Hopefully, someone would be able to help me understand the difference between yad2k.py vs keras_yolo.py.
On the onset, it seems like yad2k.py reads Darknet19 config and weights and creates Keras model with TF backend whereas keras_yolo.py creates the YOLO_v2 Model Defined in Keras.
In keras_yolo.py the library is using partial, func_tools to extend Kera's Conv2D and define convenient methods that return a compose layers equivalent to Darkent config sections like DarknetConv2D, DarknetConv2D_BN_Leaky, darknet_body.
Anything that adds more to understand will be very helpful. It's interesting that both files were initially checked in on Feb 14th, 2017 (must be a lucky day :-))
Greetings!
Hopefully, someone would be able to help me understand the difference between yad2k.py vs keras_yolo.py.
On the onset, it seems like yad2k.py reads Darknet19 config and weights and creates Keras model with TF backend whereas keras_yolo.py creates the YOLO_v2 Model Defined in Keras.
In keras_yolo.py the library is using partial, func_tools to extend Kera's Conv2D and define convenient methods that return a compose layers equivalent to Darkent config sections like DarknetConv2D, DarknetConv2D_BN_Leaky, darknet_body.
Anything that adds more to understand will be very helpful. It's interesting that both files were initially checked in on Feb 14th, 2017 (must be a lucky day :-))
Thanks