[6/N] pass whole config to inner model (#10205)

Signed-off-by: youkaichao <youkaichao@gmail.com>

vllm/model_executor/models/commandr.py

@@ -28,7 +28,7 @@ from transformers import CohereConfig
 
 from vllm.attention import Attention, AttentionMetadata
 from vllm.compilation.decorators import support_torch_compile
-from vllm.config import CacheConfig, LoRAConfig, VllmConfig
+from vllm.config import CacheConfig, VllmConfig
 from vllm.distributed import get_pp_group, get_tensor_model_parallel_world_size
 from vllm.model_executor.layers.activation import SiluAndMul
 from vllm.model_executor.layers.linear import (MergedColumnParallelLinear,
@@ -49,7 +49,8 @@ from vllm.sequence import IntermediateTensors
 
 from .interfaces import SupportsLoRA, SupportsPP
 from .utils import (is_pp_missing_parameter,
-                    make_empty_intermediate_tensors_factory, make_layers)
+                    make_empty_intermediate_tensors_factory, make_layers,
+                    maybe_prefix)
 
 
 @torch.compile
@@ -253,15 +254,14 @@ class CohereDecoderLayer(nn.Module):
 @support_torch_compile
 class CohereModel(nn.Module):
 
-    def __init__(
-        self,
-        config: CohereConfig,
-        cache_config: Optional[CacheConfig] = None,
-        quant_config: Optional[QuantizationConfig] = None,
-        lora_config: Optional[LoRAConfig] = None,
-        prefix: str = "",
-    ):
+    def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
         super().__init__()
+
+        config = vllm_config.model_config.hf_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
@@ -332,14 +332,9 @@ class CohereForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
     embedding_modules = {"embed_tokens": "input_embeddings"}
     embedding_padding_modules = []
 
-    def __init__(
-        self,
-        vllm_config: VllmConfig,
-        prefix: str = "",
-    ) -> None:
+    def __init__(self, *, vllm_config: VllmConfig, prefix: str = ""):
         super().__init__()
         config = vllm_config.model_config.hf_config
-        cache_config = vllm_config.cache_config
         quant_config = vllm_config.quant_config
         lora_config = vllm_config.lora_config
         self.config = config
@@ -353,10 +348,8 @@ class CohereForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
         self.logits_processor = LogitsProcessor(self.unpadded_vocab_size,
                                                 config.vocab_size,
                                                 scale=config.logit_scale)
-        self.model = CohereModel(config,
-                                 cache_config,
-                                 quant_config,
-                                 lora_config=lora_config)
+        self.model = CohereModel(vllm_config=vllm_config,
+                                 prefix=maybe_prefix(prefix, "model"))
         self.sampler = get_sampler()
         self.make_empty_intermediate_tensors = (
             self.model.make_empty_intermediate_tensors)