Input shape of pretrained 3D resnets

I’m using pretrained pytorch video classification model, as,
https://pytorch.org/vision/stable/models.html#video-classification

I loaded r3d_18 model, but I can’t found the dimension of inputs.
It requires 5 inputs so I tried [batch X frames X filters(R,G,B) X W X H] but not worked… And the docs do not shows the input shape.

Anybody knows the input shape of r3d_18?(Resnet 3d - 18 layers)

I was having the same issue. The error shows that its [batch x channels x frames x w x h].