Aug 24, 2017 · 1 min read
Hey Kurt Winter,
I remembered that you had this problem with the bounding boxes that it doesn’t show at all in Tensorboard. I figured out that I was using a different codebase for the Udacity dataset and refactored it for the Raccoon dataset where I deleted some important stuff when we have multiple bounding boxes on an image. I fixed the code and also added the tests for it so that it’s clearer now what to expect. You should try it again and it should work! Sorry again!
