fix: kv cache support in builtin backend

src/pages/llmodels/config/backend-parameters/index.ts

@@ -7,7 +7,8 @@ export const backendOptionsMap = {
   vllm: 'vLLM',
   voxBox: 'VoxBox',
   ascendMindie: 'MindIE',
-  custom: 'Custom'
+  custom: 'Custom',
+  SGLang: 'SGLang'
 };
 
 export interface BackendParameter {
@@ -36,5 +37,6 @@ export default {
   [backendOptionsMap.vllm]: generateBackendParameters(vllmParameters),
   [backendOptionsMap.ascendMindie]: generateBackendParameters(mindieParameters),
   [backendOptionsMap.voxBox]: [],
-  [backendOptionsMap.custom]: []
+  [backendOptionsMap.custom]: [],
+  [backendOptionsMap.SGLang]: []
 };

src/pages/llmodels/config/types.ts

@@ -260,6 +260,7 @@ export interface BackendOption {
   label: string;
   default_backend_param: string[];
   default_version: string;
+  isBuiltIn: boolean;
   versions: { label: string; value: string }[];
 }
 
@@ -273,6 +274,7 @@ export interface BackendItem {
   from_config: boolean;
   default_version: string;
   default_backend_param: string[];
+  is_built_in: boolean;
   versions: {
     version: string;
     is_deprecated: boolean;