You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
All the cores are being used and at times, 100% of the cpu is being used (AMD laptop and desktop is the same). This isn't a problem by itself, but i would like to control this behavior given that I want to run this in a server. When i look at this onnxruntime issue , I see the last comment where a user mentions that sess.run() always uses multiple threads when using the python bindings (I haven't used the onnxruntime directly, only read some code of this library), and here, the official docs mention some option to control the number of threads used.
Here is the profiling of one of my runs in a amd laptop with 4 cores/8 threads.
The text was updated successfully, but these errors were encountered:
When running this code with 100+ images:
All the cores are being used and at times, 100% of the cpu is being used (AMD laptop and desktop is the same). This isn't a problem by itself, but i would like to control this behavior given that I want to run this in a server. When i look at this onnxruntime issue , I see the last comment where a user mentions that
sess.run()
always uses multiple threads when using the python bindings (I haven't used theonnxruntime
directly, only read some code of this library), and here, the official docs mention some option to control the number of threads used.Here is the profiling of one of my runs in a amd laptop with 4 cores/8 threads.
The text was updated successfully, but these errors were encountered: