配置TensorFlow2. GPU。解决Could not load dynamic library ‘cudart64_101.dll‘； dlerror: cudart64_101.dl
2020-10-21 16:29:19.467814: tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library cudart64_101.dll
笔者之前一直使用1x版本的TensorFlow。
在GITHUB上下载了一些代码，要求在TensorFlow2.版本运行。pip install tensorflow==2.2以后，
运行tf.test.is_gpu_available() 返回False。
2020-10-21 16:07:39.297448: W tensorflow/stream_executor/platform/default/dso_loader.cc:59] Could not load dynamic library ‘cudart64_101.dll’; dlerror: cudart64_101.dll not found
2020-10-21 16:07:39.299448: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.
如何正确调用GPU？
解决：
- 去官网下载cudart64_101.dll文件，解压后有两个文件：
- 放置cudart64_101.dll文件
官网中已经写得很清楚了：
记得要重启！
再次运行程序，显示：
2020-10-21 16:29:19.467814: tensorflow/stream_executor/platform/default/dso_loader.cc:48] Successfully opened dynamic library cudart64_101.dll
成功！
转自：https://blog.csdn.net/weixin_43786241/article/details/109203995，作者给力