[CUDA] Switch to `at::empty_like` in `adaptive_avg_pool3d_backward_cuda` (#100202)
Same as #100138, `gradInput` is already zero'd out. Also clean up includes after #100138.
CC @ngimel @ptrblck
Pull Request resolved: https://github.com/pytorch/pytorch/pull/100202
Approved by: https://github.com/ngimel
