A bug about the latest update

[shadyatscu]

Thanks for your teams great work! Ive been followed your work on yolov3 from a month, it really help me a lot! In recent time, i find you update repo almost everyday and i always git the latest commit it all works very well. However when i git the latest update(2019/04/19), the loss compute seems meet a bug. I dont know its my problem or the latest update`s, i post the training process below:

Epoch Batch xy wh conf cls total nTargets time 0/272 0/1383 0 0 1.05e+03 0 1.05e+03 0 27.7 0/272 1/1383 0 0 1.05e+03 0 1.05e+03 0 3.23 0/272 2/1383 0 0 1.05e+03 0 1.05e+03 0 0.738 0/272 3/1383 0 0 1.05e+03 0 1.05e+03 0 0.599 0/272 4/1383 0 0 1.05e+03 0 1.05e+03 0 0.61 0/272 5/1383 0 0 1.05e+03 0 1.05e+03 0 0.718 0/272 6/1383 0 0 1.05e+03 0 1.05e+03 0 0.72 0/272 7/1383 0 0 1.05e+03 0 1.05e+03 0 0.648 0/272 8/1383 0 0 1.05e+03 0 1.05e+03 0 0.731 0/272 9/1383 0 0 1.05e+03 0 1.05e+03 0 0.802 0/272 10/1383 0 0 1.05e+03 0 1.05e+03 0 0.693 0/272 11/1383 0 0 1.05e+03 0 1.05e+03 0 0.634 0/272 12/1383 0 0 1.05e+03 0 1.05e+03 0 0.74 0/272 13/1383 0 0 1.05e+03 0 1.05e+03 0 0.644 0/272 14/1383 0 0 1.05e+03 0 1.05e+03 0 0.682 0/272 15/1383 0 0 1.05e+03 0 1.05e+03 0 0.713 0/272 16/1383 0 0 1.05e+03 0 1.05e+03 0 0.789 0/272 17/1383 0 0 1.05e+03 0 1.05e+03 0 0.741 0/272 18/1383 0 0 1.05e+03 0 1.05e+03 0 0.761 0/272 19/1383 0 0 1.05e+03 0 1.05e+03 0 0.787 0/272 20/1383 0 0 1.05e+03 0 1.05e+03 0 0.665 0/272 21/1383 0 0 1.05e+03 0 1.05e+03 0 0.725 0/272 22/1383 0 0 1.05e+03 0 1.05e+03 0 0.67 0/272 23/1383 0 0 1.05e+03 0 1.05e+03 0 0.659 0/272 24/1383 0 0 1.05e+03 0 1.05e+03 0 0.776 0/272 25/1383 0 0 1.05e+03 0 1.05e+03 0 0.648 0/272 26/1383 0 0 1.05e+03 0 1.05e+03 0 0.655 0/272 27/1383 0 0 1.05e+03 0 1.05e+03 0 0.683 0/272 28/1383 0 0 1.05e+03 0 1.05e+03 0 0.651 0/272 29/1383 0 0 1.05e+03 0 1.05e+03 0 0.643

the loss of xy wh and cls always be zero, and i didn`t modify any code except the input data. sorry to bother and thanks a lot!

[glenn-jocher]

Hello, thank you for your interest in our work! This is an automated response. Please note that most technical problems are due to:

• Your changes to the default repository. If your issue is not reproducible in a fresh git clone version of this repository we can not debug it. Before going further run this code and ensure your issue persists:

  sudo rm -rf yolov3  # remove exising repo
  git clone https://github.com/ultralytics/yolov3 && cd yolov3 # git clone latest
  python3 detect.py  # verify detection
  python3 train.py  # verify training (a few batches only)
  # CODE TO REPRODUCE YOUR ISSUE HERE

• Your custom data. If your issue is not reproducible with COCO data we can not debug it. Visit our Custom Training Tutorial for exact details on how to format your custom data. Examine train_batch0.jpg and test_batch0.jpg for a sanity check of training and testing data.
• Your environment. If your issue is not reproducible in a GCP Quickstart Guide VM we can not debug it. Ensure you meet the requirements specified in the README: Unix, MacOS, or Windows with Python >= 3.7, Pytorch >= 1.0, etc.

If none of these apply to you, we suggest you close this issue and raise a new one using the Bug Report template, providing screenshots and minimum viable code to reproduce your issue. Thank you!

[gongjing1996]

Hello, I have met the same problem with you. Do you solve this problem?

[glenn-jocher]

@gongjing1996 hello! We have not received further information on this specific issue. As mentioned in our previous response, please ensure to follow the suggested steps and provide all the necessary details when reporting a bug. Feel free to open a new issue using the Bug Report template if the problem persists. Thank you for your understanding!