ppantidis/I-FENN-with-TCNs-nonlocal-gradient-damage — reverse-engineered prompt
Reverse engineered prompt
Build me a runnable version of this research code so I can reproduce the paper examples without having to figure out the whole codebase myself.
I want a simple workflow where I can open the main script, pick one of the included test cases from the text files, and run either the regular finite element analysis or the faster I FENN mode that uses the Temporal Convolutional Network. It should read the provided mesh and parameter files, use the included training data and saved model where available, and produce the main results in a clear way, including saved outputs and plots of the damage evolution and load history so I can compare cases.
Please keep the original MATLAB plus Python setup, but make the project easy to run end to end with sensible defaults and short instructions in the README for training, prediction, and reproducing at least one example from the paper. If anything is unclear, look up the current docs online and fill in the missing run steps.
Want more depth? Deep Reverse