# Long Text Understanding

Dolphin 2.7 Mixtral 8X7B is a large language model based on the Mixtral architecture, focusing on code generation and instruction-following tasks.

BERT-large model pre-trained on Japanese Wikipedia data, fine-tuned for UD head parsing tasks

A pre-trained Albert large-scale Chinese Q&A model based on Baidu WebQA and Baidu DuReader datasets, suitable for Chinese Q&A tasks.

News summarization model based on ALBERT and GPT2 architectures, fine-tuned on the CNN/DailyMail dataset