[ROCm][CI] Fix test_token_classification.py::test_bert_models (#31993)

Signed-off-by: Divakar Verma <divakar.verma@amd.com>

tests/models/language/pooling/test_token_classification.py

@@ -5,6 +5,7 @@ import torch
 from transformers import AutoModelForTokenClassification
 
 from tests.models.utils import softmax
+from vllm.platforms import current_platform
 
 
 @pytest.mark.parametrize("model", ["boltuix/NeuroBERT-NER"])
@@ -21,8 +22,17 @@ def test_bert_models(
     with vllm_runner(model, max_model_len=None, dtype=dtype) as vllm_model:
         vllm_outputs = vllm_model.token_classify(example_prompts)
 
+    # Use eager attention on ROCm to avoid HF Transformers flash attention
+    # accuracy issues: https://github.com/vllm-project/vllm/issues/30167
+    hf_model_kwargs = {}
+    if current_platform.is_rocm():
+        hf_model_kwargs["attn_implementation"] = "eager"
+
     with hf_runner(
-        model, dtype=dtype, auto_cls=AutoModelForTokenClassification
+        model,
+        dtype=dtype,
+        auto_cls=AutoModelForTokenClassification,
+        model_kwargs=hf_model_kwargs,
     ) as hf_model:
         tokenizer = hf_model.tokenizer
         hf_outputs = []
@@ -36,7 +46,7 @@ def test_bert_models(
     for hf_output, vllm_output in zip(hf_outputs, vllm_outputs):
         hf_output = hf_output.detach().clone().cpu().float()
         vllm_output = vllm_output.detach().clone().cpu().float()
-        assert torch.allclose(hf_output, vllm_output, 1e-2)
+        torch.testing.assert_close(hf_output, vllm_output, atol=1.2e-2, rtol=1e-3)
 
 
 @pytest.mark.parametrize("model", ["disham993/electrical-ner-ModernBERT-base"])
@@ -49,8 +59,6 @@ def test_modernbert_models(
     model: str,
     dtype: str,
 ) -> None:
-    from vllm.platforms import current_platform
-
     with vllm_runner(model, max_model_len=None, dtype=dtype) as vllm_model:
         vllm_outputs = vllm_model.token_classify(example_prompts)