工业界还是在用 1.0,所以有必要了解静态图机制 1.0时代,还可以用另一个 API:tf.get_variable b/W/x 都只是:对象、节点 而不是:Tensor ...
分类:
其他好文 时间:
2020-11-02 09:59:40
阅读次数:
19
后面还有图灵和安培架构 大厂使用桌面卡去做商业使用,签协议,会交很多罚款 python 目录还有 自动反向求导 ...
分类:
其他好文 时间:
2020-11-02 09:51:08
阅读次数:
18
tf git view /collection:http://tfs.mycompany.com.cn /teamproject:devops /repository:hellomaven /path:src/main/java/com/mycompany/hellomaven/App.java / ...
分类:
其他好文 时间:
2020-11-01 10:18:19
阅读次数:
16
variable_names = [v.name for v in tf.all_variables()] values = sess.run(variable_names) for k,v in zip(variable_names, values): print("Variable: ", k) ...
分类:
其他好文 时间:
2020-11-01 09:47:51
阅读次数:
15
##import import tensorflow as tf from tensorflow.keras.models import Sequential from tensorflow.keras import layers from tensorflow.keras.layers impor ...
分类:
其他好文 时间:
2020-10-31 01:42:28
阅读次数:
13
希望实现图片上的功能 import tensorflow as tfa = tf.range(10,dtype=float)b = aa = tf.reshape(a,[-1,1])a = tf.tile(a,[1,3]) sess = tf.Session()print(sess.run(b))p ...
分类:
其他好文 时间:
2020-10-29 10:39:17
阅读次数:
23
本来是只用Tenorflow的,但是因为TF有些Numpy特性并不支持,比如对数组使用列表进行切片,所以只能转战Pytorch了(pytorch是支持的)。还好Pytorch比较容易上手,几乎完美复制了Numpy的特性(但还有一些特性不支持),怪不得热度上升得这么快。 模型定义 和TF很像,Pyto ...
分类:
其他好文 时间:
2020-10-26 11:15:09
阅读次数:
21
作者|Ivan Rala?i? 编译|VK 来源|Analytics Vidhya Tensorflow目标检测API(TF OD API)刚刚变得更好。最近,Google发布了tf od api的新版本,它现在支持Tensorflow 2.x,这是一个我们一直在等待的巨大改进! 简介 最近目标检测 ...
mnist = tf.keras.datasets.mnist (x_train, y_train), (x_test, y_test) = mnist.load_data() x_train, x_test = x_train / 255.0, x_test / 255.0 model = tf. ...
分类:
其他好文 时间:
2020-10-16 10:58:02
阅读次数:
22
tf.gather:用一个一维的索引数组,将张量中对应索引的向量提取出来 import tensorflow as tf a = tf.Variable([[1,2,3,4,5], [6,7,8,9,10], [11,12,13,14,15]]) index_a = tf.Variable([0,2 ...
分类:
其他好文 时间:
2020-10-13 17:50:02
阅读次数:
39