**
待解决
**
今天在Kaggle上试一下TextCNN,涉及到TensorFlow。
但是训练模型时发现GPU利用率为0,而且整个过程非常慢。
隔壁的CPU都到顶了…
查了很多方法,说是和cuda版本不对应
- https://stackoverflow.com/questions/60208936/cannot-dlopen-some-gpu-libraries-skipping-registering-gpu-devices
- https://www.codenong.com/cs108869930/
- https://blog.csdn.net/u010165147/article/details/106354671
But,cuda也不是pip能装上的
官方说需要配置,区分Ubuntu和Windows,解决方法不一样,但是都要CUDA。
软件包版本:
软件要求:
试了TensorFlow-gpu 1.15.0 和TensorFlow 2.x 都有一下信息…
2022-02-19 14:27:11.312590: W tensorflow/core/common_runtime/gpu/gpu_device.cc:1641] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at https://www.tensorflow.org/install/gpu for how to download and setup the required libraries for your platform.
Skipping registering GPU devices...
2022-02-19 14:27:11.312782: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1159] Device interconnect StreamExecutor with strength 1 edge matrix:
2022-02-19 14:27:11.312820: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1165] 0
2022-02-19 14:27:11.312836: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1178] 0: N
整体信息是这样的:
不知道怎么肥四>_>
下面这位朋友也是这样:
kaggle中使用低版本的tensorflow时GPU占用率很低
Kaggle社区的Feedback