datasets: - EdinburghNLP/xsum language: - en metrics: - bertscore - bleu - rouge base_model: - google-t5/t5-base - meta-llama/Llama-3.2-1B-Instruct