I keep running into this error if I use my custom TF Records when training the Model in TensorFlow Object Detection API V2

146 views Asked by At

Error: tensorflow.python.framework.errors_impl.InvalidArgumentError: assertion failed: [[0.956730783]] [[0.896634638]] [[{{node Assert/AssertGuard/else/_25/Assert/AssertGuard/Assert}}]] [[MultiDeviceIteratorGetNextFromShard]] [[RemoteCall]]

I am using Vehicles Data from RoboFlow and I converted it's PASCAL VOC XML format to TFRecords through TensorFlow's official script and every time I use my data, this error shows up but if use the ones RoboFlow directly gives you then the error doesn't show up at all. A similar error is already reported on TensorFlow issues but it got closed and I tried that solution but It doesn't work for me which involved making sure all my values of label heights and widths were positive

0

There are 0 answers