[Refactor] Add expert processed token count output for DispatchFFNCombine/DispatchFFNCombineBF16 (#6402)

### What this PR does / why we need it?
Add New Output for Expert Token Count
An additional output tensor expert_token_nums is added to both operators
to meet the requirement of tracking token distribution among experts:

Tensor Name: expert_token_nums
Dimension: 1D tensor
Shape: (local_expert_num,)
Data Type: int32
Semantics: Represents the number of tokens actually received by each
expert on the current card.
### Does this PR introduce _any_ user-facing change?

### How was this patch tested?

- vLLM version: v0.14.1
- vLLM main:
dc917cceb8

---------

Signed-off-by: guanguan0308 <1546542263@qq.com>
Signed-off-by: guanguan0308 <162653673+guanguan0308@users.noreply.github.com>

csrc/torch_binding_meta.cpp

@@ -194,7 +194,7 @@ void batch_matmul_transpose(const at::Tensor &tensor_a, const at::Tensor &tensor
     return;
 }
 
-at::Tensor& dispatch_ffn_combine_meta(
+std::tuple<at::Tensor&, at::Tensor&> dispatch_ffn_combine_meta(
     const at::Tensor& x,
     const at::TensorList& weight1,
     const at::TensorList& weight2,
@@ -204,9 +204,10 @@ at::Tensor& dispatch_ffn_combine_meta(
     const at::Tensor& probs,
     c10::string_view group,
     int64_t max_output_size,
-    at::Tensor& out
+    at::Tensor& out,
+    at::Tensor& expert_token_nums
 ) {
-    return out;
+    return {out, expert_token_nums};
 }
 
 at::Tensor npu_lightning_indexer_meta(