Fix non-reproducible step sizes

[devmotion]

Might be sufficient to fix #1923.

[torfjelde]

Sorry, didn't see this before making my own 😅

Should we maybe add a short test? E.g.

    @turing_testset "reproducibility" begin
        alg = NUTS(1000, 0.8)
        res1 = sample(StableRNG(123), gdemo_default, alg, 1000)
        res2 = sample(StableRNG(123), gdemo_default, alg, 1000)
        @test Array(res1) == Array(res2)
    end

to test/inference/hmc.jl?

[torfjelde]

LGTM but a test would be nice:)

[coveralls]

Pull Request Test Coverage Report for Build 3770976411

• 0 of 0 changed or added relevant lines in 0 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage remained the same at 0.0%

---

Totals
Change from base Build 3663908235:	0.0%
Covered Lines:	0
Relevant Lines:	1432

---

💛 - Coveralls

[codecov]

Codecov Report

Base: 0.00% // Head: 0.00% // No change to project coverage 👍

Coverage data is based on head (9e6494f) compared to base (61b06f6).
Patch coverage: 0.00% of modified lines in pull request are covered.

Additional details and impacted files

@@          Coverage Diff           @@
##           master   #1924   +/-   ##
======================================
  Coverage    0.00%   0.00%           
======================================
  Files          21      21           
  Lines        1432    1432           
======================================
  Misses       1432    1432           

Impacted Files	Coverage Δ
src/inference/hmc.jl	0.00% <0.00%> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[yebai]

thanks @devmotion!