![](https://img-blog.csdnimg.cn/20201014180756927.png?x-oss-process=image/resize,m_fixed,h_64,w_64)
tensorflow
神经学徒小周
这个作者很懒,什么都没留下…
展开
-
GRU新闻分类baseline
import pandas as pdimport numpy as npfrom gensim.models import Word2Vecimport tensorflow as tf1.读取数据 pd.read_csvtrain = pd.read_csv('/content/train.csv',sep='\t',header=None,names=['label','content'])val = pd.read_csv('/content/val.csv',sep='\t',hea原创 2020-12-17 21:43:18 · 270 阅读 · 3 评论 -
几种提升Dataset读取性能的方法
1.prefetch: 数据准备和参数迭代并行执行train_dataset.map( map_func=_decode_and_resize, num_parallel_calls=tf.data.experimental.AUTOTUNE), prefetch(tf.data.experimental.AUTOTUNE), num_epochs=12.interleave:读取数据多进程执行filenames = ["./interleave_data/trai原创 2020-12-16 11:01:18 · 1046 阅读 · 0 评论 -
Tensorflow自定义训练(不使用compile,fit)
#------省略了准备数据步骤# 优化器optimizer = tf.keras.optimizers.SGD(learning_rate=1e-3)# 损失函数loss_fn = tf.keras.losses.CategoricalCrossentropy(from_logits=True)# 准备metrics函数train_acc_metric = tf.keras.metrics.CategoricalAccuracy()val_acc_metric = tf.keras.met原创 2020-12-15 13:43:15 · 1058 阅读 · 3 评论 -
Tensorflow-回调函数
modelcheckpoint:定期保存模型earlystopping:测试指标不在上升时停止原创 2020-12-15 10:56:13 · 246 阅读 · 0 评论 -
(Baseline)类定义LSTM识别imdb
#数据导入以及处理batches = 128(x_train,y_train),(x_test,y_test) = keras.datasets.imdb.load_data(num_words=10000)x_train = keras.preprocessing.sequence.pad_sequences(x_train,maxlen=200)x_test = keras.preprocessing.sequence.pad_sequences(x_test,maxlen=200)db_原创 2020-12-14 21:04:36 · 134 阅读 · 0 评论 -
Tensorflow学习笔记(三)(compile与fit)
使用内置方法进行训练和评估利用minist数据举例:#定义模型inputs = keras.Input(shape=(784,), name="digits")x = layers.Dense(64, activation="relu", name="dense_1")(inputs)x = layers.Dense(64, activation="relu", name="dense_2")(x)outputs = layers.Dense(10, activation="softmax",原创 2020-12-14 13:15:11 · 857 阅读 · 2 评论 -
TensorFlow学习笔记(二)(函数式搭建模型-可解决sequential单入单出问题)
keras API一.具有多个输入和输出的模型:1.文本数据num_words = 10000 # 文本数据Embedding数量num_departments = 4 # Number of departments for predictions#定义三个输入title_input = keras.Input(shape=(None,), name="title") body_input = keras.Input(shape=(None,), name="body")tags_inp原创 2020-12-14 10:33:01 · 214 阅读 · 0 评论 -
Tensorflow中Sequential model
sequential模型适用于简单堆叠网络层(a plain stack of layers),及每一层只有一个输入和一个输出:几种创建sequential模型的方法:# 1.model = keras.Sequential( [ layers.Dense(2, activation="relu", name="layer1"), layers.Dense(3, activation="relu", name="layer2"), layers原创 2020-12-14 09:56:49 · 2320 阅读 · 0 评论