A Singularity definition file for a python-centric data science container built on Ubuntu 18.04 that is useful in GPU-accelerated HPC environments for machine learning and deep learning applications.
Installed libraries and packages include:
(if version number not specified, latest version available or default base OS package is used)
- CUDA 10.0
- Python 3.6
- nano / vim / emacs
- graphviz
- Spark 2.4.3
- OpenCV 4.1.0 (with contrib modules and GPU support enabled)
- NumPy
- jupyter / JupyterLab
- Apache Toree
- SciPy
- Pandas
- Beautiful Soup 4
- Scrapy
- matplotlib
- seaborn
- Plotly
- Bokeh
- Altair
- Pygal
- pydot
- TensorFlow 2.0.0-beta1 (GPU version)
- TensorFlow Hub
- Tensor2Tensor
- Sonnet
- CNTK (GPU version)
- Apache MXNet (GPU version)
- GluonCV
- statsmodels
- XGBoost
- scikit-learn
- NLTK
- spaCy
- Gensim
- FastText
- GluonNLP
- PyTorch (GPU version)