Hanna ChemBERT fix

[se-schmitt]

Deviations between the original HANNA model (python) and ClapeyronHANNA originated from a slightly different activation function in Transformers.jl / NNlib.jl. This was fixed in FluxML/NNlib.jl/pull/629 and chengchingwen/Transformers.jl/pull/209 (see below). This PR updates the model and the tests. Closes #346

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 1 line in your changes missing coverage. Please review.

Project coverage is 89.79%. Comparing base (3c65bee) to head (899eba9).
Report is 81 commits behind head on master.

Files with missing lines	Patch %	Lines
lib/ClapeyronHANNA/src/ClapeyronHANNA.jl	0.00%	1 Missing ⚠️

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #347   +/-   ##
=======================================
  Coverage   89.79%   89.79%           
=======================================
  Files           2        2           
  Lines          98       98           
=======================================
  Hits           88       88           
  Misses         10       10           

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

🚀 New features to boost your workflow:

• ❄ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.