re-add "unnecessary" normalization to fix UTF-8 errors.

[UsernamesLame]

I forgot to remove this line yesterday, I'm not sure why it's not causing any issues having it there referencing a nonexistent variable samples but it seems like normalization wasn't required based on transcription testing I've done so far.

[abdeladim-s]

Oh, I didn't notice yesterday, and I didn't run any tests.
Next time please run the tests in the test directory before committing!

As for this, I think normalization is important, without it sometimes I get utf-8 decode problems.
Please fix the code to arr /= np.iinfo(np.int16).max and commit it so I cam merge the fix asap.

Thanks!

[UsernamesLame]

Sorry about that. Slipped my mind to do the tests. Thankfully it didn't seem to actually break anything. That's the craziest part.

[UsernamesLame]

@abdeladim-s I re-added the normalization. I haven't had a chance to run the test suite. For now I'm relying on the CI pipeline to catch anything.

I re-added arr /= np.iinfo(np.int16).max

[abdeladim-s]

It's okey, no worries!
I, indeed, need to include the tests in CI, for now I am just running them locally 😄

I'll merge this for now.
Thanks a lot!