CanPredict: Training Corpora and Trained Model Artifacts for Covering-Array Size Prediction

This record contains the large research artifacts accompanying the CanPredict project: CanPredict: Predicting the Sizes of Covering Arrays! The artifacts are intended to support reproducible training and inference for the 2-way and 3-way covering-array size prediction experiments. Contents: - t2_corpus.pt: Serialized 2-way LCG training corpus.- t3_corpus.pt: Serialized 3-way LCG training corpus.- CanPredict-training-model-v1.0.0.zip: Trained model artifacts used by the prediction and ablation experiments. The archive contains the top-level training-model/ directory.- SHA256SUMS.txt: SHA-256 checksums for the uploaded artifacts.- ZENODO_UPLOAD_README.md: Instructions for restoring the artifacts in the CanPredict repository. The corresponding source code and documentation are available at: https://github.com/dwnie/CanPredict After downloading the files, place t2_corpus.pt and t3_corpus.pt in the repository's corpus/ directory. Extract CanPredict-training-model-v1.0.0.zip at the repository root to restore the training-model/ directory. This record corresponds to CanPredict software release v1.0.0.