Machine Learning Tools
Preinstalled packages
NICADD cluster provides machine learning (ML) core packages ( tensorflow, keras, torch , e.t.c ) prepared to use with Nvidia graphics cards. For a quick start:
- login to cms1.nicadd.niu.edu
- create a work directory on /xdata or /bdata disks
- mkdir -p /xdata/$USER/mldev
- cd /xdata/$USER/mldev
- load and activate the machine learning python environment
- module load tf/tf130-cuda75-ompi181
- source /opt/nicadd/contrib/tf/tf130-cuda75-ompi181/bin/activate
- At this point one can start using python modules of installed ML packages
- (tf130-cuda75-ompi181)[user@pcms1 mldev]$ python
- To exit the ML environment and return to the parent shell session , enter "deactivate"
- (tf130-cuda75-ompi181)[user@pcms1 mldev]$ deactivate
Tutorials
Python3 compatibility
To work with python36 just use module tf/tf130-cuda75-python3.
List of installed packages (python 2.7)
(tf130-cuda75-ompi181)[user@pcms1 mldev]$ python -m pip list Package Version ---------------------- ----------- backports.weakref 1.0.post1 bleach 1.5.0 funcsigs 1.0.2 html5lib 0.9999999 Keras 2.1.3 Markdown 2.6.9 mock 2.0.0 numpy 1.13.3 pbr 3.1.1 Pillow 5.1.0 pip 18.1 protobuf 3.4.0 PyYAML 3.12 scipy 1.0.0 setuptools 36.5.0 six 1.11.0 tensorflow 1.3.1 tensorflow-tensorboard 0.1.8 Theano 1.0.1 torch 0.3.0.post4 torchvision 0.2.1 Werkzeug 0.12.2 wheel 0.30.0
List of installed packages (python 3.6)
(tf130-cuda75-python3)[user@pcms1 mldev]$ python3 -m pip list python3 -m pip list Package Version ---------------------- ----------- bleach 1.5.0 h5py 2.9.0 html5lib 0.9999999 Keras 2.2.4 Keras-Applications 1.0.7 Keras-Preprocessing 1.0.9 Markdown 3.1 mock 2.0.0 numpy 1.16.2 pbr 5.1.3 Pillow 6.0.0 pip 19.0.3 protobuf 3.7.1 PyYAML 5.1 scipy 1.2.1 setuptools 41.0.0 six 1.12.0 tensorflow 1.3.1 tensorflow-tensorboard 0.1.8 Theano 1.0.4 torch 0.3.0.post4 torchvision 0.2.2.post3 Werkzeug 0.15.2 wheel 0.29.0