I use PyTorch, bauces AllenNLP is built on it and good libraries are for it. But can AutoKeras be used for PyTorch based ML pipelines, or am I required to switch to Keras? Google is quite silent when asked for tutorials for this combination. Maybe PyTorch is lacking automated ML framework?
"Keras is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano."
But I found a useful tool which collects different deep learning model convertors.
Hope this will be useful for you :D