search

qfgaohao / pytorch-ssd

MobileNetV1, MobileNetV2, VGG based SSD/SSD-lite implementation in Pytorch 1.0 / Pytorch 0.4. Out-of-box support for retraining on Open Images dataset. ONNX and Caffe2 support. Experiment Ideas like CoordConv.
https://medium.com/@smallfishbigsea/understand-ssd-and-implement-your-own-caa3232cd6ad
MIT License
1.39k stars 529 forks source link

NAN values in Boxes #181

Closed ishoud closed 2 years ago

ishoud commented 2 years ago

Hi, I am trying to train the mobilenet based model on my custom dataset. Capture Above image shows the data format. I modeified the open images dataloader to create my dataloader.

But when I iterate through the dataloader , I find nan values in the boxes. It comes when we apply "target_transform" to the boxes. Please let me know how I can rectify the issue. Any help will be highly appreciated

ishoud commented 2 years ago

The error was in format of data, Some conversion error from converting from VOTT format and Open image format. Documenting it here.