一小時(shí)學(xué)會(huì)TensorFlow2之Fashion Mnist
描述
Fashion Mnist 是一個(gè)類(lèi)似于 Mnist 的圖像數(shù)據(jù)集. 涵蓋 10 種類(lèi)別的 7 萬(wàn) (6 萬(wàn)訓(xùn)練集 + 1 萬(wàn)測(cè)試集) 個(gè)不同商品的圖片.
Tensorboard
Tensorboard 是 tensorflow 的一個(gè)可視化工具.
創(chuàng)建 summary
我們可以通過(guò)tf.summary.create_file_writer(file_path)來(lái)創(chuàng)建一個(gè)新的 summary 實(shí)例.
例子:
# 將當(dāng)前時(shí)間作為子文件名
current_time = datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
# 監(jiān)聽(tīng)的文件的路徑
log_dir = 'logs/' + current_time
# 創(chuàng)建writer
summary_writer = tf.summary.create_file_writer(log_dir)
存入數(shù)據(jù)
通過(guò)tf.summary.scalar我們可以向 summary 對(duì)象存入數(shù)據(jù).
格式:
tf.summary.scalar(
name, data, step=None, description=None
)
例子:
with summary_writer.as_default():
tf.summary.scalar("train-loss", float(Cross_Entropy), step=step)
metrics
metrics.Mean()
metrics.Mean()可以幫助我們計(jì)算平均數(shù).
格式:
tf.keras.metrics.Mean(
name='mean', dtype=None
)
例子:
# 準(zhǔn)確率表 loss_meter = tf.keras.metrics.Mean()
metrics.Accuracy()
格式:
tf.keras.metrics.Accuracy(
name='accuracy', dtype=None
)
例子:
# 損失表 acc_meter = tf.keras.metrics.Accuracy()
變量更新 &重置
我們可以通過(guò)update_state來(lái)實(shí)現(xiàn)變量更新, 通過(guò)rest_state來(lái)實(shí)現(xiàn)變量重置.
例如:
# 跟新?lián)p失 loss_meter.update_state(Cross_Entropy) # 重置 loss_meter.reset_state()
案例
pre_process 函數(shù)
def pre_process(x, y):
"""
數(shù)據(jù)預(yù)處理
:param x: 特征值
:param y: 目標(biāo)值
:return: 返回處理好的x, y
"""
# 轉(zhuǎn)換x
x = tf.cast(x, tf.float32) / 255
x = tf.reshape(x, [-1, 784])
# 轉(zhuǎn)換y
y = tf.cast(y, dtype=tf.int32)
y = tf.one_hot(y, depth=10)
return x, y
get_data 函數(shù)
def get_data():
"""
獲取數(shù)據(jù)
:return: 返回分批完的訓(xùn)練集和測(cè)試集
"""
# 獲取數(shù)據(jù)
(X_train, y_train), (X_test, y_test) = tf.keras.datasets.fashion_mnist.load_data()
# 分割訓(xùn)練集
train_db = tf.data.Dataset.from_tensor_slices((X_train, y_train)).shuffle(60000, seed=0)
train_db = train_db.batch(batch_size).map(pre_process)
# 分割測(cè)試集
test_db = tf.data.Dataset.from_tensor_slices((X_test, y_test)).shuffle(10000, seed=0)
test_db = test_db.batch(batch_size).map(pre_process)
# 返回
return train_db, test_db
train 函數(shù)
def train(epoch, train_db):
"""
訓(xùn)練數(shù)據(jù)
:param train_db: 分批的數(shù)據(jù)集
:return: 無(wú)返回值
"""
for step, (x, y) in enumerate(train_db):
with tf.GradientTape() as tape:
# 獲取模型輸出結(jié)果
logits = model(x)
# 計(jì)算交叉熵
Cross_Entropy = tf.losses.categorical_crossentropy(y, logits, from_logits=True)
Cross_Entropy = tf.reduce_sum(Cross_Entropy)
# 跟新?lián)p失
loss_meter.update_state(Cross_Entropy)
# 計(jì)算梯度
grads = tape.gradient(Cross_Entropy, model.trainable_variables)
# 跟新參數(shù)
optimizer.apply_gradients(zip(grads, model.trainable_variables))
# 每100批調(diào)試輸出一下誤差
if step % 100 == 0:
print("step:", step, "Cross_Entropy:", loss_meter.result().numpy())
# 重置
loss_meter.reset_state()
# 可視化
with summary_writer.as_default():
tf.summary.scalar("train-loss", float(Cross_Entropy), step= epoch * 235 + step)
test 函數(shù)
def test(epoch, test_db):
"""
測(cè)試模型
:param epoch: 輪數(shù)
:param test_db: 分批的測(cè)試集
:return: 無(wú)返回值
"""
# 重置
acc_meter.reset_state()
for x, y in test_db:
# 獲取模型輸出結(jié)果
logits = model(x)
# 預(yù)測(cè)結(jié)果
pred = tf.argmax(logits, axis=1)
# 從one_hot編碼變回來(lái)
y = tf.argmax(y, axis=1)
# 計(jì)算準(zhǔn)確率
acc_meter.update_state(y, pred)
# 調(diào)試輸出
print("epoch:", epoch + 1, "Accuracy:", acc_meter.result().numpy() * 100, "%", )
# 可視化
with summary_writer.as_default():
tf.summary.scalar("val-acc", acc_meter.result().numpy(), step=epoch * 235)
main 函數(shù)
def main():
"""
主函數(shù)
:return: 無(wú)返回值
"""
# 獲取數(shù)據(jù)
train_db, test_db = get_data()
# 輪期
for epoch in range(iteration_num):
train(epoch, train_db)
test(epoch, test_db)
完整代碼
import datetime
import tensorflow as tf
# 定義超參數(shù)
batch_size = 256 # 一次訓(xùn)練的樣本數(shù)目
learning_rate = 0.001 # 學(xué)習(xí)率
iteration_num = 20 # 迭代次數(shù)
# 優(yōu)化器
optimizer = tf.keras.optimizers.Adam(learning_rate=learning_rate)
# 準(zhǔn)確率表
loss_meter = tf.keras.metrics.Mean()
# 損失表
acc_meter = tf.keras.metrics.Accuracy()
# 可視化
current_time = datetime.datetime.now().strftime("%Y%m%d-%H%M%S")
log_dir = 'logs/' + current_time
summary_writer = tf.summary.create_file_writer(log_dir) # 創(chuàng)建writer
# 模型
model = tf.keras.Sequential([
tf.keras.layers.Dense(256, activation=tf.nn.relu),
tf.keras.layers.Dense(128, activation=tf.nn.relu),
tf.keras.layers.Dense(64, activation=tf.nn.relu),
tf.keras.layers.Dense(32, activation=tf.nn.relu),
tf.keras.layers.Dense(10)
])
# 調(diào)試輸出summary
model.build(input_shape=[None, 28 * 28])
print(model.summary())
def pre_process(x, y):
"""
數(shù)據(jù)預(yù)處理
:param x: 特征值
:param y: 目標(biāo)值
:return: 返回處理好的x, y
"""
# 轉(zhuǎn)換x
x = tf.cast(x, tf.float32) / 255
x = tf.reshape(x, [-1, 784])
# 轉(zhuǎn)換y
y = tf.cast(y, dtype=tf.int32)
y = tf.one_hot(y, depth=10)
return x, y
def get_data():
"""
獲取數(shù)據(jù)
:return: 返回分批完的訓(xùn)練集和測(cè)試集
"""
# 獲取數(shù)據(jù)
(X_train, y_train), (X_test, y_test) = tf.keras.datasets.fashion_mnist.load_data()
# 分割訓(xùn)練集
train_db = tf.data.Dataset.from_tensor_slices((X_train, y_train)).shuffle(60000, seed=0)
train_db = train_db.batch(batch_size).map(pre_process)
# 分割測(cè)試集
test_db = tf.data.Dataset.from_tensor_slices((X_test, y_test)).shuffle(10000, seed=0)
test_db = test_db.batch(batch_size).map(pre_process)
# 返回
return train_db, test_db
def train(epoch, train_db):
"""
訓(xùn)練數(shù)據(jù)
:param train_db: 分批的數(shù)據(jù)集
:return: 無(wú)返回值
"""
for step, (x, y) in enumerate(train_db):
with tf.GradientTape() as tape:
# 獲取模型輸出結(jié)果
logits = model(x)
# 計(jì)算交叉熵
Cross_Entropy = tf.losses.categorical_crossentropy(y, logits, from_logits=True)
Cross_Entropy = tf.reduce_sum(Cross_Entropy)
# 跟新?lián)p失
loss_meter.update_state(Cross_Entropy)
# 計(jì)算梯度
grads = tape.gradient(Cross_Entropy, model.trainable_variables)
# 跟新參數(shù)
optimizer.apply_gradients(zip(grads, model.trainable_variables))
# 每100批調(diào)試輸出一下誤差
if step % 100 == 0:
print("step:", step, "Cross_Entropy:", loss_meter.result().numpy())
# 重置
loss_meter.reset_state()
# 可視化
with summary_writer.as_default():
tf.summary.scalar("train-loss", float(Cross_Entropy), step=epoch * 235 + step)
def test(epoch, test_db):
"""
測(cè)試模型
:param epoch: 輪數(shù)
:param test_db: 分批的測(cè)試集
:return: 無(wú)返回值
"""
# 重置
acc_meter.reset_state()
for x, y in test_db:
# 獲取模型輸出結(jié)果
logits = model(x)
# 預(yù)測(cè)結(jié)果
pred = tf.argmax(logits, axis=1)
# 從one_hot編碼變回來(lái)
y = tf.argmax(y, axis=1)
# 計(jì)算準(zhǔn)確率
acc_meter.update_state(y, pred)
# 調(diào)試輸出
print("epoch:", epoch + 1, "Accuracy:", acc_meter.result().numpy() * 100, "%", )
# 可視化
with summary_writer.as_default():
tf.summary.scalar("val-acc", acc_meter.result().numpy(), step=epoch * 235)
def main():
"""
主函數(shù)
:return: 無(wú)返回值
"""
# 獲取數(shù)據(jù)
train_db, test_db = get_data()
# 輪期
for epoch in range(iteration_num):
train(epoch, train_db)
test(epoch, test_db)
if __name__ == "__main__":
main()
輸出結(jié)果:
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense (Dense) (None, 256) 200960
_________________________________________________________________
dense_1 (Dense) (None, 128) 32896
_________________________________________________________________
dense_2 (Dense) (None, 64) 8256
_________________________________________________________________
dense_3 (Dense) (None, 32) 2080
_________________________________________________________________
dense_4 (Dense) (None, 10) 330
=================================================================
Total params: 244,522
Trainable params: 244,522
Non-trainable params: 0
_________________________________________________________________
None
2021-06-14 18:01:27.399812: I tensorflow/compiler/mlir/mlir_graph_optimization_pass.cc:176] None of the MLIR Optimization Passes are enabled (registered 2)
step: 0 Cross_Entropy: 591.5974
step: 100 Cross_Entropy: 196.49309
step: 200 Cross_Entropy: 125.2562
epoch: 1 Accuracy: 84.72999930381775 %
step: 0 Cross_Entropy: 107.64579
step: 100 Cross_Entropy: 105.854385
step: 200 Cross_Entropy: 99.545975
epoch: 2 Accuracy: 85.83999872207642 %
step: 0 Cross_Entropy: 95.42945
step: 100 Cross_Entropy: 91.366234
step: 200 Cross_Entropy: 90.84072
epoch: 3 Accuracy: 86.69999837875366 %
step: 0 Cross_Entropy: 82.03317
step: 100 Cross_Entropy: 83.20552
step: 200 Cross_Entropy: 81.57012
epoch: 4 Accuracy: 86.11000180244446 %
step: 0 Cross_Entropy: 82.94046
step: 100 Cross_Entropy: 77.56677
step: 200 Cross_Entropy: 76.996346
epoch: 5 Accuracy: 87.27999925613403 %
step: 0 Cross_Entropy: 75.59219
step: 100 Cross_Entropy: 71.70899
step: 200 Cross_Entropy: 74.15144
epoch: 6 Accuracy: 87.29000091552734 %
step: 0 Cross_Entropy: 76.65844
step: 100 Cross_Entropy: 70.09151
step: 200 Cross_Entropy: 70.84446
epoch: 7 Accuracy: 88.27999830245972 %
step: 0 Cross_Entropy: 67.50707
step: 100 Cross_Entropy: 64.85907
step: 200 Cross_Entropy: 68.63099
epoch: 8 Accuracy: 88.41999769210815 %
step: 0 Cross_Entropy: 65.50318
step: 100 Cross_Entropy: 62.2706
step: 200 Cross_Entropy: 63.80803
epoch: 9 Accuracy: 86.21000051498413 %
step: 0 Cross_Entropy: 66.95486
step: 100 Cross_Entropy: 61.84385
step: 200 Cross_Entropy: 62.18851
epoch: 10 Accuracy: 88.45999836921692 %
step: 0 Cross_Entropy: 59.779297
step: 100 Cross_Entropy: 58.602314
step: 200 Cross_Entropy: 59.837025
epoch: 11 Accuracy: 88.66000175476074 %
step: 0 Cross_Entropy: 58.10068
step: 100 Cross_Entropy: 55.097878
step: 200 Cross_Entropy: 59.906315
epoch: 12 Accuracy: 88.70999813079834 %
step: 0 Cross_Entropy: 57.584858
step: 100 Cross_Entropy: 54.95376
step: 200 Cross_Entropy: 55.797752
epoch: 13 Accuracy: 88.44000101089478 %
step: 0 Cross_Entropy: 53.54782
step: 100 Cross_Entropy: 53.62939
step: 200 Cross_Entropy: 54.632828
epoch: 14 Accuracy: 87.02999949455261 %
step: 0 Cross_Entropy: 54.387398
step: 100 Cross_Entropy: 52.323734
step: 200 Cross_Entropy: 53.968185
epoch: 15 Accuracy: 88.98000121116638 %
step: 0 Cross_Entropy: 50.468914
step: 100 Cross_Entropy: 50.79311
step: 200 Cross_Entropy: 51.296227
epoch: 16 Accuracy: 88.67999911308289 %
step: 0 Cross_Entropy: 48.753258
step: 100 Cross_Entropy: 46.809692
step: 200 Cross_Entropy: 48.08208
epoch: 17 Accuracy: 89.10999894142151 %
step: 0 Cross_Entropy: 46.830627
step: 100 Cross_Entropy: 47.208813
step: 200 Cross_Entropy: 48.671318
epoch: 18 Accuracy: 88.77999782562256 %
step: 0 Cross_Entropy: 46.15514
step: 100 Cross_Entropy: 45.026627
step: 200 Cross_Entropy: 45.371685
epoch: 19 Accuracy: 88.7399971485138 %
step: 0 Cross_Entropy: 47.696465
step: 100 Cross_Entropy: 41.52749
step: 200 Cross_Entropy: 46.71362
epoch: 20 Accuracy: 89.56000208854675 %
可視化
到此這篇關(guān)于一小時(shí)學(xué)會(huì)TensorFlow2之Fashion Mnist的文章就介紹到這了,更多相關(guān)TensorFlow2 Fashion Mnist內(nèi)容請(qǐng)搜索腳本之家以前的文章或繼續(xù)瀏覽下面的相關(guān)文章希望大家以后多多支持腳本之家！
相關(guān)文章
大家好，本篇文章主要講的是用Python實(shí)現(xiàn)群發(fā)郵件，感興趣的同學(xué)趕快來(lái)看一看吧，對(duì)你有幫助的話(huà)記得收藏一下2022-02-02
python無(wú)法識(shí)別vim中文代碼的解決方案
這篇文章主要介紹了python無(wú)法識(shí)別vim中文代碼的解決方案，具有很好的參考價(jià)值，希望對(duì)大家有所幫助。一起跟隨小編過(guò)來(lái)看看吧2021-03-03
用Python(Tornado)模擬登錄小米帳號(hào),搶小米手機(jī)2013-11-11
在本篇文章里小編給大家整理的是關(guān)于python給視頻添加背景音樂(lè)并改變音量的具體方法，需要的朋友們可以參考下。2020-07-07
這篇文章主要介紹了Python安裝教程全過(guò)程(2022最新版本)，具有很好的參考價(jià)值，希望對(duì)大家有所幫助。如有錯(cuò)誤或未考慮完全的地方，望不吝賜教2022-12-12
Python的數(shù)據(jù)類(lèi)型與標(biāo)識(shí)符和判斷語(yǔ)句詳解
在本篇文章里小編給大家整理了一篇關(guān)于python數(shù)據(jù)類(lèi)型與標(biāo)識(shí)符和判斷語(yǔ)句的介紹，有需要的朋友們可以學(xué)習(xí)下，希望能夠給你帶來(lái)幫助2021-09-09
Python+?Flask實(shí)現(xiàn)Mock?Server詳情
這篇文章主要介紹了Python+?Flask實(shí)現(xiàn)Mock?Server詳情，文章圍繞主題展開(kāi)詳細(xì)的內(nèi)容介紹，具有一定的參考價(jià)值，需要的小伙伴可以參考一下2022-09-09
基于Python的一個(gè)自動(dòng)錄入表格的小程序
這篇文章主要介紹了基于Python的一個(gè)自動(dòng)錄入表格的小程序，文中通過(guò)示例代碼介紹的非常詳細(xì)，對(duì)大家的學(xué)習(xí)或者工作具有一定的參考學(xué)習(xí)價(jià)值，需要的朋友們下面隨著小編來(lái)一起學(xué)習(xí)學(xué)習(xí)吧2020-08-08
本文主要介紹了OpenCV黑帽運(yùn)算(BLACKHAT)的使用，文中通過(guò)示例代碼介紹的非常詳細(xì)，對(duì)大家的學(xué)習(xí)或者工作具有一定的參考學(xué)習(xí)價(jià)值，需要的朋友們下面隨著小編來(lái)一起學(xué)習(xí)學(xué)習(xí)吧2022-08-08
OpenCV中Canny邊緣檢測(cè)的實(shí)現(xiàn)
本文主要介紹了OpenCV中Canny邊緣檢測(cè)的實(shí)現(xiàn)，邊緣檢測(cè)一般是識(shí)別目標(biāo)圖像中亮度變化明顯的像素點(diǎn)，具有一定的參考價(jià)值，感興趣的小伙伴們可以參考一下2021-07-07