Bounding boxes not drawn on inference image

I am working with DIGITS6. I am able to run the training with a custom dataset and get an mAP of 37. on testing with single image, I do get the bounding boxes when the raw option is chosen. However when the bounding box option is chosen for visualization method, it says that one bounding box is detected but no inference image gets displayed.

I had a similar problem with DIGITS 6.1 noticed there was an error on the client side with angular 1.5.3

Argument ‘bbox_controller_1’ is not a function, got undefined

The issue was both in Google Chrome and Firefox.

Couldn’t figure out the issue so I ended up going version release DIGITS 6.0 which resolved that issue but I needed to manually update Flask extensions and vulnerability issues through a hand full of the Python scripts. As well as, adjusting the versions of dependencies to get it working. This was a manual DIGITS install on Ubuntu 16.04.

I did inference in jetson and bounding boxes are displayed…

Please see https://devtalk.nvidia.com/default/topic/1055119/container-digits/digits-failed-to-draw-box-after-training-process/. If it solves the issue, please let us know.