Given transposed=1, weight of size [3, 1, 2, 2], expected input[8, 1024, 8, 10] to have 3 channels, but got 1024 channels instead

vision
1

I am merging the two by referring to the fast-depth and sparse-to-dense of girhub. (I want to use sparse-to-dense on tx2)
However, if insert fast-depth’s model.py into sparse-to-dense’s model.py and run it, the following error occurs:
Is there a way to reduce the channel?

image
image917×623 54 KB

2

I’ve got a solution, thank you everyone!!