[Bugfix] Fix lm_head weights tying with lora for llama (vllm-project#…

…9227)

vllm/model_executor/layers/vocab_parallel_embedding.py

@@ -443,7 +443,7 @@ def __init__(self,
         super().__init__(num_embeddings, embedding_dim, params_dtype,
                          org_num_embeddings, padding_size, quant_config,
                          prefix)
-
+        self.quant_config = quant_config
         if bias:
             self.bias = Parameter(
                 torch.empty(self.num_embeddings_per_partition,
@@ -455,6 +455,15 @@ def __init__(self,
         else:
             self.register_parameter("bias", None)
 
+    def tie_weights(self, embed_tokens: VocabParallelEmbedding):
+        """Tie the weights with word embeddings."""
+        # GGUF quantized embed_tokens.
+        if self.quant_config and self.quant_config.get_name() == "gguf":
+            return embed_tokens
+        else:
+            self.weight = embed_tokens.weight
+            return self
+
     def forward(self, input_):
         del input_
         raise RuntimeError("LMHead's weights should be used in the sampler.")

vllm/model_executor/models/llama.py

@@ -524,7 +524,8 @@ def __init__(
                 quant_config=quant_config,
             )
             if config.tie_word_embeddings:
-                self.lm_head = self.model.embed_tokens
+                self.lm_head = self.lm_head.tie_weights(
+                    self.model.embed_tokens)
 
             logit_scale = getattr(config, "logit_scale", 1.0)
             self.logits_processor = LogitsProcessor(self.unpadded_vocab_size,