`sinkhorn_lpl1_mm` performs unnecessary computations
kachayev opened this issue · 0 comments
kachayev commented
Describe the bug
It appears that the part of the computation within the sinkhorn_lpl1_mm function may be redundant, as its result is not included in the function's return.
Expected behavior
The code could either be safely removed or the resulting matrix W should be included in the log, if deemed necessary. What do you think?