In C++ and Python, Tensors can be serialized into a `TensorProto` using functions like [`Tensor::AsProtoTensorContents`](https://github.com/tensorflow/tensorflow/blob/f7ec99516ce0e0937e0b865e90aa02c748cd36c6/tensorflow/core/framework/tensor.h#L205) and [`make_tensor_proto`](https://www.tensorflow.org/versions/r0.12/api_docs/python/contrib.util.html#make_tensor_proto) This is particularly useful when communicating tensors across process boundaries such as [PredictRequest](https://github.com/tensorflow/serving/blob/8c89d72259a7c15f0a9711a67eecf284d9460cca/tensorflow_serving/apis/predict.proto#L22) RPCs to a [TensorFlow Model Server](https://tensorflow.github.io/serving/) Filing this to track any changes to the [C API](https://github.com/tensorflow/tensorflow/blob/master/tensorflow/c/c_api.h) and language bindings. Interested parties: @asimshankar @jhseu @nfiedel @josh11b