Implements color classifier training/testing/freezing in python & tensorflow & keras

• Clone the project:

   git clone https://github.com/borisef/cm.git

Most simple and intuitive data: rgb images in folders named as names of classes: "red", "white", "green" , etc. Supports folder-in-folder structure (for example for augmented images)

Entry points are those two scripts: ColorNet_TestGeneral.py, ColorNet_TrainGeneral.py The output folder after train and test will contain:

• saved models - of several types (we used the on in folder k2tf_dir)
• statistics - (confusion matrices, histograms etc)
• checkpoints along training