[Bugfix][sleepmode][fp8 kv cache]: Fix FP8 KV cache + sleep(level=2) gibberish output (#28783)

Signed-off-by: vensen <vensenmu@gmail.com>
Co-authored-by: TJian <tunjian.tan@embeddedllm.com>

tests/utils.py

@@ -1075,6 +1075,13 @@ def large_gpu_mark(min_gb: int) -> pytest.MarkDecorator:
     )
 
 
+requires_fp8 = pytest.mark.skipif(
+    not current_platform.supports_fp8(),
+    reason="FP8 is not supported on this GPU (requires Hopper or "
+    "Ada architecture, compute capability 8.9+)",
+)
+
+
 def large_gpu_test(*, min_gb: int):
     """
     Decorate a test to be skipped if no GPU is available or it does not have