------------------------------------ XLNet模型汇总 ------------------------------------ 下表汇总介绍了目前PaddleNLP支持的XLNet模型对应预训练权重。 关于模型的具体细节可以参考对应链接。 +----------------------------------------------------------------------------------+--------------+----------------------------------------------------------------------------------+ | Pretrained Weight | Language | Details of the model | +==================================================================================+==============+==================================================================================+ |``xlnet-base-cased`` | English | 12-layer, 768-hidden, | | | | 12-heads, 110M parameters. | | | | XLNet English model. | +----------------------------------------------------------------------------------+--------------+----------------------------------------------------------------------------------+ |``xlnet-large-cased`` | English | 24-layer, 1024-hidden, | | | | 16-heads, 340M parameters. | | | | XLNet Large English model. | +----------------------------------------------------------------------------------+--------------+----------------------------------------------------------------------------------+ |``chinese-xlnet-base`` | Chinese | 12-layer, 768-hidden, | | | | 12-heads, 117M parameters. | | | | XLNet Chinese model. | +----------------------------------------------------------------------------------+--------------+----------------------------------------------------------------------------------+ |``chinese-xlnet-mid`` | Chinese | 24-layer, 768-hidden, | | | | 12-heads, 209M parameters. | | | | XLNet Medium Chinese model. | +----------------------------------------------------------------------------------+--------------+----------------------------------------------------------------------------------+ |``chinese-xlnet-large`` | Chinese | 24-layer, 1024-hidden, | | | | 16-heads, _M parameters. | | | | XLNet Large Chinese model. | +----------------------------------------------------------------------------------+--------------+----------------------------------------------------------------------------------+