[Bugfix][Speculative Decoding] Extend Eagle quantization config fix to llama_eagle.py

vllm/model_executor/models/llama_eagle.py

@@ -12,6 +12,7 @@
 from vllm.distributed.parallel_state import get_pp_group
 from vllm.logger import init_logger
 from vllm.model_executor.layers.logits_processor import LogitsProcessor
+from vllm.model_executor.layers.quantization.base_config import QuantizationConfig
 from vllm.model_executor.layers.vocab_parallel_embedding import VocabParallelEmbedding
 from vllm.model_executor.model_loader.weight_utils import default_weight_loader
 from vllm.model_executor.models.llama import LlamaDecoderLayer, LlamaForCausalLM
@@ -37,6 +38,17 @@ def __init__(
             del self.input_layernorm
             self.input_layernorm = nn.Identity()
 
+    def get_quant_config(self, vllm_config: VllmConfig) -> QuantizationConfig | None:
+        """Use drafter's quantization config instead of verifier's."""
+        draft_model_config = vllm_config.speculative_config.draft_model_config
+        draft_load_config = vllm_config.load_config
+
+        return (
+            VllmConfig.get_quantization_config(draft_model_config, draft_load_config)
+            if draft_model_config
+            else None
+        )
+
 
 @support_torch_compile
 class LlamaModel(nn.Module):