diff --git a/paddlex/cv/datasets/coco.py b/paddlex/cv/datasets/coco.py
index dc9946ac784a296dfc2751ef0a46bc2496aca2e4..0c0dda0c2c667de4e79408befc686024c988ceb3 100644
--- a/paddlex/cv/datasets/coco.py
+++ b/paddlex/cv/datasets/coco.py
@@ -44,7 +44,7 @@ class CocoDetection(VOCDetection):
                  transforms=None,
                  num_workers='auto',
                  buffer_size=100,
-                 parallel_method='process',
+                 parallel_method='thread',
                  shuffle=False):
         from pycocotools.coco import COCO
 
diff --git a/paddlex/cv/datasets/easydata_cls.py b/paddlex/cv/datasets/easydata_cls.py
index c947206045d32ee2c22e3218aaa1e3dfc81db4c0..a23061e9368ee8f42f94003ffb3f4fc75fc268c7 100644
--- a/paddlex/cv/datasets/easydata_cls.py
+++ b/paddlex/cv/datasets/easydata_cls.py
@@ -47,7 +47,7 @@ class EasyDataCls(ImageNet):
                  transforms=None,
                  num_workers='auto',
                  buffer_size=100,
-                 parallel_method='process',
+                 parallel_method='thread',
                  shuffle=False):
         super(ImageNet, self).__init__(
             transforms=transforms,
diff --git a/paddlex/cv/datasets/easydata_det.py b/paddlex/cv/datasets/easydata_det.py
index 6424806148d970e8ee868eef089f7b025c333bd7..5db31481ed3397e1c3e62484b8e29ede731b431e 100644
--- a/paddlex/cv/datasets/easydata_det.py
+++ b/paddlex/cv/datasets/easydata_det.py
@@ -48,7 +48,7 @@ class EasyDataDet(VOCDetection):
                  transforms=None,
                  num_workers='auto',
                  buffer_size=100,
-                 parallel_method='process',
+                 parallel_method='thread',
                  shuffle=False):
         super(VOCDetection, self).__init__(
             transforms=transforms,
diff --git a/paddlex/cv/datasets/easydata_seg.py b/paddlex/cv/datasets/easydata_seg.py
index 56f801721d435e9c0c92809c2aac35f43c074b7e..09b434ffcc96a8b071d75a6536c3e8c84b39cb28 100644
--- a/paddlex/cv/datasets/easydata_seg.py
+++ b/paddlex/cv/datasets/easydata_seg.py
@@ -46,7 +46,7 @@ class EasyDataSeg(Dataset):
                  transforms=None,
                  num_workers='auto',
                  buffer_size=100,
-                 parallel_method='process',
+                 parallel_method='thread',
                  shuffle=False):
         super(EasyDataSeg, self).__init__(
             transforms=transforms,
diff --git a/paddlex/cv/datasets/imagenet.py b/paddlex/cv/datasets/imagenet.py
index 06ca1cedae8bc5e0bf67c6068cd8b75376f87bfd..2d44643e93347ea4e786a31cf2e8fc32c0d6e7b3 100644
--- a/paddlex/cv/datasets/imagenet.py
+++ b/paddlex/cv/datasets/imagenet.py
@@ -46,7 +46,7 @@ class ImageNet(Dataset):
                  transforms=None,
                  num_workers='auto',
                  buffer_size=100,
-                 parallel_method='process',
+                 parallel_method='thread',
                  shuffle=False):
         super(ImageNet, self).__init__(
             transforms=transforms,
diff --git a/paddlex/cv/datasets/seg_dataset.py b/paddlex/cv/datasets/seg_dataset.py
index 2c56c5d001135abe90415444baf17877012f9541..7b78080c773111775c97d5754def603da303d06a 100644
--- a/paddlex/cv/datasets/seg_dataset.py
+++ b/paddlex/cv/datasets/seg_dataset.py
@@ -44,7 +44,7 @@ class SegDataset(Dataset):
                  transforms=None,
                  num_workers='auto',
                  buffer_size=100,
-                 parallel_method='process',
+                 parallel_method='thread',
                  shuffle=False):
         super(SegDataset, self).__init__(
             transforms=transforms,
diff --git a/paddlex/cv/datasets/voc.py b/paddlex/cv/datasets/voc.py
index db63d9828d03b95569382f7d059df5567a9ce59d..a4c73519b523eccdeb2b34774e4fc1d0b959feaf 100644
--- a/paddlex/cv/datasets/voc.py
+++ b/paddlex/cv/datasets/voc.py
@@ -48,7 +48,7 @@ class VOCDetection(Dataset):
                  transforms=None,
                  num_workers='auto',
                  buffer_size=100,
-                 parallel_method='process',
+                 parallel_method='thread',
                  shuffle=False):
         from pycocotools.coco import COCO
         super(VOCDetection, self).__init__(