Fix bug with tl.clip for the PyTorch and TensorFlow backends

[MarieRoald]

For both PyTorch and TensorFlow, the values for a_max and a_min are computed by TensorLy whenever their values are None. However, the way these values are computed leads to erroneous behaviour. With PyTorch, this occurs when a_min > max(tensor) and a_max=None. I noticed this for PyTorch when I tried to clip a negative tensor to be non-negative. Here is an example:

tl.set_backend("pytorch")
x = tl.tensor([-1, -1, -1])
tl.clip(x, 0, None)

tensor([-1., -1., -1.])

For TensorFlow, the error occurs when a_min=None and a_max<min(tensor). Here is an example:

tl.set_backend("tensorflow")
x = tl.tensor([1, 1, 1])
tl.clip(x, None, 0)

<tf.Tensor: shape=(3,), dtype=float32, numpy=array([1., 1., 1.], dtype=float32)>

[codecov]

Codecov Report

Merging #355 (c4e2615) into main (87b435b) will increase coverage by 0.03%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##             main     #355      +/-   ##
==========================================
+ Coverage   88.16%   88.19%   +0.03%     
==========================================
  Files         105      105              
  Lines        6058     6058              
==========================================
+ Hits         5341     5343       +2     
+ Misses        717      715       -2     

Impacted Files	Coverage Δ
tensorly/decomposition/_cp.py	87.59% <0.00%> (-0.37%)	⬇️
tensorly/tenalg/proximal.py	67.38% <0.00%> (+0.23%)	⬆️
tensorly/tenalg/tests/test_proximal.py	100.00% <0.00%> (+0.70%)	⬆️
tensorly/decomposition/_cmtf_als.py	92.59% <0.00%> (+3.70%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 87b435b...c4e2615. Read the comment docs.

[JeanKossaifi]

Thanks @MarieRoald, I always love your PRs: clear problem, clear fix and matching unit-tests! 🌠
Merging.