[LoRA][1/N]Remove LoRA extra vocab (#28382)

Signed-off-by: Jee Jee Li <pandaleefree@gmail.com>

vllm/model_executor/models/jamba.py

@@ -30,7 +30,6 @@ from vllm.model_executor.layers.mamba.mamba_utils import (
 from vllm.model_executor.layers.pooler import DispatchPooler, Pooler
 from vllm.model_executor.layers.quantization import QuantizationConfig
 from vllm.model_executor.layers.vocab_parallel_embedding import (
-    DEFAULT_VOCAB_PADDING_SIZE,
     ParallelLMHead,
     VocabParallelEmbedding,
 )
@@ -307,21 +306,14 @@ class JambaModel(nn.Module):
         model_config = vllm_config.model_config
         cache_config = vllm_config.cache_config
         quant_config = vllm_config.quant_config
-        lora_config = vllm_config.lora_config
 
         self.config = config
-        lora_vocab = (
-            (lora_config.lora_extra_vocab_size * (lora_config.max_loras or 1))
-            if lora_config
-            else 0
-        )
-        self.vocab_size = config.vocab_size + lora_vocab
-        self.org_vocab_size = config.vocab_size
+
+        self.vocab_size = config.vocab_size
 
         self.embed_tokens = VocabParallelEmbedding(
             self.vocab_size,
             config.hidden_size,
-            org_num_embeddings=config.vocab_size,
         )
 
         extra_kwargs = {"is_lora_enabled": bool(vllm_config.lora_config)}
@@ -492,7 +484,7 @@ class JambaForCausalLM(
 
     def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
         config = vllm_config.model_config.hf_config
-        lora_config = vllm_config.lora_config
+
         scheduler_config = vllm_config.scheduler_config
 
         super().__init__()
@@ -503,24 +495,14 @@ class JambaForCausalLM(
         self.model = JambaModel(
             vllm_config=vllm_config, prefix=maybe_prefix(prefix, "model")
         )
-        self.unpadded_vocab_size = config.vocab_size
-        if lora_config:
-            self.unpadded_vocab_size += lora_config.lora_extra_vocab_size
+
         self.lm_head = ParallelLMHead(
-            self.unpadded_vocab_size,
+            config.vocab_size,
             config.hidden_size,
-            org_num_embeddings=config.vocab_size,
-            padding_size=DEFAULT_VOCAB_PADDING_SIZE
-            # We need bigger padding if using lora for kernel
-            # compatibility
-            if not lora_config
-            else lora_config.lora_vocab_padding_size,
             prefix=maybe_prefix(prefix, "lm_head"),
         )
 
-        self.logits_processor = LogitsProcessor(
-            self.unpadded_vocab_size, config.vocab_size
-        )
+        self.logits_processor = LogitsProcessor(config.vocab_size)
 
         self.make_empty_intermediate_tensors = (
             self.model.make_empty_intermediate_tensors