restored comment

torch_harmonics/_disco_convolution.py

@@ -139,6 +139,12 @@ def _disco_s2_transpose_contraction_cuda(x: torch.Tensor, roff_idx: torch.Tensor
 
 
 def _disco_s2_contraction_torch(x: torch.Tensor, psi: torch.Tensor, nlon_out: int):
+    """
+    Reference implementation of the custom contraction as described in [1]. This requires repeated
+    shifting of the input tensor, which can potentially be costly. For an efficient implementation
+    on GPU, make sure to use the custom kernel written in CUDA.
+    """
+    
     assert len(psi.shape) == 3
     assert len(x.shape) == 4
     psi = psi.to(x.device)