error, "einops.EinopsError: Expected 2 dimensions, got 3" in line "x = self.spatial2temporal(x)"

[wxde]

VTN class , the code in model.py,
"x = self.spatial_transformer.forward_features(x)"
the shape of output x is [batchsize*f, 197, 768]
therefore, the x is 3 dimension.

x = self.spatial2temporal(x)
expect x shape is 2 dimension,

so the dimension doesn't coincident.

i don't know where is the error.

[Paulkie99]

I have the same problem.