[None][fix] Prevent YAML partial kv_cache_config from incorrectly overriding the complete kv_cache_config (#9262)

Signed-off-by: Yuening Li <[email protected]>

Author: Yuening-wa

tensorrt_llm/llmapi/llm_args.py

@@ -3032,6 +3032,15 @@ def update_llm_args_with_extra_dict(
         llm_args_dict: Dict,
         extra_llm_api_options: Optional[str] = None) -> Dict:
 
+    # Deep merge kv_cache_config to prevent partial YAML kv_cache_config from replacing the complete kv_cache_config
+    if 'kv_cache_config' in llm_args and 'kv_cache_config' in llm_args_dict:
+        # Convert KvCacheConfig object to dict if necessary
+        base_kv_config = llm_args['kv_cache_config']
+        if isinstance(base_kv_config, KvCacheConfig):
+            base_kv_config = base_kv_config.model_dump(exclude_unset=True)
+        llm_args_dict['kv_cache_config'] = base_kv_config | llm_args_dict[
+            'kv_cache_config']
+
     field_mapping = {
         "quant_config": QuantConfig,
         "calib_config": CalibConfig,
@@ -3043,6 +3052,7 @@ def update_llm_args_with_extra_dict(
         "moe_config": MoeConfig,
         "attention_dp_config": AttentionDpConfig,
         "sparse_attention_config": BaseSparseAttentionConfig,
+        "kv_cache_config": KvCacheConfig,
     }
     for field_name, field_type in field_mapping.items():
         if field_name in llm_args_dict: