Tensorflow-Tutorial icon indicating copy to clipboard operation
Tensorflow-Tutorial copied to clipboard

Some interesting TensorFlow tutorials for beginners.

Results 4 Tensorflow-Tutorial issues
Sort by recently updated
recently updated
newest added

Bi-directional lstm中文分词里,报错tensorflow.python.framework.errors_impl.InvalidArgumentError: Received a label value of -2147483648 which is outside the valid range of [0, 5). Label values: -2147483648 -2147483648 2 3 -2147483648 0 0 0 0 0 0...

我是导出成py文件用ipython这个命令执行的文件 acc, _cost, _ = sess.run(fetches, feed_dict) `--------------------------------------------------------------------------- ValueError Traceback (most recent call last) /Users/finup/gitlab/poll-parrot/demo/a.py in () 511 X_batch, y_batch = data_train.next_batch(tr_batch_size) 512 feed_dict = {X_inputs:X_batch, y_inputs:y_batch, lr:_lr, batch_size:tr_batch_size, keep_prob:0.5}...

Notebook “Tutorial_05 - An understandable example to implement Multi-LSTM for MNIST”有这样一段代码。 ``` with tf.variable_scope('RNN'): for timestep in range(timestep_size): if timestep > 0: tf.get_variable_scope().reuse_variables() ``` 这个issue是关于`tf.get_variable_scope().reuse_variables()`合理性的猜测。希望同博主一起讨论。 首先,我发现tensorflow新旧版本在定义RNNcell的`__call__`方法时有不同的处理。旧版本直接定义`__call__`方法,新版本则要先继承`_LayerRNNCell`再定义`call` 和`build` 方法(而非直接定义`__call__`)。 为何这么处理?个人认为,使用RNNcell分为两个步骤:第一,实例化一个RNNcell;第二,调用声明的RNNcell实例进行计算。定义`__call__`方法就是为了简化用RNNcell的实例进行运算时的API调用。另外,大部分关于variable sharing的考虑和决策都发生在第一步。...