Add better vectorizations of jacobian and hessian in the autodiff backends

[ninamiolane]

This PR will unify the shapes of jacobians and hessians computed across backends, to enhance the readability while using these functions in the pullback metric file.

Checklist

• My pull request has a clear and explanatory title.
• If neccessary, my code is vectorized.
• I have added apropriate unit tests.
• I have made sure the code passes all unit tests. (refer to comment below)
• My PR follows PEP8 guidelines. (refer to comment below)
• My PR follows geomstats coding style and API.
• My code is properly documented and I made sure the documentation renders properly. (Link)
• I have linked to issues and PRs that are relevant to this PR.

Description

Issue

Additional context

[codecov]

Codecov Report

Merging #1724 (f6765b0) into master (2598d50) will decrease coverage by 2.91%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master    #1724      +/-   ##
==========================================
- Coverage   90.12%   87.21%   -2.90%     
==========================================
  Files         121      115       -6     
  Lines       12060    11390     -670     
==========================================
- Hits        10868     9933     -935     
- Misses       1192     1457     +265     

Flag	Coverage Δ
numpy	87.21% <ø> (+0.05%)	⬆️
pytorch	?

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
geomstats/geometry/pullback_metric.py	26.52% <0.00%> (-64.08%)	⬇️
geomstats/geometry/sub_riemannian_metric.py	43.40% <0.00%> (-50.94%)	⬇️
geomstats/learning/geodesic_regression.py	31.17% <0.00%> (-48.70%)	⬇️
...eomstats/information_geometry/fisher_rao_metric.py	23.34% <0.00%> (-46.66%)	⬇️
geomstats/geometry/fiber_bundle.py	66.67% <0.00%> (-28.20%)	⬇️
geomstats/geometry/connection.py	75.49% <0.00%> (-17.41%)	⬇️
geomstats/geometry/riemannian_metric.py	78.27% <0.00%> (-6.52%)	⬇️
geomstats/information_geometry/gamma.py	53.34% <0.00%> (-4.12%)	⬇️
geomstats/learning/geometric_median.py	92.31% <0.00%> (-2.56%)	⬇️
geomstats/information_geometry/normal.py	91.53% <0.00%> (-1.69%)	⬇️
... and 8 more

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