infinite time execution (noob)

Dear community,

I am new in PyCharm (noob) and have problem that I can't solve by myself. I have defined environment with python 3.7 for machine learning. It works fine through all my steps but once I connect notebook to the laptop GPU I stack during model fitting step. Process shows infinite execution without any output in console or error. Help me please to solve it 

I will be thankful for any help

0
Hi, how do you connect the notebook to GPU? Could you please try running your notebook through the native Jupyter web app and let me know if the issue persists?
0

Hi Andrey,

 

thank you for replay. I think I have conflicting between libraries. When I check GPU it says me "True" ([PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]). Unlikely I found out that I am actually not using GPU or using only 2% of the GPU during calculation. Jupyter web app was able to finish my calculation without using GPU while PyCharm crashed if I apply the following code:

 

try:
# Disable all GPUS
tf.config.set_visible_devices([], 'GPU')
visible_devices = tf.config.get_visible_devices()
for device in visible_devices:
assert device.device_type != 'GPU'
except:
# Invalid device or cannot modify virtual devices once initialized.
pass
 
I also tested different code (mask RCNN TF2) using the same environment and Jupyter web app reaches "died kernel" =) I found in internet that it could be because of the libraries conflict and this is really painful problem =)

 

 

0
Please note that PyCharm doesn't control access to GPU -- it's controlled by the Jupyter kernel. The only thing PyCharm can affect is the environment, i.e. if GPU is configured using environment variables.

Normally, we can find the problem by comparing the environment between the web app and PyCharm, but if the web app is not using the GPU as well, then the problem is somewhere with the kernel/server configuration. Unfortunately, I'm not sure I can help in such case.
0

请先登录再写评论。