BERT Embeddings
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/imgarylai/bert-embedding
下载链接
链接失效反馈官方服务：
资源简介：
该数据集由预训练的768维隐藏状态转换器模型组成，该模型具有12层和12头注意力机制。在训练期间，该模型被冻结，未进行微调。
This dataset consists of a pre-trained Transformer model with 768-dimensional hidden states, 12 layers and 12-head attention mechanisms. During the training process, the model was frozen and no fine-tuning was performed.
提供机构：
Google
