Go: Simplify installation.

Utilize the TensorFlow C library release archive to
avoid building TensorFlow from source to get going
much faster when using Linux/Mac OS X on x86_64.
Change: 146805287

1 files changed, 87 insertions, 30 deletions

diff --git a/tensorflow/go/README.md b/tensorflow/go/README.md
index 7696fd4fe4..26e3927afe 100644
--- a/tensorflow/go/README.md
+++ b/tensorflow/go/README.md
@@ -1,4 +1,4 @@
-# TensorFlow Go API
+# TensorFlow in Go
 
 Construct and execute TensorFlow graphs in Go.
 
@@ -8,39 +8,86 @@ Construct and execute TensorFlow graphs in Go.
 > without notice. The same goes for the awkward package path
 > (`github.com/tensorflow/tensorflow/tensorflow/go`).
 
-## Requirements
+## Quickstart
 
--   Go version 1.7+
--   [bazel](https://www.bazel.build/versions/master/docs/install.html)
--   Environment to build TensorFlow from source code
-    ([Linux](https://www.tensorflow.org/versions/master/get_started/os_setup.html#prepare-environment-for-linux)
-    or [Mac OS
-    X](https://www.tensorflow.org/versions/master/get_started/os_setup.html#prepare-environment-for-mac-os-x)).
-    If you'd like to skip reading those details and do not care about GPU
-    support, try the following:
+1.  Download and extract the TensorFlow C library, preferably into `/usr/local`.
+    GPU-enabled versions require CUDA 8.0 and cuDNN 5.1. For other versions, the
+    TensorFlow C library will have to be built from source (see below).
+
+    -   Linux:
+        [CPU-only](https://storage.googleapis.com/tensorflow/libtensorflow/libtensorflow-cpu-linux-x86_64-1.0.0rc1.tar.gz),
+        [GPU-enabled](https://storage.googleapis.com/tensorflow/libtensorflow/libtensorflow-gpu-linux-x86_64-1.0.0rc1.tar.gz)
+    -   OS X
+        [CPU-only](https://storage.googleapis.com/tensorflow/libtensorflow/libtensorflow-cpu-darwin-x86_64-1.0.0rc1.tar.gz),
+        [GPU-enabled](https://storage.googleapis.com/tensorflow/libtensorflow/libtensorflow-gpu-darwin-x86_64-1.0.0rc1.tar.gz)
+
+    The following shell snippet downloads and extracts into `/usr/local`:
 
     ```sh
-    # On Linux
-    sudo apt-get install python swig python-numpy
+    TF_TYPE="cpu" # Set to "gpu" for GPU support
+    curl -L \
+      "https://storage.googleapis.com/tensorflow/libtensorflow/libtensorflow-${TF_TYPE}-$(go env GOOS)-x86_64-1.0.0rc1.tar.gz" |
+    sudo tar -C /usr/local -xz
+    ```
 
-    # On Mac OS X with homebrew
-    brew install swig
+2.  `go get` this package (and run tests):
+
+    ```sh
+    go get github.com/tensorflow/tensorflow/tensorflow/go
+    go test github.com/tensorflow/tensorflow/tensorflow/go
     ```
 
-## Installation
+3.  Done!
+
+### Installing into locations other than `/usr/local`
+
+The TensorFlow C library (`libtensorflow.so`) needs to be available at build
+time (e.g., `go build`) and run time (`go test` or executing binaries). If the
+library has not been extracted into `/usr/local`, then it needs to be made
+available through the `LIBRARY_PATH` environment variable at build time and the
+`LD_LIBRARY_PATH` environment variable (`DYLD_LIBRARY_PATH` on OS X) at run
+time.
+
+For example, if the TensorFlow C library was extracted into `/dir`, then:
+
+```sh
+export LIBRARY_PATH=/dir/lib
+export LD_LIBRARY_PATH=/dir/lib   # For Linux
+export DYLD_LIBRARY_PATH=/dir/lib # For OS X
+```
+
+## Building the TensorFlow C library from source
+
+If the "Quickstart" instructions above do not work (perhaps the release archives
+are not available for your operating system or architecture, or you're using a
+different version of CUDA/cuDNN), then the TensorFlow C library must be built
+from source.
 
-1.  Download the TensorFlow source code:
+### Prerequisites
+
+-   [bazel](https://www.bazel.build/versions/master/docs/install.html)
+-   Environment to build TensorFlow from source code
+    ([Linux](https://www.tensorflow.org/versions/master/get_started/os_setup.html#prepare-environment-for-linux)
+    or [OS
+    X](https://www.tensorflow.org/versions/master/get_started/os_setup.html#prepare-environment-for-mac-os-x)).
+    If you don't need GPU support, then try the following: `sh # Linux sudo
+    apt-get install python swig python-numpy # OS X with homebrew brew install
+    swig`
+
+### Build
+
+1.  Download the source code
 
     ```sh
     go get -d github.com/tensorflow/tensorflow/tensorflow/go
     ```
 
-2.  Build the TensorFlow library (`libtensorflow.so`):
+2.  Build the TensorFlow C library:
 
     ```sh
     cd ${GOPATH}/src/github.com/tensorflow/tensorflow
     ./configure
-    bazel build -c opt //tensorflow:libtensorflow.so
+    bazel build --config opt //tensorflow:libtensorflow.so
     ```
 
     This can take a while (tens of minutes, more if also building for GPU).
@@ -50,24 +97,34 @@ Construct and execute TensorFlow graphs in Go.
     a. Copying it to a system location, e.g.,
 
     ```sh
-    cp ${GOPATH}/src/github.com/tensorflow/tensorflow/bazel-bin/tensorflow/libtensorflow.so /usr/local/lib
+    sudo cp ${GOPATH}/src/github.com/tensorflow/tensorflow/bazel-bin/tensorflow/libtensorflow.so /usr/local/lib
     ```
 
     OR
 
-    b. Setting the
-    `LD_LIBRARY_PATH=${GOPATH}/src/github.com/tensorflow/tensorflow/bazel-bin/tensorflow`
-    environment variable (`DYLD_LIBRARY_PATH` on Mac OS X).
+    b. Setting environment variables:
+
+    ```sh
+    export LIBRARY_PATH=${GOPATH}/src/github.com/tensorflow/tensorflow/bazel-bin/tensorflow
+    # Linux
+    export LD_LIBRARY_PATH=${GOPATH}/src/github.com/tensorflow/tensorflow/bazel-bin/tensorflow
+    # OS X
+    export DYLD_LIBRARY_PATH=${GOPATH}/src/github.com/tensorflow/tensorflow/bazel-bin/tensorflow
+    ```
 
-After this, the `go` tool should be usable as normal. For example:
+4.  Build and test:
 
-```sh
-go test -v github.com/tensorflow/tensorflow/tensorflow/go
-```
+    ```sh
+    go test github.com/tensorflow/tensorflow/tensorflow/go
+    ```
+
+## Support
+
+Use [stackoverflow](http://stackoverflow.com/questions/tagged/tensorflow) and/or
+[Github issues](https://github.com/tensorflow/tensorflow/issues).
 
 ## Contributions
 
-This API has been built on top of the [C
-API](https://www.tensorflow.org/code/tensorflow/c/c_api.h),
-which is intended for building language bindings for TensorFlow functionality.
-However, this is far from complete. Contributions are welcome.
+Contributions are welcome. If making any signification changes, probably best to
+discuss on a [Github issue](https://github.com/tensorflow/tensorflow/issues)
+before investing too much time. Github pull requests are used for contributions.