Retry transient errors when downloading the Vale binary

[adamtheturtle]

The first vale invocation downloads the binary with a single urlopen() call. A GitHub Actions job received a transient HTTP 500 from the release asset while another concurrent job downloaded the same asset successfully: https://github.com/adamtheturtle/mypy-strict-kwargs/actions/runs/34820208110/job/103899833567

Please retry transient failures (such as 408, 429, 5xx, and network errors) a small number of times with bounded backoff.

[adamtheturtle]

Another occurrence of this failure mode:

https://github.com/VWS-Python/vws-python-mock/actions/runs/34809058861/job/103866556038?pr=3642

The vale sync hook failed while downloading Vale 3.20.0 with urllib.error.URLError: <urlopen error [Errno 104] Connection reset by peer>. A subsequent vale invocation downloaded the binary successfully, but then failed because synchronization had aborted and the ai-tells style had not been installed.