How to create a DL Model with inputs having varying dimensions

Technical Discussions
#1

I have images of various dimensions. I want to create a classification model using ResNet50. But how to make the model accept input of various dimensions?

#2

Hi, Rajtilak.

  1. You have to first know in what format i:e dimension the ResNet50 model is taking the input of the images.
  2. Then you should compare with your input images and preprocess you image into their format and resize the image into resnet format.
  3. You can the train it by neural network algorithms to get the .xml file(architecture) and .bin file(weights).
  4. Then you can write code to test it.

All the best!

1 Like