resnet18

paddle.vision.models. resnet18 ( pretrained=False, **kwargs ) [源代码]

18层的resnet模型,来自论文 “Deep Residual Learning for Image Recognition”

参数

  • pretrained (bool,可选) - 是否加载在imagenet数据集上的预训练权重。默认值:False。

返回

resnet18模型,Layer的实例。

代码示例

  1. import paddle
  2. from paddle.vision.models import resnet18
  3. # build model
  4. model = resnet18()
  5. # build model and load imagenet pretrained weight
  6. # model = resnet18(pretrained=True)
  7. x = paddle.rand([1, 3, 224, 224])
  8. out = model(x)
  9. print(out.shape)

使用本API的教程文档