-
-
Notifications
You must be signed in to change notification settings - Fork 1.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Incorrect in assigning box GT (?!), and suggest the new one #44
Comments
As it said in the original paper:
Thanks for your distinctive idea, but I can't agree with you. As we know, |
Oh, I see. But, I mean, the consistency is about which grid we assign for the object label. A good common sense is: (i) assigning to the grid which has biggest IoU to the box GT, right? See another illustration here, if we use Anyway, do you ever run this code to train MS COCO dataset from the scratch (not using pre-training network)? I wonder how many days are needed. I'm currectly still running this code to train MS COCO from the scratch. Now is in epoc 4 (3 days), it looks converged, but seems it needs a lot of epoch, i.e., a lot of days (in the original paper is stated using 160 epoch). |
Oh, I got it. Your idea is very impressive ! But, since YOLO is a regression problem, which means the regression objection must be certain. So we need a particular grid cell location to be regressor. For training MS COCO dataset from scratch, I have not done it yet. I will appreciate it very much if you would have shared your result with us. |
Yes sure, later I can share the result. Now is still in epochs 6, the recall is still Btw, do you know why running this training code constantly increases the used RAM memory? Keeping running the code will encouter running out of (ram) memory issue and eventually kills the process. |
Good news
Putting |
Can you share the code ?? Thanks a lot! |
Hi,
First, thanks a lot for making this project.
I spot a potential issue in assigning box GT in this code (line 99 and 100), rewritten as:
Using code above, the label assigment may encounter wrong assignment illustrated in this attachment pic, where the label should be in the grid of (1,1), but code above puts the label in grid (0,0) due to
floor operation
. So, I suggest to useround operation
instead. I know, the network can still learn via the training data, but I think, usinground operation
, in this case, will be more consistent and makes the network easier to learn. What do you think?Thanks.
The text was updated successfully, but these errors were encountered: