transformer中的cls
时间: 2023-09-25 09:09:31 浏览: 53
在中,"cls"代表"classification",它是一个特殊的标记符号。在ViT算法中,ViT提出了一个可学习的嵌入向量Class Token,将它与其他输入向量一起输入到Transformer结构中进行编码。然后,使用这个Class Token进行分类预测,它在所有Token的第一个位置,即编码中的x[:,0]。通过这个Class Token进行分类预测可以将Transformer应用于图像分类任务。<span class="em">1</span><span class="em">2</span><span class="em">3</span>
#### 引用[.reference_title]
- *1* [ViT:实现Vi(sion)T(transformer)](https://download.csdn.net/download/weixin_42134537/19081997)[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: 50%"]
- *2* *3* [vit transformer中的cls_token](https://blog.csdn.net/weixin_41803874/article/details/125729668)[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: 50%"]
[ .reference_list ]