Fix the tutorial example that crashes on GPU, and the PyPI logo

[gomezzz]

Three documentation defects found while validating the 0.6.0 release on real
hardware. Docs only — no library code changes.

• The vectorized-integrand tutorial example crashed on every CUDA machine.
  It built its helper tensors with the legacy torch.Tensor(...) constructor,
  which ignores the torch.set_default_device("cuda") that set_up_backend
  performs, while torch.tensor(...) honours it. The helpers stayed on the CPU,
  the sample points were on the GPU, and the example died in torch.einsum with
  RuntimeError: Expected all tensors to be on the same device. Reproduced on an
  RTX 4060 Ti, then confirmed fixed — the example now prints the True it claims.
  A note was added because this bites in user code exactly as readily.
• The tutorial's import block failed after the documented install. It imports
  matplotlib.pyplot, but matplotlib is only in the dev extra, so a user who
  ran the README's pip install torchquad torch hit ModuleNotFoundError on the
  first code block. Called out as a prerequisite rather than promoted to a runtime
  dependency — 0.6.0 deliberately dropped it as one.
• The README logo did not render on the PyPI project page. It used a
  repository-relative <img src="logos/...">; PyPI cannot resolve relative paths
  in a long description. Now an absolute ?raw=true URL, matching what the seven
  performance figures already do. Pre-existing — 0.5.0's project page has the same
  broken image.

Related to #258.

Numerical behaviour

None. The example's computed values are unchanged; only where its helper tensors
are allocated changed.

Test plan

• Reproduced the GPU crash with the example exactly as published, then ran the
  corrected version on the same machine: all close -> True, with grid,
  result and result_vectorized all on cuda:0 in float64.
• sphinx-build -W -b html docs/source — build succeeded, no warnings.
• Verified the new logo URL returns HTTP 200.

Note for merging

This touches CHANGELOG.md's [Unreleased] section, as does #262, so whichever
lands second will have a small conflict there. The two entries are independent.

What changed after review

• The logo was not the only broken link.environment.yml,
  environment_all_backends.yml (twice), benchmarking/genz_functions.py and
  CONTRIBUTING.md are all repository-relative and 404 on the PyPI project page
  for the same reason. All five are now absolute; fixing only the logo would have
  shipped the same class of bug this PR is named after.
• **:func:set_up_backend was unqualified**, and `tutorial.rst` has no `py:module` context, so Sphinx silently rendered it as plain text — unresolved Python xrefs are not warnings, which is why the `-W` gate stayed green. Now `:func:`torchquad.set_up_backend, and verified to emit a real anchor in the
  built HTML.
• The device note moved into reach. It sat at the end of section 7 while the
  trap is hit from section 1. It now carries a label, and the set_up_backend
  block near the top links to it. Both the label target and the :ref: link were
  confirmed present in the generated HTML.

Not addressed here

Nothing executes the tutorial's code blocks, which is why this survived several
releases, and the failure is GPU-only so no CI job would catch it either. A
two-line version of this example belongs in release_testing/, which already
runs on a GPU box during release validation. Worth its own issue rather than
widening a docs PR.

[github-actions]

Coverage Report

File	Stmts	Miss	Cover	Missing
torchquad
__init__.py	26	2	92%	69–70
torchquad/integration
base_integrator.py	49	7	86%	28, 71–74, 78, 113, 117, 122
boole.py	30	3	90%	54, 73–76
gaussian.py	56	3	95%	97, 121, 143
grid_integrator.py	97	3	97%	50, 164, 268
integration_grid.py	49	4	92%	64–65, 132, 135
monte_carlo.py	93	3	97%	79, 101, 230
qmc.py	20	1	95%	67
rng.py	60	7	88%	71–74, 93–94, 102
simpson.py	29	3	90%	52, 71–74
utils.py	131	11	92%	35–36, 102, 173, 177, 184, 196, 198, 204, 299, 306
vegas.py	167	4	98%	106–107, 212, 359
vegas_map.py	103	8	92%	247–257
vegas_result.py	25	1	96%	22
torchquad/utils
deployment_test.py	192	63	67%	53–55, 65–66, 69–70, 78–79, 91–93, 164, 175–176, 180–181, 183–184, 196–199, 211–213, 221, 232–234, 237, 247, 250, 253, 256–259, 267–275, 280–288, 297–306, 316–323
enable_cuda.py	15	6	60%	17–22
set_log_level.py	18	4	78%	56–57, 72–73
set_precision.py	34	9	74%	35–38, 50–51, 70–72
TOTAL	1290	142	89%

Tests	Skipped	Failures	Errors	Time
193	0 💤	0 ❌	0 🔥	2m 25s ⏱️