feat: E4M3fnuz FP8 format added

[maktukmak]

Currently, the E4M3 fP8 format implemented is ARM-Intel-Nvidia style. However, there is another style, IEEE 754 (torch name is float8_e4m3fnuz), which has different bit configuration and min-max values. This pull request aims to incorporate this style. The unit tests currently pass on CPU because it supports both styles. However, they will fail when tested on other devices. I need guidance on how to design the tests so that they only run a specific style based on the device. Once I have this information, I can complete this PR.

[dacorvo]

@maktukmak thank you for this pull-request.
As a first comment, please amend your commit to make it conventional (feat: e4m3fnuz added). This is a bit tedious but I want people to use meaningful commit messages and the simplest way was to use a predefined CI workflow that enforces conventional commit messages (I may improve this in the future to make it more flexible).
To exclude e4m3fnuz tests for a specific device, you have two options:

• duplicate the tests for these types and use the pytest.mark.skip_device decorator,
• inside each test, start by testing the device and qtype, and skip the test explicitly with pytest.skip.
  You will find examples of both solutions in the existing test files (the first solution is used to skip float8 tests for MPS, and the second for tinygemm if CUDA version is less than 2.1).

[maktukmak]

@dacorvo, I excluded CUDA for e4m3fnuz in tests using the second option, and changed the commit names.

[github-actions]

This PR is stale because it has been open 15 days with no activity. Remove stale label or comment or this will be closed in 5 days.

[maktukmak]

@dacorvo , I fixed the style so it may pass the checks now.

[dacorvo]

Thanks for this pull-request !

[dacorvo]

Rebased and merged as #310