 2ms/step - loss: 2.2886 - sparse_categorical_accuracy: 0.1236 - val_loss: 2.2737 - val_sparse_categorical_accuracy: 0.1301
Epoch 3/160
118/118 [==============================] - 0s 3ms/step - loss: 2.2576 - sparse_categorical_accuracy: 0.1471 - val_loss: 2.2461 - val_sparse_categorical_accuracy: 0.1534
Epoch 4/160
118/118 [==============================] - 0s 2ms/step - loss: 2.2314 - sparse_categorical_accuracy: 0.1738 - val_loss: 2.2202 - val_sparse_categorical_accuracy: 0.1830
Epoch 5/160
118/118 [==============================] - 0s 2ms/step - loss: 2.2053 - sparse_categorical_accuracy: 0.2046 - val_loss: 2.1932 - val_sparse_categorical_accuracy: 0.2137
Epoch 6/160
118/118 [==============================] - 0s 2ms/step - loss: 2.1778 - sparse_categorical_accuracy: 0.2369 - val_loss: 2.1644 - val_sparse_categorical_accuracy: 0.2512
Epoch 7/160
118/118 [==============================] - 0s 3ms/step - loss: 2.1485 - sparse_categorical_accuracy: 0.2714 - val_loss: 2.1337 - val_sparse_categorical_accuracy: 0.2874
Epoch 8/160
118/118 [==============================] - 0s 3ms/step - loss: 2.1173 - sparse_categorical_accuracy: 0.3032 - val_loss: 2.1013 - val_sparse_categorical_accuracy: 0.3162
Epoch 9/160
118/118 [==============================] - 0s 3ms/step - loss: 2.0847 - sparse_categorical_accuracy: 0.3284 - val_loss: 2.0678 - val_sparse_categorical_accuracy: 0.3446
Epoch 10/160
118/118 [==============================] - 0s 3ms/step - loss: 2.0511 - sparse_categorical_accuracy: 0.3503 - val_loss: 2.0332 - val_sparse_categorical_accuracy: 0.3654
Epoch 11/160
118/118 [==============================] - 0s 3ms/step - loss: 2.0166 - sparse_categorical_accuracy: 0.3687 - val_loss: 1.9976 - val_sparse_categorical_accuracy: 0.3843
Epoch 12/160
118/118 [==============================] - 0s 3ms/step - loss: 1.9812 - sparse_categorical_accuracy: 0.3861 - val_loss: 1.9612 - val_sparse_categorical_accuracy: 0.4039
Epoch 13/160
118/118 [==============================] - 0s 3ms/step - loss: 1.9451 - sparse_categorical_accuracy: 0.4027 - val_loss: 1.9239 - val_sparse_categorical_accuracy: 0.4177
Epoch 14/160
118/118 [==============================] - 0s 3ms/step - loss: 1.9081 - sparse_categorical_accuracy: 0.4186 - val_loss: 1.8858 - val_sparse_categorical_accuracy: 0.4357
Epoch 15/160
118/118 [==============================] - 0s 3ms/step - loss: 1.8704 - sparse_categorical_accuracy: 0.4367 - val_loss: 1.8471 - val_sparse_categorical_accuracy: 0.4526
Epoch 16/160
118/118 [==============================] - 0s 3ms/step - loss: 1.8318 - sparse_categorical_accuracy: 0.4556 - val_loss: 1.8075 - val_sparse_categorical_accuracy: 0.4681
Epoch 17/160
118/118 [==============================] - 0s 3ms/step - loss: 1.7924 - sparse_categorical_accuracy: 0.4744 - val_loss: 1.7672 - val_sparse_categorical_accuracy: 0.4880
Epoch 18/160
118/118 [==============================] - 0s 3ms/step - loss: 1.7522 - sparse_categorical_accuracy: 0.4949 - val_loss: 1.7259 - val_sparse_categorical_accuracy: 0.5057
Epoch 19/160
118/118 [==============================] - 0s 3ms/step - loss: 1.7110 - sparse_categorical_accuracy: 0.5152 - val_loss: 1.6837 - val_sparse_categorical_accuracy: 0.5252
Epoch 20/160
118/118 [==============================] - 0s 3ms/step - loss: 1.6691 - sparse_categorical_accuracy: 0.5355 - val_loss: 1.6409 - val_sparse_categorical_accuracy: 0.5473
Epoch 21/160
118/118 [==============================] - 0s 3ms/step - loss: 1.6267 - sparse_categorical_accuracy: 0.5579 - val_loss: 1.5978 - val_sparse_categorical_accuracy: 0.5675
Epoch 22/160
118/118 [==============================] - 0s 2ms/step - loss: 1.5840 - sparse_categorical_accuracy: 0.5785 - val_loss: 1.5546 - val_sparse_categorical_accuracy: 0.5889
Epoch 23/160
118/118 [==============================] - 0s 3ms/step - loss: 1.5413 - sparse_categorical_accuracy: 0.5994 - val_loss: 1.5115 - val_sparse_categorical_accuracy: 0.6073
Epoch 24/160
118/118 [==============================] - 0s 3ms/step - loss: 1.4988 - sparse_categorical_accuracy: 0.6158 - val_loss: 1.4687 - val_sparse_categorical_accuracy: 0.6238
Epoch 25/160
118/118 [==============================] - 0s 3ms/step - loss: 1.4566 - sparse_categorical_accuracy: 0.6315 - val_loss: 1.4262 - val_sparse_categorical_accuracy: 0.6374
Epoch 26/160
118/118 [==============================] - 0s 3ms/step - loss: 1.4146 - sparse_categorical_accuracy: 0.6450 - val_loss: 1.3842 - val_sparse_categorical_accuracy: 0.6512
Epoch 27/160
118/118 [==============================] - 0s 2ms/step - loss: 1.3732 - sparse_categorical_accuracy: 0.6564 - val_loss: 1.3426 - val_sparse_categorical_accuracy: 0.6629
Epoch 28/160
118/118 [==============================] - 0s 2ms/step - loss: 1.3323 - sparse_categorical_accuracy: 0.6666 - val_loss: 1.3017 - val_sparse_categorical_accuracy: 0.6720
Epoch 29/160
118/118 [==============================] - 0s 3ms/step - loss: 1.2922 - sparse_categorical_accuracy: 0.6763 - val_loss: 1.2615 - val_sparse_categorical_accuracy: 0.6826
Epoch 30/160
118/118 [==============================] - 0s 2ms/step - loss: 1.2528 - sparse_categorical_accuracy: 0.6847 - val_loss: 1.2224 - val_sparse_categorical_accuracy: 0.6909
Epoch 31/160
118/118 [==============================] - 0s 2ms/step - loss: 1.2146 - sparse_categorical_accuracy: 0.6924 - val_loss: 1.1842 - val_sparse_categorical_accuracy: 0.6995
Epoch 32/160
118/118 [==============================] - 0s 2ms/step - loss: 1.1775 - sparse_categorical_accuracy: 0.6997 - val_loss: 1.1472 - val_sparse_categorical_accuracy: 0.7074
Epoch 33/160
118/118 [==============================] - 0s 3ms/step - loss: 1.1416 - sparse_categorical_accuracy: 0.7074 - val_loss: 1.1116 - val_sparse_categorical_accuracy: 0.7140
Epoch 34/160
118/118 [==============================] - 0s 3ms/step - loss: 1.1070 - sparse_categorical_accuracy: 0.7154 - val_loss: 1.0773 - val_sparse_categorical_accuracy: 0.7221
Epoch 35/160
118/118 [==============================] - 0s 3ms/step - loss: 1.0739 - sparse_categorical_accuracy: 0.7236 - val_loss: 1.0445 - val_sparse_categorical_accuracy: 0.7305
Epoch 36/160
118/118 [==============================] - 0s 3ms/step - loss: 1.0422 - sparse_categorical_accuracy: 0.7308 - val_loss: 1.0131 - val_sparse_categorical_accuracy: 0.7377
Epoch 37/160
118/118 [==============================] - 0s 3ms/step - loss: 1.0120 - sparse_categorical_accuracy: 0.7387 - val_loss: 0.9833 - val_sparse_categorical_accuracy: 0.7476
Epoch 38/160
118/118 [==============================] - 0s 3ms/step - loss: 0.9833 - sparse_categorical_accuracy: 0.7468 - val_loss: 0.9549 - val_sparse_categorical_accuracy: 0.7545
Epoch 39/160
118/118 [==============================] - 0s 3ms/step - loss: 0.9560 - sparse_categorical_accuracy: 0.7544 - val_loss: 0.9280 - val_sparse_categorical_accuracy: 0.7615
Epoch 40/160
118/118 [==============================] - 0s 3ms/step - loss: 0.9302 - sparse_categorical_accuracy: 0.7609 - val_loss: 0.9027 - val_sparse_categorical_accuracy: 0.7673
Epoch 41/160
118/118 [==============================] - 0s 3ms/step - loss: 0.9059 - sparse_categorical_accuracy: 0.7675 - val_loss: 0.8788 - val_sparse_categorical_accuracy: 0.7747
Epoch 42/160
118/118 [==============================] - 0s 3ms/step - loss: 0.8830 - sparse_categorical_accuracy: 0.7745 - val_loss: 0.8564 - val_sparse_categorical_accuracy: 0.7804
Epoch 43/160
118/118 [==============================] - 0s 3ms/step - loss: 0.8614 - sparse_categorical_accuracy: 0.7794 - val_loss: 0.8354 - val_sparse_categorical_accuracy: 0.7858
Epoch 44/160
118/118 [==============================] - 0s 3ms/step - loss: 0.8412 - sparse_categorical_accuracy: 0.7837 - val_loss: 0.8157 - val_sparse_categorical_accuracy: 0.7903
Epoch 45/160
118/118 [==============================] - 0s 3ms/step - loss: 0.8222 - sparse_categorical_accuracy: 0.7888 - val_loss: 0.7972 - val_sparse_categorical_accuracy: 0.7953
Epoch 46/160
118/118 [==============================] - 0s 3ms/step - loss: 0.8043 - sparse_categorical_accuracy: 0.7930 - val_loss: 0.7798 - val_sparse_categorical_accuracy: 0.7986
Epoch 47/160
118/118 [==============================] - 0s 3ms/step - loss: 0.7874 - sparse_categorical_accuracy: 0.7962 - val_loss: 0.7635 - val_sparse_categorical_accuracy: 0.8011
Epoch 48/160
118/118 [==============================] - 0s 3ms/step - loss: 0.7716 - sparse_categorical_accuracy: 0.7994 - val_loss: 0.7482 - val_sparse_categorical_accuracy: 0.8043
Epoch 49/160
118/118 [==============================] - 0s 3ms/step - loss: 0.7566 - sparse_categorical_accuracy: 0.8025 - val_loss: 0.7337 - val_sparse_categorical_accuracy: 0.8078
Epoch 50/160
118/118 [==============================] - 0s 3ms/step - loss: 0.7425 - sparse_categorical_accuracy: 0.8054 - val_loss: 0.7200 - val_sparse_categorical_accuracy: 0.8110
Epoch 51/160
118/118 [==============================] - 0s 3ms/step - loss: 0.7291 - sparse_categorical_accuracy: 0.8077 - val_loss: 0.7071 - val_sparse_categorical_accuracy: 0.8130
Epoch 52/160
118/118 [==============================] - 0s 3ms/step - loss: 0.7164 - sparse_categorical_accuracy: 0.8104 - val_loss: 0.6948 - val_sparse_categorical_accuracy: 0.8161
Epoch 53/160
118/118 [==============================] - 0s 3ms/step - loss: 0.7044 - sparse_categorical_accuracy: 0.8127 - val_loss: 0.6832 - val_sparse_categorical_accuracy: 0.8194
Epoch 54/160
118/118 [==============================] - 0s 3ms/step - loss: 0.6929 - sparse_categorical_accuracy: 0.8158 - val_loss: 0.6723 - val_sparse_categorical_accuracy: 0.8212
Epoch 55/160
118/118 [==============================] - 0s 3ms/step - loss: 0.6821 - sparse_categorical_accuracy: 0.8175 - val_loss: 0.6617 - val_sparse_categorical_accuracy: 0.8242
Epoch 56/160
118/118 [==============================] - 0s 3ms/step - loss: 0.6717 - sparse_categorical_accuracy: 0.8199 - val_loss: 0.6518 - val_sparse_categorical_accuracy: 0.8263
Epoch 57/160
118/118 [==============================] - 0s 3ms/step - loss: 0.6619 - sparse_categorical_accuracy: 0.8217 - val_loss: 0.6423 - val_sparse_categorical_accuracy: 0.8287
Epoch 58/160
118/118 [==============================] - 0s 3ms/step - loss: 0.6525 - sparse_categorical_accuracy: 0.8236 - val_loss: 0.6332 - val_sparse_categorical_accuracy: 0.8304
Epoch 59/160
118/118 [==============================] - 0s 3ms/step - loss: 0.6435 - sparse_categorical_accuracy: 0.8259 - val_loss: 0.6246 - val_sparse_categorical_accuracy: 0.8321
Epoch 60/160
118/118 [==============================] - 0s 3ms/step - loss: 0.6349 - sparse_categorical_accuracy: 0.8275 - val_loss: 0.6164 - val_sparse_categorical_accuracy: 0.8345
Epoch 61/160
118/118 [==============================] - 0s 3ms/step - loss: 0.6267 - sparse_categorical_accuracy: 0.8295 - val_loss: 0.6085 - val_sparse_categorical_accuracy: 0.8362
Epoch 62/160
118/118 [==============================] - 0s 3ms/step - loss: 0.6188 - sparse_categorical_accuracy: 0.8313 - val_loss: 0.6009 - val_sparse_categorical_accuracy: 0.8378
Epoch 63/160
118/118 [==============================] - 0s 3ms/step - loss: 0.6113 - sparse_categorical_accuracy: 0.8334 - val_loss: 0.5936 - val_sparse_categorical_accuracy: 0.8396
Epoch 64/160
118/118 [==============================] - 0s 3ms/step - loss: 0.6040 - sparse_categorical_accuracy: 0.8350 - val_loss: 0.5867 - val_sparse_categorical_accuracy: 0.8415
Epoch 65/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5970 - sparse_categorical_accuracy: 0.8367 - val_loss: 0.5800 - val_sparse_categorical_accuracy: 0.8427
Epoch 66/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5904 - sparse_categorical_accuracy: 0.8382 - val_loss: 0.5736 - val_sparse_categorical_accuracy: 0.8446
Epoch 67/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5839 - sparse_categorical_accuracy: 0.8398 - val_loss: 0.5674 - val_sparse_categorical_accuracy: 0.8463
Epoch 68/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5777 - sparse_categorical_accuracy: 0.8413 - val_loss: 0.5615 - val_sparse_categorical_accuracy: 0.8476
Epoch 69/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5718 - sparse_categorical_accuracy: 0.8424 - val_loss: 0.5558 - val_sparse_categorical_accuracy: 0.8490
Epoch 70/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5660 - sparse_categorical_accuracy: 0.8438 - val_loss: 0.5503 - val_sparse_categorical_accuracy: 0.8500
Epoch 71/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5605 - sparse_categorical_accuracy: 0.8450 - val_loss: 0.5449 - val_sparse_categorical_accuracy: 0.8512
Epoch 72/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5551 - sparse_categorical_accuracy: 0.8466 - val_loss: 0.5398 - val_sparse_categorical_accuracy: 0.8523
Epoch 73/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5499 - sparse_categorical_accuracy: 0.8480 - val_loss: 0.5348 - val_sparse_categorical_accuracy: 0.8545
Epoch 74/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5449 - sparse_categorical_accuracy: 0.8493 - val_loss: 0.5300 - val_sparse_categorical_accuracy: 0.8550
Epoch 75/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5400 - sparse_categorical_accuracy: 0.8504 - val_loss: 0.5254 - val_sparse_categorical_accuracy: 0.8560
Epoch 76/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5353 - sparse_categorical_accuracy: 0.8515 - val_loss: 0.5210 - val_sparse_categorical_accuracy: 0.8578
Epoch 77/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5308 - sparse_categorical_accuracy: 0.8523 - val_loss: 0.5166 - val_sparse_categorical_accuracy: 0.8584
Epoch 78/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5264 - sparse_categorical_accuracy: 0.8536 - val_loss: 0.5124 - val_sparse_categorical_accuracy: 0.8598
Epoch 79/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5221 - sparse_categorical_accuracy: 0.8548 - val_loss: 0.5083 - val_sparse_categorical_accuracy: 0.8614
Epoch 80/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5179 - sparse_categorical_accuracy: 0.8561 - val_loss: 0.5044 - val_sparse_categorical_accuracy: 0.8612
Epoch 81/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5139 - sparse_categorical_accuracy: 0.8572 - val_loss: 0.5005 - val_sparse_categorical_accuracy: 0.8625
Epoch 82/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5100 - sparse_categorical_accuracy: 0.8585 - val_loss: 0.4968 - val_sparse_categorical_accuracy: 0.8635
Epoch 83/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5062 - sparse_categorical_accuracy: 0.8595 - val_loss: 0.4932 - val_sparse_categorical_accuracy: 0.8640
Epoch 84/160
118/118 [==============================] - 0s 3ms/step - loss: 0.5025 - sparse_categorical_accuracy: 0.8602 - val_loss: 0.4897 - val_sparse_categorical_accuracy: 0.8650
Epoch 85/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4989 - sparse_categorical_accuracy: 0.8614 - val_loss: 0.4862 - val_sparse_categorical_accuracy: 0.8655
Epoch 86/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4954 - sparse_categorical_accuracy: 0.8620 - val_loss: 0.4829 - val_sparse_categorical_accuracy: 0.8661
Epoch 87/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4920 - sparse_categorical_accuracy: 0.8632 - val_loss: 0.4797 - val_sparse_categorical_accuracy: 0.8670
Epoch 88/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4887 - sparse_categorical_accuracy: 0.8638 - val_loss: 0.4766 - val_sparse_categorical_accuracy: 0.8680
Epoch 89/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4855 - sparse_categorical_accuracy: 0.8648 - val_loss: 0.4735 - val_sparse_categorical_accuracy: 0.8683
Epoch 90/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4823 - sparse_categorical_accuracy: 0.8656 - val_loss: 0.4705 - val_sparse_categorical_accuracy: 0.8692
Epoch 91/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4793 - sparse_categorical_accuracy: 0.8665 - val_loss: 0.4676 - val_sparse_categorical_accuracy: 0.8697
Epoch 92/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4763 - sparse_categorical_accuracy: 0.8673 - val_loss: 0.4648 - val_sparse_categorical_accuracy: 0.8706
Epoch 93/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4734 - sparse_categorical_accuracy: 0.8680 - val_loss: 0.4621 - val_sparse_categorical_accuracy: 0.8709
Epoch 94/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4705 - sparse_categorical_accuracy: 0.8689 - val_loss: 0.4594 - val_sparse_categorical_accuracy: 0.8714
Epoch 95/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4678 - sparse_categorical_accuracy: 0.8695 - val_loss: 0.4568 - val_sparse_categorical_accuracy: 0.8723
Epoch 96/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4651 - sparse_categorical_accuracy: 0.8702 - val_loss: 0.4542 - val_sparse_categorical_accuracy: 0.8733
Epoch 97/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4624 - sparse_categorical_accuracy: 0.8710 - val_loss: 0.4517 - val_sparse_categorical_accuracy: 0.8741
Epoch 98/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4598 - sparse_categorical_accuracy: 0.8713 - val_loss: 0.4492 - val_sparse_categorical_accuracy: 0.8743
Epoch 99/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4573 - sparse_categorical_accuracy: 0.8720 - val_loss: 0.4469 - val_sparse_categorical_accuracy: 0.8751
Epoch 100/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4548 - sparse_categorical_accuracy: 0.8730 - val_loss: 0.4445 - val_sparse_categorical_accuracy: 0.8756
Epoch 101/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4524 - sparse_categorical_accuracy: 0.8735 - val_loss: 0.4422 - val_sparse_categorical_accuracy: 0.8763
Epoch 102/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4501 - sparse_categorical_accuracy: 0.8741 - val_loss: 0.4400 - val_sparse_categorical_accuracy: 0.8768
Epoch 103/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4478 - sparse_categorical_accuracy: 0.8748 - val_loss: 0.4378 - val_sparse_categorical_accuracy: 0.8781
Epoch 104/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4455 - sparse_categorical_accuracy: 0.8755 - val_loss: 0.4358 - val_sparse_categorical_accuracy: 0.8787
Epoch 105/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4433 - sparse_categorical_accuracy: 0.8758 - val_loss: 0.4336 - val_sparse_categorical_accuracy: 0.8791
Epoch 106/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4411 - sparse_categorical_accuracy: 0.8768 - val_loss: 0.4316 - val_sparse_categorical_accuracy: 0.8793
Epoch 107/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4390 - sparse_categorical_accuracy: 0.8770 - val_loss: 0.4297 - val_sparse_categorical_accuracy: 0.8796
Epoch 108/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4369 - sparse_categorical_accuracy: 0.8777 - val_loss: 0.4277 - val_sparse_categorical_accuracy: 0.8804
Epoch 109/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4349 - sparse_categorical_accuracy: 0.8780 - val_loss: 0.4258 - val_sparse_categorical_accuracy: 0.8806
Epoch 110/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4329 - sparse_categorical_accuracy: 0.8787 - val_loss: 0.4239 - val_sparse_categorical_accuracy: 0.8811
Epoch 111/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4310 - sparse_categorical_accuracy: 0.8795 - val_loss: 0.4221 - val_sparse_categorical_accuracy: 0.8818
Epoch 112/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4290 - sparse_categorical_accuracy: 0.8799 - val_loss: 0.4203 - val_sparse_categorical_accuracy: 0.8826
Epoch 113/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4272 - sparse_categorical_accuracy: 0.8805 - val_loss: 0.4185 - val_sparse_categorical_accuracy: 0.8829
Epoch 114/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4253 - sparse_categorical_accuracy: 0.8808 - val_loss: 0.4167 - val_sparse_categorical_accuracy: 0.8834
Epoch 115/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4235 - sparse_categorical_accuracy: 0.8812 - val_loss: 0.4151 - val_sparse_categorical_accuracy: 0.8835
Epoch 116/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4217 - sparse_categorical_accuracy: 0.8818 - val_loss: 0.4134 - val_sparse_categorical_accuracy: 0.8843
Epoch 117/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4200 - sparse_categorical_accuracy: 0.8822 - val_loss: 0.4117 - val_sparse_categorical_accuracy: 0.8847
Epoch 118/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4183 - sparse_categorical_accuracy: 0.8828 - val_loss: 0.4102 - val_sparse_categorical_accuracy: 0.8852
Epoch 119/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4166 - sparse_categorical_accuracy: 0.8831 - val_loss: 0.4087 - val_sparse_categorical_accuracy: 0.8854
Epoch 120/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4150 - sparse_categorical_accuracy: 0.8838 - val_loss: 0.4071 - val_sparse_categorical_accuracy: 0.8858
Epoch 121/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4133 - sparse_categorical_accuracy: 0.8839 - val_loss: 0.4056 - val_sparse_categorical_accuracy: 0.8861
Epoch 122/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4117 - sparse_categorical_accuracy: 0.8844 - val_loss: 0.4040 - val_sparse_categorical_accuracy: 0.8864
Epoch 123/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4102 - sparse_categorical_accuracy: 0.8849 - val_loss: 0.4026 - val_sparse_categorical_accuracy: 0.8867
Epoch 124/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4086 - sparse_categorical_accuracy: 0.8852 - val_loss: 0.4012 - val_sparse_categorical_accuracy: 0.8867
Epoch 125/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4071 - sparse_categorical_accuracy: 0.8853 - val_loss: 0.3997 - val_sparse_categorical_accuracy: 0.8869
Epoch 126/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4056 - sparse_categorical_accuracy: 0.8859 - val_loss: 0.3984 - val_sparse_categorical_accuracy: 0.8869
Epoch 127/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4042 - sparse_categorical_accuracy: 0.8861 - val_loss: 0.3970 - val_sparse_categorical_accuracy: 0.8870
Epoch 128/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4027 - sparse_categorical_accuracy: 0.8862 - val_loss: 0.3956 - val_sparse_categorical_accuracy: 0.8875
Epoch 129/160
118/118 [==============================] - 0s 3ms/step - loss: 0.4013 - sparse_categorical_accuracy: 0.8870 - val_loss: 0.3943 - val_sparse_categorical_accuracy: 0.8879
Epoch 130/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3999 - sparse_categorical_accuracy: 0.8873 - val_loss: 0.3930 - val_sparse_categorical_accuracy: 0.8883
Epoch 131/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3985 - sparse_categorical_accuracy: 0.8877 - val_loss: 0.3917 - val_sparse_categorical_accuracy: 0.8882
Epoch 132/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3972 - sparse_categorical_accuracy: 0.8880 - val_loss: 0.3904 - val_sparse_categorical_accuracy: 0.8891
Epoch 133/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3958 - sparse_categorical_accuracy: 0.8884 - val_loss: 0.3891 - val_sparse_categorical_accuracy: 0.8890
Epoch 134/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3945 - sparse_categorical_accuracy: 0.8886 - val_loss: 0.3880 - val_sparse_categorical_accuracy: 0.8895
Epoch 135/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3932 - sparse_categorical_accuracy: 0.8889 - val_loss: 0.3868 - val_sparse_categorical_accuracy: 0.8898
Epoch 136/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3919 - sparse_categorical_accuracy: 0.8892 - val_loss: 0.3856 - val_sparse_categorical_accuracy: 0.8899
Epoch 137/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3907 - sparse_categorical_accuracy: 0.8896 - val_loss: 0.3844 - val_sparse_categorical_accuracy: 0.8900
Epoch 138/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3894 - sparse_categorical_accuracy: 0.8900 - val_loss: 0.3832 - val_sparse_categorical_accuracy: 0.8902
Epoch 139/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3882 - sparse_categorical_accuracy: 0.8902 - val_loss: 0.3821 - val_sparse_categorical_accuracy: 0.8906
Epoch 140/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3870 - sparse_categorical_accuracy: 0.8906 - val_loss: 0.3810 - val_sparse_categorical_accuracy: 0.8905
Epoch 141/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3858 - sparse_categorical_accuracy: 0.8907 - val_loss: 0.3799 - val_sparse_categorical_accuracy: 0.8907
Epoch 142/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3847 - sparse_categorical_accuracy: 0.8912 - val_loss: 0.3788 - val_sparse_categorical_accuracy: 0.8909
Epoch 143/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3835 - sparse_categorical_accuracy: 0.8917 - val_loss: 0.3777 - val_sparse_categorical_accuracy: 0.8918
Epoch 144/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3824 - sparse_categorical_accuracy: 0.8921 - val_loss: 0.3766 - val_sparse_categorical_accuracy: 0.8924
Epoch 145/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3812 - sparse_categorical_accuracy: 0.8923 - val_loss: 0.3757 - val_sparse_categorical_accuracy: 0.8924
Epoch 146/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3801 - sparse_categorical_accuracy: 0.8929 - val_loss: 0.3745 - val_sparse_categorical_accuracy: 0.8931
Epoch 147/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3790 - sparse_categorical_accuracy: 0.8930 - val_loss: 0.3736 - val_sparse_categorical_accuracy: 0.8937
Epoch 148/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3780 - sparse_categorical_accuracy: 0.8934 - val_loss: 0.3726 - val_sparse_categorical_accuracy: 0.8939
Epoch 149/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3769 - sparse_categorical_accuracy: 0.8934 - val_loss: 0.3716 - val_sparse_categorical_accuracy: 0.8939
Epoch 150/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3759 - sparse_categorical_accuracy: 0.8938 - val_loss: 0.3706 - val_sparse_categorical_accuracy: 0.8941
Epoch 151/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3748 - sparse_categorical_accuracy: 0.8939 - val_loss: 0.3696 - val_sparse_categorical_accuracy: 0.8945
Epoch 152/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3738 - sparse_categorical_accuracy: 0.8942 - val_loss: 0.3686 - val_sparse_categorical_accuracy: 0.8943
Epoch 153/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3728 - sparse_categorical_accuracy: 0.8947 - val_loss: 0.3676 - val_sparse_categorical_accuracy: 0.8947
Epoch 154/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3718 - sparse_categorical_accuracy: 0.8948 - val_loss: 0.3668 - val_sparse_categorical_accuracy: 0.8952
Epoch 155/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3708 - sparse_categorical_accuracy: 0.8951 - val_loss: 0.3658 - val_sparse_categorical_accuracy: 0.8953
Epoch 156/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3698 - sparse_categorical_accuracy: 0.8953 - val_loss: 0.3650 - val_sparse_categorical_accuracy: 0.8960
Epoch 157/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3689 - sparse_categorical_accuracy: 0.8956 - val_loss: 0.3641 - val_sparse_categorical_accuracy: 0.8962
Epoch 158/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3679 - sparse_categorical_accuracy: 0.8959 - val_loss: 0.3632 - val_sparse_categorical_accuracy: 0.8967
Epoch 159/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3670 - sparse_categorical_accuracy: 0.8962 - val_loss: 0.3623 - val_sparse_categorical_accuracy: 0.8966
Epoch 160/160
118/118 [==============================] - 0s 3ms/step - loss: 0.3660 - sparse_categorical_accuracy: 0.8963 - val_loss: 0.3614 - val_sparse_categorical_accuracy: 0.8971





<tensorflow.python.keras.callbacks.History at 0x21359af4160>
搬砖大法师
关注
  • 0
    点赞
  • 踩
  • 0
    收藏
    觉得还不错? 一键收藏
  • 1
    评论
专栏目录
在tensorflow中训练mnist_data模型
qq_41821067的博客
03-24 381
import tensorflow as tf from tensorflow.examples.tutorials.mnist import input_data mnist_data = input_data.read_data_sets('MNIST_data', one_hot=True) input_size = 784 #数字从0-9 no_classes = 10 batch_size = 100 total_batches = 200 x_input = tf.placeholder(
示如何构建和训练一个简单的神经网络来进行手写数字识别任务，使用经典的 MNIST 数据集
最新发布
05-27
这个示例将展示如何构建和训练一个简单的神经网络来进行手写数字识别任务，使用经典的 MNIST 数据集。 首先，确保您已安装 TensorFlow 和 Keras。您可以使用以下命令进行安装： pip install tensorflow 代码说明 导...
【TensorFlow2.0】经典数据集加载与数据集处理
qq_36926037的博客
04-28 2307
  利用tensorflow提供的工具便捷的加载经典数据集。自定义的数据集以后再讲。   tensorflow中keras.Datasets数据集对象，方便实现多线程、预处理、随机打散、批训练等常用数据集的功能。   常用经典数据集：   （1）Boston housing：波士顿房价趋势数据集   （2）CIFAR10/100：真实图片数据集，用于图片分类   （3）MINIST/Fashion...
tensorflow Keras的搭建训练步骤
老邹的博客
12-28 1866
keras搭建训练模型步骤
使用tensorflow中的keras(自定义model)
wdxgj123的博客
07-03 2702
使用tensorflow中的keras(自定义model) import tensorflow as tf from tensorflow.keras import datasets, layers, optimizers, Sequential, metrics from tensorflow import keras def preprocess(x, y): """ x is a simple image, not a batch """ x = tf.cast(x
tensorflow2.x自建数据集训练
smallworldxyl的博客
10-19 1074
1.导入库 import tensorflow as tf from tensorflow import keras from tensorflow.keras import datasets, layers, optimizers, Sequential, metrics from tensorflow.python.framework.convert_to_constants import convert_variables_to_constants_v2 import os import pathli
PGD_Towards deep learning models resistant to adversarial attacks_CSDN
qq_32925101的博客
12-17 2362
Towards Deep Learning Models Resistant to Adversarial Attacks Towards Deep Learning Models Resistant to Adversarial Attacks (PGD)，ICLR2018，涉及PGD和对抗训练。 **Abstract：**本文从优化的角度研究了神经网络的对抗鲁棒性问题。本文提出的方法提供了一个广阔、统一的观点来看待对抗样本的问题。本文提出方法的自然性质使得我们可以可靠的选择训练和攻击神经网络的方法，并且
【Keras】30 秒上手 Keras+实例对mnist手写数字进行识别准确率达99%以上
Allen . Liu
12-18 1279
本文我们将学习使用Keras一步一步搭建一个卷积神经网络。具体来说，我们将使用卷积神经网络对手写数字(MNIST数据集)进行识别，并达到99%以上的正确率。 @为什么选择Keras呢？ 主要是因为简单方便。更多细节请看：https://keras.io/ @什么卷积神经网络？ 简单地说，卷积神经网络(CNNs)是一种多层神经网络，它可以有效地减少全连接神经网络参数量太大的问题。 下面就直接进入...
DAY1-深度学习100例-卷积神经网络（CNN）实现mnist手写数字识别
weixin_44336912的博客
08-06 327
Mnist数据集分为两部分，分别含有60000张训练图片和10000张测试图片。每一张图片包含28*28个像素。Mnist数据集把代表一张图片的二维数据转换成一个向量，长度为28 * 28=784。因此在Mnist的训练数据集中mnist.train.images是一个形状为[60000, 784]的张量，第一个维度数字用来索引图片，第二个维度数字用来索引每张图片中的像素点，图片里的某个像素的强度值介于0-1之间。...
TensorFlow(Keras) 一步步实现MNIST手写体图片分类 (1) Coursera深度学习教程分享
dongfangxiaozi_的专栏
04-11 458
Coursera Tensorflow和ML/DL 机器学习/深度学习教程，分享：TensorFlow一步步手写体Mnist识别<一> TensorFlow和Andrew Ng 的 DeepLearning.AI合作推出了一个简单的TensorFlow入门教程。但是看了一下，这简直是Keras入门教程，对于笔者来说简直就像回头看小学数学，但温故而知新可以为师矣。因为Keras的作者加...
mnist 识别率达98%以上，学习率lr越来越小，优化器AdamOptimizer
weixin_30617797的博客
11-23 194
import tensorflow as tffrom tensorflow.examples.tutorials.mnist import input_data#载入数据集mnist = input_data.read_data_sets("MNIST_data",one_hot=True)#每个批次的大小batch_size = 100#计算一共有多少个批次n_batch = mni...
CNN卷积神经网络训练并测试Mnist数据集
04-21
CNN卷积神经网络训练并测试Mnist数据集，准确率99.07%,环境是pytorch+GPU+pycharm(使用cuda),可直接下载使用，适合想入门深度学习或者神经网络的初学者，代码可直接跑，并且代码中含有大量注释以及个人见解。...
基于Keras的Mnist手写体识别
09-05
构建模型：详细解释如何使用Keras构建一个简单的卷积神经网络（CNN）模型。包括构建模型结构，定义优化器，设置损失函数等步骤。 模型训练：说明如何使用构建的模型对MNIST数据集进行训练，并展示训练过程中的准确率...
tensoflow2.0学习笔记——手写识别（MNIST数据）的两种方法，附代码。
01-20
一、用tensorflow构建一个基本的神经网络，用于预测手写数字。  本例为3层神经网，包括一个输入层，两个隐层，一个输出层。  输入数据大小：28*28dpi的手写图片，即（28*28）的二维数组，展平后为长度784的数组。 ...
基于tensorflow深度学习VGG-19图像风格迁移+自动去噪（MNIST数据集）机器学习+人工智能+神经网络
05-25
1.项目基于 MNIST 数据集，使用 VGG-19 网络模型，将图像进行风格迁移，实现去噪功能。 2.项目运行环境：Python 和 TensorFlow 运行环境。需要 Python 3.6 及以上配置，使用conda安装环境 conda create -n ...
keras对minist数据集识别
qq_27328197的博客
02-23 380
文章目录前言步骤读取数据数据探索数据处理搭建网络模型模型训练总结 前言 什么是AI？ The theory and development of computer systems able to perform tasks normally requiring human intelligence.(–Oxford Dictionary) Using data to solve problems.(–cy) 步骤 先导包 import tensorflow as tf #print(tf.__version
TensorFlow2.0 Keras CNN识别Mnist数据集
Poor - Because you have no ambition
08-22 2360
from keras.datasets import mnist from keras.utils import np_utils import numpy as np np.random.seed(10) (x_train, y_train), (x_test, y_test) = mnist.load_data() x_train40 = x_train.reshape(x_train.s...
解决 from tensorflow.keras import datasets,layers, optimizers报错的问题， SGD没有apply_gradients
热门推荐
qq_44644355的博客
05-05 1万+
首先我的tensorflow-gpu版本是2.0.0 解决方法： from tensorflow_core.python.keras import datasets, layers 问题描述： 我是条看视频的时候，有这样一句代码： from tensorflow.keras import datasets, layers 但是我自己输入的时候会报错，然后我看了一下代码所在环境下的包site-pac...
Tensorflow2实现常用的卷积神经网络
FriendshipTang的博客
02-05 632
Tensorflow2实现常用的卷积神经网络LeNetAlexNetVGGVGG16VGG19 LeNet import tensorflow as tf from tensorflow.keras import datasets, layers, models # LeNet-5 model = models.Sequential(name='LeNet-5') model.add(layers.Conv2D(6, (5, 5), activation='sigmoid', input_shape=(3
keras实现cnn:手写数字识别准确率99.6%
02-07
Keras是一种用户友好的深度学习库，能够简化卷积神经网络（CNN）的实现过程。通过使用Keras库，我们可以非常轻松地实现手写数字识别的模型，并达到99.6%的准确率。 首先，我们需要准备一个适用于手写数字识别的数据...

“相关推荐”对你有帮助么？

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论 1
添加红包

请填写红包祝福语或标题

个

红包个数最小为10个

元

红包金额最低5元

当前余额3.43元 前往充值 >
需支付：10.00元
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付元
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明：

1.余额是钱包充值的虚拟货币，按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载，可以购买VIP、付费专栏及课程。

余额充值