diff --git a/python/sglang/srt/models/deepseek_v2.py b/python/sglang/srt/models/deepseek_v2.py index 2160a9e6c..0239736c7 100644 --- a/python/sglang/srt/models/deepseek_v2.py +++ b/python/sglang/srt/models/deepseek_v2.py @@ -1786,8 +1786,7 @@ class DeepseekV2ForCausalLM(nn.Module): for name in weight_names: if "kv_b_proj" in name: layer_id = int(name.split(".")[2]) - # filter the nextn layer. - if layer_id != self.config.num_hidden_layers: + if layer_id < self.config.num_hidden_layers: layer_ids.add(layer_id) for layer_id in layer_ids: