指定target_size后所有图像都变为相同大小gen_data=datagen.flow_from_directory(in_path,batch_size=1,shuffle=Fals
简单的分类任务importtensorflowastffromtensorflow.examples.tutorials.mnistimportinput_datamnist=input_data.read_data_sets('data/',one_hot=True)#tobeabletorerunthemodelwithoutoverwritingtfvariablestf.reset_def
简单的只有一层隐藏层importtensorflowastffromtensorflow.examples.tutorials.mnistimportinput_datamnist=input_data.read_data_sets('data/',one_hot=True)num_classes=10batch_size=64hidden_units=50input_size=784train_
写代码真的要小心的,小问题调试半天。。。importtensorflowastffromtensorflow.examples.tutorials.mnistimportinput_datamnist=input_data.read_data_sets('data/',one_hot=True)num_classes=10input_size=784train_iter=50000batch_si
Tensorflow自带的Mnist数据集相关情况importtensorflowastfimportnumpyasnpimportmatplotlib.pyplotaspltfromtensorflow.examples.tutorials.mnistimportinput_datamnist=input_data.read_data_sets('data/',one_hot=True)prin
构造数据,实现简单的线性回归importtensorflowastfimportmatplotlib.pyplotaspltimportnumpyasnp#模拟数据num_points=1000vec_points=[]foriinrange(num_points):x=np.random.normal(0.0,0.5)y=x*0.1+0.3+np.random.normal(0.0,0.06)v
Placeholder可以用来取数据importtensorflowastfinput_1=tf.placeholder(tf.float32)#先把坑占了,不用提前指定具体值input_2=tf.placeholder(tf.float32)#比如每个batchoutput_op=tf.multiply(input_1,input_2)withtf.Session()assess:sess.ru
numpy数据转成Tensor小Demoimportnumpyasnpimporttensorflowastftemp_np=np.zeros((3,3))print(type(temp_np))#<class'numpy.ndarray'>tensor_temp=tf.convert_to_tensor(temp_np)#<class'tensorflow.python.fra
tensorflow变量定义和赋值没有python那么简单,需要在session中run才能拿到结果importtensorflowastfw=tf.Variable([[1.0,2.0]])print(w)#<tf.Variable'Variable:0'shape=(1,2)dtype=float32_ref>x=tf.Variable([[1.0],[0.5]])print(x)
tensorflow的环境配置主要分为两步:anaconda安装和配置tensorflow包下载和安装然后就可以开始写tensorflow小程序了
Copyright © 2005-2024 51CTO.COM 版权所有 京ICP证060544号