Adding types on the public API surface would allow us to do some runtime type checking later on and would allow user's IDE to have more info for static analysis.

The functions/signatures to type are the ones listed here https://github.com/keras-team/autokeras/blob/master/autokeras/__init__.py

For the context, see #856 where I add some type information on a ImageClassifier.

This issue can

We would like to forward a particular 'key' column which is part of the features to appear alongside the predictions - this is to be able to identify to which set of features a particular prediction belongs to. Here is an example of predictions output using the tensorflow.contrib.estimator.multi_class_head:

{"classes": ["0", "1", "2", "3", "4", "5", "6", "7", "8", "9"],
 "scores": [0.068196

The download buttons for the examples on readthedocs are not working. I cloned the repo and built the docs locally and it works fine. I guess just something in the build process went wrong and it should be working when you build it again.

In the code, it seems that you use the model stored in ema to do the validation and get the final result.Here i have two questions:
1.Why do you use the ema model to do the validation, but do not use the model_wrapper which is directely preuned?
2.Is it supposed that ema model will get better result than model_wrapper?And why?
Thanks very much for your reply!

We need one. Either we list out a py virtualenv with dependencies or we assume a simple non distributed TF or PyTorch and have instructions