Toby Roseman

A minimal example to reproduce the problem would be good. Looking at the code I'm not really sure what's going on. Each process loads its' own model? Or is the...

I'm not really understanding the table you've shared. How does your side information contain both a person id and a product id? I would expect two separate SFrames: one with...

Please share more information. It sounds like you're doing image classification, is that correct? `squeezenet_v1.1` and `VisionFeaturePrint_Scene` work well but `resnet-50` does not work, is this correct?

> when i try to get prediction it says for every image not predicted. I've never seen this behavior. Can you share the output you are seeing?

@kld123509945 - MXNet has been removed as a dependency. Please try using TuriCreate again with Jupyter Notebook and WSL. If you get an error, please share the **entire** stack trace.

I'm not able to replicate this issue. I can successfully run the [Introductory Example](https://apple.github.io/turicreate/docs/userguide/sound_classifier/#introductory-example) for the Sound Classifier using your versions of TuriCreate and TensorFlow. A few questions: 1 -...

This is a TensorFlow issue. On Linux (and older versions of macOS) TuriCreate uses TensorFlow to generate the deep feature. To reproduce this issue without TuriCreate, first [download the keras...

This is related to #933.

Do you get consistent results by setting the `initial_centers` parameter?

Using the same value for `initial_centers` does indeed seem to always yield the same results. So that can effectively be used as a random seed. I still think having an...