Tensor Mode Object Detection Killed Before Launch

Question

I am running dorker image tensorflow: 1.1.0. I added detection of the tensorflow api github object, cloning it locally and giving my docker a connection to this folder. I am trying to recreate their favorite example.

I believe that I have the code and the code in the right places. However, when I try to relearn, the tensor flow kills itself before the start of the training, but does not give any problems or errors.

INFO:tensorflow:Starting Session.
INFO:tensorflow:Starting Queues.
INFO:tensorflow:global_step/sec: 0
Killed

I guess I have something or the other, but without any errors or conclusion, I don't know where to look!

I follow the guide here to run locally: link . Pet data was obtained from the same GitHub: link . I got the model configuration from the same GitHub. link

I choose inception_v2.

Answer 1

Now I understand that I simply did not have enough memory as soon as I started. When I switched to ssd_mobilenet_v1_coco_2017_11_17and changed batch_size = 1, everything worked.

However, this still does not completely fix my problem. I need to figure out how to provide more memory for the call.