mBART

该数据集是一个多语言编码器-解码器模型，能够在一个模型内支持多种语言的机器翻译和文本生成任务。该模型的输入序列长度设置为1024，并拥有50万的词汇量，确保了在处理多种语言时的广泛适用性和高效性能。

This dataset is a multilingual encoder-decoder model capable of supporting machine translation and text generation tasks for multiple languages within a single model. Its input sequence length is configured to 1024, and it has a vocabulary size of 500,000, which ensures broad applicability and efficient performance when handling multilingual tasks.