🚧
| Platform | CI Status |
|---|---|
| Linux, Python 3.6 |  |
A PyTorch application for next-generation decline curve modelling. The package is named luibeal (Lewis and Beal, 1918).
- Conventional decline curve models are not intended for changing conditions
- Recurrent neural networks (RNN) can process temporal data
- These temporal data may be used to represent changes in conditions during a well’s lifetime
- This provides means for a new decline curve modeling framework
- The new framework is able to account for refracking, well interactions and more
PyTorch has chosen a somewhat tight alignment with Anaconda, so this project falls in line given the strong dependency. So standard procedure using a virtual environment would then be
conda create -n dcenv
source activate dcenv
conda env update -f environment.yml -n dcenv
python setup.py
which creates an environment, activates it, installs dependecies and gets you out of the environment. Optionally you can specify python setup.py develop Alternatively, you can create the environment in one go and step into it later, all set up.
conda env create -f environment.yml -n dcenv
python setup.py
There is also a dockerfile which provided the required development environment - once in the container, cd into the mapped directory and python setup.py. Also, no need for GPU, this is kindergarden.