代码

# -*- coding: utf-8 -*-
"""
Created on  2020/11/20 14:04
@Author: CY
@email: 5844104706@qq.com
"""
# Hyperparameters
## Model hyperparameters
## Algorithm hyperparameters
import tensorflow as tf
from tensorflow import kerasimport IPython
# !pip install -q -U keras-tuner
import kerastuner as ktprint("#1. 数据集 Fashion MNIST dataset.")
(img_train, label_train), (img_test, label_test) = keras.datasets.fashion_mnist.load_data()
# Normalize pixel values between 0 and 1
img_train = img_train.astype('float32') / 255.0
img_test = img_test.astype('float32') / 255.0print("#2. 模型定义")def model_builder(hp):model = keras.Sequential()model.add(keras.layers.Flatten(input_shape=(28, 28)))# Tune the number of units in the first Dense layer# Choose an optimal value between 32-512hp_units = hp.Int('units', min_value=32, max_value=512, step=32)model.add(keras.layers.Dense(units=hp_units, activation='relu'))model.add(keras.layers.Dense(10))# Tune the learning rate for the optimizer# Choose an optimal value from 0.01, 0.001, or 0.0001hp_learning_rate = hp.Choice('learning_rate', values=[1e-2, 1e-3, 1e-4])model.compile(optimizer=keras.optimizers.Adam(learning_rate=hp_learning_rate),loss=keras.losses.SparseCategoricalCrossentropy(from_logits=True),metrics=['accuracy'])return modeltuner = kt.Hyperband(model_builder,objective='val_accuracy',max_epochs=10,factor=3,directory='my_dir',project_name='intro_to_kt')
class ClearTrainingOutput(tf.keras.callbacks.Callback):def on_train_end(*args, **kwargs):IPython.display.clear_output(wait= True)tuner.search(img_train, label_train, epochs = 10, validation_data = (img_test, label_test), callbacks = [ClearTrainingOutput()])# Get the optimal hyperparameters
best_hps = tuner.get_best_hyperparameters(num_trials = 1)[0]print(f"""
The hyperparameter search is complete. The optimal number of units in the first densely-connected
layer is {best_hps.get('units')} and the optimal learning rate for the optimizer
is {best_hps.get('learning_rate')}.
""")# Build the model with the optimal hyperparameters and train it on the data
model = tuner.hypermodel.build(best_hps)
model.fit(img_train, label_train, epochs = 10, validation_data = (img_test, label_test))

运行结果

Epoch 1/2
1875/1875 [==============================] - 4s 2ms/step - loss: 0.6146 - accuracy: 0.7994 - val_loss: 0.4803 - val_accuracy: 0.8346
Epoch 2/2
1875/1875 [==============================] - 4s 2ms/step - loss: 0.4323 - accuracy: 0.8503 - val_loss: 0.4366 - val_accuracy: 0.8492
[Trial complete]
[Trial summary]|-Trial ID: 191820c4c1db0f9e0887d266e14cbd00|-Score: 0.8492000102996826|-Best step: 0> Hyperparameters:|-learning_rate: 0.0001|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 384
Epoch 1/2
1875/1875 [==============================] - 2s 875us/step - loss: 0.7257 - accuracy: 0.7698 - val_loss: 0.5402 - val_accuracy: 0.8172
Epoch 2/2
1875/1875 [==============================] - 2s 942us/step - loss: 0.4874 - accuracy: 0.8370 - val_loss: 0.4871 - val_accuracy: 0.8330
[Trial complete]
[Trial summary]|-Trial ID: 779b0ec8dbfceee7e4b7166b20aa3d8f|-Score: 0.8330000042915344|-Best step: 0> Hyperparameters:|-learning_rate: 0.0001|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 96
Epoch 1/2
1875/1875 [==============================] - 5s 3ms/step - loss: 0.5946 - accuracy: 0.8058 - val_loss: 0.4797 - val_accuracy: 0.8346
Epoch 2/2
1875/1875 [==============================] - 5s 3ms/step - loss: 0.4261 - accuracy: 0.8531 - val_loss: 0.4372 - val_accuracy: 0.8464
[Trial complete]
[Trial summary]|-Trial ID: d3a96eb83cc740ec1e5db373d9fd55c3|-Score: 0.8464000225067139|-Best step: 0> Hyperparameters:|-learning_rate: 0.0001|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 448
Epoch 1/2
1875/1875 [==============================] - 3s 2ms/step - loss: 0.6247 - accuracy: 0.7956 - val_loss: 0.4990 - val_accuracy: 0.8292
Epoch 2/2
1875/1875 [==============================] - 3s 1ms/step - loss: 0.4386 - accuracy: 0.8506 - val_loss: 0.4458 - val_accuracy: 0.8466
[Trial complete]
[Trial summary]|-Trial ID: 5e8585e74ca9cba4e8a4064cbc7bd668|-Score: 0.8465999960899353|-Best step: 0> Hyperparameters:|-learning_rate: 0.0001|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 320
Epoch 1/2
1875/1875 [==============================] - 6s 3ms/step - loss: 0.5458 - accuracy: 0.8098 - val_loss: 0.4780 - val_accuracy: 0.8282
Epoch 2/2
1875/1875 [==============================] - 5s 3ms/step - loss: 0.4339 - accuracy: 0.8444 - val_loss: 0.4672 - val_accuracy: 0.8367
[Trial complete]
[Trial summary]|-Trial ID: 0d8ec7608c018b20fd9b4cad6e66a9b6|-Score: 0.8367000222206116|-Best step: 0> Hyperparameters:|-learning_rate: 0.01|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 480
Epoch 1/2
1875/1875 [==============================] - 2s 983us/step - loss: 0.5315 - accuracy: 0.8114 - val_loss: 0.4696 - val_accuracy: 0.8292
Epoch 2/2
1875/1875 [==============================] - 2s 983us/step - loss: 0.4341 - accuracy: 0.8454 - val_loss: 0.4831 - val_accuracy: 0.8317
[Trial complete]
[Trial summary]|-Trial ID: e53c8d4b661b5520e392b34a65f0ca1d|-Score: 0.8317000269889832|-Best step: 0> Hyperparameters:|-learning_rate: 0.01|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 160
Epoch 1/2
1875/1875 [==============================] - 2s 883us/step - loss: 0.5238 - accuracy: 0.8195 - val_loss: 0.4598 - val_accuracy: 0.8344
Epoch 2/2
1875/1875 [==============================] - 2s 842us/step - loss: 0.3932 - accuracy: 0.8602 - val_loss: 0.3968 - val_accuracy: 0.8580
[Trial complete]
[Trial summary]|-Trial ID: 57f46babcb66e2f2829177035445f65c|-Score: 0.8579999804496765|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 64
Epoch 1/2
1875/1875 [==============================] - 2s 1ms/step - loss: 0.4923 - accuracy: 0.8249 - val_loss: 0.4501 - val_accuracy: 0.8431
Epoch 2/2
1875/1875 [==============================] - 2s 983us/step - loss: 0.3672 - accuracy: 0.8670 - val_loss: 0.3996 - val_accuracy: 0.8564
[Trial complete]
[Trial summary]|-Trial ID: 59551e7ff2ab521be9ae98c3c76f764b|-Score: 0.8564000129699707|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 192
Epoch 1/2
1875/1875 [==============================] - 2s 942us/step - loss: 0.5265 - accuracy: 0.8126 - val_loss: 0.6236 - val_accuracy: 0.7760
Epoch 2/2
1875/1875 [==============================] - 2s 917us/step - loss: 0.4402 - accuracy: 0.8414 - val_loss: 0.4718 - val_accuracy: 0.8286
[Trial complete]
[Trial summary]|-Trial ID: b566a3dd369847ae7e6064563cb3d444|-Score: 0.8285999894142151|-Best step: 0> Hyperparameters:|-learning_rate: 0.01|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 128
Epoch 1/2
1875/1875 [==============================] - 2s 825us/step - loss: 0.5264 - accuracy: 0.8130 - val_loss: 0.4499 - val_accuracy: 0.8422
Epoch 2/2
1875/1875 [==============================] - 2s 843us/step - loss: 0.4450 - accuracy: 0.8419 - val_loss: 0.4610 - val_accuracy: 0.8329
[Trial complete]
[Trial summary]|-Trial ID: a29d6687b5dd7f074f320d273567ff01|-Score: 0.842199981212616|-Best step: 0> Hyperparameters:|-learning_rate: 0.01|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 64
Epoch 1/2
1875/1875 [==============================] - 3s 1ms/step - loss: 0.6550 - accuracy: 0.7891 - val_loss: 0.5078 - val_accuracy: 0.8273
Epoch 2/2
1875/1875 [==============================] - 2s 1ms/step - loss: 0.4537 - accuracy: 0.8454 - val_loss: 0.4598 - val_accuracy: 0.8418
[Trial complete]
[Trial summary]|-Trial ID: d64f04c0c7f461eabfca4f7c97a28f43|-Score: 0.8417999744415283|-Best step: 0> Hyperparameters:|-learning_rate: 0.0001|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 224
Epoch 1/2
1875/1875 [==============================] - 2s 945us/step - loss: 0.4904 - accuracy: 0.8273 - val_loss: 0.4384 - val_accuracy: 0.8435
Epoch 2/2
1875/1875 [==============================] - 2s 946us/step - loss: 0.3706 - accuracy: 0.8664 - val_loss: 0.3780 - val_accuracy: 0.8651
[Trial complete]
[Trial summary]|-Trial ID: ff8dd848cb342726b8552ff47ec287f1|-Score: 0.8651000261306763|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 2|-tuner/epochs: 2|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 160
Epoch 3/4
1875/1875 [==============================] - 2s 998us/step - loss: 0.4912 - accuracy: 0.8274 - val_loss: 0.4180 - val_accuracy: 0.8528
Epoch 4/4
1875/1875 [==============================] - 2s 917us/step - loss: 0.3691 - accuracy: 0.8672 - val_loss: 0.3944 - val_accuracy: 0.8602
[Trial complete]
[Trial summary]|-Trial ID: 6ecbc23cf3d4d28965a809d247ed8cd8|-Score: 0.8601999878883362|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 2|-tuner/epochs: 4|-tuner/initial_epoch: 2|-tuner/round: 1|-tuner/trial_id: ff8dd848cb342726b8552ff47ec287f1|-units: 160
Epoch 3/4
1875/1875 [==============================] - 2s 816us/step - loss: 0.5227 - accuracy: 0.8191 - val_loss: 0.4512 - val_accuracy: 0.8399
Epoch 4/4
1875/1875 [==============================] - 2s 851us/step - loss: 0.3965 - accuracy: 0.8584 - val_loss: 0.4245 - val_accuracy: 0.8439
[Trial complete]
[Trial summary]|-Trial ID: 002b7310fa5aa135e0b011b5c74053a9|-Score: 0.8439000248908997|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 2|-tuner/epochs: 4|-tuner/initial_epoch: 2|-tuner/round: 1|-tuner/trial_id: 57f46babcb66e2f2829177035445f65c|-units: 64
Epoch 3/4
1875/1875 [==============================] - 2s 1ms/step - loss: 0.4887 - accuracy: 0.8270 - val_loss: 0.4204 - val_accuracy: 0.8476
Epoch 4/4
1875/1875 [==============================] - 2s 1ms/step - loss: 0.3679 - accuracy: 0.8658 - val_loss: 0.3753 - val_accuracy: 0.8646
[Trial complete]
[Trial summary]|-Trial ID: 4fffdb277afe57a86bbac8e5c6988256|-Score: 0.8646000027656555|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 2|-tuner/epochs: 4|-tuner/initial_epoch: 2|-tuner/round: 1|-tuner/trial_id: 59551e7ff2ab521be9ae98c3c76f764b|-units: 192
Epoch 3/4
1875/1875 [==============================] - 5s 2ms/step - loss: 0.6099 - accuracy: 0.7977 - val_loss: 0.4864 - val_accuracy: 0.8295
Epoch 4/4
1875/1875 [==============================] - 5s 2ms/step - loss: 0.4346 - accuracy: 0.8508 - val_loss: 0.4445 - val_accuracy: 0.8406
[Trial complete]
[Trial summary]|-Trial ID: e34528c7b96b72621522837e3050c77d|-Score: 0.8406000137329102|-Best step: 0> Hyperparameters:|-learning_rate: 0.0001|-tuner/bracket: 2|-tuner/epochs: 4|-tuner/initial_epoch: 2|-tuner/round: 1|-tuner/trial_id: 191820c4c1db0f9e0887d266e14cbd00|-units: 384
Epoch 5/10
1875/1875 [==============================] - 2s 1ms/step - loss: 0.4848 - accuracy: 0.8283 - val_loss: 0.4595 - val_accuracy: 0.8349
Epoch 6/10
1875/1875 [==============================] - 2s 983us/step - loss: 0.3695 - accuracy: 0.8666 - val_loss: 0.4070 - val_accuracy: 0.8504
Epoch 7/10
1875/1875 [==============================] - 2s 992us/step - loss: 0.3314 - accuracy: 0.8776 - val_loss: 0.3604 - val_accuracy: 0.8675
Epoch 8/10
1875/1875 [==============================] - 2s 1ms/step - loss: 0.3057 - accuracy: 0.8881 - val_loss: 0.3576 - val_accuracy: 0.8721
Epoch 9/10
1875/1875 [==============================] - 2s 1ms/step - loss: 0.2885 - accuracy: 0.8935 - val_loss: 0.3380 - val_accuracy: 0.8804
Epoch 10/10
1875/1875 [==============================] - 2s 1ms/step - loss: 0.2744 - accuracy: 0.8990 - val_loss: 0.3404 - val_accuracy: 0.8789
[Trial complete]
[Trial summary]|-Trial ID: ce7c97b6b272a033a21f8740aefb0fee|-Score: 0.8804000020027161|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 2|-tuner/epochs: 10|-tuner/initial_epoch: 4|-tuner/round: 2|-tuner/trial_id: 4fffdb277afe57a86bbac8e5c6988256|-units: 192
Epoch 5/10
1875/1875 [==============================] - 2s 975us/step - loss: 0.4869 - accuracy: 0.8296 - val_loss: 0.4113 - val_accuracy: 0.8551
Epoch 6/10
1875/1875 [==============================] - 2s 942us/step - loss: 0.3654 - accuracy: 0.8677 - val_loss: 0.3796 - val_accuracy: 0.8647
Epoch 7/10
1875/1875 [==============================] - 2s 925us/step - loss: 0.3290 - accuracy: 0.8804 - val_loss: 0.3623 - val_accuracy: 0.8699
Epoch 8/10
1875/1875 [==============================] - 2s 983us/step - loss: 0.3056 - accuracy: 0.8869 - val_loss: 0.3561 - val_accuracy: 0.8705
Epoch 9/10
1875/1875 [==============================] - 2s 958us/step - loss: 0.2868 - accuracy: 0.8945 - val_loss: 0.3568 - val_accuracy: 0.8735
Epoch 10/10
1875/1875 [==============================] - 2s 975us/step - loss: 0.2736 - accuracy: 0.8979 - val_loss: 0.3435 - val_accuracy: 0.8739
[Trial complete]
[Trial summary]|-Trial ID: 4984706cf32ca96cad85570be7030a65|-Score: 0.8738999962806702|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 2|-tuner/epochs: 10|-tuner/initial_epoch: 4|-tuner/round: 2|-tuner/trial_id: 6ecbc23cf3d4d28965a809d247ed8cd8|-units: 160
Epoch 1/4
1875/1875 [==============================] - 2s 850us/step - loss: 0.5127 - accuracy: 0.8209 - val_loss: 0.4297 - val_accuracy: 0.8448
Epoch 2/4
1875/1875 [==============================] - 2s 867us/step - loss: 0.3832 - accuracy: 0.8624 - val_loss: 0.3865 - val_accuracy: 0.8606
Epoch 3/4
1875/1875 [==============================] - 1s 784us/step - loss: 0.3428 - accuracy: 0.8755 - val_loss: 0.4015 - val_accuracy: 0.8521
Epoch 4/4
1875/1875 [==============================] - 1s 792us/step - loss: 0.3193 - accuracy: 0.8836 - val_loss: 0.3597 - val_accuracy: 0.8709
[Trial complete]
[Trial summary]|-Trial ID: 4d4cbfecb20eba350c19ef24eb93b5fe|-Score: 0.8708999752998352|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 1|-tuner/epochs: 4|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 96
Epoch 1/4
1875/1875 [==============================] - 5s 3ms/step - loss: 0.4795 - accuracy: 0.8296 - val_loss: 0.4015 - val_accuracy: 0.8555
Epoch 2/4
1875/1875 [==============================] - 5s 2ms/step - loss: 0.3593 - accuracy: 0.8692 - val_loss: 0.3843 - val_accuracy: 0.8551
Epoch 3/4
1875/1875 [==============================] - 5s 2ms/step - loss: 0.3249 - accuracy: 0.8806 - val_loss: 0.3704 - val_accuracy: 0.8662
Epoch 4/4
1875/1875 [==============================] - 5s 2ms/step - loss: 0.2993 - accuracy: 0.8898 - val_loss: 0.3475 - val_accuracy: 0.8764
[Trial complete]
[Trial summary]|-Trial ID: 88acc74da2222e8e3065b173d0d7a29e|-Score: 0.8763999938964844|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 1|-tuner/epochs: 4|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 416
Epoch 1/4
1875/1875 [==============================] - 2s 1ms/step - loss: 0.5466 - accuracy: 0.8072 - val_loss: 0.4805 - val_accuracy: 0.8332
Epoch 2/4
1875/1875 [==============================] - 2s 1ms/step - loss: 0.4400 - accuracy: 0.8423 - val_loss: 0.4656 - val_accuracy: 0.8398
Epoch 3/4
1875/1875 [==============================] - 2s 1ms/step - loss: 0.4175 - accuracy: 0.8500 - val_loss: 0.4765 - val_accuracy: 0.8367
Epoch 4/4
1875/1875 [==============================] - 2s 1ms/step - loss: 0.4046 - accuracy: 0.8546 - val_loss: 0.4595 - val_accuracy: 0.8384
[Trial complete]
[Trial summary]|-Trial ID: 2b1c17016314b1718e50f25dcec92d7b|-Score: 0.8398000001907349|-Best step: 0> Hyperparameters:|-learning_rate: 0.01|-tuner/bracket: 1|-tuner/epochs: 4|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 256
Epoch 1/4
1875/1875 [==============================] - 5s 3ms/step - loss: 0.5373 - accuracy: 0.8137 - val_loss: 0.4528 - val_accuracy: 0.8397
Epoch 2/4
1875/1875 [==============================] - 5s 2ms/step - loss: 0.4338 - accuracy: 0.8440 - val_loss: 0.4934 - val_accuracy: 0.8261
Epoch 3/4
1875/1875 [==============================] - 4s 2ms/step - loss: 0.4134 - accuracy: 0.8540 - val_loss: 0.4633 - val_accuracy: 0.8330
Epoch 4/4
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3976 - accuracy: 0.8587 - val_loss: 0.4684 - val_accuracy: 0.8372
[Trial complete]
[Trial summary]|-Trial ID: 6da41f58d68b1e0b4716461f158371d3|-Score: 0.8396999835968018|-Best step: 0> Hyperparameters:|-learning_rate: 0.01|-tuner/bracket: 1|-tuner/epochs: 4|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 416
Epoch 1/4
1875/1875 [==============================] - 2s 800us/step - loss: 0.5390 - accuracy: 0.8118 - val_loss: 0.4523 - val_accuracy: 0.8395
Epoch 2/4
1875/1875 [==============================] - 2s 1ms/step - loss: 0.4122 - accuracy: 0.8546 - val_loss: 0.4267 - val_accuracy: 0.8498
Epoch 3/4
1875/1875 [==============================] - 2s 992us/step - loss: 0.3762 - accuracy: 0.8647 - val_loss: 0.4199 - val_accuracy: 0.8429
Epoch 4/4
1875/1875 [==============================] - 1s 667us/step - loss: 0.3536 - accuracy: 0.8725 - val_loss: 0.3799 - val_accuracy: 0.8678
[Trial complete]
[Trial summary]|-Trial ID: 9f027191b67b74ce4794e5affaa42a49|-Score: 0.8677999973297119|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 1|-tuner/epochs: 4|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 32
Epoch 1/4
1875/1875 [==============================] - 4s 2ms/step - loss: 0.4789 - accuracy: 0.8303 - val_loss: 0.4434 - val_accuracy: 0.8387
Epoch 2/4
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3612 - accuracy: 0.8680 - val_loss: 0.3609 - val_accuracy: 0.8679
Epoch 3/4
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3224 - accuracy: 0.8812 - val_loss: 0.3557 - val_accuracy: 0.8750
Epoch 4/4
1875/1875 [==============================] - 3s 2ms/step - loss: 0.2992 - accuracy: 0.8892 - val_loss: 0.3677 - val_accuracy: 0.8670
[Trial complete]
[Trial summary]|-Trial ID: 8b44a5bbd769e9fd3da6efc318b10647|-Score: 0.875|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 1|-tuner/epochs: 4|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 352
Epoch 5/10
1875/1875 [==============================] - 5s 3ms/step - loss: 0.4782 - accuracy: 0.8289 - val_loss: 0.4330 - val_accuracy: 0.8473
Epoch 6/10
1875/1875 [==============================] - 5s 2ms/step - loss: 0.3622 - accuracy: 0.8679 - val_loss: 0.3994 - val_accuracy: 0.8564
Epoch 7/10
1875/1875 [==============================] - 5s 2ms/step - loss: 0.3250 - accuracy: 0.8797 - val_loss: 0.3655 - val_accuracy: 0.8673
Epoch 8/10
1875/1875 [==============================] - 5s 2ms/step - loss: 0.3014 - accuracy: 0.8880 - val_loss: 0.3417 - val_accuracy: 0.8759
Epoch 9/10
1875/1875 [==============================] - 5s 2ms/step - loss: 0.2825 - accuracy: 0.8954 - val_loss: 0.3363 - val_accuracy: 0.8777
Epoch 10/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.2682 - accuracy: 0.8992 - val_loss: 0.3413 - val_accuracy: 0.8767
[Trial complete]
[Trial summary]|-Trial ID: c5909f061215673b01109e1a65f5676d|-Score: 0.8776999711990356|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 1|-tuner/epochs: 10|-tuner/initial_epoch: 4|-tuner/round: 1|-tuner/trial_id: 88acc74da2222e8e3065b173d0d7a29e|-units: 416
Epoch 5/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.4758 - accuracy: 0.8317 - val_loss: 0.4028 - val_accuracy: 0.8563
Epoch 6/10
1875/1875 [==============================] - 3s 2ms/step - loss: 0.3600 - accuracy: 0.8681 - val_loss: 0.3870 - val_accuracy: 0.8608
Epoch 7/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3208 - accuracy: 0.8810 - val_loss: 0.3579 - val_accuracy: 0.8703
Epoch 8/10
1875/1875 [==============================] - 3s 2ms/step - loss: 0.3006 - accuracy: 0.8868 - val_loss: 0.3434 - val_accuracy: 0.8761
Epoch 9/10
1875/1875 [==============================] - 3s 2ms/step - loss: 0.2815 - accuracy: 0.8952 - val_loss: 0.3626 - val_accuracy: 0.8712
Epoch 10/10
1875/1875 [==============================] - 3s 2ms/step - loss: 0.2678 - accuracy: 0.8998 - val_loss: 0.3343 - val_accuracy: 0.8820
[Trial complete]
[Trial summary]|-Trial ID: 4333cf5ce3c7e1004becdc45a151f7e9|-Score: 0.8820000290870667|-Best step: 0> Hyperparameters:|-learning_rate: 0.001|-tuner/bracket: 1|-tuner/epochs: 10|-tuner/initial_epoch: 4|-tuner/round: 1|-tuner/trial_id: 8b44a5bbd769e9fd3da6efc318b10647|-units: 352
Epoch 1/10
1875/1875 [==============================] - 2s 917us/step - loss: 0.7114 - accuracy: 0.7708 - val_loss: 0.5391 - val_accuracy: 0.8180
Epoch 2/10
1875/1875 [==============================] - 2s 950us/step - loss: 0.4786 - accuracy: 0.8386 - val_loss: 0.4967 - val_accuracy: 0.8257
Epoch 3/10
1875/1875 [==============================] - 2s 933us/step - loss: 0.4343 - accuracy: 0.8505 - val_loss: 0.4532 - val_accuracy: 0.8405
Epoch 4/10
1875/1875 [==============================] - 2s 925us/step - loss: 0.4087 - accuracy: 0.8582 - val_loss: 0.4308 - val_accuracy: 0.8474
Epoch 5/10
1875/1875 [==============================] - 2s 942us/step - loss: 0.3899 - accuracy: 0.8651 - val_loss: 0.4148 - val_accuracy: 0.8524
Epoch 6/10
1875/1875 [==============================] - 2s 967us/step - loss: 0.3752 - accuracy: 0.8688 - val_loss: 0.4058 - val_accuracy: 0.8561
Epoch 7/10
1875/1875 [==============================] - 2s 950us/step - loss: 0.3629 - accuracy: 0.8734 - val_loss: 0.3968 - val_accuracy: 0.8604
Epoch 8/10
1875/1875 [==============================] - 2s 917us/step - loss: 0.3526 - accuracy: 0.8770 - val_loss: 0.3923 - val_accuracy: 0.8607
Epoch 9/10
1875/1875 [==============================] - 2s 925us/step - loss: 0.3435 - accuracy: 0.8795 - val_loss: 0.3870 - val_accuracy: 0.8609
Epoch 10/10
1875/1875 [==============================] - 2s 892us/step - loss: 0.3354 - accuracy: 0.8820 - val_loss: 0.3771 - val_accuracy: 0.8657
[Trial complete]
[Trial summary]|-Trial ID: af4ca9415caf6a4e05807e452f15b095|-Score: 0.8657000064849854|-Best step: 0> Hyperparameters:|-learning_rate: 0.0001|-tuner/bracket: 0|-tuner/epochs: 10|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 128
Epoch 1/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.6247 - accuracy: 0.7961 - val_loss: 0.4982 - val_accuracy: 0.8314
Epoch 2/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.4378 - accuracy: 0.8496 - val_loss: 0.4520 - val_accuracy: 0.8397
Epoch 3/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3965 - accuracy: 0.8641 - val_loss: 0.4244 - val_accuracy: 0.8521
Epoch 4/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3709 - accuracy: 0.8712 - val_loss: 0.4001 - val_accuracy: 0.8611
Epoch 5/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3512 - accuracy: 0.8766 - val_loss: 0.3927 - val_accuracy: 0.8656
Epoch 6/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3368 - accuracy: 0.8821 - val_loss: 0.3743 - val_accuracy: 0.8685
Epoch 7/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3242 - accuracy: 0.8858 - val_loss: 0.3671 - val_accuracy: 0.8716
Epoch 8/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3129 - accuracy: 0.8892 - val_loss: 0.3654 - val_accuracy: 0.8715
Epoch 9/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3029 - accuracy: 0.8929 - val_loss: 0.3587 - val_accuracy: 0.8701
Epoch 10/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.2933 - accuracy: 0.8952 - val_loss: 0.3430 - val_accuracy: 0.8787
[Trial complete]
[Trial summary]|-Trial ID: 482b6110f8977c221413c42f85c52c60|-Score: 0.8787000179290771|-Best step: 0> Hyperparameters:|-learning_rate: 0.0001|-tuner/bracket: 0|-tuner/epochs: 10|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 352
Epoch 1/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.5354 - accuracy: 0.8099 - val_loss: 0.4774 - val_accuracy: 0.8261
Epoch 2/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.4370 - accuracy: 0.8432 - val_loss: 0.5032 - val_accuracy: 0.8288
Epoch 3/10
1875/1875 [==============================] - 3s 2ms/step - loss: 0.4183 - accuracy: 0.8498 - val_loss: 0.5043 - val_accuracy: 0.8263
Epoch 4/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.4032 - accuracy: 0.8550 - val_loss: 0.4821 - val_accuracy: 0.8313
Epoch 5/10
1875/1875 [==============================] - 5s 2ms/step - loss: 0.3887 - accuracy: 0.8585 - val_loss: 0.4602 - val_accuracy: 0.8339
Epoch 6/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3865 - accuracy: 0.8603 - val_loss: 0.4423 - val_accuracy: 0.8467
Epoch 7/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3771 - accuracy: 0.8646 - val_loss: 0.4956 - val_accuracy: 0.8329
Epoch 8/10
1875/1875 [==============================] - 5s 3ms/step - loss: 0.3710 - accuracy: 0.8644 - val_loss: 0.4489 - val_accuracy: 0.8509
Epoch 9/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3698 - accuracy: 0.8662 - val_loss: 0.4763 - val_accuracy: 0.8322
Epoch 10/10
1875/1875 [==============================] - 3s 2ms/step - loss: 0.3630 - accuracy: 0.8680 - val_loss: 0.4581 - val_accuracy: 0.8453
[Trial complete]
[Trial summary]|-Trial ID: f2959fbb15e860be5f80588c0516b81f|-Score: 0.8508999943733215|-Best step: 0> Hyperparameters:|-learning_rate: 0.01|-tuner/bracket: 0|-tuner/epochs: 10|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 352
Epoch 1/10
1875/1875 [==============================] - 3s 2ms/step - loss: 0.5399 - accuracy: 0.8069 - val_loss: 0.4596 - val_accuracy: 0.8269
Epoch 2/10
1875/1875 [==============================] - 3s 1ms/step - loss: 0.4343 - accuracy: 0.8462 - val_loss: 0.4763 - val_accuracy: 0.8384
Epoch 3/10
1875/1875 [==============================] - 3s 1ms/step - loss: 0.4208 - accuracy: 0.8496 - val_loss: 0.4611 - val_accuracy: 0.8395
Epoch 4/10
1875/1875 [==============================] - 3s 1ms/step - loss: 0.4026 - accuracy: 0.8576 - val_loss: 0.4485 - val_accuracy: 0.8433
Epoch 5/10
1875/1875 [==============================] - 3s 1ms/step - loss: 0.3954 - accuracy: 0.8572 - val_loss: 0.4652 - val_accuracy: 0.8392
Epoch 6/10
1875/1875 [==============================] - 3s 1ms/step - loss: 0.3883 - accuracy: 0.8609 - val_loss: 0.4564 - val_accuracy: 0.8350
Epoch 7/10
1875/1875 [==============================] - 3s 1ms/step - loss: 0.3829 - accuracy: 0.8606 - val_loss: 0.4661 - val_accuracy: 0.8345
Epoch 8/10
1875/1875 [==============================] - 2s 1ms/step - loss: 0.3813 - accuracy: 0.8623 - val_loss: 0.4922 - val_accuracy: 0.8371
Epoch 9/10
1875/1875 [==============================] - 3s 2ms/step - loss: 0.3765 - accuracy: 0.8653 - val_loss: 0.4553 - val_accuracy: 0.8551
Epoch 10/10
1875/1875 [==============================] - 3s 2ms/step - loss: 0.3689 - accuracy: 0.8671 - val_loss: 0.4418 - val_accuracy: 0.8561
[Trial complete]
[Trial summary]|-Trial ID: ebf7d288ffbf54d692fcab93a0eaf54a|-Score: 0.8561000227928162|-Best step: 0> Hyperparameters:|-learning_rate: 0.01|-tuner/bracket: 0|-tuner/epochs: 10|-tuner/initial_epoch: 0|-tuner/round: 0|-units: 320The hyperparameter search is complete. The optimal number of units in the first densely-connected
layer is 352 and the optimal learning rate for the optimizer
is 0.001.Epoch 1/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.4782 - accuracy: 0.8300 - val_loss: 0.4132 - val_accuracy: 0.8549
Epoch 2/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3597 - accuracy: 0.8688 - val_loss: 0.4138 - val_accuracy: 0.8470
Epoch 3/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3246 - accuracy: 0.8811 - val_loss: 0.3640 - val_accuracy: 0.8633
Epoch 4/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.3008 - accuracy: 0.8893 - val_loss: 0.3614 - val_accuracy: 0.8719
Epoch 5/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.2839 - accuracy: 0.8945 - val_loss: 0.3623 - val_accuracy: 0.8747
Epoch 6/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.2673 - accuracy: 0.9007 - val_loss: 0.3451 - val_accuracy: 0.8749
Epoch 7/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.2538 - accuracy: 0.9045 - val_loss: 0.3264 - val_accuracy: 0.8851
Epoch 8/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.2436 - accuracy: 0.9089 - val_loss: 0.3344 - val_accuracy: 0.8831
Epoch 9/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.2339 - accuracy: 0.9123 - val_loss: 0.3476 - val_accuracy: 0.8797
Epoch 10/10
1875/1875 [==============================] - 4s 2ms/step - loss: 0.2261 - accuracy: 0.9155 - val_loss: 0.3461 - val_accuracy: 0.8780Process finished with exit code 0

Keras Tuner 超参数优化相关推荐

1. tensorflow官方Blog-使用Keras Tuner超参数优化框架 进行超参数调整 ,具体实现版本

   文章目录 进入正题,keras tuner超参数优化框架 模型构建def build_model(hp): 实例化tuner 加载数据集,进行超参数搜索tuner.search() 找到最佳的模型tu ...

2. 使用贝叶斯优化进行深度神经网络超参数优化

   在本文中,我们将深入研究超参数优化. 为了方便起见本文将使用 Tensorflow 中包含的 Fashion MNIST[1] 数据集.该数据集在训练集中包含 60,000 张灰度图像,在测试集中包含 ...

3. 贝叶斯优化神经网络参数_贝叶斯超参数优化：神经网络，TensorFlow，相预测示例

   贝叶斯优化神经网络参数 The purpose of this work is to optimize the neural network model hyper-parameters to est ...

4. 常见的10大超参数优化库

   Skopt https://scikit-optimize.github.io/ 是一个超参数优化库,包括随机搜索.贝叶斯搜索.决策森林和梯度提升树.这个库包含一些理论成熟且可靠的优化方法,但是这些模 ...

5. 超参数优化：网格搜索法

   文章目录 网格搜索法在机器学习和深度学习中的使用 1.项目简介 2.机器学习案例 2.1导入相关库 2.2导入数据 2.3拆分数据集 2.4网格搜索法 2.5使用最优参数重新训练模型 3.深度学习案例 ...

6. 超参数优化 贝叶斯优化框架_10个超参数优化框架

   超参数优化 贝叶斯优化框架 Tune your Machine Learning models with open-source optimization libraries 使用开源优化库调整机器学 ...

7. 使用TensorBoard进行超参数优化

   在本文中,我们将介绍超参数优化,然后使用TensorBoard显示超参数优化的结果. 深度神经网络的超参数是什么? 深度学习神经网络的目标是找到节点的权重,这将帮助我们理解图像.文本或语音中的数据模式 ...

8. 超参数优化专题之工具—microsoft/nni(1)

   超参数优化专题之工具-microsoft/nni 这篇博客主要讲述模型的调参的一些基本知识,主要两个方面,调参的工具和相应的算法. 工具我比较推荐的是微软的nni框架,以及weight & b ...

9. 机器学习模型的超参数优化 | 原力计划

   作者 | deephub 责编 | 王晓曼 出品 | CSDN博客 头图 | CSDN付费下载自东方IC 引言 模型优化是机器学习算法实现中最困难的挑战之一.机器学习和深度学习理论的所有分支都致力于模 ...

最新文章

1. java调用kettle连hive_使用java连接hive，并执行hive语句详解
2. Zabbix监控Mysql数据库性能
3. CentOS/RHEL安装oracle 11G
4. C++学习之路 | PTA乙级—— 1024 科学计数法 (20 分)（精简）
5. php怎么上传函数,php上传函数怎么封装
6. IDEA常用快捷键大全
7. dock run mysql v3_Docker入门（三） - 搭建mysql
8. Asp.Net MVC 体验 2 持久层的构建
9. 计算机兴趣小组的意义,计算机兴趣小组总结
10. SharePoint 2010 PowerShell 系列 之 Create List and Field --Lookup
11. python变量赋值给数组_python 变量，数组，字符串
12. Python官方中文开发文档
13. 数学建模写作指导20篇（一）-如何写好数学建模论文？
14. 动听百年：音乐播放器发展沉浮史
15. Revisiting Time Series Outlier Detection: Definitions and Benchmarks
16. 阿里云同步gcr.io的镜像
17. 安卓手机便签怎么快速修改重复提醒设置?
18. Linux——万字总结用户与组相关知识！建议收藏！
19. python如何给某列数据打标签_Python map, apply, transform 打标签方法汇总(初阶到高阶)...
20. 【pytorch】时间序列预测 —— 同时预测多个分位点

热门文章

1. C++无法输出txt文件全部内容以及输出出现乱码的解决方法
2. VBA中字典的写入技巧
3. modx注册码_MODX：您从未使用过的最佳CMS？
4. 如何用python画出一般函数图_python如何画函数图像
5. 在中断函数和MAIN函数中同时调用同一个函数的有趣现象
6. 友盟 点完登陆后无反应_免登陆、免安装、不限速，打开网页直接下载百度云！...
7. WPF中MaterialDesignThemes前端框架的使用（超详细图解步骤）
8. R语言-查看加载包、卸除加载包及安装包与卸载包
9. c++ sprintf()详解
10. 什么是servlet ？简述servlet执行原理？生命周期是怎样的？