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
Thank you for the work, this repository looks great. I just have one questions regarding recreating the results in the original CenterNet paper. I've been training the model using CenterNet/centernet_detection.py with the default parameters, but change batch size to 16 instead of 32 (my GPU cannot train with batch size of 32). But I'm struggling to reproduce the results. I'm wondering if there are any additional parameters that I need to change?
The text was updated successfully, but these errors were encountered:
Hi,
Thank you for the work, this repository looks great. I just have one questions regarding recreating the results in the original CenterNet paper. I've been training the model using
CenterNet/centernet_detection.pywith the default parameters, but change batch size to 16 instead of 32 (my GPU cannot train with batch size of 32). But I'm struggling to reproduce the results. I'm wondering if there are any additional parameters that I need to change?The text was updated successfully, but these errors were encountered: