/rust

Rust language bindings for TensorFlow

Primary LanguageRustApache License 2.0Apache-2.0

#TensorFlow Rust Version Status

TensorFlow Rust provides idiomatic Rust language bindings for TensorFlow.

This project is still under active development and not guaranteed to have a stable API. This is especially true because the TensorFlow C API used by this project has not yet stabilized.

Building

If you only intend to use TensorFlow from within Rust, then you don't need to build TensorFlow manually and can follow the automatic steps. If you do need to use TensorFlow outside of Rust, the manual steps will provide you with a TensorFlow header file and shared library that can be used by other languages.

Automatically building TensorFlow

Install SWIG and NumPy. The version from your distro's package manager should be fine for these two. Also install Bazel, which you may need to do from source. Then run cargo build -j 1. Since TensorFlow is built during this process, and the TensorFlow build is very memory intensive, we recommend using the -j 1 flag which tells cargo to use only one task, which in turn tells TensorFlow to build with only on task. Of course, if you have a lot of RAM, you can use a higher value. To include the especially unstable API (which is currently the expr module), use --features tensorflow_unstable.

Manually building TensorFlow

Install TensorFlow from source. The Python/pip steps are not necessary, but building tensorflow:libtensorflow_c.so is.

In short:

  1. Install SWIG and NumPy. The version from your distro's package manager should be fine for these two.

  2. Install Bazel, which you may need to do from source.

  3. git clone --recurse-submodules https://github.com/tensorflow/tensorflow

  4. cd tensorflow

  5. ./configure

  6. bazel build -c opt --jobs=1 tensorflow:libtensorflow_c.so

    If you are building TensorFlow version 0.9.0 or earlier, use libtensorflow.so instead of libtensorflow_c.so.

    Using --jobs=1 is recommended unless you have a lot of RAM, because TensorFlow's build is very memory intensive.

Copy $TENSORFLOW_SRC/bazel-bin/tensorflow/libtensorflow_c.so to /usr/local/lib. If this is not possible, add $TENSORFLOW_SRC/bazel-bin/tensorflow to LD_LIBRARY_PATH.

If you are building TensorFlow version 0.9.0 or earlier, use $TENSORFLOW_SRC/bazel-bin/tensorflow/libtensorflow.so instead.

You may need to run ldconfig to reset ld's cache after copying libtensorflow.so.

Now run cargo build as usual. To include the especially unstable API (which is currently the expr module), use --features tensorflow_unstable.

RFCs

RFCs are issues tagged with RFC. Check them out and comment. Discussions are welcome. After all, thats what a Request For Comment is for!

FAQs

Why does the compiler say that parts of the API don't exist?

The especially unstable parts of the API (which is currently the expr modul) are feature-gated behind the feature tensorflow_unstable to prevent accidental use. See http://doc.crates.io/manifest.html#the-features-section. (We would prefer using an #[unstable] attribute, but that doesn't exist yet.)

Other

This project is not directly affiliated with the TensorFlow project, although we do intend to communicate and cooperate with them.

Developers and users are welcome to join #tensorflow-rust on irc.mozilla.org.

See CONTRIBUTING.md for information on how to contribute code.

This is not an official Google product.

##For more information