Replace new_tensor

[marvinfriede]

Create tensors from floats with

torch.tensor(<float>, device=<tensor>.device, dtype=<tensor>.dtype) 

instead of using

<tensor>.new_tensor(<float>) 

The latter has side effects (copies additional data beside dtype and device) that become apparent in more advanced AD applications.

[codecov]

Codecov Report

Merging #17 (fb0a3c6) into main (588fa74) will increase coverage by 0.06%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main      #17      +/-   ##
==========================================
+ Coverage   96.09%   96.16%   +0.06%     
==========================================
  Files          14       14              
  Lines         282      287       +5     
  Branches       36       36              
==========================================
+ Hits          271      276       +5     
  Misses          2        2              
  Partials        9        9              

Impacted Files	Coverage Δ
src/tad_dftd3/model.py	100.00% <ø> (ø)
src/tad_dftd3/damping/atm.py	100.00% <100.00%> (ø)
src/tad_dftd3/damping/rational.py	100.00% <100.00%> (ø)
src/tad_dftd3/disp.py	97.56% <100.00%> (+0.19%)	⬆️
src/tad_dftd3/ncoord.py	95.45% <100.00%> (ø)

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more