fix: aclFinalize repeat

RELEASE.md

@@ -14,7 +14,7 @@
 #### DataSet
 
 - The TFRecordDataset API can directly read TFRecord files compressed by GZIP or ZLIB.
-- The NumpySliceDataset API can process data of different dimensions at the same time.
+- The NumpySlicesDataset API can process data of different dimensions at the same time.
 - Optimize the structure of error log to display more clear call stack information for debugging.
 - Fixed `mindspore.dataset.config.set_seed` does not take effect for random seeds in distributed training scenarios.
 

RELEASE_CN.md

@@ -14,7 +14,7 @@
 #### DataSet
 
 - TFRecordDataset API支持直接读取通过GZIP或ZLIB压缩后的TFRecord文件。
-- NumpySliceDataset API支持同时处理不同维度的数据。
+- NumpySlicesDataset API支持同时处理不同维度的数据。
 - 优化错误日志信息的结构，展示更清晰的调用栈信息便于调试、定位问题。
 - 修复分布式训练场景下 `mindspore.dataset.config.set_seed` 对随机种子设置不生效的问题。
 

mindspore/ccsrc/minddata/dataset/kernels/image/dvpp/utils/ResourceManager.cc

@@ -66,7 +66,7 @@ void ResourceManager::Release() {
   }
 
   // finalize the acl when the process exit
-  ret = mindspore::AclInitAdapter::GetInstance().ForceFinalize();
+  ret = mindspore::AclInitAdapter::GetInstance().AclFinalize();
   if (ret != APP_ERR_OK) {
     MS_LOG(DEBUG) << "Failed to finalize acl, ret = " << ret << ".";
   }