Use assert <expr> directly.

Signed-off-by: Zheyu Fu <zheyuf@NVIDIA.com>

Author: zheyuf

tests/unittest/_torch/speculative/test_dynamic_spec_decode.py

@@ -107,42 +107,40 @@ def prepare_draft_tokens(self,
     assert drafter.should_use_spec_decode(active_requests,
                                           max_batch_size=8,
                                           max_num_tokens=4096 * 8,
-                                          max_draft_len=4) is True
+                                          max_draft_len=4)
 
     # Small batch size ON case: num_effective_requests = min(12, 5, very_large) = 5 <= 6 → True
     active_requests = [object()] * 12
     assert drafter.should_use_spec_decode(active_requests,
                                           max_batch_size=5,
                                           max_num_tokens=4096 * 8,
-                                          max_draft_len=4) is True
+                                          max_draft_len=4)
 
     # Small token budget ON case: token_cap = 28 // (1+4) = 5 → min(8, 12, 5) = 5 <= 6 → True
     active_requests = [object()] * 12
     assert drafter.should_use_spec_decode(active_requests,
                                           max_batch_size=8,
                                           max_num_tokens=28,
-                                          max_draft_len=4) is True
+                                          max_draft_len=4)
 
     # Generic OFF case: num_effective_requests = min(12, 8, very_large) = 8 > 6 → False
     active_requests = [object()] * 12
-    assert drafter.should_use_spec_decode(active_requests,
-                                          max_batch_size=8,
-                                          max_num_tokens=4096 * 8,
-                                          max_draft_len=4) is False
+    assert not drafter.should_use_spec_decode(active_requests,
+                                              max_batch_size=8,
+                                              max_num_tokens=4096 * 8,
+                                              max_draft_len=4)
 
     # Edge case - None active requests OFF case
     active_requests = []
-    assert drafter.should_use_spec_decode(active_requests,
-                                          max_batch_size=8,
-                                          max_num_tokens=4096 * 8,
-                                          max_draft_len=4) is False
+    assert not drafter.should_use_spec_decode(active_requests,
+                                              max_batch_size=8,
+                                              max_num_tokens=4096 * 8,
+                                              max_draft_len=4)
 
     # Edge case - Token cap equals 0 OFF case: token_cap = 4 // (1+4) = 0 → min(12, 8, 0) = 0 <= 6 → False
     active_requests = [object()] * 12
-    assert drafter.should_use_spec_decode(active_requests,
-                                          max_batch_size=8,
-                                          max_num_tokens=4,
-                                          max_draft_len=4) is False
+    assert not drafter.should_use_spec_decode(
+        active_requests, max_batch_size=8, max_num_tokens=4, max_draft_len=4)
 
 
 if __name__ == "__main__":