pytorch-3d-point-cloud-generation icon indicating copy to clipboard operation
pytorch-3d-point-cloud-generation copied to clipboard

Published 20 hours ago • verified publisher icon lkhphuc

Question about model

Open marcvoegele-fhnw opened this issue 5 years ago • 0 comments

After conv4 (encoder) we have an output shape of XY with 256 channels. But the input in fc1 is 4096, shouldn't the input be XY256 or can this model only work with input images of size 6464 (that's the one case when the input is 4096)

marcvoegele-fhnw avatar Mar 30 '20 02:03 marcvoegele-fhnw