Cause: No module named ‘tensorflow_core.estimator‘（解决方法）

1. 问题描述：准备在VS code执行以下命令：

x_input, x_output = gcn_lstm.in_out_tensors()

结果报错：Cause: No module named 'tensorflow_core.estimator'

WARNING: AutoGraph could not transform <bound method FixedAdjacencyGraphConvolution.call of <stellargraph.layer.gcn_lstm.FixedAdjacencyGraphConvolution object at 0x0000015F1D634F28>> and will run it as-is.
Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output.
Cause: No module named 'tensorflow_core.estimator'

2. 解决方法：打开‪ Cmd，查看该运行环境的conda list

tensorflow                2.1.0           eigen_py36hdbbabfe_0    https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main
tensorflow-base           2.1.0           eigen_py36h49b2757_0    https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main
tensorflow-estimator      2.6.0

如果tensorflow-estimator版本和tensorflow不一致，就需要换成一致的版本：

conda install tensorflow-estimator==2.1.0

3.重启VS code，看看是否还有报错。

 done.

---

参考链接：

ModuleNotFoundError: No module named ‘tensorflow_core.estimator‘ 解决办法，已解决，可参考