[Sync] Update to latest code (#2679)

* [Sync] Update to latest code

* Add new code files

* Add new code files

* update code

* Try to fix build.sh

* Try to fix build.sh

* Update code

* Update requirements.txt

* Update code

---------

Co-authored-by: Jiang-Jia-Jun <jiangjiajun@baidu.com>

fastdeploy/platforms/cuda.py

@@ -46,7 +46,7 @@ class CUDAPlatform(Platform):
             return False
 
     @classmethod
-    def get_attention_backend_cls(cls, selected_backend):
+    def get_attention_backend_cls(cls, selected_backend: _Backend):
         """
         get_attention_backend_cls
         """
@@ -60,5 +60,13 @@ class CUDAPlatform(Platform):
             return (
                 "fastdeploy.model_executor.layers.attention.AppendAttentionBackend"
             )
+        elif selected_backend == _Backend.MLA_ATTN:
+            logger.info("Using MLA ATTN backend.")
+            return (
+                "fastdeploy.model_executor.layers.attention.MLAAttentionBackend"
+            )
         else:
-            logger.warning("Other backends are not supported for now.")
+            raise ValueError(
+                "Invalid attention backend you specified.\n"
+                "Now only support [NATIVE_ATTN, MLA_ATTN, APPEND_ATTN] in cuda place."
+            )