E tensorflow/core/grappler/clusters/utils.cc:87] Failed to get device properties, error code: 30

我win10+anaconda5.3.1环境下conda安装包之后报错，我卸载了anaconda5.3.1转而安装anaconda5.2.0之后成功解决安装包报错问题，详见我博客https://blog.csdn.net/Msjiangmei/article/details/100925060。

转换为anaconda5.2.0+python3.6+tensorflow-gup1.13.1，但是训练yolov3时，有如下报错？或者说警告？

注意：虽然报错如下但是程序运行能出现结果（图中find 0 box是因为训练太烂，置信度放低就出现box了）

 

2019-09-19 11:16:10.344857: E tensorflow/core/grappler/clusters/utils.cc:87] Failed to get device properties, error code: 30

2019-09-19 11:16:10.597394: E tensorflow/core/grappler/optimizers/meta_optimizer.cc:502] remapper failed: Invalid argument: Subshape must have computed start >= end since stride is negative, but is 0 and 2 (computed from start 0 and end 9223372036854775807 over shape with rank 2 and stride-1)

2019-09-19 11:16:10.848319: E tensorflow/core/grappler/optimizers/meta_optimizer.cc:502] remapper failed: Invalid argument: Subshape must have computed start >= end since stride is negative, but is 0 and 2 (computed from start 0 and end 9223372036854775807 over shape with rank 2 and stride-1)

期间我还安装anaconda5.2.0+ tensorflow-gpu1.12.0+python3.6也报Failed to get device properties, error code: 30

anaconda5.2.0+ tensorflow-gpu1.14.0+python3.6也报Failed to get device properties, error code: 30

开始默认安装的是tf1.12.0，报错后百度，查询到可能是GPU驱动版本的问题，那个博客说降低驱动版本可是驱动版本我之前升到最高了，驱动版本可以降低吗？我知道tf-gpu1.13.1以上版本是cuda10驱动版本也很高(400以上)，所以我又安装了tf-gpu1.13.1，1.14.0。但是报错是一样的，而且1.12.0报错(真错)最多。

目前尚不清楚为啥“Failed to get device properties, error code: 30”，但因代码能出运行结果而且已折腾良久暂且不管了。

最新更新：找到解决办法，更新最新gpu驱动版本，我更新到436.48（2019.10.6发布）成功解决！！！