Experimental: Intel CPU Software Installation¶

Experimental Release¶

Intel® Extension for TensorFlow* provides experimental support for 2nd Generation Intel® Xeon® Scalable Processors and newer.

Stock TensorFlow and Intel® Optimization for TensorFlow* are recommended if product quality is required.

Verified Hardware Platforms:

Ubuntu 20.04 (64-bit), Ubuntu 22.04 (64-bit) or CentOS Linux 8 (64-bit), and Sapphire Rapids requires Ubuntu 22.04 or CentOS Linux 8 with kernel version >= 5.16
Python 3.8-3.10
pip 19.0 or later (requires manylinux2014 support)

Run the following Dockerfile build procedure to build the pip based deployment container.

Pre-built docker images are available at DockerHub. Run the following command to pull the CPU Docker container image to your local machine.

$ docker pull intel/intel-extension-for-tensorflow:cpu
$ docker run -it -p 8888:8888 intel/intel-extension-for-tensorflow:cpu

Then go to your browser on http://localhost:8888/

The Python development and virtual environment setup recommendation by TensorFlow to isolate package installation from the system.

The Intel® Extension for TensorFlow* requires stock TensorFlow, and the version should be == 2.12.0.

You can follow the instructions in stock tensorflow install to activate the virtual environment.

On Linux, it is often necessary to first update pip to a version that supports manylinux2014 wheels.

(tf)$ pip install --upgrade pip

To install in virtual environment, you can run

(tf)$ pip install tensorflow==2.12.0

If you prefer install tensorflow in $HOME, append --user to the commands.

$ pip3 install --user tensorflow==2.12.0

And the following system environment install for Intel® Extension for TensorFlow* will also append --user to the command.

To install a CPU-only version in virtual environment, you can run

(tf)$ pip install --upgrade intel-extension-for-tensorflow[cpu]

python -c "import intel_extension_for_tensorflow as itex; print(itex.__version__)"