VLLM_USE_TRITON_FLASH_ATTN V0 variable deprecation (#27611)

Signed-off-by: Andreas Karatzas <akaratza@amd.com>
Signed-off-by: Andreas Karatzas <Andreas.Karatzas@amd.com>

tests/quantization/test_quark.py

@@ -8,7 +8,6 @@ See also `tests/kernels/moe/test_ocp_mx_moe.py`.
 """
 
 import importlib.metadata
-import os
 from dataclasses import dataclass
 from importlib.util import find_spec
 
@@ -246,8 +245,6 @@ def test_mxfp4_gsm8k_correctness(config: AccuracyTestConfig):
     task = "gsm8k"
     rtol = 0.03
 
-    os.environ["VLLM_USE_TRITON_FLASH_ATTN"] = "0"
-
     results = lm_eval.simple_evaluate(
         model="vllm",
         model_args=config.get_model_args(tp_size=8, model_max_len=38768),
@@ -263,8 +260,6 @@ def test_mxfp4_gsm8k_correctness(config: AccuracyTestConfig):
         and measured_value + rtol > EXPECTED_VALUE
     ), f"Expected: {EXPECTED_VALUE} |  Measured: {measured_value}"
 
-    del os.environ["VLLM_USE_TRITON_FLASH_ATTN"]
-
 
 @pytest.mark.skipif(not QUARK_MXFP4_AVAILABLE, reason="amd-quark>=0.9 is not available")
 @pytest.mark.parametrize("float_dtype", [torch.bfloat16, torch.float16])