Add support for qwen3_vl and qwen3_vl_moe

Signed-off-by: booker123456 <945658361@qq.com>

Author: booker123456

vllm_ascend/models/qwen3_vl.py

@@ -40,7 +40,7 @@
 
 
 class AscendQwen3_VisionPatchEmbed(Qwen3_VisionPatchEmbed):
-    
+
     def forward(self, x: torch.Tensor) -> torch.Tensor:
         x = x.matmul(
             self.proj.weight.data.view(self.hidden_size, -1).transpose(0, 1))
@@ -71,14 +71,16 @@ def __init__(
             use_data_parallel=use_data_parallel)
 
     def forward(self, x: torch.Tensor, cu_seqlens: torch.Tensor,
-            cos: torch.Tensor, sin: torch.Tensor) -> torch.Tensor:
+                cos: torch.Tensor, sin: torch.Tensor) -> torch.Tensor:
         x = x + self.attn(
             self.norm1(x), cu_seqlens=cu_seqlens, cos=cos, sin=sin)
 
         x = x + self.mlp(self.norm2(x))
         return x
 
+
 class AscendQwen3_VisionTransformer(Qwen3_VisionTransformer):
+
     def __init__(
         self,
         vision_config: Qwen3VLVisionConfig,
@@ -198,4 +200,4 @@ def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
             norm_eps=getattr(config, "rms_norm_eps", 1e-6),
             quant_config=self._maybe_ignore_quant_config(quant_config),
             prefix=maybe_prefix(prefix, "visual"),
-            use_data_parallel=self.use_data_parallel)
+            use_data_parallel=self.use_data_parallel)

vllm_ascend/models/qwen3_vl_moe.py

@@ -40,7 +40,7 @@
                                                   Qwen3MoeSparseMoeBlock)
 from vllm.model_executor.models.qwen3_vl_moe import (
     Qwen3MoeLLMForCausalLM, Qwen3MoeLLMModel, Qwen3VLDummyInputsBuilder,
-    Qwen3VLMoeForConditionalGeneration, Qwen3VLMoeProcessingInfo, 
+    Qwen3VLMoeForConditionalGeneration, Qwen3VLMoeProcessingInfo,
     Qwen3VLMultiModalProcessor)
 from vllm.model_executor.models.utils import (
     PPMissingLayer, WeightsMapper, make_empty_intermediate_tensors_factory,
@@ -62,6 +62,7 @@
         "deepstack_input_embeds": 0
     })
 class AscendQwen3MoeLLMModel(Qwen3MoeLLMModel):
+
     def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
         nn.Module.__init__(self)
         config = vllm_config.model_config.hf_config.get_text_config()