[quant] Fixing the hypothesis test for topk (#66057)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/66057
The current test is creating the sets that are too slow.
This will cause either "Filtering too much" or "Timeout" errors in the future versions of hypothesis.
This PR preemptively fixes the issue.
Test Plan: `python test/test_quantization.py`
Reviewed By: vkuzo
Differential Revision: D31366065
Pulled By: z-a-f
fbshipit-source-id: deaab4da8ee02a5dee8943cabdd30fc53d894a34