Merge branch 'v0.15.1-blockshape' into 'v0.15.1-dev'

fix: resolve block_shape conflicts between DeepEP MoE and non-DeepEP quantization

See merge request dcutoolkit/deeplearing/vllm!507

vllm/model_executor/layers/fused_moe/config.py

@@ -94,6 +94,9 @@ def _quant_flags_to_group_shape(
         # dim should be 1.
         a_shape = GroupShape(row=block_shape[0], col=block_shape[1])
         w_shape = GroupShape(row=block_shape[0], col=block_shape[1])
+    elif block_shape is not None and quant_dtype == torch.int8:
+        a_shape = GroupShape(row=block_shape[0], col=block_shape[1])
+        w_shape = GroupShape(row=block_shape[0], col=block_shape[1])
     else:
         w_shape = None
         a_shape = None if quant_dtype is None else GroupShape.PER_TENSOR
@@ -248,7 +251,6 @@ class FusedMoEQuantConfig:
     def block_shape(self) -> list[int] | None:
         # if self.use_int8_w8a8:
         #     return [256, 256]
-
         if (
             self._a1.shape is not None
             and self._a1.shape != GroupShape.PER_TENSOR
@@ -516,6 +518,7 @@ class FusedMoEQuantConfig:
                 weight_dtype, w_shape, w2_scale, g2_alphas, w2_zp, w2_bias
             ),
         )
+        if quant_dtype != torch.int8:
             assert quant_config.per_act_token_quant == per_act_token_quant
             assert quant_config.per_out_ch_quant == per_out_ch_quant
             assert quant_config.block_shape == block_shape
@@ -560,6 +563,7 @@ def int8_w8a8_moe_quant_config(
     a1_scale: torch.Tensor | None,
     a2_scale: torch.Tensor | None,
     per_act_token_quant: bool = False,
+    block_shape: list[int] | None = None,
 ) -> FusedMoEQuantConfig:
     """
     Construct a quant config for int8 activations and int8 weights.
@@ -572,7 +576,7 @@ def int8_w8a8_moe_quant_config(
         a2_scale=a2_scale,
         per_act_token_quant=per_act_token_quant,
         per_out_ch_quant=False,
-        block_shape=None,
+        block_shape=block_shape,
     )
 
 

vllm/model_executor/layers/quantization/compressed_tensors/compressed_tensors_moe_marlin.py

@@ -311,6 +311,7 @@ class CompressedTensorsW8A8Int8MarlinMoEMethod(CompressedTensorsMarlinMoEMethod)
             a1_scale=layer.w13_input_scale,
             a2_scale=layer.w2_input_scale,
             per_act_token_quant=True,
+            block_shape=[256, 256] if self.use_deepep else None,
         )
         
     def create_weights(self, layer: torch.nn.Module, num_experts: int,