ALMA-13B-R

该数据集是通过对比偏好优化方法训练的解码器模型，它使用了高质量的数据，专门用于翻译任务。在提出的集成方法中，该模型作为一个排名器，用以提升翻译质量。该数据集涉及的任务是机器翻译。This dataset is designed for training decoder models optimized via Contrastive Preference Optimization (CPO) with high-quality data, and is specifically tailored for translation tasks. In the proposed ensemble method, this model acts as a ranker to improve translation quality. The task associated with this dataset is machine translation.