This folder contains the solution developed by ZJUKLAB for the SemEval 2025 Task 4 competition.

conda create -n semeval_unlearn python=3.12
conda activate semeval_unlearn
pip install -r requirements.txt

• --forget_dataset: Specifies the dataset to forget (must be a valid dataset path or identifier).
• --retain_dataset: Specifies the dataset to retain.
• --model_path: Path to the pre-trained model.
• --output_dir: Directory where results and logs will be saved.

torchrun --nproc_per_node=1 --master_port=29500 unlearn-merging.py --forget_dataset /path/to/forget_data --retain_dataset /path/to/retain_data --model_path /path/to/model --output_dir /path/to/output