Causal Transformer
时间: 2023-10-15 22:00:58 浏览: 399
Causal Transformer是一种使用自回归方法进行分类的Transformer模型。在样本输入时,首先通过卷积层进行嵌入,然后将嵌入的图像与图像类别标签配对形成序列输入到仅有解码器的Transformer中。训练目标是预测Omniglot类别标签。研究者通过在长度为L的序列上测量图像分类性能来评估模型的性能,使用子集准确度来度量。当所有L个图像都被正确分类时,子集准确度为1,否则为0。在分类Omniglot手写字符的任务中,Causal Transformer展现出涌现能力,即在预测中出现了意外的结果。这个结果符合之前发表论文中关于涌现能力的描述,比如在大规模多任务语言理解中的情况。<span class="em">1</span><span class="em">2</span><span class="em">3</span>
#### 引用[.reference_title]
- *1* [大模型涌现能力被质疑!斯坦福最新研究,“涌现”只是度量选择的结果](https://blog.csdn.net/xixiaoyaoww/article/details/130498066)[target="_blank" data-report-click={"spm":"1018.2226.3001.9630","extra":{"utm_source":"vip_chatgpt_common_search_pc_result","utm_medium":"distribute.pc_search_result.none-task-cask-2~all~insert_cask~default-1-null.142^v93^chatsearchT3_1"}}] [.reference_item style="max-width: 33.333333333333336%"]
- *2* [routing-transformer:路由变压器的全功能实现](https://download.csdn.net/download/weixin_42109639/16016872)[target="_blank" data-report-click={"spm":"1018.2226.3001.9630","extra":{"utm_source":"vip_chatgpt_common_search_pc_result","utm_medium":"distribute.pc_search_result.none-task-cask-2~all~insert_cask~default-1-null.142^v93^chatsearchT3_1"}}] [.reference_item style="max-width: 33.333333333333336%"]
- *3* [dalle2:hierarchical text-conditional image generation with clip](https://blog.csdn.net/u012193416/article/details/126162618)[target="_blank" data-report-click={"spm":"1018.2226.3001.9630","extra":{"utm_source":"vip_chatgpt_common_search_pc_result","utm_medium":"distribute.pc_search_result.none-task-cask-2~all~insert_cask~default-1-null.142^v93^chatsearchT3_1"}}] [.reference_item style="max-width: 33.333333333333336%"]
[ .reference_list ]
阅读全文
相关推荐


















