I followed the whole TensorFlow image captioning tutorial (https://www.tensorflow.org/tutorials/text/image_captioning) and everything works. Now that my model is trained I wanted to know how to export it and reuse it, I couldn't find this anywhere on the internet.
This is a custom auto-encoder model, I tried to save the encoder and decoder separately (I'm not against saving the whole thing) but I have an error:
'RNN_Decoder' object has no attribute 'reset_state'
I try to load the model in this gist: https://gist.github.com/AntoninBeaufort/4f2ddc54772dc51e437d4a4a6ef19b52
And to save the model I use encoder.save() and decoder.save() like in this file: https://github.com/Variares1/Projet_DataScience_A5/blob/master/final/Livrable%203.ipynb
Thanks in advance for your help!