fix: fix the failed sampling unittest on 5090 (#1886)

<!-- .github/pull_request_template.md -->

## 📌 Description

Applying softmax followed by top_k_renorm does not guarantee
bitwise-identical results compared to top_k_mask followed by softmax.
This may cause slight differences in subsequent top-p sampling.

In this PR we relax the condition to up to a 1% mismatch rate.

## 🔍 Related Issues

<!-- Link any related issues here -->

## 🚀 Pull Request Checklist

Thank you for contributing to FlashInfer! Before we review your pull
request, please make sure the following items are complete.

### ✅ Pre-commit Checks

- [x] I have installed `pre-commit` by running `pip install pre-commit`
(or used your preferred method).
- [x] I have installed the hooks with `pre-commit install`.
- [x] I have run the hooks manually with `pre-commit run --all-files`
and fixed any reported issues.

> If you are unsure about how to set up `pre-commit`, see [the
pre-commit documentation](https://pre-commit.com/).

## 🧪 Tests

- [x] Tests have been added or updated as needed.
- [ ] All tests are passing (`unittest`, etc.).

## Reviewer Notes

<!-- Optional: anything you'd like reviewers to focus on, concerns, etc.
-->

Author: yzh119

tests/utils/test_sampling.py

@@ -377,7 +377,18 @@ def test_top_k_top_p_sampling_from_probs_logits_alignment(batch_size, vocab_size
         filter_apply_order="top_k_first",
         generator=generator_probs,
     )
-    assert torch.all(samples == samples_ref)
+
+    num_matches = (samples == samples_ref).sum().item()
+    match_rate = num_matches / samples.numel()
+
+    # NOTE(Zihao): Applying softmax followed by top_k_renorm (softmax -> top_k_renorm)
+    # does not guarantee bitwise-identical results compared to top_k_mask followed by softmax (top_k_mask -> softmax).
+    # This may cause slight differences in subsequent top-p sampling.
+    # We tolerate up to a 1% mismatch rate.
+    assert match_rate >= 0.99, (
+        f"Sample match rate {match_rate:.2%} is below threshold "
+        f"({batch_size - num_matches}/{batch_size} mismatches, expected <=1%)"
+    )
 
 
 @pytest.mark.parametrize("batch_size", [1, 99, 989])