Input dimension for resnet pre trained model

According to the pytorch tutorial documentation, the pre trained model accepts input dimensions of only, [batch_sizex3x224x224]. Is there a way I could input an image of size less than 224x224. I tried to resize the image using scikit-image but since the image is not made of pixel values rather float values, im getting an error.

You can try one of the upsampling algorithms

1 Like