Building NCA for accelerating solidification microstructure modeling
Folder MUlti-GPU-Code contains codes for training and evaluating the NCA. The file names and corresponding functions are as follow:
File | Function |
---|---|
MUlti-GPU-Code/CA.py | CA simulation for generating training and validation data |
MUlti-GPU-Code/CA_test_run.py | CA simulation for testing |
MUlti-GPU-Code/Evaluate_model.py | Perform test run |
MUlti-GPU-Code/LoadFunc_train.py | Load useful function and python package |
MUlti-GPU-Code/Loadmodel_T_train.py | Define and load the NCA model |
MUlti-GPU-Code/generate_larger_grad.py | Setting for generating training and validation data |
MUlti-GPU-Code/mgpu_ppre_t_grad.py | Train the NCA |
The files of testing results and trained models for various cases are in Folder Paper_review/Cases. Details see Paper_review/Cases/README.md.