Official PyTorch implementation
Reconstructing reservoir states from multimodal data via score-based generative models
This code is developed based on the paper Shiqin Zeng, Haoyun Li, Abhinav Prakash Gahlot, Felix J. Herrmann. “Well2Flow: Reconstruction of reservoir states from sparse wells using score-based generative models.”
Below is the workflow diagram illustrating the training and inference process:
Python libraries: See environment.yml for library dependencies. The conda environment can be set up using these commands:
conda env create -f environment.yml
conda activate DiffusionPDE_seismic
