tensorflow: Failed to get device attribute 13 for device 0

When I’am trying to run Yolo detection examples, I got that error:

2020-02-02 21:39:00.821721: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_101.dll                                                                                                                  
WARNING:tensorflow:From C:\Users\Dominux\AppData\Local\Programs\Python\Python37\lib\site-packages\tensorflow_core\python\ops\resource_variable_ops.py:1635: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version.                                                                        
Instructions for updating:                                                                                                       
If using Keras pass *_constraint arguments to layers.                                                                            
2020-02-02 21:39:03.863436: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library nvcuda.dll                                                                                                                        
2020-02-02 21:39:04.431694: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1555] Found device 0 with properties:             
pciBusID: 0000:02:00.0 name: GeForce MX230 computeCapability: 6.1                                                                
coreClock: 1.531GHz coreCount: 2 deviceMemorySize: 2.00GiB deviceMemoryBandwidth: 44.76GiB/s                                     
2020-02-02 21:39:04.437212: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudart64_101.dll                                                                                                                  
2020-02-02 21:39:04.444498: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cublas64_10.dll                                                                                                                   
2020-02-02 21:39:04.450110: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cufft64_10.dll                                                                                                                    
2020-02-02 21:39:04.453997: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library curand64_10.dll                                                                                                                   
2020-02-02 21:39:04.459404: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cusolver64_10.dll                                                                                                                 
2020-02-02 21:39:04.464501: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cusparse64_10.dll                                                                                                                 
2020-02-02 21:39:04.477818: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library cudnn64_7.dll                                                                                                                     
2020-02-02 21:39:04.480586: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1697] Adding visible gpu devices: 0               
2020-02-02 21:39:09.674559: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2                                                                                         
2020-02-02 21:39:09.678508: F tensorflow/stream_executor/lib/statusor.cc:34] Attempting to fetch value instead of handling error 
Internal: failed to get device attribute 13 for device 0: CUDA_ERROR_UNKNOWN: unknown error

My stack:

Win 10,
Tensoflow 2.1,
Intel Core I5,
Nvidia GeForce MX230 2GB,
8GD DDR4

I checked similar issues, but they don’t have solutions. Just despaired people… But it problem was talked about by them only with TF 1.14 - I didn’t find other But how you could notice at me stack above I’m using TF 2.1 already

Please, can you help me! Maybe I’ve got problems with drivers or CUDA software? Any ideas?

About this issue

Original URL
State: closed
Created 4 years ago
Comments: 39 (6 by maintainers)

Most upvoted comments

Had the same issue, but resolved by updating NVIDIA device drivers to the latest (v446.14). https://www.nvidia.com/download/index.aspx?lang=en-us

razmik on Jun 23, 2020

So … my friend … I had the same problem and it was solved this way click right in any empty place in Desktop and chose Nvidia Control Panel and the picture explain ddd

mobajjal on Jul 22, 2020

Hello, I got a similar error with Pytorch. I am new to TensorFlow but have experience with Pytorch.

I have…

Windows 10
NVIDIA GTX 1050 4G and CUDA v10.1
PyCharm as IDE
TensorFlow 2.1.0

In my cases, I could solve those errors by simply generating a small tensor on the GPU. In this TensorFlow case, I run the following python code on the terminal.

import tensorflow as tf
t = tf.random.uniform([2,2])

Unfortunately, I cannot explain why this happens but hope this helps someone.

takakihagen on Apr 17, 2020

Thanks for the reply. Can you edit the detect.py script by limiting gpu memory growth. Kill all python sessions/jupyter notebook and just add following lines on top after you import modules in detect.py.

import tensorflow as tf
gpus = tf.config.experimental.list_physical_devices('GPU')
tf.config.experimental.set_memory_growth(gpus[0], True)
##    Put rest of the detect.py code  ##

ymodak on Feb 7, 2020