Remove reference_video_tensor() fixture

[NicolasHug]

This PR removes reference_video_tensor() because its unnecessary.

I assume the use of this fixture was used to cache the corresponding returned value. This is unnecessary for 2 reasons:

• The fixture wasn't even cached anyway, because its scope wasn't set to "session". You can verify that by printing something within the fixture, and running tests with pytest -s. You'll see the print statement being executed each time, proving that no caching was involved.
• Getting the return value is dead cheap and doesn't need to be cached in the first place. Proof:

[ins] In [1]: import test_utils
tes
[nav] In [2]: %timeit test_utils.NASA_VIDEO.to_tensor()
55.8 µs ± 3.76 µs per loop (mean ± std. dev. of 7 runs, 10,000 loops each)

[facebook-github-bot]

@NicolasHug has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@NicolasHug has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@NicolasHug has imported this pull request. If you are a Meta employee, you can view this diff on Phabricator.

[facebook-github-bot]

@NicolasHug merged this pull request in 508eee1.