F.conv2d,dilation=dilation
时间: 2024-04-28 22:24:54 浏览: 79
当使用F.conv2d函数进行卷积操作时,参数dilation控制着卷积核的空洞大小。根据引用中提到的,默认情况下,dilation的值为1。根据引用中的解释,当dilation=1时,卷积操作不会跳过任何输入像素。因此,对应的是图1,而不是图2。<span class="em">1</span><span class="em">2</span><span class="em">3</span>
#### 引用[.reference_title]
- *1* [Pytorch nn.conv2d 的 参数 dilation =1 究竟是?](https://blog.csdn.net/OrdinaryMatthew/article/details/121249881)[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* [torch.conv2d stride,dilation,groups](https://blog.csdn.net/wuhao1205/article/details/124677878)[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* [Pytorch.nn.conv2d 过程验证方式(单,多通道卷积过程)](https://download.csdn.net/download/weixin_38590456/13760595)[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