[Bugfix/CI] Turn test_compressed_tensors_2of4_sparse back on (vllm-project#13250)

Signed-off-by: Louis Ulmer <ulmerlouis@gmail.com>

Author: tlrmchlsmth

tests/quantization/test_compressed_tensors.py

@@ -495,7 +495,6 @@ def check_model(model):
         assert output
 
 
-@pytest.mark.skip(reason="2of4 sparse w16a16 CUTLASS produces bad output.")
 @pytest.mark.skipif(
     not sparse_cutlass_supported(),
     reason="2of4 Sparse is not yet supported on this GPU type.",