Installation¶
Requirements¶
NengoDL works with Python 3.7 or later. pip will do its best to install
all of NengoDL’s requirements when it installs NengoDL. However, if anything
goes wrong during this process you can install the requirements manually and
then try to pip install nengo-dl again.
See the Nengo documentation
for instructions on installing numpy and nengo, and the tensorflow
installation instructions below.
Developer installation¶
If you want to modify NengoDL, or get the very latest updates, you will need to perform a developer installation:
git clone https://github.com/nengo/nengo-dl.git
pip install -e ./nengo-dl
Installing TensorFlow¶
Use pip install tensorflow to install the latest version of TensorFlow. GPU support
is included in this package as of version 2.1.0.
In order to use TensorFlow with GPU support you will need to install the appropriate
Nvidia drivers. The precise steps for accomplishing this will depend
on your system. On Linux the Nvidia drivers can be installed via
sudo apt install nvidia-driver-<version> (check the Nvidia/TensorFlow documentation
for the supported driver <version> for your version of TensorFlow/CUDA).
On Windows simply using the most up-to-date drivers should work.
For GPU installations we recommend using
conda to
simplify the process. conda install -c conda-forge tensorflow will install
TensorFlow as well as all the CUDA/cuDNN requirements. If you run into any problems,
see the
TensorFlow GPU installation instructions
for more details.
It is also possible to build TensorFlow from source. This is significantly more complicated but allows you to customize the installation to your computer, which can improve simulation speeds.