Fix grammatical error in MoE variable name: expert_hitted → expert_hit, hitted_experts → hit_experts (#39959)

* Fix grammatical error: expert_hitted -> expert_hit in MoE implementations

* Fix grammatical error: hitted_experts -> hit_experts in MoE implementation

Author: Mihonarium

src/transformers/integrations/mxfp4.py

@@ -264,8 +264,8 @@ def topk(vals, k):
 
         expt_data = compute_expt_data_torch(hist, n_local_experts, n_gates_pad)
 
-        hitted_experts = n_expts_act
-    return RoutingData(gate_scal, hist, n_local_experts, hitted_experts, expt_data), gather_indx, scatter_indx
+        hit_experts = n_expts_act
+    return RoutingData(gate_scal, hist, n_local_experts, hit_experts, expt_data), gather_indx, scatter_indx
 
 
 def mlp_forward(self, hidden_states):

src/transformers/models/ernie4_5_moe/modeling_ernie4_5_moe.py

@@ -356,8 +356,8 @@ def forward(
         expert_mask = torch.nn.functional.one_hot(selected_experts, num_classes=self.num_experts).permute(2, 1, 0)
 
         # Loop over all available experts in the model and perform the computation on each expert
-        expert_hitted = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
-        for expert_idx in expert_hitted:
+        expert_hit = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
+        for expert_idx in expert_hit:
             expert_layer = self.experts[expert_idx]
             idx, top_x = torch.where(expert_mask[expert_idx].squeeze(0))
 

src/transformers/models/ernie4_5_moe/modular_ernie4_5_moe.py

@@ -167,8 +167,8 @@ def forward(
         expert_mask = torch.nn.functional.one_hot(selected_experts, num_classes=self.num_experts).permute(2, 1, 0)
 
         # Loop over all available experts in the model and perform the computation on each expert
-        expert_hitted = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
-        for expert_idx in expert_hitted:
+        expert_hit = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
+        for expert_idx in expert_hit:
             expert_layer = self.experts[expert_idx]
             idx, top_x = torch.where(expert_mask[expert_idx].squeeze(0))
 

src/transformers/models/gpt_oss/modeling_gpt_oss.py

@@ -97,8 +97,8 @@ def forward(self, hidden_states: torch.Tensor, router_indices=None, routing_weig
                 expert_mask = expert_mask.permute(2, 1, 0)
                 # we sum on the top_k and on the sequence lenght to get which experts
                 # are hit this time around
-                expert_hitted = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
-            for expert_idx in expert_hitted[:]:
+                expert_hit = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
+            for expert_idx in expert_hit[:]:
                 with torch.no_grad():
                     _, token_idx = torch.where(expert_mask[expert_idx[0]])
                 current_state = hidden_states[token_idx]

src/transformers/models/gpt_oss/modular_gpt_oss.py

@@ -95,8 +95,8 @@ def forward(self, hidden_states: torch.Tensor, router_indices=None, routing_weig
                 expert_mask = expert_mask.permute(2, 1, 0)
                 # we sum on the top_k and on the sequence lenght to get which experts
                 # are hit this time around
-                expert_hitted = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
-            for expert_idx in expert_hitted[:]:
+                expert_hit = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
+            for expert_idx in expert_hit[:]:
                 with torch.no_grad():
                     _, token_idx = torch.where(expert_mask[expert_idx[0]])
                 current_state = hidden_states[token_idx]

src/transformers/models/minimax/modeling_minimax.py

@@ -465,8 +465,8 @@ def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         # this will be used to easily index which expert is going to be sollicitated
         expert_mask = torch.nn.functional.one_hot(selected_experts, num_classes=self.num_experts).permute(2, 1, 0)
 
-        expert_hitted = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
-        for expert_idx in expert_hitted:
+        expert_hit = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
+        for expert_idx in expert_hit:
             expert_layer = self.experts[expert_idx]
             idx, top_x = torch.where(expert_mask[expert_idx].squeeze(0))
             # Index the correct hidden states and compute the expert hidden state for

src/transformers/models/mixtral/modeling_mixtral.py

@@ -121,8 +121,8 @@ def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         # this will be used to easily index which expert is going to be sollicitated
         expert_mask = torch.nn.functional.one_hot(selected_experts, num_classes=self.num_experts).permute(2, 1, 0)
 
-        expert_hitted = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
-        for expert_idx in expert_hitted:
+        expert_hit = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
+        for expert_idx in expert_hit:
             expert_layer = self.experts[expert_idx]
             idx, top_x = torch.where(expert_mask[expert_idx].squeeze(0))
             # Index the correct hidden states and compute the expert hidden state for

src/transformers/models/mixtral/modular_mixtral.py

@@ -201,8 +201,8 @@ def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         # this will be used to easily index which expert is going to be sollicitated
         expert_mask = torch.nn.functional.one_hot(selected_experts, num_classes=self.num_experts).permute(2, 1, 0)
 
-        expert_hitted = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
-        for expert_idx in expert_hitted:
+        expert_hit = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
+        for expert_idx in expert_hit:
             expert_layer = self.experts[expert_idx]
             idx, top_x = torch.where(expert_mask[expert_idx].squeeze(0))
             # Index the correct hidden states and compute the expert hidden state for

src/transformers/models/qwen2_moe/modeling_qwen2_moe.py

@@ -621,8 +621,8 @@ def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         expert_mask = torch.nn.functional.one_hot(selected_experts, num_classes=self.num_experts).permute(2, 1, 0)
 
         # Loop over all available experts in the model and perform the computation on each expert
-        expert_hitted = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
-        for expert_idx in expert_hitted:
+        expert_hit = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
+        for expert_idx in expert_hit:
             expert_layer = self.experts[expert_idx]
             idx, top_x = torch.where(expert_mask[expert_idx].squeeze(0))
 

src/transformers/models/qwen3_moe/modeling_qwen3_moe.py

@@ -244,8 +244,8 @@ def forward(self, hidden_states: torch.Tensor) -> torch.Tensor:
         expert_mask = torch.nn.functional.one_hot(selected_experts, num_classes=self.num_experts).permute(2, 1, 0)
 
         # Loop over all available experts in the model and perform the computation on each expert
-        expert_hitted = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
-        for expert_idx in expert_hitted:
+        expert_hit = torch.greater(expert_mask.sum(dim=(-1, -2)), 0).nonzero()
+        for expert_idx in expert_hit:
             expert_layer = self.experts[expert_idx]
             idx, top_x = torch.where(expert_mask[expert_idx].squeeze(0))