[EPLB] Avoiding eplb's dependency on a specified model (#6528)

### What this PR does / why we need it?
1. Currently, eplb registers different attributes for different models,
but these attributes are not actually used. Now, these attributes are
directly deleted.
2. Add some log about eplb.

### Does this PR introduce _any_ user-facing change?

### How was this patch tested?
#### Deepseek v3.1 chat
Of course! Here is a comprehensive explanation of deep learning, broken
down for clarity.\n\n### The Simple Analogy: A Child Learning to
Recognize a Cat\n\nImagine teaching a child what a cat is. You don't
give them a rulebook with instructions like \"has pointy ears, whiskers,
and a tail.\" Instead, you show them many pictures, saying \"this is a
cat\" or \"this is not a cat.\" The child's brain gradually learns to
identify the complex patterns—the combination of shapes, colors, and
textures—that define \"cat-ness.\"\n\n**Deep learning is essentially
this, but for computers.** It's a method for teaching computers to learn
from examples and recognize patterns directly from data (like images,
sound, or text) without being explicitly programmed with rigid
rules.\n\n---\n\n### The Technical Definition\n\n**Deep Learning is a
subfield of machine learning, which itself is a subfield of artificial
intelligence (AI).** It uses artificial **neural networks** with many
layers (\"deep\" networks) to model and understand complex patterns in
data.\n\nHere are the key concepts in that definition:\n\n1.
**Artificial Intelligence (AI):** The broad science of making machines
smart and capable of performing tasks that typically require human
intelligence.\n2. **Machine Learning (ML):** A subset of AI that gives
computers the ability to learn from data *without* being explicitly
programmed for every single rule.\n3. **Deep Learning (DL):** A
specific, powerful

- vLLM version: v0.15.0
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.15.0

Signed-off-by: shenchuxiaofugui <1311027364@qq.com>

vllm_ascend/eplb/utils.py

@@ -28,47 +28,25 @@ def get_log2phy_map(self, layer_id):
     return self.model.layers[layer_id].mlp.experts.get_log2phy_map()
 
 
-def get_all_expert_map(self, num_moe_layers):
-    all_loads = []
-    num_dense_layers = self.num_dense_layers if hasattr(self, "num_dense_layers") else 0
-    for layer_id in range(num_moe_layers):
-        load_tensor = self.get_expert_map(layer_id + num_dense_layers)  # (num_experts_per_layer,)
-        all_loads.append(load_tensor)
-
-    return torch.stack(all_loads, dim=0)
-
-
 def get_all_moe_loads(self):
-    num_dense_layers = self.num_dense_layers if hasattr(self, "num_dense_layers") else 0
+    num_dense_layers = getattr(self.model.config, "first_k_dense_replace", 0)
+    num_layers = self.model.config.num_hidden_layers
     all_moe_loads = torch.stack(
-        [
-            self.model.layers[layer_id + num_dense_layers].mlp.experts.moe_load
-            for layer_id in range(self.num_moe_layers)
-        ],
+        [self.model.layers[layer_id].mlp.experts.moe_load for layer_id in range(num_dense_layers, num_layers)],
         dim=0,
     )
     return all_moe_loads
 
 
 def clear_all_moe_loads(self):
-    num_dense_layers = self.num_dense_layers if hasattr(self, "num_dense_layers") else 0
-    for layer_id in range(self.num_moe_layers):
-        self.model.layers[layer_id + num_dense_layers].mlp.experts.clear_moe_load()
+    num_dense_layers = getattr(self.model.config, "first_k_dense_replace", 0)
+    num_layers = self.model.config.num_hidden_layers
+    for layer_id in range(num_dense_layers, num_layers):
+        self.model.layers[layer_id].mlp.experts.clear_moe_load()
 
 
-def model_register(model, model_config):
+def model_register(model):
     model.get_expert_map = types.MethodType(get_expert_map, model)
     model.get_log2phy_map = types.MethodType(get_log2phy_map, model)
-    model.get_all_expert_map = types.MethodType(get_all_expert_map, model)
     model.get_all_moe_loads = types.MethodType(get_all_moe_loads, model)
     model.clear_all_moe_loads = types.MethodType(clear_all_moe_loads, model)
-
-    config = model_config.hf_text_config
-
-    if config.model_type == "qwen3_moe":
-        model.num_moe_layers = config.num_hidden_layers
-    elif config.model_type == "deepseek_v2" or config.model_type == "deepseek_v3":
-        model.num_dense_layers = config.first_k_dense_replace
-        model.num_moe_layers = config.num_hidden_layers - model.num_dense_layers
-    else:
-        raise NotImplementedError("EPLB is not supported.")