denseNet169Prediction

This examples demonstrates the inference concept on DenseNet169 model:

• Model configuration, model weights and labels are obtained from TFModelHub.

• Weights are loaded from .h5 file, configuration is loaded from .json file.

• Model predicts on a few images located in resources.

• Special preprocessing (used in DenseNet169 during training on ImageNet dataset) is applied to images before prediction.

NOTE: Input resolution is 224*224