ValueError: Variable rnn/basic_lstm_cell/kernel already exists, disallowed.

转载于 2020-05-07 16:45:17 发布 · 1k 阅读

2 ·

CC 4.0 BY-SA版权

原文链接：https://www.jianshu.com/p/c2e5bf82d37b

tensorflow 专栏收录该内容

77 篇文章

订阅专栏

本文详细介绍了在使用TensorFlow进行深度学习项目时遇到的变量重复定义错误，特别是ValueError: Variable rnn/basic_lstm_cell/kernel already exists的解决办法。通过添加tf.reset_default_graph()函数重置计算图，避免了因多次运行会话导致的变量冲突问题。

ValueError: Variable rnn/basic_lstm_cell/kernel already exists, disallowed.

用spyder运行mnist_lstm.py

第一次没问题，修改n_hidden_units后运行报错

ValueError: Variable rnn/basic_lstm_cell/kernel already exists, disallowed. Did you mean to set reuse=True or reuse=tf.AUTO_REUSE in VarScope? Originally defined at:

File "/media/oliver/工具/deep-60/day4/mnist_lstm.py", line 44, in RNN outputs, states = tf.nn.dynamic_rnn(lstm_cell, x_in, initial_state=_init_state, time_major=False)

File "/media/oliver/工具/deep-60/day4/mnist_lstm.py", line 55, in pred = RNN(x, weights, biases)

File "/home/oliver/.conda/envs/tf-gpu/lib/python3.6/site-packages/spyder/utils/site/sitecustomize.py", line 101, in execfile exec(compile(f.read(), filename, 'exec'), namespace)

解决方案：

添加函数：tf.reset_default_graph()

重置graph就可以正常运行