Decision Transformer 复现
时间: 2023-10-14 16:58:27 浏览: 51
Decision Transformer是一种以自回归方式生成动作的网络结构,它是对Transformer的Decoder部分进行修改而来的,主要采用了masked multi-head self-attention的机制\[1\]。Transformer是一种能够整合各种Prior知识、进行信息表达、廉价训练和推理的理想架构,能够更好地处理各种以"set of units"存在的数据,包括计算机视觉、语音、自然语言处理等领域的数据\[2\]。关于Decision Transformer的复现,可以参考相关的论文和代码实现\[3\]。
#### 引用[.reference_title]
- *1* [NeurIPS论文解读|Decision Transformer: 通过序列建模解决离线强化学习问题](https://blog.csdn.net/m0_55289267/article/details/125871135)[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^v91^insertT0,239^v3^insert_chatgpt"}} ] [.reference_item]
- *2* *3* [Transformer课程 第48章Transformer模型RealFormer架构](https://blog.csdn.net/duan_zhihua/article/details/121573075)[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^v91^insertT0,239^v3^insert_chatgpt"}} ] [.reference_item]
[ .reference_list ]