Justin Patriquin

Hey @arlyon that looks like a good solution but one thing I can't figure out is how to get the upload_url from the create_release job to the upload jobs. Thanks!

Close! ``` from tf_encrypted.utils import unwrap_fetches @tf.function def func(): with tfe.protocol.SecureNN() as prot x_in = np.array([1, 2, 3, 4]) y_in = np.array([5, 6, 7, 8]) x_p = prot.define_private_tensor(x_in) y_p =...

you got it.

If `tf.function` works the way we expect it too I think it would be best encourage use of that because that is what users would expect when using tf 2.0...

> One thing I don't like is that we're leaking implementation details by doing it that way. I suggest introducing a tfe.encrypted_computation decorator instead, along the lines of what TF...

If we implement these changes the pip package deployment system will become quite a bit more complicated because we will have to upload a different pip package for each python...

> It would be nice to kick the training process with model.fit(data_sources, epochs=10) and see the training process. @yanndupis how do you think we could accomplish the above? Would we...

This is being worked on [here](https://github.com/tf-encrypted/tf-encrypted/pull/684). Feedback welcome!

In https://github.com/tf-encrypted/tf-encrypted/pull/695 we added the flag to support simulating the splitting the data into the parties. It currently only works with local computations we should see if we can get...

Hmmm probably something like that. I'm confused why anything is running on the 5th CPU player though especially see as that CPU doesn't even exist. Do we run any tests...