Learn2Zinc-Augmented: Dataset for MiniZinc Generation and Error Correction
Overview
Learn2Zinc-Augmented is a supervised fine-tuning dataset that combines two complementary tasks for training large language models on MiniZinc: code generation (translating natural-language problems into MiniZinc) and code correction (fixing buggy MiniZinc models). The augmented dataset is designed to improve both the generative and debugging capabilities of fine-tuned models.
This dataset… See the full description on the dataset page: https://huggingface.co/datasets/skadio/learn2zinc-augmented.
