Init mode == "random" does not return the correct shape in initialize_tucker

[hello-fri-end]

Modes are not considered when init is random:

tensorly/tensorly/decomposition/_tucker.py

Lines 78 to 86 in a982cf9

	elif init == "random":
	rng = tl.check_random_state(random_state)
	core = tl.tensor(
	rng.random_sample(rank) + 0.01, **tl.context(tensor)
	) # Check this
	factors = [
	tl.tensor(rng.random_sample(s), **tl.context(tensor))
	for s in zip(tl.shape(tensor), rank)
	]

When init is random, do we always assume that all modes should be considered? init == svd allows to initialize along specific modes.

Morever, in the following test init is svd while it should be random
https://github.com/tensorly/tensorly/blob/main/tensorly/decomposition/tests/test_tucker.py#L59-L64

[JeanKossaifi]

You are absolutely right, thanks for highlighting @hello-fri-end! Would be great if you could open a small PR to fix it!

[hello-fri-end]

You are absolutely right, thanks for highlighting @hello-fri-end! Would be great if you could open a small PR to fix it!

Sure, created a PR here to fix the same: #519

[hello-fri-end]

Fixed in #519