Starting from version 6.6 OpendTect-specific Python environments can be installed. OpendTect uses these Python environments for the free Presentation Maker and OpendTect Pro can make use of the CPU MKL and CUDA 11 environments for the Machine Learning plugin. Currently the following Python environments are provided:

Miniconda3

Basic Python environment. Allows using the Presentation Maker plugin without having to install Python yourself.

Provides libraries, including:

  • Python 3.9
  • pandas
  • python-pptx

Math Kernel Library (MKL)

Complete Python environment for Machine Learning using CPU only (with Intel™ Math Kernel Library - MKL).

Provides libraries, including:

  • Python 3.9
  • intel-tensorflow
  • keras
  • scikit-learn
  • pandas
  • bokeh
  • numpy
  • h5py
  • joblib
  • psutil
  • xgboost
  • python-pptx
  • jupyterlab
  • onnxruntime
  • pytorch
  • matplotlib

Python CUDA 11

Complete Python environment for Machine Learning on the GPU with CUDA 11.3.

Requires minimum driver version 466.XX, for specifics see this table on the Nvidia CUDA Toolkit documentation page.

Will fallback automatically on the CPU if the hardware/driver requirements are not met.

Provides libraries, including:

  • Python 3.9
  • tensorflow
  • keras
  • scikit-learn
  • pandas
  • bokeh
  • numpy
  • h5py
  • joblib
  • psutil
  • xgboost
  • python-pptx
  • jupyterlab
  • onnxruntime
  • pytorch
  • matplotlib

Please note: The Python CUDA 10 environment is now obsolete and will no longer receive security updates. Also note that OpendTect does not support the Python CUDA 10 environment from 6.6.6 and onwards.