yolov5 GSconv

YOLOv5中的GSConv是指使用了轻量级卷积方法GSConv来代替传统的SC（Spatial Convolutional）卷积。GSConv的计算成本约为SC的60%~70%，但在模型学习能力方面与SC不相上下。同时，为了进一步优化模型结构，还引入了GSbottleneck模块。GSbottleneck模块的结构如图5（a）所示。此外，在配置文件中还需要修改Neck模块中的Conv为GSConv，以及将C3模块换为VoVGSCSP。<span class="em">1</span><span class="em">2</span><span class="em">3</span>
#### 引用[.reference_title]
- *1* *3* [YOLOv5改进之YOLOv5+GSConv+Slim Neck](https://blog.csdn.net/m0_56247038/article/details/126546033)[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_2"}}] [.reference_item style="max-width: 50%"]
- *2* [改进Yolov5 | 用 GSConv+Slim Neck 一步步把 Yolov5 提升到极致！！！](https://blog.csdn.net/qq_29462849/article/details/125326062)[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_2"}}] [.reference_item style="max-width: 50%"]
[ .reference_list ]