Feature: Allow Noise2Noise with different in and out channels

[jdeschamps]

Description

Until now, Noise2Noise was only possible with the same number of input and output channels. But one can envision have multiple channels in (e.g. time frames) and supervising with a single frame.

This PR relaxes the constraints on Noise2Noise: channels in and out can now be different, and UNet is not mandatory!

• What: Remove algorithm_model constraint on N2N, and change the configuration convenience function.
• Why: We might try to run Noise2Noise with different in and out channels.
• How: See what.

Changes Made

• Modified: algorithm_model and create_n2n_configuration, and the corresponding tests.

---

Please ensure your PR meets the following requirements:

• Code builds and passes tests locally, including doctests
• New tests have been added (for bug fixes/features)
• Pre-commit passes
• PR to the documentation exists (for bug fixes / features)
  • Add changes from #152 careamics.github.io#9
  • Add changes from #152 careamics-examples#3

[jdeschamps]

Requires #149 to be merged for tests to pass.