Faster rcnn training aborts after few epochs

Hi, Im trying to train a faster rcnn model using tao toolkit on a single gpu. After few epochs the process terminates automatically. Im attaching the error image below. Help me with it.