Each frame in the sequences is saved as a .npy array of shape 128×128.

• Training Set Files: training_set_0-99.zip, ..., training_set_1500-1599.zip

Content: 1600 sequences of simulated microstructure evolution used to train the neural network models.

Each zip file contains 100 sequences.

• Validation Set Files: validation_set_1600-1699.zip, ..., validation_set_1900-1999.zip

Content: 400 sequences used for model validation and hyperparameter tuning.

Each zip file contains 100 sequences.

• Test Set – Parameter Inference Files: testset_par_0-99.zip, ..., testset_par_400-499.zip

Content: 500 sequences used to evaluate the model’s ability to infer the lattice mismatch parameter (η).

Each file contains 100 sequences.

• Test Set – Evolution Prediction Files: testset_evo_0-99.zip, ..., testset_evo_400-499.zip

Content: 500 sequences used to test the model's ability to predict long-term microstructure evolution under known parameters.

Each file contains 100 sequences. In this test set, frames are saved every 10 time steps to reduce size.