Training completes with errors but testing outputs huge number of bounding boxes for test images

[maatteenn]

During training every epoch has the error : TypeError: 'y' of 'Sub' Op has type float32 that doesn't match type int64 of argument 'x'
Also when I test the images, a huge number of bounding boxes appear all over the image with confidence greater than 97.
I had to change resize_images() to resize() and image_dim_ordering() to image_data_format() otherwise code doesn't work.
I'm using tensorflow version 2.3.0 on Colab, and testing model with stock classes or mobile phone, car and person.
Thank-you

[shubham-cpp]

Is there any updates on how to fix this

[Pola8976]

You should check Issue #71