[v0.18.0][Bugfix]Fix Error "AttributeError: 'AscendCompressedTensorsConfig' obiect has no attribute 'enabling_fa_quant'" (#7748)

### What this PR does / why we need it?
cherry-pick from https://github.com/vllm-project/vllm-ascend/pull/7736

**Error information**
When the quantized weights in CompressedTensors format of the kimi-k2
model are used, the following error is reported:
`AttributeError: 'AscendCompressedTensorsConfig' obiect has no attribute
'enabling_fa_quant'`

**Error Cause**
Currently, FA3 quantization supports only the weights of modelslim
quantization. The added methods are not defined in
AscendCompressedTensorsConfig.

**Solution**
Before invoking related methods, check whether the FA3 feature is
enabled.
Additionally, the unused `get_scaled_act_names` method and its
corresponding unit test have been removed.

### Does this PR introduce _any_ user-facing change?
No.

### How was this patch tested?
Existing unit tests were updated by removing a deprecated test case, and
the refactored logic was reviewed for correctness.

Signed-off-by: Wang Kunpeng <1289706727@qq.com>

tests/ut/quantization/test_modelslim_config.py

@@ -253,6 +253,3 @@ class TestAscendModelSlimConfig(TestBase):
         self.assertIn("model.layers.0.weight", config.quant_description)
         self.assertEqual(config.quant_description["model.layers.0.weight"],
                          "INT8")
-
-    def test_get_scaled_act_names(self):
-        self.assertEqual(self.ascend_config.get_scaled_act_names(), [])

vllm_ascend/attention/mla_v1.py

@@ -49,6 +49,7 @@ from vllm_ascend.ops.layer_shard_linear import (
 )
 from vllm_ascend.ops.rotary_embedding import get_cos_and_sin_mla
 from vllm_ascend.quantization.methods.w8a8_static import AscendW8A8LinearMethod
+from vllm_ascend.quantization.utils import enable_fa_quant
 from vllm_ascend.utils import ACL_FORMAT_FRACTAL_ND, get_weight_prefetch_method, maybe_trans_nz, weak_ref_tensors
 from vllm_ascend.worker.npu_input_batch import NPUInputBatch
 
@@ -730,10 +731,7 @@ class AscendMLAImpl(MLAAttentionImpl):
             self.vllm_config.kv_transfer_config is not None and self.vllm_config.kv_transfer_config.is_kv_producer
         )
         self.layer_name = kwargs.get("layer_name")
-        quant_config = self.vllm_config.quant_config
+        self.fa_quant_layer = enable_fa_quant(self.vllm_config, self.layer_name)
-        self.fa_quant_layer = (
-            quant_config.enabling_fa_quant(self.vllm_config, self.layer_name) if quant_config is not None else False
-        )
         self.dtype = torch.int8 if self.fa_quant_layer else self.vllm_config.model_config.dtype
         self.layer_sharding_kwargs = []
         for layer_name in get_ascend_config().layer_sharding or []:

vllm_ascend/quantization/modelslim_config.py

@@ -660,9 +660,6 @@ class AscendModelSlimConfig(QuantizationConfig):
                 extra_quant_dict[new_k] = self.quant_description[k]
         self.quant_description.update(extra_quant_dict)
 
-    def get_scaled_act_names(self) -> list[str]:
-        return []
-
     def _add_kvcache_quant_metadata(self):
         fa_quant_type = self.quant_description.get("fa_quant_type", "")
         self.enable_fa_quant = fa_quant_type != ""

vllm_ascend/quantization/utils.py

@@ -197,3 +197,12 @@ def maybe_auto_detect_quantization(vllm_config) -> None:
     from vllm.config import VllmConfig as _VllmConfig
 
     vllm_config.quant_config = _VllmConfig._get_quantization_config(model_config, vllm_config.load_config)
+
+
+def enable_fa_quant(vllm_config, layer_name=None) -> bool:
+    if vllm_config.quant_config is not None and getattr(vllm_config.quant_config, "enable_fa_quant", False):
+        if layer_name is not None:
+            return vllm_config.quant_config.enabling_fa_quant(vllm_config, layer_name)
+        else:
+            return True
+    return False

vllm_ascend/worker/model_runner_v1.py

@@ -109,6 +109,7 @@ from vllm_ascend.eplb.utils import model_register
 from vllm_ascend.ops.rotary_embedding import set_cos_and_sin, update_cos_sin
 from vllm_ascend.patch.worker.patch_draft_quarot import patch_load_weights
 from vllm_ascend.patch.worker.patch_module import patch_torch_npu_argsort
+from vllm_ascend.quantization.utils import enable_fa_quant
 from vllm_ascend.sample.sampler import AscendSampler
 from vllm_ascend.spec_decode import get_spec_decode_method
 from vllm_ascend.spec_decode.draft_proposer import AscendDraftModelProposer
@@ -2763,7 +2764,7 @@ class NPUModelRunner(GPUModelRunner):
                             k_dim,
                             v_dim,
                         ]
-                        if self.is_kv_consumer and self.vllm_config.quant_config is not None:
+                        if self.is_kv_consumer and enable_fa_quant(self.vllm_config):
                             k_tensor_split_factor, v_tensor_split_factor = (
                                 self.vllm_config.quant_config.get_kv_quant_split_factor(layer_name, kv_head_dim_list)
                             )
@@ -2950,7 +2951,7 @@ class NPUModelRunner(GPUModelRunner):
                             v_dim,
                         )
                     k_cache_dtype = v_cache_dtype = current_kv_cache_spec.dtype
-                    if self.is_kv_consumer and self.vllm_config.quant_config is not None:
+                    if self.is_kv_consumer and enable_fa_quant(self.vllm_config):
                         k_cache_dtype, v_cache_dtype = self.vllm_config.quant_config.get_kv_quant_dtype(
                             layer_name, current_kv_cache_spec.dtype, self.model_config
                         )