I am trying to develop an image captioning network in Keras, but after training, the network outputs the same caption for every image.
This is my model:
