diff --git a/python/sglang/srt/models/exaone.py b/python/sglang/srt/models/exaone.py
index 9cddcb34f..bb077f2c8 100644
--- a/python/sglang/srt/models/exaone.py
+++ b/python/sglang/srt/models/exaone.py
@@ -323,27 +323,6 @@ class ExaoneForCausalLM(nn.Module):
         sample_output = self.sampler(logits_output, input_metadata.sampling_info)
         return sample_output, logits_output
 
-    def get_module_name(self, name):
-        stacked_params_mapping = [
-            # (param_name, shard_name, shard_id, num_shard)
-            ("qkv_proj", "q_proj", "q", 3),
-            ("qkv_proj", "k_proj", "k", 3),
-            ("qkv_proj", "v_proj", "v", 3),
-            ("gate_up_proj", "c_fc_0", 0, 2),
-            ("gate_up_proj", "c_fc_1", 1, 2),
-        ]
-        for param_name, weight_name, shard_id, num_shard in stacked_params_mapping:
-            if weight_name in name:
-                return (
-                    name.replace(weight_name, param_name)[: -len(".weight")],
-                    num_shard,
-                )
-        return name[: -len(".weight")], 1
-
-    def get_num_params(self):
-        params_dict = dict(self.named_parameters())
-        return len(params_dict)
-
     def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):
         stacked_params_mapping = [
             # (param_name, shard_name, shard_id)
@@ -357,13 +336,13 @@ class ExaoneForCausalLM(nn.Module):
 
         for name, loaded_weight in weights:
             if "rotary_emb.inv_freq" in name or "projector" in name:
-                return
+                continue
             if "rotary_emb.cos_cached" in name or "rotary_emb.sin_cached" in name:
                 # Models trained using ColossalAI may include these tensors in
                 # the checkpoint. Skip them.
-                return
+                continue
             if name.startswith("model.vision_tower") and name not in params_dict:
-                return
+                continue
 
             name = name.replace("attn.attention", "self_attn")
             for param_name, weight_name, shard_id in stacked_params_mapping:
@@ -380,7 +359,7 @@ class ExaoneForCausalLM(nn.Module):
             else:
                 # Skip loading extra bias for GPTQ models.
                 if name.endswith(".bias") and name not in params_dict:
-                    return
+                    continue
                 param = params_dict[name]
                 weight_loader = getattr(param, "weight_loader", default_weight_loader)
                 weight_loader(param, loaded_weight)
diff --git a/python/sglang/srt/models/llama.py b/python/sglang/srt/models/llama.py
index b875e0c98..926d87db8 100644
--- a/python/sglang/srt/models/llama.py
+++ b/python/sglang/srt/models/llama.py
@@ -334,13 +334,13 @@ class LlamaForCausalLM(nn.Module):
 
         for name, loaded_weight in weights:
             if "rotary_emb.inv_freq" in name or "projector" in name:
-                return
+                continue
             if "rotary_emb.cos_cached" in name or "rotary_emb.sin_cached" in name:
                 # Models trained using ColossalAI may include these tensors in
                 # the checkpoint. Skip them.
-                return
+                continue
             if name.startswith("model.vision_tower") and name not in params_dict:
-                return
+                continue
 
             for param_name, weight_name, shard_id in stacked_params_mapping:
                 if weight_name not in name:
@@ -356,7 +356,7 @@ class LlamaForCausalLM(nn.Module):
             else:
                 # Skip loading extra bias for GPTQ models.
                 if name.endswith(".bias") and name not in params_dict:
-                    return
+                    continue
                 param = params_dict[name]
                 weight_loader = getattr(param, "weight_loader", default_weight_loader)
                 weight_loader(param, loaded_weight)