Use custom isnan for IPU

[callumm-graphcore]

Changelogs

• enumerate the changes of that PR.

---

Checklist:

• Was this PR discussed in an issue? It is recommended to first discuss a new feature into a GitHub issue before opening a PR.
• Add tests to cover the fixed bug(s) or the new introduced feature(s) (if appropriate).
• Update the API documentation is a new function is added, or an existing one is deleted.
• Write concise and explanatory changelogs above.
• If possible, assign one of the following labels to the PR: feature, fix or test (or ask a maintainer to do it for you).

---

discussion related to that PR

[DomInvivo]

Although the name of the class ends with IPU, this class might be used on CPU or GPU, for reproducibility, debugging, or laziness. So I would recommend creating a new function

from graphium.ipu import is_running_on_ipu
def device_is_nan(tensor):
    if is_running_on_ipu():
        return ipu_isnan(tensor)
    else:
        return torch.isnan(tensor)

[DomInvivo]

@callumm-graphcore this PR was also never merged. Should we close it?