Add xxh128 version of hashes for model

[hmaarrfk]

Closes: #283

Computes speed of all hashes in approximately 1.8 seconds vs 6.3.

See https://github.com/Cyan4973/xxHash for more info

[constantinpape]

Thanks @hmaarrfk !
(Gonna merge as soon as tests pass.)

[codecov]

Codecov Report

Attention: 25 lines in your changes are missing coverage. Please review.

Comparison is base (71181d4) 39.72% compared to head (212e948) 40.60%.
Report is 28 commits behind head on dev.

Files	Patch %	Lines
micro_sam/sample_data.py	6.66%	14 Missing ⚠️
micro_sam/util.py	84.44%	7 Missing ⚠️
micro_sam/sam_annotator/_widgets.py	57.14%	3 Missing ⚠️
micro_sam/precompute_state.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##              dev     #284      +/-   ##
==========================================
+ Coverage   39.72%   40.60%   +0.88%     
==========================================
  Files          33       33              
  Lines        4179     4174       -5     
==========================================
+ Hits         1660     1695      +35     
+ Misses       2519     2479      -40     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.