Adam Crume

The title doesn't, really. What exactly are you proposing?

I've tried to convert a rank-1 tensor of dtype=uint8 to a rank-0 tensor of dtype=string using `Cast`, `ReduceJoin`, and `DecodeRaw`, and they all fail. There doesn't seem to be any...

I can't review this change until the CLA is signed.

The problem is that the generated code (which is checked in) is only compatible with the exact same version of `protobuf` that generated it. This is a very annoying property...

I haven't been able to reproduce this, either. (I only have an x86 Linux machine.) Can you try running your binary under `valgrind`? That'll be something like ``` cargo build...

I just asked the TensorFlow folks to fix the missing download: https://github.com/tensorflow/tensorflow/issues/59828

Addressing #381 seems reasonable, as long as we guard it behind a feature flag that makes it clear it's experimental. It looks like building from source or running under WSL...

Are you pointing Rust to the TensorFlow library you built? There are instructions on how to do that at https://github.com/tensorflow/rust/blob/master/tensorflow-sys/README.md#manual-tensorflow-compilation.