Speed optimizations

[ajinkya-kulkarni]

Using slicing and indexing to extract the L, a, and b channels from the input array, and perform the necessary operations using broadcasting. This allows eliminating the need for explicit loops, improving the performance of the function.

[carloalbertobarbano]

Hi, thanks for your contribution. I updated the workflows in the main branch (they were using a deprecated version of ubuntu). Can you update your branches with the correct workflows so that I can run the unit tests?

[andreped]

@carloalbertobarbano this PR now runs CIs with ubuntu-20.04. Can you approve the workflow to be run, such that we can test that everything is in order? If yes, we can add the same fix to the two remaining PRs @ajinkya-kulkarni made.

[andreped]

I had forgotten to use ubuntu-20.04 for a pytorch test. However, I have done that now. Also applied the same fixes to all three PRs @ajinkya-kulkarni made. However, just waiting for my PRs to his fork to be merged. If so, then I believe tests could be ran.

[ajinkya-kulkarni]

I have synced my fork! Sorry for the delay.

[andreped]

I have synced my fork! Sorry for the delay.

@ajinkya-kulkarni No worries. You have to sync the individual branches as well, not just the main branch. Thats what my PRs to your fork do.

[carloalbertobarbano]

All checks passed. Merging into development