fix(qwen3_moe): correct return type annotation on Qwen3MoeSparseMoeBlock.forward (#45352)

* fix(qwen3_moe): correct return type annotation on Qwen3MoeSparseMoeBlock.forward

* fix: propagate Qwen3MoeSparseMoeBlock forward return type fix to generated vl_moe and omni_moe files

Built by Rudrendu Paul, developed with Claude Code

---------

Co-authored-by: Rudrendu <RudrenduPaul@users.noreply.github.com>

Author: RudrenduPaul

src/transformers/models/qwen3_moe/modeling_qwen3_moe.py

@@ -278,7 +278,7 @@ def __init__(self, config: Qwen3MoeConfig):
         self.experts = Qwen3MoeExperts(config)
         self.gate = Qwen3MoeTopKRouter(config)
 
-    def forward(self, hidden_states: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         batch_size, sequence_length, hidden_dim = hidden_states.shape
         hidden_states_reshaped = hidden_states.view(-1, hidden_dim)
         _, routing_weights, selected_experts = self.gate(hidden_states_reshaped)

src/transformers/models/qwen3_moe/modular_qwen3_moe.py

@@ -66,7 +66,7 @@ def __init__(self, config: Qwen3MoeConfig):
         self.experts = Qwen3MoeExperts(config)
         self.gate = Qwen3MoeTopKRouter(config)
 
-    def forward(self, hidden_states: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         batch_size, sequence_length, hidden_dim = hidden_states.shape
         hidden_states_reshaped = hidden_states.view(-1, hidden_dim)
         _, routing_weights, selected_experts = self.gate(hidden_states_reshaped)

src/transformers/models/qwen3_omni_moe/modeling_qwen3_omni_moe.py

@@ -1415,7 +1415,7 @@ def __init__(self, config: Qwen3OmniMoeThinkerConfig):
         self.experts = Qwen3OmniMoeThinkerTextExperts(config)
         self.gate = Qwen3OmniMoeThinkerTextTopKRouter(config)
 
-    def forward(self, hidden_states: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         batch_size, sequence_length, hidden_dim = hidden_states.shape
         hidden_states_reshaped = hidden_states.view(-1, hidden_dim)
         _, routing_weights, selected_experts = self.gate(hidden_states_reshaped)

src/transformers/models/qwen3_vl_moe/modeling_qwen3_vl_moe.py

@@ -136,7 +136,7 @@ def __init__(self, config: Qwen3VLMoeTextConfig):
         self.experts = Qwen3VLMoeTextExperts(config)
         self.gate = Qwen3VLMoeTextTopKRouter(config)
 
-    def forward(self, hidden_states: torch.Tensor) -> tuple[torch.Tensor, torch.Tensor]:
+    def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         batch_size, sequence_length, hidden_dim = hidden_states.shape
         hidden_states_reshaped = hidden_states.view(-1, hidden_dim)
         _, routing_weights, selected_experts = self.gate(hidden_states_reshaped)