Hello there everyone. I am really new to all of this and trying to figure things out one by one. I’ve been able to put together a Faster R-CNN model and I’m tracking loss over eachiteration of the batch. What I’m puzzled by and what I can’t seem to find any information on is the unit of measure of the loss I keep printing out. I know this might seem trivial to everyone but it’s not exactly self evident at all. What exactly does a loss of 0.006 mean. Just trying to make sense of this all. Thanks for any help you can offer me.
If you are following the tutorial, you should get multiple loss values.
Which loss are you referring to?
So as I understand it, each epoch is a complete cycle through my dataset and each batch is a smaller subset of that to help manage memory. The loss I am getting looks like it might be an average loss over the batch that was just processed. The thing I don’t seem to quite understand is the unit of measure of that loss. What exactly is that loss.