[Feature] block sparse attention (#3668)

* 支持稀疏attn

* fix bug

* code style

* fix moba attn get kv shape

* 修复a100编译

* codestyle

* code style

* code style

* code style

* fix conflict

* 增加单侧

* code style

* 增加eblite 加载时间

* fix bug

* for ci

* for ci

* for ci

* for ci

* 支持mlp block size 128

* 增加小算子单测

* fix 单测 mlp

* 将环境变量加入到config里面

* fix rollout config

* 修复显存

* add test server

* add test server

* fix mlp  最后一层使用full attn

fastdeploy/model_executor/layers/attention/__init__.py

@@ -20,6 +20,7 @@ from .block_multihead_attn_backend import BlockAttentionBackend
 from .flash_attn_backend import FlashAttentionBackend
 from .iluvatar_attn_backend import IluvatarAttnBackend
 from .mla_attention_backend import MLAAttentionBackend
+from .moba_attention_backend import MobaAttentionBackend
 from .native_paddle_backend import PaddleNativeAttnBackend
 from .xpu_attn_backend import XPUAttentionBackend
 
@@ -34,4 +35,5 @@ __all__ = [
     "IluvatarAttnBackend",
     "BlockAttentionBackend",
     "Attention",
+    "MobaAttentionBackend",
 ]