找不到：未注册的操作类型“ CountExtremelyRandomStats” [英] Not found: Op type not registered 'CountExtremelyRandomStats'

问题描述

在docker容器中运行bazel命令时遇到以下错误。

I am getting the following error while running bazel command in the docker container.

命令ran：

-bazel-bin / tensorflow_serving / model_servers / tensorflow_model_server --port = 9000 --model_name = rf_quora --model_base_path = / serving / rf_model&> rf_log&

错误：

• 未找到：Op类型未在运行在864822af1c6c上的二进制文件中注册为'CountExtremelyRandomStats'。确保操作和内核已在此过程中运行的二进制文件中注册了
  。

我尝试使用以下链接进行搜索，但没有成功。
https://github.com/tensorflow/tensorflow/issues/11847

I tried my search with the following link but in vain. https://github.com/tensorflow/tensorflow/issues/11847

我正在尝试使用tensorflow张量服务进行推理，但是我被上述错误所阻塞。
使用Tensorflow 1.3并使用 tf.contrib.tensor_forest_python 中的 tensor_forest api。

I am trying to do inference using tensorflow tensorserving, but I am getting blocked by the above error. Using Tensorflow 1.3 and using tensor_forest api present in tf.contrib.tensor_forest_python.

有人可以帮助我解决此错误，因为它阻止了我的测试。

Can anyone help me with this error as it is blocking my testing.

注意：-我已经成功运行了 tensorserving 推断 mnist 和初始示例模型。

Note:- I have successfully ran tensorserving inference for mnist and inception examples models.

推荐答案

升级到最新的tensorflow版本，即从1.3.0到1.4版本。

Issue got resolved after upgrading to latest tensorflow version i.e. from 1.3.0 to 1.4 version.

参考：-
> https://github.com/tensorflow/tensorflow/issues/14358

这篇关于找不到：未注册的操作类型“ CountExtremelyRandomStats”的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！