baz-net
A Deep Neural Network for Confident Three-component Backazimuth Prediction
Setup the Environment:
Creating the Python Environment:
Ensure that Anaconda is installed on your system, then create a new conda virtual environment named SeisTF2, and activate it:
conda create -n SeisTF2
conda activate SeisTF2
Next, install the basic anaconda packages, geoscience packages, xgboost and gpu acceleration packages:
conda install -c anaconda anaconda
conda install -c conda-forge obspy cartopy geographiclib
conda install -c conda-forge xgboost eli5
conda install cudnn cupti cudatoolkit=10.0
Finally, use pip to install tensorflow 2.0 with gpu support:
pip install tensorflow-gpu
Running the Notebook:
Save the contents of this git repository to your local computer, open a terminal in the folder where it resides, then type the following:
conda activate SeisTF2
jupyter notebook
Your web browser should now open to show the contents of the folder from which you activated the notebook. Simply click on the BazNet.ipynb file and the notebook should open.