Merge branch 'PaddlePaddle:develop_reg' into develop_reg

docs/zh_CN/models/LeViT.md

+# LeViT
+
+## 概述
+LeViT是一种快速推理的、用于图像分类任务的混合神经网络。其设计之初考虑了网络模型在不同的硬件平台上的性能，因此能够更好地反映普遍应用的真实场景。通过大量实验，作者找到了卷积神经网络与Transformer体系更好的结合方式，并且提出了attention-based方法，用于整合Transformer中的位置信息编码。[论文地址](https://arxiv.org/abs/2104.01136)。
+
+## 精度、FLOPS和参数量
+
+| Models           | Top1 | Top5 | Reference<br>top1 | Reference<br>top5 | FLOPS<br>(M) | Params<br>(M) |
+|:--:|:--:|:--:|:--:|:--:|:--:|:--:|
+| LeViT-128S | 0.7598 | 0.9269 | 0.766 | 0.929 | 305  | 7.8 |
+| LeViT-128  | 0.7810 | 0.9371 | 0.786 | 0.940 | 406  | 9.2 |
+| LeViT-192  | 0.7934 | 0.9446 | 0.800 | 0.947 | 658  | 11 |
+| LeViT-256  | 0.8085 | 0.9497 | 0.816 | 0.954 | 1120 | 19 |
+| LeViT-384  | 0.8191 | 0.9551 | 0.826 | 0.960 | 2353 | 39 |
+
+
+**注**：与Reference的精度差异源于数据预处理不同及未使用蒸馏的head作为输出。

docs/zh_CN/models/Twins.md

+# Twins
+
+## 概述
+Twins网络包括Twins-PCPVT和Twins-SVT，其重点对空间注意力机制进行了精心设计，得到了简单却更为有效的方案。由于该体系结构仅涉及矩阵乘法，而目前的深度学习框架中对矩阵乘法有较高的优化程度，因此该体系结构十分高效且易于实现。并且，该体系结构在图像分类、目标检测和语义分割等多种下游视觉任务中都能够取得优异的性能。[论文地址](https://arxiv.org/abs/2104.13840)。
+
+## 精度、FLOPS和参数量
+
+| Models        | Top1 | Top5 | Reference<br>top1 | Reference<br>top5 | FLOPS<br>(G) | Params<br>(M) |
+|:--:|:--:|:--:|:--:|:--:|:--:|:--:|
+| pcpvt_small   | 0.8082 | 0.9552 | 0.812 | - | 3.7 | 24.1   |
+| pcpvt_base    | 0.8242 | 0.9619 | 0.827 | - | 6.4 | 43.8   |
+| pcpvt_large   | 0.8273 | 0.9650 | 0.831 | - | 9.5 | 60.9   |
+| alt_gvt_small | 0.8140 | 0.9546 | 0.817 | - | 2.8  | 24   |
+| alt_gvt_base  | 0.8294 | 0.9621 | 0.832 | - | 8.3  | 56   |
+| alt_gvt_large | 0.8331 | 0.9642 | 0.837 | - | 14.8 | 99.2   |
+
+**注**：与Reference的精度差异源于数据预处理不同。

ppcls/arch/backbone/__init__.py

-# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
+# copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -19,11 +19,12 @@ from ppcls.arch.backbone.legendary_models.vgg import VGG11, VGG13, VGG16, VGG19
 from ppcls.arch.backbone.legendary_models.inception_v3 import InceptionV3
 from ppcls.arch.backbone.legendary_models.hrnet import HRNet_W18_C, HRNet_W30_C, HRNet_W32_C, HRNet_W40_C, HRNet_W44_C, HRNet_W48_C, HRNet_W60_C, HRNet_W64_C, SE_HRNet_W64_C
 
-from ppcls.arch.backbone.model_zoo.resnet_vc import ResNet18_vc, ResNet34_vc, ResNet50_vc, ResNet101_vc, ResNet152_vc
+from ppcls.arch.backbone.model_zoo.resnet_vc import ResNet50_vc
 from ppcls.arch.backbone.model_zoo.resnext import ResNeXt50_32x4d, ResNeXt50_64x4d, ResNeXt101_32x4d, ResNeXt101_64x4d, ResNeXt152_32x4d, ResNeXt152_64x4d
-from ppcls.arch.backbone.model_zoo.res2net import Res2Net50_48w_2s, Res2Net50_26w_4s, Res2Net50_14w_8s, Res2Net50_48w_2s, Res2Net50_26w_6s, Res2Net50_26w_8s, Res2Net101_26w_4s, Res2Net152_26w_4s, Res2Net200_26w_4s
-from ppcls.arch.backbone.model_zoo.res2net_vd import Res2Net50_vd_48w_2s, Res2Net50_vd_26w_4s, Res2Net50_vd_14w_8s, Res2Net50_vd_48w_2s, Res2Net50_vd_26w_6s, Res2Net50_vd_26w_8s, Res2Net101_vd_26w_4s, Res2Net152_vd_26w_4s, Res2Net200_vd_26w_4s
-from ppcls.arch.backbone.model_zoo.se_resnet_vd import SE_ResNet18_vd, SE_ResNet34_vd, SE_ResNet50_vd, SE_ResNet101_vd, SE_ResNet152_vd, SE_ResNet200_vd
+from ppcls.arch.backbone.model_zoo.resnext_vd import ResNeXt50_vd_32x4d, ResNeXt50_vd_64x4d, ResNeXt101_vd_32x4d, ResNeXt101_vd_64x4d, ResNeXt152_vd_32x4d, ResNeXt152_vd_64x4d
+from ppcls.arch.backbone.model_zoo.res2net import Res2Net50_26w_4s, Res2Net50_14w_8s
+from ppcls.arch.backbone.model_zoo.res2net_vd import Res2Net50_vd_26w_4s, Res2Net101_vd_26w_4s, Res2Net200_vd_26w_4s
+from ppcls.arch.backbone.model_zoo.se_resnet_vd import SE_ResNet18_vd, SE_ResNet34_vd, SE_ResNet50_vd
 from ppcls.arch.backbone.model_zoo.se_resnext_vd import SE_ResNeXt50_vd_32x4d, SE_ResNeXt50_vd_32x4d, SENet154_vd
 from ppcls.arch.backbone.model_zoo.se_resnext import SE_ResNeXt50_32x4d, SE_ResNeXt101_32x4d, SE_ResNeXt152_64x4d
 from ppcls.arch.backbone.model_zoo.dpn import DPN68, DPN92, DPN98, DPN107, DPN131
@@ -33,10 +34,11 @@ from ppcls.arch.backbone.model_zoo.resnest import ResNeSt50_fast_1s1x64d, ResNeS
 from ppcls.arch.backbone.model_zoo.googlenet import GoogLeNet
 from ppcls.arch.backbone.model_zoo.mobilenet_v2 import MobileNetV2_x0_25, MobileNetV2_x0_5, MobileNetV2_x0_75, MobileNetV2, MobileNetV2_x1_5, MobileNetV2_x2_0
 from ppcls.arch.backbone.model_zoo.shufflenet_v2 import ShuffleNetV2_x0_25, ShuffleNetV2_x0_33, ShuffleNetV2_x0_5, ShuffleNetV2_x1_0, ShuffleNetV2_x1_5, ShuffleNetV2_x2_0, ShuffleNetV2_swish
+from ppcls.arch.backbone.model_zoo.ghostnet import GhostNet_x0_5, GhostNet_x1_0, GhostNet_x1_3
 from ppcls.arch.backbone.model_zoo.alexnet import AlexNet
 from ppcls.arch.backbone.model_zoo.inception_v4 import InceptionV4
 from ppcls.arch.backbone.model_zoo.xception import Xception41, Xception65, Xception71
-from ppcls.arch.backbone.model_zoo.xception_deeplab import Xception41_deeplab, Xception65_deeplab, Xception71_deeplab
+from ppcls.arch.backbone.model_zoo.xception_deeplab import Xception41_deeplab, Xception65_deeplab
 from ppcls.arch.backbone.model_zoo.resnext101_wsl import ResNeXt101_32x8d_wsl, ResNeXt101_32x16d_wsl, ResNeXt101_32x32d_wsl, ResNeXt101_32x48d_wsl
 from ppcls.arch.backbone.model_zoo.squeezenet import SqueezeNet1_0, SqueezeNet1_1
 from ppcls.arch.backbone.model_zoo.darknet import DarkNet53
@@ -47,6 +49,8 @@ from ppcls.arch.backbone.model_zoo.distillation_models import ResNet50_vd_distil
 from ppcls.arch.backbone.model_zoo.swin_transformer import SwinTransformer_tiny_patch4_window7_224, SwinTransformer_small_patch4_window7_224, SwinTransformer_base_patch4_window7_224, SwinTransformer_base_patch4_window12_384, SwinTransformer_large_patch4_window7_224, SwinTransformer_large_patch4_window12_384
 from ppcls.arch.backbone.model_zoo.mixnet import MixNet_S, MixNet_M, MixNet_L
 from ppcls.arch.backbone.model_zoo.rexnet import ReXNet_1_0, ReXNet_1_3, ReXNet_1_5, ReXNet_2_0, ReXNet_3_0
+from ppcls.arch.backbone.model_zoo.gvt import pcpvt_small, pcpvt_base, pcpvt_large, alt_gvt_small, alt_gvt_base, alt_gvt_large
+from ppcls.arch.backbone.model_zoo.levit import LeViT_128S, LeViT_128, LeViT_192, LeViT_256, LeViT_384
 from ppcls.arch.backbone.model_zoo.dla import DLA34, DLA46_c, DLA46x_c, DLA60, DLA60x, DLA60x_c, DLA102, DLA102x, DLA102x2, DLA169
 from ppcls.arch.backbone.model_zoo.rednet import RedNet26, RedNet38, RedNet50, RedNet101, RedNet152 
 from ppcls.arch.backbone.model_zoo.tnt import TNT_small

ppcls/arch/backbone/model_zoo/alexnet.py

+# copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import paddle
 from paddle import ParamAttr
 import paddle.nn as nn
@@ -7,8 +21,11 @@ from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 from paddle.nn.initializer import Uniform
 import math
 
-__all__ = ["AlexNet"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"AlexNet": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/AlexNet_pretrained.pdparams"}
 
+__all__ = list(MODEL_URLS.keys())
 
 class ConvPoolLayer(nn.Layer):
     def __init__(self,
@@ -126,7 +143,19 @@ class AlexNetDY(nn.Layer):
         x = self._fc8(x)
         return x
 
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
 
-def AlexNet(**args):
-    model = AlexNetDY(**args)
+def AlexNet(pretrained=False, use_ssld=False, **kwargs):
+    model = AlexNetDY(**kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["AlexNet"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/darknet.py

+# copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import paddle
 from paddle import ParamAttr
 import paddle.nn as nn
@@ -7,8 +21,11 @@ from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 from paddle.nn.initializer import Uniform
 import math
 
-__all__ = ["DarkNet53"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"DarkNet53": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DarkNet53_pretrained.pdparams"}
 
+__all__ = list(MODEL_URLS.keys())
 
 class ConvBNLayer(nn.Layer):
     def __init__(self,
@@ -155,7 +172,19 @@ class DarkNet(nn.Layer):
         x = self._out(x)
         return x
 
-
-def DarkNet53(**args):
-    model = DarkNet(**args)
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+def DarkNet53(pretrained=False, use_ssld=False, **kwargs):
+    model = DarkNet(**kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DarkNet53"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/densenet.py

-# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
+# copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -26,9 +26,16 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = [
-    "DenseNet121", "DenseNet161", "DenseNet169", "DenseNet201", "DenseNet264"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"DenseNet121": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DenseNet121_pretrained.pdparams",
+              "DenseNet161": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DenseNet161_pretrained.pdparams",
+              "DenseNet169": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DenseNet169_pretrained.pdparams",
+              "DenseNet201": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DenseNet201_pretrained.pdparams",
+              "DenseNet264": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DenseNet264_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class BNACConvLayer(nn.Layer):
@@ -282,27 +289,43 @@ class DenseNet(nn.Layer):
         y = self.out(y)
         return y
 
-
-def DenseNet121(**args):
-    model = DenseNet(layers=121, **args)
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+def DenseNet121(pretrained=False, use_ssld=False, **kwargs):
+    model = DenseNet(layers=121, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DenseNet121"], use_ssld=use_ssld)
     return model
 
 
-def DenseNet161(**args):
-    model = DenseNet(layers=161, **args)
+def DenseNet161(pretrained=False, use_ssld=False, **kwargs):
+    model = DenseNet(layers=161, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DenseNet161"], use_ssld=use_ssld)
     return model
 
 
-def DenseNet169(**args):
-    model = DenseNet(layers=169, **args)
+def DenseNet169(pretrained=False, use_ssld=False, **kwargs):
+    model = DenseNet(layers=169, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DenseNet169"], use_ssld=use_ssld)
     return model
 
 
-def DenseNet201(**args):
-    model = DenseNet(layers=201, **args)
+def DenseNet201(pretrained=False, use_ssld=False, **kwargs):
+    model = DenseNet(layers=201, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DenseNet201"], use_ssld=use_ssld)
     return model
 
 
-def DenseNet264(**args):
-    model = DenseNet(layers=264, **args)
+def DenseNet264(pretrained=False, use_ssld=False, **kwargs):
+    model = DenseNet(layers=264, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DenseNet264"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/distilled_vision_transformer.py

@@ -16,12 +16,20 @@ import paddle
 import paddle.nn as nn
 from .vision_transformer import VisionTransformer, Identity, trunc_normal_, zeros_
 
-__all__ = [
-    'DeiT_tiny_patch16_224', 'DeiT_small_patch16_224', 'DeiT_base_patch16_224',
-    'DeiT_tiny_distilled_patch16_224', 'DeiT_small_distilled_patch16_224',
-    'DeiT_base_distilled_patch16_224', 'DeiT_base_patch16_384',
-    'DeiT_base_distilled_patch16_384'
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "DeiT_tiny_patch16_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DeiT_tiny_patch16_224_pretrained.pdparams",
+              "DeiT_small_patch16_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DeiT_small_patch16_224_pretrained.pdparams",
+              "DeiT_base_patch16_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DeiT_base_patch16_224_pretrained.pdparams",
+              "DeiT_tiny_distilled_patch16_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DeiT_tiny_distilled_patch16_224_pretrained.pdparams",
+              "DeiT_small_distilled_patch16_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DeiT_small_distilled_patch16_224_pretrained.pdparams",
+              "DeiT_base_distilled_patch16_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DeiT_base_distilled_patch16_224_pretrained.pdparams", 
+              "DeiT_base_patch16_384": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DeiT_base_patch16_384_pretrained.pdparams",
+              "DeiT_base_distilled_patch16_384": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DeiT_base_distilled_patch16_384_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class DistilledVisionTransformer(VisionTransformer):
@@ -90,7 +98,20 @@ class DistilledVisionTransformer(VisionTransformer):
         return (x + x_dist) / 2
 
 
-def DeiT_tiny_patch16_224(**kwargs):
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+
+def DeiT_tiny_patch16_224(pretrained=False, use_ssld=False, **kwargs):
     model = VisionTransformer(
         patch_size=16,
         embed_dim=192,
@@ -100,10 +121,11 @@ def DeiT_tiny_patch16_224(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DeiT_tiny_patch16_224"], use_ssld=use_ssld)
     return model
 
 
-def DeiT_small_patch16_224(**kwargs):
+def DeiT_small_patch16_224(pretrained=False, use_ssld=False, **kwargs):
     model = VisionTransformer(
         patch_size=16,
         embed_dim=384,
@@ -113,10 +135,11 @@ def DeiT_small_patch16_224(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DeiT_small_patch16_224"], use_ssld=use_ssld)
     return model
 
 
-def DeiT_base_patch16_224(**kwargs):
+def DeiT_base_patch16_224(pretrained=False, use_ssld=False, **kwargs):
     model = VisionTransformer(
         patch_size=16,
         embed_dim=768,
@@ -126,10 +149,11 @@ def DeiT_base_patch16_224(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DeiT_base_patch16_224"], use_ssld=use_ssld)
     return model
 
 
-def DeiT_tiny_distilled_patch16_224(**kwargs):
+def DeiT_tiny_distilled_patch16_224(pretrained=False, use_ssld=False, **kwargs):
     model = DistilledVisionTransformer(
         patch_size=16,
         embed_dim=192,
@@ -139,10 +163,11 @@ def DeiT_tiny_distilled_patch16_224(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DeiT_tiny_distilled_patch16_224"], use_ssld=use_ssld)
     return model
 
 
-def DeiT_small_distilled_patch16_224(**kwargs):
+def DeiT_small_distilled_patch16_224(pretrained=False, use_ssld=False, **kwargs):
     model = DistilledVisionTransformer(
         patch_size=16,
         embed_dim=384,
@@ -152,10 +177,11 @@ def DeiT_small_distilled_patch16_224(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DeiT_small_distilled_patch16_224"], use_ssld=use_ssld)
     return model
 
 
-def DeiT_base_distilled_patch16_224(**kwargs):
+def DeiT_base_distilled_patch16_224(pretrained=False, use_ssld=False, **kwargs):
     model = DistilledVisionTransformer(
         patch_size=16,
         embed_dim=768,
@@ -165,10 +191,11 @@ def DeiT_base_distilled_patch16_224(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DeiT_base_distilled_patch16_224"], use_ssld=use_ssld)
     return model
 
 
-def DeiT_base_patch16_384(**kwargs):
+def DeiT_base_patch16_384(pretrained=False, use_ssld=False, **kwargs):
     model = VisionTransformer(
         img_size=384,
         patch_size=16,
@@ -179,10 +206,11 @@ def DeiT_base_patch16_384(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DeiT_base_patch16_384"], use_ssld=use_ssld)
     return model
 
 
-def DeiT_base_distilled_patch16_384(**kwargs):
+def DeiT_base_distilled_patch16_384(pretrained=False, use_ssld=False, **kwargs):
     model = DistilledVisionTransformer(
         img_size=384,
         patch_size=16,
@@ -193,4 +221,5 @@ def DeiT_base_distilled_patch16_384(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DeiT_base_distilled_patch16_384"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/dla.py

@@ -26,25 +26,25 @@ from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_f
 
 MODEL_URLS = {
     "DLA34":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/DLA34_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DLA34_pretrained.pdparams",
     "DLA46_c":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/DLA46_c_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DLA46_c_pretrained.pdparams",
     "DLA46x_c":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/DLA46x_c_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DLA46x_c_pretrained.pdparams",
     "DLA60":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/DLA60_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DLA60_pretrained.pdparams",
     "DLA60x":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/DLA60x_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DLA60x_pretrained.pdparams",
     "DLA60x_c":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/DLA60x_c_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DLA60x_c_pretrained.pdparams",
     "DLA102":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/DLA102_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DLA102_pretrained.pdparams",
     "DLA102x":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/DLA102x_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DLA102x_pretrained.pdparams",
     "DLA102x2":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/DLA102x2_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DLA102x2_pretrained.pdparams",
     "DLA169":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/DLA169_pretrained.pdparams"
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DLA169_pretrained.pdparams"
 }
 
 

ppcls/arch/backbone/model_zoo/dpn.py

@@ -27,14 +27,16 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = [
-    "DPN",
-    "DPN68",
-    "DPN92",
-    "DPN98",
-    "DPN107",
-    "DPN131",
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"DPN68": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DPN68_pretrained.pdparams",
+              "DPN92": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DPN92_pretrained.pdparams",
+              "DPN98": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DPN98_pretrained.pdparams",
+              "DPN107": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DPN107_pretrained.pdparams",
+              "DPN131": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/DPN131_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -398,28 +400,45 @@ class DPN(nn.Layer):
         net_arg['init_padding'] = init_padding
 
         return net_arg
-
-
-def DPN68(**args):
-    model = DPN(layers=68, **args)
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )  
+
+
+def DPN68(pretrained=False, use_ssld=False, **kwargs):
+    model = DPN(layers=68, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DPN68"])
     return model
 
 
-def DPN92(**args):
-    model = DPN(layers=92, **args)
+def DPN92(pretrained=False, use_ssld=False, **kwargs):
+    model = DPN(layers=92, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DPN92"])
     return model
 
 
-def DPN98(**args):
-    model = DPN(layers=98, **args)
+def DPN98(pretrained=False, use_ssld=False, **kwargs):
+    model = DPN(layers=98, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DPN98"])
     return model
 
 
-def DPN107(**args):
-    model = DPN(layers=107, **args)
+def DPN107(pretrained=False, use_ssld=False, **kwargs):
+    model = DPN(layers=107, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DPN107"])
     return model
 
 
-def DPN131(**args):
-    model = DPN(layers=131, **args)
-    return model
+def DPN131(pretrained=False, use_ssld=False, **kwargs):
+    model = DPN(layers=131, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["DPN131"])
+    return model
\ No newline at end of file

ppcls/arch/backbone/model_zoo/efficientnet.py

@@ -9,11 +9,20 @@ import collections
 import re
 import copy
 
-__all__ = [
-    'EfficientNet', 'EfficientNetB0_small', 'EfficientNetB0', 'EfficientNetB1',
-    'EfficientNetB2', 'EfficientNetB3', 'EfficientNetB4', 'EfficientNetB5',
-    'EfficientNetB6', 'EfficientNetB7'
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"EfficientNetB0_small":  "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/EfficientNetB0_small_pretrained.pdparams",
+              "EfficientNetB0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/EfficientNetB0_pretrained.pdparams",
+              "EfficientNetB1": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/EfficientNetB1_pretrained.pdparams",
+              "EfficientNetB2": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/EfficientNetB2_pretrained.pdparams",
+              "EfficientNetB3": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/EfficientNetB3_pretrained.pdparams",
+              "EfficientNetB4": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/EfficientNetB4_pretrained.pdparams",
+              "EfficientNetB5": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/EfficientNetB5_pretrained.pdparams",
+              "EfficientNetB6": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/EfficientNetB6_pretrained.pdparams",
+              "EfficientNetB7": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/EfficientNetB7_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 GlobalParams = collections.namedtuple('GlobalParams', [
     'batch_norm_momentum',
@@ -783,119 +792,159 @@ class EfficientNet(nn.Layer):
         x = self._fc(x)
         return x
 
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )  
+
 
 def EfficientNetB0_small(padding_type='DYNAMIC',
                          override_params=None,
                          use_se=False,
-                         **args):
+                         pretrained=False, 
+                         use_ssld=False, 
+                         **kwargs):
     model = EfficientNet(
         name='b0',
         padding_type=padding_type,
         override_params=override_params,
         use_se=use_se,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["EfficientNetB0_small"])
     return model
 
 
 def EfficientNetB0(padding_type='SAME',
                    override_params=None,
                    use_se=True,
-                   **args):
+                   pretrained=False, 
+                   use_ssld=False, 
+                   **kwargs):
     model = EfficientNet(
         name='b0',
         padding_type=padding_type,
         override_params=override_params,
         use_se=use_se,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["EfficientNetB0"])
     return model
 
 
 def EfficientNetB1(padding_type='SAME',
                    override_params=None,
                    use_se=True,
-                   **args):
+                   pretrained=False, 
+                   use_ssld=False, 
+                   **kwargs):
     model = EfficientNet(
         name='b1',
         padding_type=padding_type,
         override_params=override_params,
         use_se=use_se,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["EfficientNetB1"])
     return model
 
 
 def EfficientNetB2(padding_type='SAME',
                    override_params=None,
                    use_se=True,
-                   **args):
+                   pretrained=False, 
+                   use_ssld=False, 
+                   **kwargs):
     model = EfficientNet(
         name='b2',
         padding_type=padding_type,
         override_params=override_params,
         use_se=use_se,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["EfficientNetB2"])
     return model
 
 
 def EfficientNetB3(padding_type='SAME',
                    override_params=None,
                    use_se=True,
-                   **args):
+                   pretrained=False, 
+                   use_ssld=False, 
+                   **kwargs):
     model = EfficientNet(
         name='b3',
         padding_type=padding_type,
         override_params=override_params,
         use_se=use_se,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["EfficientNetB3"])
     return model
 
 
 def EfficientNetB4(padding_type='SAME',
                    override_params=None,
                    use_se=True,
-                   **args):
+                   pretrained=False, 
+                   use_ssld=False, 
+                   **kwargs):
     model = EfficientNet(
         name='b4',
         padding_type=padding_type,
         override_params=override_params,
         use_se=use_se,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["EfficientNetB4"])
     return model
 
 
 def EfficientNetB5(padding_type='SAME',
                    override_params=None,
                    use_se=True,
-                   **args):
+                   pretrained=False, 
+                   use_ssld=False, 
+                   **kwargs):
     model = EfficientNet(
         name='b5',
         padding_type=padding_type,
         override_params=override_params,
         use_se=use_se,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["EfficientNetB5"])
     return model
 
 
 def EfficientNetB6(padding_type='SAME',
                    override_params=None,
                    use_se=True,
-                   **args):
+                   pretrained=False, 
+                   use_ssld=False, 
+                   **kwargs):
     model = EfficientNet(
         name='b6',
         padding_type=padding_type,
         override_params=override_params,
         use_se=use_se,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["EfficientNetB6"])
     return model
 
 
 def EfficientNetB7(padding_type='SAME',
                    override_params=None,
                    use_se=True,
-                   **args):
+                   pretrained=False, 
+                   use_ssld=False, 
+                   **kwargs):
     model = EfficientNet(
         name='b7',
         padding_type=padding_type,
         override_params=override_params,
         use_se=use_se,
-        **args)
-    return model
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["EfficientNetB7"])
+    return model
\ No newline at end of file

ppcls/arch/backbone/model_zoo/ghostnet.py

-# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
+# copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -21,7 +21,14 @@ from paddle.nn import Conv2D, BatchNorm, AdaptiveAvgPool2D, Linear
 from paddle.regularizer import L2Decay
 from paddle.nn.initializer import Uniform, KaimingNormal
 
-__all__ = ["GhostNet_x0_5", "GhostNet_x1_0", "GhostNet_x1_3"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"GhostNet_x0_5": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/GhostNet_x0_5_pretrained.pdparams",
+              "GhostNet_x1_0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/GhostNet_x1_0_pretrained.pdparams",
+              "GhostNet_x1_3": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/GhostNet_x1_3_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -315,17 +322,33 @@ class GhostNet(nn.Layer):
             new_v += divisor
         return new_v
 
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
 
-def GhostNet_x0_5(**args):
-    model = GhostNet(scale=0.5)
+def GhostNet_x0_5(pretrained=False, use_ssld=False, **kwargs):
+    model = GhostNet(scale=0.5, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["GhostNet_x0_5"], use_ssld=use_ssld)
     return model
 
 
-def GhostNet_x1_0(**args):
-    model = GhostNet(scale=1.0)
+def GhostNet_x1_0(pretrained=False, use_ssld=False, **kwargs):
+    model = GhostNet(scale=1.0, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["GhostNet_x1_0"], use_ssld=use_ssld)
     return model
 
 
-def GhostNet_x1_3(**args):
-    model = GhostNet(scale=1.3)
+def GhostNet_x1_3(pretrained=False, use_ssld=False, **kwargs):
+    model = GhostNet(scale=1.3, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["GhostNet_x1_3"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/googlenet.py

@@ -8,7 +8,12 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = ['GoogLeNet']
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"GoogLeNet": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/GoogLeNet_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 def xavier(channels, filter_size, name):
@@ -200,8 +205,22 @@ class GoogLeNetDY(nn.Layer):
         x = self._drop_o2(x)
         out2 = self._out2(x)
         return [out, out1, out2]
-
-
-def GoogLeNet(**args):
-    model = GoogLeNetDY(**args)
+    
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+
+def GoogLeNet(pretrained=False, use_ssld=False, **kwargs):
+    model = GoogLeNetDY(**kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["GoogLeNet"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/hardnet.py

@@ -20,13 +20,13 @@ from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_f
 
 MODEL_URLS = {
     'HarDNet39_ds':
-    'https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/HarDNet39_ds_pretrained.pdparams',
+    'https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/HarDNet39_ds_pretrained.pdparams',
     'HarDNet68_ds':
-    'https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/HarDNet68_ds_pretrained.pdparams',
+    'https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/HarDNet68_ds_pretrained.pdparams',
     'HarDNet68':
-    'https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/HarDNet68_pretrained.pdparams',
+    'https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/HarDNet68_pretrained.pdparams',
     'HarDNet85':
-    'https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/HarDNet85_pretrained.pdparams'
+    'https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/HarDNet85_pretrained.pdparams'
 }
 
 

ppcls/arch/backbone/model_zoo/hrnet.py

@@ -27,24 +27,18 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = [
-    "HRNet_W18_C",
-    "HRNet_W30_C",
-    "HRNet_W32_C",
-    "HRNet_W40_C",
-    "HRNet_W44_C",
-    "HRNet_W48_C",
-    "HRNet_W60_C",
-    "HRNet_W64_C",
-    "SE_HRNet_W18_C",
-    "SE_HRNet_W30_C",
-    "SE_HRNet_W32_C",
-    "SE_HRNet_W40_C",
-    "SE_HRNet_W44_C",
-    "SE_HRNet_W48_C",
-    "SE_HRNet_W60_C",
-    "SE_HRNet_W64_C",
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"HRNet_W18_C": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/HRNet_W18_C_pretrained.pdparams",
+              "HRNet_W30_C": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/HRNet_W30_C_pretrained.pdparams",
+              "HRNet_W32_C": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/HRNet_W32_C_pretrained.pdparams",
+              "HRNet_W40_C": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/HRNet_W40_C_pretrained.pdparams",
+              "HRNet_W44_C": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/HRNet_W44_C_pretrained.pdparams",
+              "HRNet_W48_C": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/HRNet_W48_C_pretrained.pdparams",
+              "HRNet_W64_C": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/HRNet_W64_C_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -661,82 +655,62 @@ class HRNet(nn.Layer):
         y = self.out(y)
         return y
 
-
-def HRNet_W18_C(**args):
-    model = HRNet(width=18, **args)
-    return model
-
-
-def HRNet_W30_C(**args):
-    model = HRNet(width=30, **args)
-    return model
-
-
-def HRNet_W32_C(**args):
-    model = HRNet(width=32, **args)
-    return model
-
-
-def HRNet_W40_C(**args):
-    model = HRNet(width=40, **args)
-    return model
-
-
-def HRNet_W44_C(**args):
-    model = HRNet(width=44, **args)
-    return model
-
-
-def HRNet_W48_C(**args):
-    model = HRNet(width=48, **args)
-    return model
-
-
-def HRNet_W60_C(**args):
-    model = HRNet(width=60, **args)
-    return model
-
-
-def HRNet_W64_C(**args):
-    model = HRNet(width=64, **args)
-    return model
-
-
-def SE_HRNet_W18_C(**args):
-    model = HRNet(width=18, has_se=True, **args)
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+
+def HRNet_W18_C(pretrained=False, use_ssld=False, **kwarg):
+    model = HRNet(width=18, **kwarg)
+    _load_pretrained(pretrained, model, MODEL_URLS["HRNet_W18_C"], use_ssld=use_ssld)
     return model
 
 
-def SE_HRNet_W30_C(**args):
-    model = HRNet(width=30, has_se=True, **args)
+def HRNet_W30_C(pretrained=False, use_ssld=False, **kwarg):
+    model = HRNet(width=30, **kwarg)
+    _load_pretrained(pretrained, model, MODEL_URLS["HRNet_W30_C"], use_ssld=use_ssld)
     return model
 
 
-def SE_HRNet_W32_C(**args):
-    model = HRNet(width=32, has_se=True, **args)
+def HRNet_W32_C(pretrained=False, use_ssld=False, **kwarg):
+    model = HRNet(width=32, **kwarg)
+    _load_pretrained(pretrained, model, MODEL_URLS["HRNet_W32_C"], use_ssld=use_ssld)
     return model
 
 
-def SE_HRNet_W40_C(**args):
-    model = HRNet(width=40, has_se=True, **args)
+def HRNet_W40_C(pretrained=False, use_ssld=False, **kwarg):
+    model = HRNet(width=40, **kwarg)
+    _load_pretrained(pretrained, model, MODEL_URLS["HRNet_W40_C"], use_ssld=use_ssld)
     return model
 
 
-def SE_HRNet_W44_C(**args):
-    model = HRNet(width=44, has_se=True, **args)
+def HRNet_W44_C(pretrained=False, use_ssld=False, **kwarg):
+    model = HRNet(width=44, **kwarg)
+    _load_pretrained(pretrained, model, MODEL_URLS["HRNet_W44_C"], use_ssld=use_ssld)
     return model
 
 
-def SE_HRNet_W48_C(**args):
-    model = HRNet(width=48, has_se=True, **args)
+def HRNet_W48_C(pretrained=False, use_ssld=False, **kwarg):
+    model = HRNet(width=48, **kwarg)
+    _load_pretrained(pretrained, model, MODEL_URLS["HRNet_W48_C"], use_ssld=use_ssld)
     return model
 
 
-def SE_HRNet_W60_C(**args):
-    model = HRNet(width=60, has_se=True, **args)
+def HRNet_W64_C(pretrained=False, use_ssld=False, **kwarg):
+    model = HRNet(width=64, **kwarg)
+    _load_pretrained(pretrained, model, MODEL_URLS["HRNet_W64_C"], use_ssld=use_ssld)
     return model
 
 
-def SE_HRNet_W64_C(**args):
-    model = HRNet(width=64, has_se=True, **args)
+def SE_HRNet_W64_C(pretrained=False, use_ssld=False, **kwarg):
+    model = HRNet(width=64, **kwarg)
+    _load_pretrained(pretrained, model, MODEL_URLS["SE_HRNet_W64_C"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/inception_v3.py

-# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
+# copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -26,7 +26,11 @@ from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 from paddle.nn.initializer import Uniform
 import math
 
-__all__ = ["InceptionV3"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"InceptionV3": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/InceptionV3_pretrained.pdparams"}
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -425,9 +429,9 @@ class InceptionE(nn.Layer):
         return outputs   
 
     
-class InceptionV3(nn.Layer):
+class Inception_V3(nn.Layer):
     def __init__(self, class_dim=1000):
-        super(InceptionV3, self).__init__()
+        super(Inception_V3, self).__init__()
         self.inception_a_list = [[192, 256, 288], [32, 64, 64]]
         self.inception_c_list = [[768, 768, 768, 768], [128, 160, 160, 192]]
         
@@ -472,10 +476,28 @@ class InceptionV3(nn.Layer):
     def forward(self, x):
         y = self.inception_stem(x)
         for inception_block in self.inception_block_list:
-           y = inception_block(y)
+            y = inception_block(y)
         y = self.gap(y)
         y = paddle.reshape(y, shape=[-1, 2048])
         y = self.drop(y)
         y = self.out(y)
         return y
     
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+def InceptionV3(pretrained=False, use_ssld=False, **kwargs):
+    model = Inception_V3(**kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["InceptionV3"], use_ssld=use_ssld)
+    return model
+

ppcls/arch/backbone/model_zoo/inception_v4.py

@@ -21,7 +21,11 @@ from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 from paddle.nn.initializer import Uniform
 import math
 
-__all__ = ["InceptionV4"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"InceptionV4": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/InceptionV4_pretrained.pdparams"}
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -450,6 +454,19 @@ class InceptionV4DY(nn.Layer):
         return x
 
 
-def InceptionV4(**args):
-    model = InceptionV4DY(**args)
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+def InceptionV4(pretrained=False, use_ssld=False, **kwargs):
+    model = InceptionV4DY(**kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["InceptionV4"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/levit.py

+# copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import itertools
+import math
+import warnings
+
+import paddle
+import paddle.nn as nn
+import paddle.nn.functional as F
+from paddle.nn.initializer import TruncatedNormal, Constant
+from paddle.regularizer import L2Decay
+
+from .vision_transformer import trunc_normal_, zeros_, ones_, Identity
+
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "LeViT_128S": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/LeViT_128S_pretrained.pdparams",
+              "LeViT_128": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/LeViT_128_pretrained.pdparams",
+              "LeViT_192": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/LeViT_192_pretrained.pdparams",
+              "LeViT_256": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/LeViT_256_pretrained.pdparams",
+              "LeViT_384": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/LeViT_384_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
+
+
+def cal_attention_biases(attention_biases, attention_bias_idxs):
+    gather_list = []
+    attention_bias_t = paddle.transpose(attention_biases, (1, 0))
+    for idx in attention_bias_idxs:
+        gather = paddle.gather(attention_bias_t, idx)
+        gather_list.append(gather)
+    shape0, shape1 = attention_bias_idxs.shape
+    return paddle.transpose(paddle.concat(gather_list), (1, 0)).reshape(
+        (0, shape0, shape1))
+
+
+class Conv2d_BN(nn.Sequential):
+    def __init__(self,
+                 a,
+                 b,
+                 ks=1,
+                 stride=1,
+                 pad=0,
+                 dilation=1,
+                 groups=1,
+                 bn_weight_init=1,
+                 resolution=-10000):
+        super().__init__()
+        self.add_sublayer(
+            'c',
+            nn.Conv2D(
+                a, b, ks, stride, pad, dilation, groups, bias_attr=False))
+        bn = nn.BatchNorm2D(b)
+        ones_(bn.weight)
+        zeros_(bn.bias)
+        self.add_sublayer('bn', bn)
+
+
+class Linear_BN(nn.Sequential):
+    def __init__(self, a, b, bn_weight_init=1):
+        super().__init__()
+        self.add_sublayer('c', nn.Linear(a, b, bias_attr=False))
+        bn = nn.BatchNorm1D(b)
+        ones_(bn.weight)
+        zeros_(bn.bias)
+        self.add_sublayer('bn', bn)
+
+    def forward(self, x):
+        l, bn = self._sub_layers.values()
+        x = l(x)
+        return paddle.reshape(bn(x.flatten(0, 1)), x.shape)
+
+
+class BN_Linear(nn.Sequential):
+    def __init__(self, a, b, bias=True, std=0.02):
+        super().__init__()
+        self.add_sublayer('bn', nn.BatchNorm1D(a))
+        l = nn.Linear(a, b, bias_attr=bias)
+        trunc_normal_(l.weight)
+        if bias:
+            zeros_(l.bias)
+        self.add_sublayer('l', l)
+
+
+def b16(n, activation, resolution=224):
+    return nn.Sequential(
+        Conv2d_BN(
+            3, n // 8, 3, 2, 1, resolution=resolution),
+        activation(),
+        Conv2d_BN(
+            n // 8, n // 4, 3, 2, 1, resolution=resolution // 2),
+        activation(),
+        Conv2d_BN(
+            n // 4, n // 2, 3, 2, 1, resolution=resolution // 4),
+        activation(),
+        Conv2d_BN(
+            n // 2, n, 3, 2, 1, resolution=resolution // 8))
+
+
+class Residual(nn.Layer):
+    def __init__(self, m, drop):
+        super().__init__()
+        self.m = m
+        self.drop = drop
+
+    def forward(self, x):
+        if self.training and self.drop > 0:
+            return x + self.m(x) * paddle.rand(
+                x.size(0), 1, 1,
+                device=x.device).ge_(self.drop).div(1 - self.drop).detach()
+        else:
+            return x + self.m(x)
+
+
+class Attention(nn.Layer):
+    def __init__(self,
+                 dim,
+                 key_dim,
+                 num_heads=8,
+                 attn_ratio=4,
+                 activation=None,
+                 resolution=14):
+        super().__init__()
+        self.num_heads = num_heads
+        self.scale = key_dim**-0.5
+        self.key_dim = key_dim
+        self.nh_kd = nh_kd = key_dim * num_heads
+        self.d = int(attn_ratio * key_dim)
+        self.dh = int(attn_ratio * key_dim) * num_heads
+        self.attn_ratio = attn_ratio
+        self.h = self.dh + nh_kd * 2
+        self.qkv = Linear_BN(dim, self.h)
+        self.proj = nn.Sequential(
+            activation(), Linear_BN(
+                self.dh, dim, bn_weight_init=0))
+        points = list(itertools.product(range(resolution), range(resolution)))
+        N = len(points)
+        attention_offsets = {}
+        idxs = []
+        for p1 in points:
+            for p2 in points:
+                offset = (abs(p1[0] - p2[0]), abs(p1[1] - p2[1]))
+                if offset not in attention_offsets:
+                    attention_offsets[offset] = len(attention_offsets)
+                idxs.append(attention_offsets[offset])
+        self.attention_biases = self.create_parameter(
+            shape=(num_heads, len(attention_offsets)),
+            default_initializer=zeros_,
+            attr=paddle.ParamAttr(regularizer=L2Decay(0.0)))
+        tensor_idxs = paddle.to_tensor(idxs, dtype='int64')
+        self.register_buffer('attention_bias_idxs',
+                             paddle.reshape(tensor_idxs, [N, N]))
+
+    @paddle.no_grad()
+    def train(self, mode=True):
+        if mode:
+            super().train()
+        else:
+            super().eval()
+        if mode and hasattr(self, 'ab'):
+            del self.ab
+        else:
+            self.ab = cal_attention_biases(self.attention_biases,
+                                           self.attention_bias_idxs)
+
+    def forward(self, x):
+        self.training = True
+        B, N, C = x.shape
+        qkv = self.qkv(x)
+        qkv = paddle.reshape(qkv,
+                             [B, N, self.num_heads, self.h // self.num_heads])
+        q, k, v = paddle.split(
+            qkv, [self.key_dim, self.key_dim, self.d], axis=3)
+        q = paddle.transpose(q, perm=[0, 2, 1, 3])
+        k = paddle.transpose(k, perm=[0, 2, 1, 3])
+        v = paddle.transpose(v, perm=[0, 2, 1, 3])
+        k_transpose = paddle.transpose(k, perm=[0, 1, 3, 2])
+
+        if self.training:
+            attention_biases = cal_attention_biases(self.attention_biases,
+                                                    self.attention_bias_idxs)
+        else:
+            attention_biases = self.ab
+        attn = ((q @k_transpose) * self.scale + attention_biases)
+        attn = F.softmax(attn)
+        x = paddle.transpose(attn @v, perm=[0, 2, 1, 3])
+        x = paddle.reshape(x, [B, N, self.dh])
+        x = self.proj(x)
+        return x
+
+
+class Subsample(nn.Layer):
+    def __init__(self, stride, resolution):
+        super().__init__()
+        self.stride = stride
+        self.resolution = resolution
+
+    def forward(self, x):
+        B, N, C = x.shape
+        x = paddle.reshape(x, [B, self.resolution, self.resolution,
+                               C])[:, ::self.stride, ::self.stride]
+        x = paddle.reshape(x, [B, -1, C])
+        return x
+
+
+class AttentionSubsample(nn.Layer):
+    def __init__(self,
+                 in_dim,
+                 out_dim,
+                 key_dim,
+                 num_heads=8,
+                 attn_ratio=2,
+                 activation=None,
+                 stride=2,
+                 resolution=14,
+                 resolution_=7):
+        super().__init__()
+        self.num_heads = num_heads
+        self.scale = key_dim**-0.5
+        self.key_dim = key_dim
+        self.nh_kd = nh_kd = key_dim * num_heads
+        self.d = int(attn_ratio * key_dim)
+        self.dh = int(attn_ratio * key_dim) * self.num_heads
+        self.attn_ratio = attn_ratio
+        self.resolution_ = resolution_
+        self.resolution_2 = resolution_**2
+        self.training = True
+        h = self.dh + nh_kd
+        self.kv = Linear_BN(in_dim, h)
+
+        self.q = nn.Sequential(
+            Subsample(stride, resolution), Linear_BN(in_dim, nh_kd))
+        self.proj = nn.Sequential(activation(), Linear_BN(self.dh, out_dim))
+
+        self.stride = stride
+        self.resolution = resolution
+        points = list(itertools.product(range(resolution), range(resolution)))
+        points_ = list(
+            itertools.product(range(resolution_), range(resolution_)))
+
+        N = len(points)
+        N_ = len(points_)
+        attention_offsets = {}
+        idxs = []
+        i = 0
+        j = 0
+        for p1 in points_:
+            i += 1
+            for p2 in points:
+                j += 1
+                size = 1
+                offset = (abs(p1[0] * stride - p2[0] + (size - 1) / 2),
+                          abs(p1[1] * stride - p2[1] + (size - 1) / 2))
+                if offset not in attention_offsets:
+                    attention_offsets[offset] = len(attention_offsets)
+                idxs.append(attention_offsets[offset])
+        self.attention_biases = self.create_parameter(
+            shape=(num_heads, len(attention_offsets)),
+            default_initializer=zeros_,
+            attr=paddle.ParamAttr(regularizer=L2Decay(0.0)))
+
+        tensor_idxs_ = paddle.to_tensor(idxs, dtype='int64')
+        self.register_buffer('attention_bias_idxs',
+                             paddle.reshape(tensor_idxs_, [N_, N]))
+
+    @paddle.no_grad()
+    def train(self, mode=True):
+        if mode:
+            super().train()
+        else:
+            super().eval()
+        if mode and hasattr(self, 'ab'):
+            del self.ab
+        else:
+            self.ab = cal_attention_biases(self.attention_biases,
+                                           self.attention_bias_idxs)
+
+    def forward(self, x):
+        self.training = True
+        B, N, C = x.shape
+        kv = self.kv(x)
+        kv = paddle.reshape(kv, [B, N, self.num_heads, -1])
+        k, v = paddle.split(kv, [self.key_dim, self.d], axis=3)
+        k = paddle.transpose(k, perm=[0, 2, 1, 3])  # BHNC
+        v = paddle.transpose(v, perm=[0, 2, 1, 3])
+        q = paddle.reshape(
+            self.q(x), [B, self.resolution_2, self.num_heads, self.key_dim])
+        q = paddle.transpose(q, perm=[0, 2, 1, 3])
+
+        if self.training:
+            attention_biases = cal_attention_biases(self.attention_biases,
+                                                    self.attention_bias_idxs)
+        else:
+            attention_biases = self.ab
+
+        attn = (q @paddle.transpose(
+            k, perm=[0, 1, 3, 2])) * self.scale + attention_biases
+        attn = F.softmax(attn)
+
+        x = paddle.reshape(
+            paddle.transpose(
+                (attn @v), perm=[0, 2, 1, 3]), [B, -1, self.dh])
+        x = self.proj(x)
+        return x
+
+
+class LeViT(nn.Layer):
+    """ Vision Transformer with support for patch or hybrid CNN input stage
+    """
+
+    def __init__(self,
+                 img_size=224,
+                 patch_size=16,
+                 in_chans=3,
+                 class_dim=1000,
+                 embed_dim=[192],
+                 key_dim=[64],
+                 depth=[12],
+                 num_heads=[3],
+                 attn_ratio=[2],
+                 mlp_ratio=[2],
+                 hybrid_backbone=None,
+                 down_ops=[],
+                 attention_activation=nn.Hardswish,
+                 mlp_activation=nn.Hardswish,
+                 distillation=True,
+                 drop_path=0):
+        super().__init__()
+
+        self.class_dim = class_dim
+        self.num_features = embed_dim[-1]
+        self.embed_dim = embed_dim
+        self.distillation = distillation
+
+        self.patch_embed = hybrid_backbone
+
+        self.blocks = []
+        down_ops.append([''])
+        resolution = img_size // patch_size
+        for i, (ed, kd, dpth, nh, ar, mr, do) in enumerate(
+                zip(embed_dim, key_dim, depth, num_heads, attn_ratio,
+                    mlp_ratio, down_ops)):
+            for _ in range(dpth):
+                self.blocks.append(
+                    Residual(
+                        Attention(
+                            ed,
+                            kd,
+                            nh,
+                            attn_ratio=ar,
+                            activation=attention_activation,
+                            resolution=resolution, ),
+                        drop_path))
+                if mr > 0:
+                    h = int(ed * mr)
+                    self.blocks.append(
+                        Residual(
+                            nn.Sequential(
+                                Linear_BN(ed, h),
+                                mlp_activation(),
+                                Linear_BN(
+                                    h, ed, bn_weight_init=0), ),
+                            drop_path))
+            if do[0] == 'Subsample':
+                #('Subsample',key_dim, num_heads, attn_ratio, mlp_ratio, stride)
+                resolution_ = (resolution - 1) // do[5] + 1
+                self.blocks.append(
+                    AttentionSubsample(
+                        *embed_dim[i:i + 2],
+                        key_dim=do[1],
+                        num_heads=do[2],
+                        attn_ratio=do[3],
+                        activation=attention_activation,
+                        stride=do[5],
+                        resolution=resolution,
+                        resolution_=resolution_))
+                resolution = resolution_
+                if do[4] > 0:  # mlp_ratio
+                    h = int(embed_dim[i + 1] * do[4])
+                    self.blocks.append(
+                        Residual(
+                            nn.Sequential(
+                                Linear_BN(embed_dim[i + 1], h),
+                                mlp_activation(),
+                                Linear_BN(
+                                    h, embed_dim[i + 1], bn_weight_init=0), ),
+                            drop_path))
+        self.blocks = nn.Sequential(*self.blocks)
+
+        # Classifier head
+        self.head = BN_Linear(embed_dim[-1],
+                              class_dim) if class_dim > 0 else Identity()
+        if distillation:
+            self.head_dist = BN_Linear(
+                embed_dim[-1], class_dim) if class_dim > 0 else Identity()
+
+    def forward(self, x):
+        x = self.patch_embed(x)
+        x = x.flatten(2)
+        x = paddle.transpose(x, perm=[0, 2, 1])
+        x = self.blocks(x)
+        x = x.mean(1)
+        if self.distillation:
+            x = self.head(x), self.head_dist(x)
+            if not self.training:
+                x = (x[0] + x[1]) / 2
+        else:
+            x = self.head(x)
+        return x
+
+
+def model_factory(C, D, X, N, drop_path, class_dim, distillation):
+    embed_dim = [int(x) for x in C.split('_')]
+    num_heads = [int(x) for x in N.split('_')]
+    depth = [int(x) for x in X.split('_')]
+    act = nn.Hardswish
+    model = LeViT(
+        patch_size=16,
+        embed_dim=embed_dim,
+        num_heads=num_heads,
+        key_dim=[D] * 3,
+        depth=depth,
+        attn_ratio=[2, 2, 2],
+        mlp_ratio=[2, 2, 2],
+        down_ops=[
+            #('Subsample',key_dim, num_heads, attn_ratio, mlp_ratio, stride)
+            ['Subsample', D, embed_dim[0] // D, 4, 2, 2],
+            ['Subsample', D, embed_dim[1] // D, 4, 2, 2],
+        ],
+        attention_activation=act,
+        mlp_activation=act,
+        hybrid_backbone=b16(embed_dim[0], activation=act),
+        class_dim=class_dim,
+        drop_path=drop_path,
+        distillation=distillation)
+
+    return model
+
+
+specification = {
+    'LeViT_128S': {
+        'C': '128_256_384',
+        'D': 16,
+        'N': '4_6_8',
+        'X': '2_3_4',
+        'drop_path': 0
+    },
+    'LeViT_128': {
+        'C': '128_256_384',
+        'D': 16,
+        'N': '4_8_12',
+        'X': '4_4_4',
+        'drop_path': 0
+    },
+    'LeViT_192': {
+        'C': '192_288_384',
+        'D': 32,
+        'N': '3_5_6',
+        'X': '4_4_4',
+        'drop_path': 0
+    },
+    'LeViT_256': {
+        'C': '256_384_512',
+        'D': 32,
+        'N': '4_6_8',
+        'X': '4_4_4',
+        'drop_path': 0
+    },
+    'LeViT_384': {
+        'C': '384_512_768',
+        'D': 32,
+        'N': '6_9_12',
+        'X': '4_4_4',
+        'drop_path': 0.1
+    },
+}
+
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+
+def LeViT_128S(pretrained=False, use_ssld=False, class_dim=1000, distillation=False, **kwargs):
+    model = model_factory(
+        **specification['LeViT_128S'],
+        class_dim=class_dim,
+        distillation=distillation)
+    _load_pretrained(pretrained, model, MODEL_URLS["LeViT_128S"], use_ssld=use_ssld)
+    return model
+
+
+def LeViT_128(pretrained=False, use_ssld=False, class_dim=1000, distillation=False, **kwargs):
+    model = model_factory(
+        **specification['LeViT_128'],
+        class_dim=class_dim,
+        distillation=distillation)
+    _load_pretrained(pretrained, model, MODEL_URLS["LeViT_128"], use_ssld=use_ssld)
+    return model
+
+
+def LeViT_192(pretrained=False, use_ssld=False, class_dim=1000, distillation=False, **kwargs):
+    model = model_factory(
+        **specification['LeViT_192'],
+        class_dim=class_dim,
+        distillation=distillation)
+    _load_pretrained(pretrained, model, MODEL_URLS["LeViT_192"], use_ssld=use_ssld)
+    return model
+
+
+def LeViT_256(pretrained=False, use_ssld=False, class_dim=1000, distillation=False, **kwargs):
+    model = model_factory(
+        **specification['LeViT_256'],
+        class_dim=class_dim,
+        distillation=distillation)
+    _load_pretrained(pretrained, model, MODEL_URLS["LeViT_256"], use_ssld=use_ssld)
+    return model
+
+
+def LeViT_384(pretrained=False, use_ssld=False, class_dim=1000, distillation=False, **kwargs):
+    model = model_factory(
+        **specification['LeViT_384'],
+        class_dim=class_dim,
+        distillation=distillation)
+    _load_pretrained(pretrained, model, MODEL_URLS["LeViT_384"], use_ssld=use_ssld)
+    return model

ppcls/arch/backbone/model_zoo/mixnet.py

@@ -17,14 +17,20 @@
     https://arxiv.org/abs/1907.09595.
 """
 
-__all__ = ['MixNet_S', 'MixNet_M', 'MixNet_L']
-
 import os
 from inspect import isfunction
 from functools import reduce
 import paddle
 import paddle.nn as nn
 
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"MixNet_S": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MixNet_S_pretrained.pdparams", 
+              "MixNet_M": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MixNet_M_pretrained.pdparams", 
+              "MixNet_L": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MixNet_L_pretrained.pdparams"}
+
+__all__ = list(MODEL_URLS.keys())
+
 
 class Identity(nn.Layer):
     """
@@ -755,13 +761,33 @@ def get_mixnet(version, width_scale, model_name=None, **kwargs):
     return net
 
 
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+def MixNet_S(pretrained=False, use_ssld=False, **kwargs):
+    model = InceptionV4DY(**kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["InceptionV4"], use_ssld=use_ssld)
+    return model
+
+
 def MixNet_S(**kwargs):
     """
     MixNet-S model from 'MixConv: Mixed Depthwise Convolutional Kernels,'
     https://arxiv.org/abs/1907.09595.
     """
-    return get_mixnet(
+    model = get_mixnet(
         version="s", width_scale=1.0, model_name="MixNet_S", **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MixNet_S"], use_ssld=use_ssld)
+    return model
 
 
 def MixNet_M(**kwargs):
@@ -769,14 +795,19 @@ def MixNet_M(**kwargs):
     MixNet-M model from 'MixConv: Mixed Depthwise Convolutional Kernels,'
     https://arxiv.org/abs/1907.09595.
     """
-    return get_mixnet(
+    model = get_mixnet(
         version="m", width_scale=1.0, model_name="MixNet_M", **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MixNet_M"], use_ssld=use_ssld)
+    return model
 
 
 def MixNet_L(**kwargs):
     """
-    MixNet-L model from 'MixConv: Mixed Depthwise Convolutional Kernels,'
+    MixNet-S model from 'MixConv: Mixed Depthwise Convolutional Kernels,'
     https://arxiv.org/abs/1907.09595.
     """
-    return get_mixnet(
+    model = get_mixnet(
         version="m", width_scale=1.3, model_name="MixNet_L", **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MixNet_L"], use_ssld=use_ssld)
+    return model
+

ppcls/arch/backbone/model_zoo/mobilenet_v1.py

@@ -26,9 +26,14 @@ from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 from paddle.nn.initializer import KaimingNormal
 import math
 
-__all__ = [
-    "MobileNetV1_x0_25", "MobileNetV1_x0_5", "MobileNetV1_x0_75", "MobileNetV1"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"MobileNetV1_x0_25": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV1_x0_25_pretrained.pdparams", 
+              "MobileNetV1_x0_5": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV1_x0_5_pretrained.pdparams", 
+              "MobileNetV1_x0_75": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV1_x0_75_pretrained.pdparams",
+              "MobileNetV1": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV1_pretrained.pdparams"}
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -245,22 +250,39 @@ class MobileNet(nn.Layer):
         y = self.out(y)
         return y
 
-
-def MobileNetV1_x0_25(**args):
-    model = MobileNet(scale=0.25, **args)
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+
+def MobileNetV1_x0_25(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNet(scale=0.25, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV1_x0_25"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV1_x0_5(**args):
-    model = MobileNet(scale=0.5, **args)
+def MobileNetV1_x0_5(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNet(scale=0.5, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV1_x0_5"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV1_x0_75(**args):
-    model = MobileNet(scale=0.75, **args)
+def MobileNetV1_x0_75(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNet(scale=0.75, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV1_x0_75"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV1(**args):
-    model = MobileNet(scale=1.0, **args)
-    return model
+def MobileNetV1(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNet(scale=1.0, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV1"], use_ssld=use_ssld)
+    return model
\ No newline at end of file

ppcls/arch/backbone/model_zoo/mobilenet_v2.py

@@ -26,10 +26,16 @@ from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 
 import math
 
-__all__ = [
-    "MobileNetV2_x0_25", "MobileNetV2_x0_5", "MobileNetV2_x0_75",
-    "MobileNetV2", "MobileNetV2_x1_5", "MobileNetV2_x2_0"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"MobileNetV2_x0_25": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV2_x0_25_pretrained.pdparams", 
+              "MobileNetV2_x0_5": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV2_x0_5_pretrained.pdparams", 
+              "MobileNetV2_x0_75": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV2_x0_75_pretrained.pdparams",
+              "MobileNetV2": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV2_pretrained.pdparams",
+              "MobileNetV2_x1_5": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV2_x1_5_pretrained.pdparams",
+              "MobileNetV2_x2_0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV2_x2_0_pretrained.pdparams"}
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -149,7 +155,7 @@ class InvresiBlocks(nn.Layer):
 
 
 class MobileNet(nn.Layer):
-    def __init__(self, class_dim=1000, scale=1.0, prefix_name="", **args):
+    def __init__(self, class_dim=1000, scale=1.0, prefix_name=""):
         super(MobileNet, self).__init__()
         self.scale = scale
         self.class_dim = class_dim
@@ -216,33 +222,52 @@ class MobileNet(nn.Layer):
         y = paddle.flatten(y, start_axis=1, stop_axis=-1)
         y = self.out(y)
         return y
-
-
-def MobileNetV2_x0_25(**args):
-    model = MobileNet(scale=0.25, **args)
+    
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+
+def MobileNetV2_x0_25(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNet(scale=0.25, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV2_x0_25"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV2_x0_5(**args):
-    model = MobileNet(scale=0.5, **args)
+def MobileNetV2_x0_5(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNet(scale=0.5, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV2_x0_5"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV2_x0_75(**args):
-    model = MobileNet(scale=0.75, **args)
+def MobileNetV2_x0_75(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNet(scale=0.75, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV2_x0_75"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV2(**args):
-    model = MobileNet(scale=1.0, **args)
+def MobileNetV2(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNet(scale=1.0, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV2"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV2_x1_5(**args):
-    model = MobileNet(scale=1.5, **args)
+def MobileNetV2_x1_5(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNet(scale=1.5, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV2_x1_5"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV2_x2_0(**args):
-    model = MobileNet(scale=2.0, **args)
+def MobileNetV2_x2_0(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNet(scale=2.0, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV2_x2_0"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/mobilenet_v3.py

@@ -28,13 +28,20 @@ from paddle.regularizer import L2Decay
 
 import math
 
-__all__ = [
-    "MobileNetV3_small_x0_35", "MobileNetV3_small_x0_5",
-    "MobileNetV3_small_x0_75", "MobileNetV3_small_x1_0",
-    "MobileNetV3_small_x1_25", "MobileNetV3_large_x0_35",
-    "MobileNetV3_large_x0_5", "MobileNetV3_large_x0_75",
-    "MobileNetV3_large_x1_0", "MobileNetV3_large_x1_25"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"MobileNetV3_small_x0_35": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV3_small_x0_35_pretrained.pdparams", 
+             "MobileNetV3_small_x0_5": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV3_small_x0_5_pretrained.pdparams", 
+             "MobileNetV3_small_x0_75": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV3_small_x0_75_pretrained.pdparams", 
+             "MobileNetV3_small_x1_0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV3_small_x1_0_pretrained.pdparams", 
+             "MobileNetV3_small_x1_25": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV3_small_x1_25_pretrained.pdparams",
+             "MobileNetV3_large_x0_35": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV3_large_x0_35_pretrained.pdparams", 
+             "MobileNetV3_large_x0_5": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV3_large_x0_5_pretrained.pdparams", 
+             "MobileNetV3_large_x0_75": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV3_large_x0_75_pretrained.pdparams", 
+             "MobileNetV3_large_x1_0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV3_large_x1_0_pretrained.pdparams", 
+             "MobileNetV3_large_x1_25": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/MobileNetV3_large_x1_25_pretrained.pdparams"}
+
+__all__ = list(MODEL_URLS.keys())
 
 
 def make_divisible(v, divisor=8, min_value=None):
@@ -308,52 +315,75 @@ class SEModule(nn.Layer):
         outputs = hardsigmoid(outputs, slope=0.2, offset=0.5)
         return paddle.multiply(x=inputs, y=outputs)
 
-
-def MobileNetV3_small_x0_35(**args):
-    model = MobileNetV3(model_name="small", scale=0.35, **args)
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+
+def MobileNetV3_small_x0_35(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNetV3(model_name="small", scale=0.35, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV3_small_x0_35"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV3_small_x0_5(**args):
-    model = MobileNetV3(model_name="small", scale=0.5, **args)
+def MobileNetV3_small_x0_5(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNetV3(model_name="small", scale=0.5, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV3_small_x0_5"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV3_small_x0_75(**args):
-    model = MobileNetV3(model_name="small", scale=0.75, **args)
+def MobileNetV3_small_x0_75(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNetV3(model_name="small", scale=0.75, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV3_small_x0_75"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV3_small_x1_0(**args):
-    model = MobileNetV3(model_name="small", scale=1.0, **args)
+def MobileNetV3_small_x1_0(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNetV3(model_name="small", scale=1.0, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV3_small_x1_0"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV3_small_x1_25(**args):
-    model = MobileNetV3(model_name="small", scale=1.25, **args)
+def MobileNetV3_small_x1_25(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNetV3(model_name="small", scale=1.25, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV3_small_x1_25"], use_ssld=use_ssld)
     return model
 
-
-def MobileNetV3_large_x0_35(**args):
-    model = MobileNetV3(model_name="large", scale=0.35, **args)
+                        
+def MobileNetV3_large_x0_35(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNetV3(model_name="large", scale=0.35, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV3_large_x0_35"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV3_large_x0_5(**args):
-    model = MobileNetV3(model_name="large", scale=0.5, **args)
+def MobileNetV3_large_x0_5(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNetV3(model_name="large", scale=0.5, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV3_large_x0_5"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV3_large_x0_75(**args):
-    model = MobileNetV3(model_name="large", scale=0.75, **args)
+def MobileNetV3_large_x0_75(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNetV3(model_name="large", scale=0.75, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV3_large_x0_75"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV3_large_x1_0(**args):
-    model = MobileNetV3(model_name="large", scale=1.0, **args)
+def MobileNetV3_large_x1_0(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNetV3(model_name="large", scale=1.0, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV3_large_x1_0"], use_ssld=use_ssld)
     return model
 
 
-def MobileNetV3_large_x1_25(**args):
-    model = MobileNetV3(model_name="large", scale=1.25, **args)
+def MobileNetV3_large_x1_25(pretrained=False, use_ssld=False, **kwargs):
+    model = MobileNetV3(model_name="large", scale=1.25, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["MobileNetV3_large_x1_25"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/rednet.py

@@ -22,15 +22,15 @@ from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_f
 
 MODEL_URLS = {
     "RedNet26":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/RedNet26_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RedNet26_pretrained.pdparams",
     "RedNet38":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/RedNet38_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RedNet38_pretrained.pdparams",
     "RedNet50":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/RedNet50_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RedNet50_pretrained.pdparams",
     "RedNet101":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/RedNet101_pretrained.pdparams",
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RedNet101_pretrained.pdparams",
     "RedNet152":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/RedNet152_pretrained.pdparams"
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RedNet152_pretrained.pdparams"
 }
 
 

ppcls/arch/backbone/model_zoo/regnet.py

@@ -26,10 +26,17 @@ from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 from paddle.nn.initializer import Uniform
 import math
 
-__all__ = [
-    "RegNetX_200MF", "RegNetX_4GF", "RegNetX_32GF", "RegNetY_200MF",
-    "RegNetY_4GF", "RegNetY_32GF"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"RegNetX_200MF": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RegNetX_200MF_pretrained.pdparams", 
+              "RegNetX_4GF": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RegNetX_4GF_pretrained.pdparams", 
+              "RegNetX_32GF": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RegNetX_32GF_pretrained.pdparams", 
+              "RegNetY_200MF": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RegNetY_200MF_pretrained.pdparams", 
+              "RegNetY_4GF": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RegNetY_4GF_pretrained.pdparams", 
+              "RegNetY_32GF": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RegNetY_32GF_pretrained.pdparams", 
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 def quantize_float(f, q):
@@ -308,14 +315,28 @@ class RegNet(nn.Layer):
         y = self.out(y)
         return y
 
-
-def RegNetX_200MF(**args):
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+    
+def RegNetX_200MF(pretrained=False, use_ssld=False, **kwargs):
     model = RegNet(
-        w_a=36.44, w_0=24, w_m=2.49, d=13, group_w=8, bot_mul=1.0, q=8, **args)
+        w_a=36.44, w_0=24, w_m=2.49, d=13, group_w=8, bot_mul=1.0, q=8, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RegNetX_200MF"], use_ssld=use_ssld)
     return model
 
 
-def RegNetX_4GF(**args):
+def RegNetX_4GF(pretrained=False, use_ssld=False, **kwargs):
     model = RegNet(
         w_a=38.65,
         w_0=96,
@@ -324,11 +345,12 @@ def RegNetX_4GF(**args):
         group_w=40,
         bot_mul=1.0,
         q=8,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RegNetX_4GF"], use_ssld=use_ssld)
     return model
 
 
-def RegNetX_32GF(**args):
+def RegNetX_32GF(pretrained=False, use_ssld=False, **kwargs):
     model = RegNet(
         w_a=69.86,
         w_0=320,
@@ -337,11 +359,12 @@ def RegNetX_32GF(**args):
         group_w=168,
         bot_mul=1.0,
         q=8,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RegNetX_32GF"], use_ssld=use_ssld)
     return model
 
 
-def RegNetY_200MF(**args):
+def RegNetY_200MF(pretrained=False, use_ssld=False, **kwargs):
     model = RegNet(
         w_a=36.44,
         w_0=24,
@@ -351,11 +374,12 @@ def RegNetY_200MF(**args):
         bot_mul=1.0,
         q=8,
         se_on=True,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RegNetX_32GF"], use_ssld=use_ssld)
     return model
 
 
-def RegNetY_4GF(**args):
+def RegNetY_4GF(pretrained=False, use_ssld=False, **kwargs):
     model = RegNet(
         w_a=31.41,
         w_0=96,
@@ -365,11 +389,12 @@ def RegNetY_4GF(**args):
         bot_mul=1.0,
         q=8,
         se_on=True,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RegNetX_32GF"], use_ssld=use_ssld)
     return model
 
 
-def RegNetY_32GF(**args):
+def RegNetY_32GF(pretrained=False, use_ssld=False, **kwargs):
     model = RegNet(
         w_a=115.89,
         w_0=232,
@@ -379,5 +404,6 @@ def RegNetY_32GF(**args):
         bot_mul=1.0,
         q=8,
         se_on=True,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RegNetX_32GF"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/repvgg.py

@@ -2,22 +2,29 @@ import paddle.nn as nn
 import paddle
 import numpy as np
 
-__all__ = [
-    'RepVGG',
-    'RepVGG_A0',
-    'RepVGG_A1',
-    'RepVGG_A2',
-    'RepVGG_B0',
-    'RepVGG_B1',
-    'RepVGG_B2',
-    'RepVGG_B3',
-    'RepVGG_B1g2',
-    'RepVGG_B1g4',
-    'RepVGG_B2g2',
-    'RepVGG_B2g4',
-    'RepVGG_B3g2',
-    'RepVGG_B3g4',
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"RepVGG_A0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_A0_pretrained.pdparams",
+              "RepVGG_A1": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_A1_pretrained.pdparams",
+              "RepVGG_A2": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_A2_pretrained.pdparams",
+              "RepVGG_B0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_B0_pretrained.pdparams",
+              "RepVGG_B1": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_B1_pretrained.pdparams",
+              "RepVGG_B2": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_B2_pretrained.pdparams",
+              "RepVGG_B3": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_B3_pretrained.pdparams",
+              "RepVGG_B1g2": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_B1g2_pretrained.pdparams",
+              "RepVGG_B1g4": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_B1g4_pretrained.pdparams",
+              "RepVGG_B2g2": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_B2g2_pretrained.pdparams",
+              "RepVGG_B2g4": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_B2g4_pretrained.pdparams",
+              "RepVGG_B3g2": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_B3g2_pretrained.pdparams",
+              "RepVGG_B3g4": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/RepVGG_B3g4_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
+
+
+optional_groupwise_layers = [2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26]
+g2_map = {l: 2 for l in optional_groupwise_layers}
+g4_map = {l: 4 for l in optional_groupwise_layers}
 
 
 class ConvBN(nn.Layer):
@@ -230,110 +237,144 @@ class RepVGG(nn.Layer):
         return out
 
 
-optional_groupwise_layers = [2, 4, 6, 8, 10, 12, 14, 16, 18, 20, 22, 24, 26]
-g2_map = {l: 2 for l in optional_groupwise_layers}
-g4_map = {l: 4 for l in optional_groupwise_layers}
-
-
-def RepVGG_A0(**kwargs):
-    return RepVGG(
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+
+def RepVGG_A0(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[2, 4, 14, 1],
         width_multiplier=[0.75, 0.75, 0.75, 2.5],
         override_groups_map=None,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_A0"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_A1(**kwargs):
-    return RepVGG(
+def RepVGG_A1(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[2, 4, 14, 1],
         width_multiplier=[1, 1, 1, 2.5],
         override_groups_map=None,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_A1"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_A2(**kwargs):
-    return RepVGG(
+def RepVGG_A2(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[2, 4, 14, 1],
         width_multiplier=[1.5, 1.5, 1.5, 2.75],
         override_groups_map=None,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_A2"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_B0(**kwargs):
-    return RepVGG(
+def RepVGG_B0(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[4, 6, 16, 1],
         width_multiplier=[1, 1, 1, 2.5],
         override_groups_map=None,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_B0"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_B1(**kwargs):
-    return RepVGG(
+def RepVGG_B1(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[4, 6, 16, 1],
         width_multiplier=[2, 2, 2, 4],
         override_groups_map=None,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_B1"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_B1g2(**kwargs):
-    return RepVGG(
+def RepVGG_B1g2(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[4, 6, 16, 1],
         width_multiplier=[2, 2, 2, 4],
         override_groups_map=g2_map,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_B1g2"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_B1g4(**kwargs):
-    return RepVGG(
+def RepVGG_B1g4(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[4, 6, 16, 1],
         width_multiplier=[2, 2, 2, 4],
         override_groups_map=g4_map,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_B1g4"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_B2(**kwargs):
-    return RepVGG(
+def RepVGG_B2(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[4, 6, 16, 1],
         width_multiplier=[2.5, 2.5, 2.5, 5],
         override_groups_map=None,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_B2"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_B2g2(**kwargs):
-    return RepVGG(
+def RepVGG_B2g2(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[4, 6, 16, 1],
         width_multiplier=[2.5, 2.5, 2.5, 5],
         override_groups_map=g2_map,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_B2g2"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_B2g4(**kwargs):
-    return RepVGG(
+def RepVGG_B2g4(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[4, 6, 16, 1],
         width_multiplier=[2.5, 2.5, 2.5, 5],
         override_groups_map=g4_map,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_B2g4"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_B3(**kwargs):
-    return RepVGG(
+def RepVGG_B3(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[4, 6, 16, 1],
         width_multiplier=[3, 3, 3, 5],
         override_groups_map=None,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_B3"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_B3g2(**kwargs):
-    return RepVGG(
+def RepVGG_B3g2(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[4, 6, 16, 1],
         width_multiplier=[3, 3, 3, 5],
         override_groups_map=g2_map,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_B3g2"], use_ssld=use_ssld)
+    return model
 
 
-def RepVGG_B3g4(**kwargs):
-    return RepVGG(
+def RepVGG_B3g4(pretrained=False, use_ssld=False, **kwargs):
+    model = RepVGG(
         num_blocks=[4, 6, 16, 1],
         width_multiplier=[3, 3, 3, 5],
         override_groups_map=g4_map,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["RepVGG_B3g4"], use_ssld=use_ssld)
+    return model

ppcls/arch/backbone/model_zoo/res2net.py

@@ -27,11 +27,13 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = [
-    "Res2Net50_48w_2s", "Res2Net50_26w_4s", "Res2Net50_14w_8s",
-    "Res2Net50_48w_2s", "Res2Net50_26w_6s", "Res2Net50_26w_8s",
-    "Res2Net101_26w_4s", "Res2Net152_26w_4s", "Res2Net200_26w_4s"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"Res2Net50_26w_4s": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/Res2Net50_26w_4s_pretrained.pdparams",
+              "Res2Net50_14w_8s": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/Res2Net50_14w_8s_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -232,41 +234,26 @@ class Res2Net(nn.Layer):
         return y
 
 
-def Res2Net50_48w_2s(**args):
-    model = Res2Net(layers=50, scales=2, width=48, **args)
-    return model
-
-
-def Res2Net50_26w_4s(**args):
-    model = Res2Net(layers=50, scales=4, width=26, **args)
-    return model
-
-
-def Res2Net50_14w_8s(**args):
-    model = Res2Net(layers=50, scales=8, width=14, **args)
-    return model
-
-
-def Res2Net50_26w_6s(**args):
-    model = Res2Net(layers=50, scales=6, width=26, **args)
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+
+def Res2Net50_26w_4s(pretrained=False, use_ssld=False, **kwargs):
+    model = Res2Net(layers=50, scales=4, width=26, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["Res2Net50_26w_4s"], use_ssld=use_ssld)
     return model
 
 
-def Res2Net50_26w_8s(**args):
-    model = Res2Net(layers=50, scales=8, width=26, **args)
-    return model
-
-
+def Res2Net50_14w_8s(pretrained=False, use_ssld=False, **kwargs):
+    model = Res2Net(layers=50, scales=8, width=14, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["Res2Net50_14w_8s"], use_ssld=use_ssld)
+    return model
\ No newline at end of file
-def Res2Net101_26w_4s(**args):
-    model = Res2Net(layers=101, scales=4, width=26, **args)
-    return model
-
-
-def Res2Net152_26w_4s(**args):
-    model = Res2Net(layers=152, scales=4, width=26, **args)
-    return model
-
-
-def Res2Net200_26w_4s(**args):
-    model = Res2Net(layers=200, scales=4, width=26, **args)
-    return model

ppcls/arch/backbone/model_zoo/res2net_vd.py

@@ -27,11 +27,14 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = [
-    "Res2Net50_vd_48w_2s", "Res2Net50_vd_26w_4s", "Res2Net50_vd_14w_8s",
-    "Res2Net50_vd_48w_2s", "Res2Net50_vd_26w_6s", "Res2Net50_vd_26w_8s",
-    "Res2Net101_vd_26w_4s", "Res2Net152_vd_26w_4s", "Res2Net200_vd_26w_4s"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"Res2Net50_vd_26w_4s": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/Res2Net50_vd_26w_4s_pretrained.pdparams",
+              "Res2Net101_vd_26w_4s": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/Res2Net101_vd_26w_4s_pretrained.pdparams",
+              "Res2Net200_vd_26w_4s": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/Res2Net200_vd_26w_4s_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -255,41 +258,32 @@ class Res2Net_vd(nn.Layer):
         return y
 
 
-def Res2Net50_vd_48w_2s(**args):
-    model = Res2Net_vd(layers=50, scales=2, width=48, **args)
-    return model
-
-
-def Res2Net50_vd_26w_4s(**args):
-    model = Res2Net_vd(layers=50, scales=4, width=26, **args)
-    return model
-
-
-def Res2Net50_vd_14w_8s(**args):
-    model = Res2Net_vd(layers=50, scales=8, width=14, **args)
-    return model
-
-
-def Res2Net50_vd_26w_6s(**args):
-    model = Res2Net_vd(layers=50, scales=6, width=26, **args)
-    return model
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
 
 
-def Res2Net50_vd_26w_8s(**args):
-    model = Res2Net_vd(layers=50, scales=8, width=26, **args)
+def Res2Net50_vd_26w_4s(pretrained=False, use_ssld=False, **kwargs):
+    model = Res2Net_vd(layers=50, scales=4, width=26, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["Res2Net50_vd_26w_4s"], use_ssld=use_ssld)
     return model
 
 
-def Res2Net101_vd_26w_4s(**args):
-    model = Res2Net_vd(layers=101, scales=4, width=26, **args)
+def Res2Net101_vd_26w_4s(pretrained=False, use_ssld=False, **kwargs):
+    model = Res2Net_vd(layers=101, scales=4, width=26, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["Res2Net101_vd_26w_4s"], use_ssld=use_ssld)
     return model
 
 
-def Res2Net152_vd_26w_4s(**args):
-    model = Res2Net_vd(layers=152, scales=4, width=26, **args)
-    return model
-
-
+def Res2Net200_vd_26w_4s(pretrained=False, use_ssld=False, **kwargs):
+    model = Res2Net_vd(layers=200, scales=4, width=26, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["Res2Net200_vd_26w_4s"], use_ssld=use_ssld)
+    return model
\ No newline at end of file
-def Res2Net200_vd_26w_4s(**args):
-    model = Res2Net_vd(layers=200, scales=4, width=26, **args)
-    return model

ppcls/arch/backbone/model_zoo/resnest.py

@@ -27,7 +27,14 @@ from paddle.nn import Conv2D, BatchNorm, Linear, Dropout
 from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 from paddle.regularizer import L2Decay
 
-__all__ = ["ResNeSt50_fast_1s1x64d", "ResNeSt50", "ResNeSt101"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"ResNeSt50_fast_1s1x64d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeSt50_fast_1s1x64d_pretrained.pdparams",
+              "ResNeSt50": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeSt50_pretrained.pdparams",
+              "ResNeSt101": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeSt101_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -656,8 +663,21 @@ class ResNeSt(nn.Layer):
         x = self.out(x)
         return x
 
-
-def ResNeSt50_fast_1s1x64d(**args):
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+    
+def ResNeSt50_fast_1s1x64d(pretrained=False, use_ssld=False, **kwargs):
     model = ResNeSt(
         layers=[3, 4, 6, 3],
         radix=1,
@@ -669,11 +689,12 @@ def ResNeSt50_fast_1s1x64d(**args):
         avd=True,
         avd_first=True,
         final_drop=0.0,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeSt50_fast_1s1x64d"], use_ssld=use_ssld)
     return model
 
 
-def ResNeSt50(**args):
+def ResNeSt50(pretrained=False, use_ssld=False, **kwargs):
     model = ResNeSt(
         layers=[3, 4, 6, 3],
         radix=2,
@@ -685,11 +706,12 @@ def ResNeSt50(**args):
         avd=True,
         avd_first=False,
         final_drop=0.0,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeSt50"], use_ssld=use_ssld)
     return model
 
 
-def ResNeSt101(**args):
+def ResNeSt101(pretrained=False, use_ssld=False, **kwargs):
     model = ResNeSt(
         layers=[3, 4, 23, 3],
         radix=2,
@@ -701,5 +723,6 @@ def ResNeSt101(**args):
         avd=True,
         avd_first=False,
         final_drop=0.0,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeSt101"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/resnet.py

@@ -27,7 +27,16 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = ["ResNet18", "ResNet34", "ResNet50", "ResNet101", "ResNet152"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"ResNet18": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet18_pretrained.pdparams",
+              "ResNet34": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet34_pretrained.pdparams",
+              "ResNet50": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet50_pretrained.pdparams",
+              "ResNet101": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet101_pretrained.pdparams",
+              "ResNet152": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet152_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -290,27 +299,45 @@ class ResNet(nn.Layer):
             y = self.out(y)
             return y
 
-
-def ResNet18(**args):
-    model = ResNet(layers=18, **args)
+        
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+        
+def ResNet18(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet(layers=18, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet18"], use_ssld=use_ssld)
     return model
 
 
-def ResNet34(**args):
-    model = ResNet(layers=34, **args)
+def ResNet34(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet(layers=34, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet34"], use_ssld=use_ssld)
     return model
 
 
-def ResNet50(**args):
-    model = ResNet(layers=50, **args)
+def ResNet50(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet(layers=50, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet50"], use_ssld=use_ssld)
     return model
 
 
-def ResNet101(**args):
-    model = ResNet(layers=101, **args)
+def ResNet101(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet(layers=101, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet101"], use_ssld=use_ssld)
     return model
 
 
-def ResNet152(**args):
-    model = ResNet(layers=152, **args)
+def ResNet152(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet(layers=152, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet152"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/resnet_vc.py

@@ -27,9 +27,13 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = [
-    "ResNet18_vc", "ResNet34_vc", "ResNet50_vc", "ResNet101_vc", "ResNet152_vc"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "ResNet50_vc": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet50_vc_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -283,27 +287,22 @@ class ResNet_vc(nn.Layer):
         y = self.out(y)
         return y
 
-
-def ResNet18_vc(**args):
-    model = ResNet_vc(layers=18, **args)
-    return model
-
-
-def ResNet34_vc(**args):
-    model = ResNet_vc(layers=34, **args)
-    return model
-
-
-def ResNet50_vc(**args):
-    model = ResNet_vc(layers=50, **args)
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+        
+def ResNet50_vc(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet_vc(layers=50, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet50_vc"], use_ssld=use_ssld)
     return model
 
-
-def ResNet101_vc(**args):
-    model = ResNet_vc(layers=101, **args)
-    return model
-
-
-def ResNet152_vc(**args):
-    model = ResNet_vc(layers=152, **args)
-    return model

ppcls/arch/backbone/model_zoo/resnet_vd.py

@@ -27,9 +27,18 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = [
-    "ResNet18_vd", "ResNet34_vd", "ResNet50_vd", "ResNet101_vd", "ResNet152_vd"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "ResNet18_vd": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet18_vd_pretrained.pdparams",
+              "ResNet34_vd": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet34_vd_pretrained.pdparams",
+              "ResNet50_vd": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet50_vd_pretrained.pdparams",
+              "ResNet101_vd": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet101_vd_pretrained.pdparams",
+              "ResNet152_vd": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet152_vd_pretrained.pdparams",
+              "ResNet200_vd": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNet200_vd_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -324,31 +333,50 @@ class ResNet_vd(nn.Layer):
         return y
 
 
-def ResNet18_vd(**args):
-    model = ResNet_vd(layers=18, **args)
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+
+def ResNet18_vd(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet_vd(layers=18, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet18_vd"], use_ssld=use_ssld)
     return model
 
 
-def ResNet34_vd(**args):
-    model = ResNet_vd(layers=34, **args)
+def ResNet34_vd(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet_vd(layers=34, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet34_vd"], use_ssld=use_ssld)
     return model
 
 
-def ResNet50_vd(**args):
-    model = ResNet_vd(layers=50, **args)
+def ResNet50_vd(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet_vd(layers=50, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet50_vd"], use_ssld=use_ssld)
     return model
 
 
-def ResNet101_vd(**args):
-    model = ResNet_vd(layers=101, **args)
+def ResNet101_vd(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet_vd(layers=101, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet101_vd"], use_ssld=use_ssld)
     return model
 
 
-def ResNet152_vd(**args):
-    model = ResNet_vd(layers=152, **args)
+def ResNet152_vd(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet_vd(layers=152, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet152_vd"], use_ssld=use_ssld)
     return model
 
 
-def ResNet200_vd(**args):
-    model = ResNet_vd(layers=200, **args)
+def ResNet200_vd(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNet_vd(layers=200, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNet200_vd"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/resnext.py

@@ -27,10 +27,18 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = [
-    "ResNeXt50_32x4d", "ResNeXt50_64x4d", "ResNeXt101_32x4d",
-    "ResNeXt101_64x4d", "ResNeXt152_32x4d", "ResNeXt152_64x4d"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "ResNeXt50_32x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt50_32x4d_pretrained.pdparams",
+              "ResNeXt50_64x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt50_64x4d_pretrained.pdparams",
+              "ResNeXt101_32x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt101_32x4d_pretrained.pdparams",
+              "ResNeXt101_64x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt101_64x4d_pretrained.pdparams",
+              "ResNeXt152_32x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt152_32x4d_pretrained.pdparams",
+              "ResNeXt152_64x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt152_64x4d_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -222,32 +230,51 @@ class ResNeXt(nn.Layer):
             y = self.out(y)
             return y
 
-
-def ResNeXt50_32x4d(**args):
-    model = ResNeXt(layers=50, cardinality=32, **args)
+        
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+
+def ResNeXt50_32x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=50, cardinality=32, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt50_32x4d"], use_ssld=use_ssld)
     return model
 
 
-def ResNeXt50_64x4d(**args):
-    model = ResNeXt(layers=50, cardinality=64, **args)
+def ResNeXt50_64x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=50, cardinality=64, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt50_64x4d"], use_ssld=use_ssld)
     return model
 
 
-def ResNeXt101_32x4d(**args):
-    model = ResNeXt(layers=101, cardinality=32, **args)
+def ResNeXt101_32x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=101, cardinality=32, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt101_32x4d"], use_ssld=use_ssld)
     return model
 
 
-def ResNeXt101_64x4d(**args):
-    model = ResNeXt(layers=101, cardinality=64, **args)
+def ResNeXt101_64x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=101, cardinality=64, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt101_64x4d"], use_ssld=use_ssld)
     return model
 
 
-def ResNeXt152_32x4d(**args):
-    model = ResNeXt(layers=152, cardinality=32, **args)
+def ResNeXt152_32x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=152, cardinality=32, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt152_32x4d"], use_ssld=use_ssld)
     return model
 
 
-def ResNeXt152_64x4d(**args):
-    model = ResNeXt(layers=152, cardinality=64, **args)
+def ResNeXt152_64x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=152, cardinality=64, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt152_64x4d"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/resnext101_wsl.py

@@ -6,10 +6,18 @@ from paddle.nn import Conv2D, BatchNorm, Linear, Dropout
 from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 from paddle.nn.initializer import Uniform
 
-__all__ = [
-    "ResNeXt101_32x8d_wsl", "ResNeXt101_32x16d_wsl", "ResNeXt101_32x32d_wsl",
-    "ResNeXt101_32x48d_wsl"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "ResNeXt101_32x8d_wsl": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt101_32x8d_wsl_pretrained.pdparams",
+              "ResNeXt101_32x16d_wsl": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt101_32x816_wsl_pretrained.pdparams",
+              "ResNeXt101_32x32d_wsl": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt101_32x32d_wsl_pretrained.pdparams",
+              "ResNeXt101_32x48d_wsl": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt101_32x48d_wsl_pretrained.pdparams",
+
+             }
+
+__all__ = list(MODEL_URLS.keys())
+
 
 
 class ConvBNLayer(nn.Layer):
@@ -426,22 +434,39 @@ class ResNeXt101WSL(nn.Layer):
         x = self._out(x)
         return x
 
-
-def ResNeXt101_32x8d_wsl(**args):
-    model = ResNeXt101WSL(cardinality=32, width=8, **args)
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+
+def ResNeXt101_32x8d_wsl(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt101WSL(cardinality=32, width=8, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt101_32x8d_wsl"], use_ssld=use_ssld)
     return model
 
 
 def ResNeXt101_32x16d_wsl(**args):
-    model = ResNeXt101WSL(cardinality=32, width=16, **args)
+    model = ResNeXt101WSL(cardinality=32, width=16, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt101_32x16d_ws"], use_ssld=use_ssld)
     return model
 
 
 def ResNeXt101_32x32d_wsl(**args):
-    model = ResNeXt101WSL(cardinality=32, width=32, **args)
+    model = ResNeXt101WSL(cardinality=32, width=32, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt101_32x32d_wsl"], use_ssld=use_ssld)
     return model
 
 
 def ResNeXt101_32x48d_wsl(**args):
-    model = ResNeXt101WSL(cardinality=32, width=48, **args)
+    model = ResNeXt101WSL(cardinality=32, width=48, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt101_32x48d_wsl"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/resnext_vd.py

@@ -27,11 +27,18 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = [
-    "ResNeXt50_vd_32x4d", "ResNeXt50_vd_64x4d", "ResNeXt101_vd_32x4d",
-    "ResNeXt101_vd_64x4d", "ResNeXt152_vd_32x4d", "ResNeXt152_vd_64x4d"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
 
+MODEL_URLS = {
+              "ResNeXt50_vd_32x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt50_vd_32x4d_pretrained.pdparams",
+              "ResNeXt50_vd_64x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt50_vd_64x4d_pretrained.pdparams",
+              "ResNeXt101_vd_32x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt101_vd_32x4d_pretrained.pdparams",
+              "ResNeXt101_vd_64x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt101_vd_64x4d_pretrained.pdparams",
+              "ResNeXt152_vd_32x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt152_vd_32x4d_pretrained.pdparams",
+              "ResNeXt152_vd_64x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ResNeXt152_vd_64x4d_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 class ConvBNLayer(nn.Layer):
     def __init__(
@@ -235,32 +242,50 @@ class ResNeXt(nn.Layer):
         y = self.out(y)
         return y
 
-
-def ResNeXt50_vd_32x4d(**args):
-    model = ResNeXt(layers=50, cardinality=32, **args)
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+
+def ResNeXt50_vd_32x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=50, cardinality=32, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt50_vd_32x4d"], use_ssld=use_ssld)
     return model
 
 
-def ResNeXt50_vd_64x4d(**args):
-    model = ResNeXt(layers=50, cardinality=64, **args)
+def ResNeXt50_vd_64x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=50, cardinality=64, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt50_vd_64x4d"], use_ssld=use_ssld)
     return model
 
 
-def ResNeXt101_vd_32x4d(**args):
-    model = ResNeXt(layers=101, cardinality=32, **args)
+def ResNeXt101_vd_32x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=101, cardinality=32, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt101_vd_32x4d"], use_ssld=use_ssld)
     return model
 
 
-def ResNeXt101_vd_64x4d(**args):
-    model = ResNeXt(layers=101, cardinality=64, **args)
+def ResNeXt101_vd_64x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=101, cardinality=64, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt101_vd_64x4d"], use_ssld=use_ssld)
     return model
 
 
-def ResNeXt152_vd_32x4d(**args):
-    model = ResNeXt(layers=152, cardinality=32, **args)
+def ResNeXt152_vd_32x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=152, cardinality=32, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt152_vd_32x4d"], use_ssld=use_ssld)
     return model
 
 
-def ResNeXt152_vd_64x4d(**args):
-    model = ResNeXt(layers=152, cardinality=64, **args)
+def ResNeXt152_vd_64x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=152, cardinality=64, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ResNeXt152_vd_64x4d"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/rexnet.py

@@ -22,9 +22,17 @@ from paddle import ParamAttr
 import paddle.nn as nn
 from math import ceil
 
-__all__ = [
-    "ReXNet_1_0", "ReXNet_1_3", "ReXNet_1_5", "ReXNet_2_0", "ReXNet_3_0"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "ReXNet_1_0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ReXNet_1_0_pretrained.pdparams",
+              "ReXNet_1_3": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ReXNet_1_3_pretrained.pdparams",
+              "ReXNet_1_5": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ReXNet_1_5_32x4d_pretrained.pdparams",
+              "ReXNet_2_0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ReXNet_2_0_pretrained.pdparams",
+              "ReXNet_3_0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ReXNet_3_0_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 def conv_bn_act(out,
@@ -220,21 +228,44 @@ class ReXNetV1(nn.Layer):
         return x
 
 
-def ReXNet_1_0(**args):
-    return ReXNetV1(width_mult=1.0, **args)
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+    
+    
+def ReXNet_1_0(pretrained=False, use_ssld=False, **kwargs):
+    model = ReXNetV1(width_mult=1.0, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ReXNet_1_0"], use_ssld=use_ssld)
+    return model
 
 
-def ReXNet_1_3(**args):
-    return ReXNetV1(width_mult=1.3, **args)
+def ReXNet_1_3(pretrained=False, use_ssld=False, **kwargs):
+    model = ReXNetV1(width_mult=1.3, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ReXNet_1_3"], use_ssld=use_ssld)
+    return model
 
 
-def ReXNet_1_5(**args):
-    return ReXNetV1(width_mult=1.5, **args)
+def ReXNet_1_5(pretrained=False, use_ssld=False, **kwargs):
+    model = ReXNetV1(width_mult=1.5, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ReXNet_1_5"], use_ssld=use_ssld)
+    return model
 
 
-def ReXNet_2_0(**args):
-    return ReXNetV1(width_mult=2.0, **args)
+def ReXNet_2_0(pretrained=False, use_ssld=False, **kwargs):
+    model = ReXNetV1(width_mult=2.0, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ReXNet_2_0"], use_ssld=use_ssld)
+    return model
 
 
-def ReXNet_3_0(**args):
-    return ReXNetV1(width_mult=3.0, **args)
+def ReXNet_3_0(pretrained=False, use_ssld=False, **kwargs):
+    model = ReXNetV1(width_mult=3.0, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ReXNet_3_0"], use_ssld=use_ssld)
+    return model
\ No newline at end of file

ppcls/arch/backbone/model_zoo/se_resnet_vd.py

@@ -26,10 +26,16 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = [
-    "SE_ResNet18_vd", "SE_ResNet34_vd", "SE_ResNet50_vd", "SE_ResNet101_vd",
-    "SE_ResNet152_vd", "SE_ResNet200_vd"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "SE_ResNet18_vd": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SE_ResNet18_vd_pretrained.pdparams",
+              "SE_ResNet34_vd": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SE_ResNet34_vd_pretrained.pdparams",
+              "SE_ResNet50_vd": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SE_ResNet50_vd_pretrained.pdparams",
+
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -347,32 +353,33 @@ class SE_ResNet_vd(nn.Layer):
         y = self.out(y)
         return y
 
-
-def SE_ResNet18_vd(**args):
-    model = SE_ResNet_vd(layers=18, **args)
-    return model
-
-
-def SE_ResNet34_vd(**args):
-    model = SE_ResNet_vd(layers=34, **args)
-    return model
-
-
-def SE_ResNet50_vd(**args):
-    model = SE_ResNet_vd(layers=50, **args)
-    return model
-
-
-def SE_ResNet101_vd(**args):
-    model = SE_ResNet_vd(layers=101, **args)
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+    
+
+def SE_ResNet18_vd(pretrained=False, use_ssld=False, **kwargs):
+    model = SE_ResNet_vd(layers=18, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SE_ResNet18_vd"], use_ssld=use_ssld)
     return model
 
 
-def SE_ResNet152_vd(**args):
-    model = SE_ResNet_vd(layers=152, **args)
+def SE_ResNet34_vd(pretrained=False, use_ssld=False, **kwargs):
+    model = SE_ResNet_vd(layers=34, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SE_ResNet34_vd"], use_ssld=use_ssld)
     return model
 
 
-def SE_ResNet200_vd(**args):
-    model = SE_ResNet_vd(layers=200, **args)
+def SE_ResNet50_vd(pretrained=False, use_ssld=False, **kwargs):
+    model = SE_ResNet_vd(layers=50, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SE_ResNet50_vd"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/se_resnext.py

@@ -27,7 +27,16 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = ["SE_ResNeXt50_32x4d", "SE_ResNeXt101_32x4d", "SE_ResNeXt152_64x4d"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "SE_ResNeXt50_32x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SE_ResNeXt50_32x4d_pretrained.pdparams",
+              "SE_ResNeXt101_32x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SE_ResNeXt101_32x4d_pretrained.pdparams",
+              "SE_ResNeXt152_64x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SE_ResNeXt152_64x4d_pretrained.pdparams",
+
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -301,17 +310,33 @@ class ResNeXt(nn.Layer):
             y = self.out(y)
             return y
 
-
-def SE_ResNeXt50_32x4d(**args):
-    model = ResNeXt(layers=50, cardinality=32, **args)
+        
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+
+def SE_ResNeXt50_32x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=50, cardinality=32, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SE_ResNeXt50_32x4d"], use_ssld=use_ssld)
     return model
 
 
-def SE_ResNeXt101_32x4d(**args):
-    model = ResNeXt(layers=101, cardinality=32, **args)
+def SE_ResNeXt101_32x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=101, cardinality=32, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SE_ResNeXt101_32x4d"], use_ssld=use_ssld)
     return model
 
 
-def SE_ResNeXt152_64x4d(**args):
-    model = ResNeXt(layers=152, cardinality=64, **args)
+def SE_ResNeXt152_64x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=152, cardinality=64, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SE_ResNeXt152_64x4d"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/se_resnext_vd.py

@@ -27,7 +27,16 @@ from paddle.nn.initializer import Uniform
 
 import math
 
-__all__ = ["SE_ResNeXt50_vd_32x4d", "SE_ResNeXt50_vd_32x4d", "SENet154_vd"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "SE_ResNeXt50_vd_32x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SE_ResNeXt50_vd_32x4d_pretrained.pdparams",
+              "SE_ResNeXt50_vd_32x4d": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SE_ResNeXt50_vd_32x4d_pretrained.pdparams",
+              "SENet154_vd": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SENet154_vd_pretrained.pdparams",
+
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -269,17 +278,33 @@ class ResNeXt(nn.Layer):
         y = self.out(y)
         return y
 
-
-def SE_ResNeXt50_vd_32x4d(**args):
-    model = ResNeXt(layers=50, cardinality=32, **args)
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+
+def SE_ResNeXt50_vd_32x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=50, cardinality=32, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SE_ResNeXt50_vd_32x4d"], use_ssld=use_ssld)
     return model
 
 
-def SE_ResNeXt101_vd_32x4d(**args):
-    model = ResNeXt(layers=101, cardinality=32, **args)
+def SE_ResNeXt101_vd_32x4d(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=101, cardinality=32, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SE_ResNeXt101_vd_32x4d"], use_ssld=use_ssld)
     return model
 
 
-def SENet154_vd(**args):
-    model = ResNeXt(layers=152, cardinality=64, **args)
+def SENet154_vd(pretrained=False, use_ssld=False, **kwargs):
+    model = ResNeXt(layers=152, cardinality=64, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SENet154_vd"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/shufflenet_v2.py

@@ -22,11 +22,19 @@ from paddle.nn import Layer, Conv2D, MaxPool2D, AdaptiveAvgPool2D, BatchNorm, Li
 from paddle.nn.initializer import KaimingNormal
 from paddle.nn.functional import swish
 
-__all__ = [
-    "ShuffleNetV2_x0_25", "ShuffleNetV2_x0_33", "ShuffleNetV2_x0_5",
-    "ShuffleNetV2_x1_0", "ShuffleNetV2_x1_5", "ShuffleNetV2_x2_0",
-    "ShuffleNetV2_swish"
-]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "ShuffleNetV2_x0_25": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ShuffleNetV2_x0_25_pretrained.pdparams",
+              "ShuffleNetV2_x0_33": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ShuffleNetV2_x0_33_pretrained.pdparams",
+              "ShuffleNetV2_x0_5": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ShuffleNetV2_x0_5_pretrained.pdparams",
+              "ShuffleNetV2_x1_0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ShuffleNetV2_x1_0_pretrained.pdparams",
+              "ShuffleNetV2_x1_5": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ShuffleNetV2_x1_5_pretrained.pdparams",
+              "ShuffleNetV2_x2_0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ShuffleNetV2_x2_0_pretrained.pdparams",
+              "ShuffleNetV2_swish": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ShuffleNetV2_swish_pretrained.pdparams"
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 def channel_shuffle(x, groups):
@@ -285,36 +293,56 @@ class ShuffleNet(Layer):
         return y
 
 
-def ShuffleNetV2_x0_25(**args):
-    model = ShuffleNet(scale=0.25, **args)
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+
+def ShuffleNetV2_x0_25(pretrained=False, use_ssld=False, **kwargs):
+    model = ShuffleNet(scale=0.25, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ShuffleNetV2_x0_25"], use_ssld=use_ssld)
     return model
 
 
-def ShuffleNetV2_x0_33(**args):
-    model = ShuffleNet(scale=0.33, **args)
+def ShuffleNetV2_x0_33(pretrained=False, use_ssld=False, **kwargs):
+    model = ShuffleNet(scale=0.33, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ShuffleNetV2_x0_33"], use_ssld=use_ssld)
     return model
 
 
-def ShuffleNetV2_x0_5(**args):
-    model = ShuffleNet(scale=0.5, **args)
+def ShuffleNetV2_x0_5(pretrained=False, use_ssld=False, **kwargs):
+    model = ShuffleNet(scale=0.5, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ShuffleNetV2_x0_5"], use_ssld=use_ssld)
     return model
 
 
-def ShuffleNetV2_x1_0(**args):
-    model = ShuffleNet(scale=1.0, **args)
+def ShuffleNetV2_x1_0(pretrained=False, use_ssld=False, **kwargs):
+    model = ShuffleNet(scale=1.0, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ShuffleNetV2_x1_0"], use_ssld=use_ssld)
     return model
 
 
-def ShuffleNetV2_x1_5(**args):
-    model = ShuffleNet(scale=1.5, **args)
+def ShuffleNetV2_x1_5(pretrained=False, use_ssld=False, **kwargs):
+    model = ShuffleNet(scale=1.5, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ShuffleNetV2_x1_5"], use_ssld=use_ssld)
     return model
 
 
-def ShuffleNetV2_x2_0(**args):
-    model = ShuffleNet(scale=2.0, **args)
+def ShuffleNetV2_x2_0(pretrained=False, use_ssld=False, **kwargs):
+    model = ShuffleNet(scale=2.0, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ShuffleNetV2_x2_0"], use_ssld=use_ssld)
     return model
 
 
-def ShuffleNetV2_swish(**args):
-    model = ShuffleNet(scale=1.0, act="swish", **args)
+def ShuffleNetV2_swish(pretrained=False, use_ssld=False, **kwargs):
+    model = ShuffleNet(scale=1.0, act="swish", **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ShuffleNetV2_swish"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/squeezenet.py

+# copyright (c) 2020 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import paddle
 from paddle import ParamAttr
 import paddle.nn as nn
@@ -5,7 +19,14 @@ import paddle.nn.functional as F
 from paddle.nn import Conv2D, BatchNorm, Linear, Dropout
 from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 
-__all__ = ["SqueezeNet1_0", "SqueezeNet1_1"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "SqueezeNet1_0": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SqueezeNet1_0_pretrained.pdparams",
+              "SqueezeNet1_1": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SqueezeNet1_1_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class MakeFireConv(nn.Layer):
@@ -143,12 +164,26 @@ class SqueezeNet(nn.Layer):
         x = paddle.squeeze(x, axis=[2, 3])
         return x
 
-
-def SqueezeNet1_0(**args):
-    model = SqueezeNet(version="1.0", **args)
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+
+def SqueezeNet1_0(pretrained=False, use_ssld=False, **kwargs):
+    model = SqueezeNet(version="1.0", **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SqueezeNet1_0"], use_ssld=use_ssld)
     return model
 
 
-def SqueezeNet1_1(**args):
-    model = SqueezeNet(version="1.1", **args)
+def SqueezeNet1_1(pretrained=False, use_ssld=False, **kwargs):
+    model = SqueezeNet(version="1.1", **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SqueezeNet1_1"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/swin_transformer.py

@@ -21,6 +21,19 @@ from paddle.nn.initializer import TruncatedNormal, Constant
 
 from .vision_transformer import trunc_normal_, zeros_, ones_, to_2tuple, DropPath, Identity
 
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "SwinTransformer_tiny_patch4_window7_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformer_tiny_patch4_window7_224_pretrained.pdparams",
+              "SwinTransformer_small_patch4_window7_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformer_small_patch4_window7_224_pretrained.pdparams",
+              "SwinTransformer_base_patch4_window7_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformer_base_patch4_window7_224_pretrained.pdparams",
+              "SwinTransformer_base_patch4_window12_384": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformer_base_patch4_window12_384_pretrained.pdparams",
+              "SwinTransformer_large_patch4_window7_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformer_large_patch4_window7_224_pretrained.pdparams",
+              "SwinTransformer_large_patch4_window12_384": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/SwinTransformer_large_patch4_window12_384_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
+
 
 class Mlp(nn.Layer):
     def __init__(self,
@@ -716,40 +729,56 @@ class SwinTransformer(nn.Layer):
         flops += self.num_features * self.num_classes
         return flops
 
-
-def SwinTransformer_tiny_patch4_window7_224(**args):
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+
+def SwinTransformer_tiny_patch4_window7_224(pretrained=False, use_ssld=False, **kwargs):
     model = SwinTransformer(
         embed_dim=96,
         depths=[2, 2, 6, 2],
         num_heads=[3, 6, 12, 24],
         window_size=7,
         drop_path_rate=0.2,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SwinTransformer_tiny_patch4_window7_224"], use_ssld=use_ssld)
     return model
 
 
-def SwinTransformer_small_patch4_window7_224(**args):
+def SwinTransformer_small_patch4_window7_224(pretrained=False, use_ssld=False, **kwargs):
     model = SwinTransformer(
         embed_dim=96,
         depths=[2, 2, 18, 2],
         num_heads=[3, 6, 12, 24],
         window_size=7,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SwinTransformer_small_patch4_window7_224"], use_ssld=use_ssld)
     return model
 
 
-def SwinTransformer_base_patch4_window7_224(**args):
+def SwinTransformer_base_patch4_window7_224(pretrained=False, use_ssld=False, **kwargs):
     model = SwinTransformer(
         embed_dim=128,
         depths=[2, 2, 18, 2],
         num_heads=[4, 8, 16, 32],
         window_size=7,
         drop_path_rate=0.5,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SwinTransformer_base_patch4_window7_224"], use_ssld=use_ssld)
     return model
 
 
-def SwinTransformer_base_patch4_window12_384(**args):
+def SwinTransformer_base_patch4_window12_384(pretrained=False, use_ssld=False, **kwargs):
     model = SwinTransformer(
         img_size=384,
         embed_dim=128,
@@ -757,26 +786,29 @@ def SwinTransformer_base_patch4_window12_384(**args):
         num_heads=[4, 8, 16, 32],
         window_size=12,
         drop_path_rate=0.5,  # NOTE: do not appear in offical code
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SwinTransformer_base_patch4_window12_384"], use_ssld=use_ssld)
     return model
 
 
-def SwinTransformer_large_patch4_window7_224(**args):
+def SwinTransformer_large_patch4_window7_224(pretrained=False, use_ssld=False, **kwargs):
     model = SwinTransformer(
         embed_dim=192,
         depths=[2, 2, 18, 2],
         num_heads=[6, 12, 24, 48],
         window_size=7,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SwinTransformer_large_patch4_window7_224"], use_ssld=use_ssld)
     return model
 
 
-def SwinTransformer_large_patch4_window12_384(**args):
+def SwinTransformer_large_patch4_window12_384(pretrained=False, use_ssld=False, **kwargs):
     model = SwinTransformer(
         img_size=384,
         embed_dim=192,
         depths=[2, 2, 18, 2],
         num_heads=[6, 12, 24, 48],
         window_size=12,
-        **args)
+        **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["SwinTransformer_large_patch4_window12_384"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/tnt.py

@@ -26,7 +26,7 @@ from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_f
 
 MODEL_URLS = {
     "TNT_small":
-    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/model_zoo/TNT_small_pretrained.pdparams"
+    "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/TNT_small_pretrained.pdparams"
 }
 
 

ppcls/arch/backbone/model_zoo/vgg.py

@@ -5,7 +5,16 @@ import paddle.nn.functional as F
 from paddle.nn import Conv2D, BatchNorm, Linear, Dropout
 from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 
-__all__ = ["VGG11", "VGG13", "VGG16", "VGG19"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "VGG11": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/VGG11_pretrained.pdparams",
+              "VGG13": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/VGG13_pretrained.pdparams",
+              "VGG16": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/VGG16_pretrained.pdparams",
+              "VGG19": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/VGG19_pretrained.pdparams",
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBlock(nn.Layer):
@@ -131,22 +140,40 @@ class VGGNet(nn.Layer):
         x = self._out(x)
         return x
 
-
-def VGG11(**args):
-    model = VGGNet(layers=11, **args)
+    
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+
+def VGG11(pretrained, model, model_url, use_ssld=False, **kwargs):
+    model = VGGNet(layers=11, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["VGG11"], use_ssld=use_ssld)
     return model
 
 
-def VGG13(**args):
-    model = VGGNet(layers=13, **args)
+def VGG13(pretrained, model, model_url, use_ssld=False, **kwargs):
+    model = VGGNet(layers=13, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["VGG13"], use_ssld=use_ssld)
     return model
 
 
-def VGG16(**args):
-    model = VGGNet(layers=16, **args)
+def VGG16(pretrained, model, model_url, use_ssld=False, **kwargs):
+    model = VGGNet(layers=16, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["VGG16"], use_ssld=use_ssld)
     return model
 
 
-def VGG19(**args):
-    model = VGGNet(layers=19, **args)
+def VGG19(pretrained, model, model_url, use_ssld=False, **kwargs):
+    model = VGGNet(layers=19, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["VGG19"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/vision_transformer.py

@@ -12,19 +12,32 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from collections import Callable
+
 import numpy as np
 import paddle
 import paddle.nn as nn
-from paddle.nn.initializer import TruncatedNormal, Constant
+from paddle.nn.initializer import TruncatedNormal, Constant, Normal
+
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "ViT_small_patch16_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ViT_small_patch16_224_pretrained.pdparams",
+              "ViT_base_patch16_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ViT_base_patch16_224_pretrained.pdparams",
+              "ViT_base_patch16_384": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ViT_base_patch16_384_pretrained.pdparams",
+              "ViT_base_patch32_384": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ViT_base_patch32_384_pretrained.pdparams",
+              "ViT_large_patch16_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ViT_large_patch16_224_pretrained.pdparams",
+              "ViT_large_patch16_384": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ViT_large_patch16_384_pretrained.pdparams", 
+              "ViT_large_patch32_384": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ViT_large_patch32_384_pretrained.pdparams",
+              "ViT_huge_patch16_224": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ViT_huge_patch16_224_pretrained.pdparams",
+              "ViT_huge_patch32_384": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/ViT_huge_patch32_384_pretrained.pdparams"
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
-__all__ = [
-    "VisionTransformer", "ViT_small_patch16_224", "ViT_base_patch16_224",
-    "ViT_base_patch16_384", "ViT_base_patch32_384", "ViT_large_patch16_224",
-    "ViT_large_patch16_384", "ViT_large_patch32_384", "ViT_huge_patch16_224",
-    "ViT_huge_patch32_384"
-]
 
 trunc_normal_ = TruncatedNormal(std=.02)
+normal_ = Normal
 zeros_ = Constant(value=0.)
 ones_ = Constant(value=1.)
 
@@ -141,7 +154,13 @@ class Block(nn.Layer):
                  norm_layer='nn.LayerNorm',
                  epsilon=1e-5):
         super().__init__()
-        self.norm1 = eval(norm_layer)(dim, epsilon=epsilon)
+        if isinstance(norm_layer, str):
+            self.norm1 = eval(norm_layer)(dim, epsilon=epsilon)
+        elif isinstance(norm_layer, Callable):
+            self.norm1 = norm_layer(dim)
+        else:
+            raise TypeError(
+                "The norm_layer must be str or paddle.nn.layer.Layer class")
         self.attn = Attention(
             dim,
             num_heads=num_heads,
@@ -151,7 +170,13 @@ class Block(nn.Layer):
             proj_drop=drop)
         # NOTE: drop path for stochastic depth, we shall see if this is better than dropout here
         self.drop_path = DropPath(drop_path) if drop_path > 0. else Identity()
-        self.norm2 = eval(norm_layer)(dim, epsilon=epsilon)
+        if isinstance(norm_layer, str):
+            self.norm2 = eval(norm_layer)(dim, epsilon=epsilon)
+        elif isinstance(norm_layer, Callable):
+            self.norm2 = norm_layer(dim)
+        else:
+            raise TypeError(
+                "The norm_layer must be str or paddle.nn.layer.Layer class")
         mlp_hidden_dim = int(dim * mlp_ratio)
         self.mlp = Mlp(in_features=dim,
                        hidden_features=mlp_hidden_dim,
@@ -285,7 +310,21 @@ class VisionTransformer(nn.Layer):
         return x
 
 
-def ViT_small_patch16_224(**kwargs):
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+
+
+        
+def ViT_small_patch16_224(pretrained, model, model_url, use_ssld=False, **kwargs):
     model = VisionTransformer(
         patch_size=16,
         embed_dim=768,
@@ -294,10 +333,12 @@ def ViT_small_patch16_224(**kwargs):
         mlp_ratio=3,
         qk_scale=768**-0.5,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ViT_small_patch16_224"], use_ssld=use_ssld)
     return model
 
 
-def ViT_base_patch16_224(**kwargs):
+
+def ViT_base_patch16_224(pretrained, model, model_url, use_ssld=False, **kwargs):
     model = VisionTransformer(
         patch_size=16,
         embed_dim=768,
@@ -307,10 +348,11 @@ def ViT_base_patch16_224(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ViT_base_patch16_224"], use_ssld=use_ssld)
     return model
 
 
-def ViT_base_patch16_384(**kwargs):
+def ViT_base_patch16_384(pretrained, model, model_url, use_ssld=False, **kwargs):
     model = VisionTransformer(
         img_size=384,
         patch_size=16,
@@ -321,10 +363,11 @@ def ViT_base_patch16_384(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ViT_base_patch16_384"], use_ssld=use_ssld)
     return model
 
 
-def ViT_base_patch32_384(**kwargs):
+def ViT_base_patch32_384(pretrained, model, model_url, use_ssld=False, **kwargs):
     model = VisionTransformer(
         img_size=384,
         patch_size=32,
@@ -335,10 +378,11 @@ def ViT_base_patch32_384(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ViT_base_patch32_384"], use_ssld=use_ssld)
     return model
 
 
-def ViT_large_patch16_224(**kwargs):
+def ViT_large_patch16_224(pretrained, model, model_url, use_ssld=False, **kwargs):
     model = VisionTransformer(
         patch_size=16,
         embed_dim=1024,
@@ -348,10 +392,11 @@ def ViT_large_patch16_224(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ViT_large_patch16_224"], use_ssld=use_ssld)
     return model
 
 
-def ViT_large_patch16_384(**kwargs):
+def ViT_large_patch16_384(pretrained, model, model_url, use_ssld=False, **kwargs):
     model = VisionTransformer(
         img_size=384,
         patch_size=16,
@@ -362,10 +407,11 @@ def ViT_large_patch16_384(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ViT_large_patch16_384"], use_ssld=use_ssld)
     return model
 
 
-def ViT_large_patch32_384(**kwargs):
+def ViT_large_patch32_384(pretrained, model, model_url, use_ssld=False, **kwargs):
     model = VisionTransformer(
         img_size=384,
         patch_size=32,
@@ -376,10 +422,11 @@ def ViT_large_patch32_384(**kwargs):
         qkv_bias=True,
         epsilon=1e-6,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ViT_large_patch32_384"], use_ssld=use_ssld)
     return model
 
 
-def ViT_huge_patch16_224(**kwargs):
+def ViT_huge_patch16_224(pretrained, model, model_url, use_ssld=False, **kwargs):
     model = VisionTransformer(
         patch_size=16,
         embed_dim=1280,
@@ -387,10 +434,11 @@ def ViT_huge_patch16_224(**kwargs):
         num_heads=16,
         mlp_ratio=4,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ViT_huge_patch16_224"], use_ssld=use_ssld)
     return model
 
 
-def ViT_huge_patch32_384(**kwargs):
+def ViT_huge_patch32_384(pretrained, model, model_url, use_ssld=False, **kwargs):
     model = VisionTransformer(
         img_size=384,
         patch_size=32,
@@ -399,4 +447,5 @@ def ViT_huge_patch32_384(**kwargs):
         num_heads=16,
         mlp_ratio=4,
         **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["ViT_huge_patch32_384"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/xception.py

@@ -8,7 +8,16 @@ from paddle.nn.initializer import Uniform
 import math
 import sys
 
-__all__ = ['Xception41', 'Xception65', 'Xception71']
+
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {
+              "Xception41": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/Xception41_pretrained.pdparams",
+              "Xception65": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/Xception65_pretrained.pdparams",
+              "Xception71": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/Xception71_pretrained.pdparams"
+             }
+
+__all__ = list(MODEL_URLS.keys())
 
 
 class ConvBNLayer(nn.Layer):
@@ -329,17 +338,32 @@ class Xception(nn.Layer):
         x = self._exit_flow(x)
         return x
 
-
-def Xception41(**args):
-    model = Xception(entry_flow_block_num=3, middle_flow_block_num=8, **args)
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
+        
+
+def Xception41(pretrained=False, use_ssld=False, **kwargs):
+    model = Xception(entry_flow_block_num=3, middle_flow_block_num=8, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["Xception41"], use_ssld=use_ssld)
     return model
 
 
-def Xception65(**args):
-    model = Xception(entry_flow_block_num=3, middle_flow_block_num=16, **args)
+def Xception65(pretrained=False, use_ssld=False, **kwargs):
+    model = Xception(entry_flow_block_num=3, middle_flow_block_num=16, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["Xception65"], use_ssld=use_ssld)
     return model
 
 
-def Xception71(**args):
-    model = Xception(entry_flow_block_num=5, middle_flow_block_num=16, **args)
+def Xception71(pretrained=False, use_ssld=False, **kwargs):
+    model = Xception(entry_flow_block_num=5, middle_flow_block_num=16, **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["Xception71"], use_ssld=use_ssld)
     return model

ppcls/arch/backbone/model_zoo/xception_deeplab.py

+# copyright (c) 2021 PaddlePaddle Authors. All Rights Reserve.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#    http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
 import paddle
 from paddle import ParamAttr
 import paddle.nn as nn
@@ -5,7 +19,12 @@ import paddle.nn.functional as F
 from paddle.nn import Conv2D, BatchNorm, Linear, Dropout
 from paddle.nn import AdaptiveAvgPool2D, MaxPool2D, AvgPool2D
 
-__all__ = ["Xception41_deeplab", "Xception65_deeplab", "Xception71_deeplab"]
+from ppcls.utils.save_load import load_dygraph_pretrain, load_dygraph_pretrain_from_url
+
+MODEL_URLS = {"Xception41_deeplab": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/Xception41_deeplab_pretrained.pdparams",
+             "Xception65_deeplab": "https://paddle-imagenet-models-name.bj.bcebos.com/dygraph/Xception41_deeplab_pretrained.pdparams"}
+
+__all__ = list(MODEL_URLS.keys())
 
 
 def check_data(data, number):
@@ -369,18 +388,28 @@ class XceptionDeeplab(nn.Layer):
         x = paddle.squeeze(x, axis=[2, 3])
         x = self._fc(x)
         return x
+    
+    
+def _load_pretrained(pretrained, model, model_url, use_ssld=False):
+    if pretrained is False:
+        pass
+    elif pretrained is True:
+        load_dygraph_pretrain_from_url(model, model_url, use_ssld=use_ssld)
+    elif isinstance(pretrained, str):
+        load_dygraph_pretrain(model, pretrained)
+    else:
+        raise RuntimeError(
+            "pretrained type is not available. Please use `string` or `boolean` type."
+        )
 
 
-def Xception41_deeplab(**args):
-    model = XceptionDeeplab('xception_41', **args)
-    return model
-
-
-def Xception65_deeplab(**args):
-    model = XceptionDeeplab("xception_65", **args)
+def Xception41_deeplab(pretrained=False, use_ssld=False, **kwargs):
+    model = XceptionDeeplab('xception_41', **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["Xception41_deeplab"], use_ssld=use_ssld)
     return model
 
 
-def Xception71_deeplab(**args):
-    model = XceptionDeeplab("xception_71", **args)
+def Xception65_deeplab(pretrained=False, use_ssld=False, **kwargs):
+    model = XceptionDeeplab("xception_65", **kwargs)
+    _load_pretrained(pretrained, model, MODEL_URLS["Xception65_deeplab"], use_ssld=use_ssld)
     return model

ppcls/configs/ImageNet/AlexNet/AlexNet.yaml

+# global configs
+Global:
+  checkpoints: null
+  pretrained_model: null
+  output_dir: ./output/
+  device: gpu
+  class_num: 1000
+  save_interval: 1
+  eval_during_train: True
+  eval_interval: 1
+  epochs: 120
+  print_batch_step: 10
+  use_visualdl: False
+  # used for static mode and model export
+  image_shape: [3, 224, 224]
+  save_inference_dir: ./inference
+
+# model architecture
+Arch:
+  name: AlexNet
+ 
+# loss function config for traing/eval process
+Loss:
+  Train:
+    - CELoss:
+        weight: 1.0
+  Eval:
+    - CELoss:
+        weight: 1.0
+
+
+Optimizer:
+  name: Momentum
+  momentum: 0.9
+  lr:
+    name: Piecewise
+    learning_rate: 0.01
+    decay_epochs: [30, 60, 90]
+    values: [0.1, 0.01, 0.001, 0.0001]
+  regularizer:
+    name: 'L2'
+    coeff: 0.0001
+
+
+# data loader for train and eval
+DataLoader:
+  Train:
+    dataset:
+      name: ImageNetDataset
+      image_root: ./dataset/ILSVRC2012/
+      cls_label_path: ./dataset/ILSVRC2012/train_list.txt
+      transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
+        - RandCropImage:
+            size: 224
+        - RandFlipImage:
+            flip_code: 1
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+            order: ''
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 64
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+  Eval:
+    # TOTO: modify to the latest trainer
+    dataset: 
+      name: ImageNetDataset
+      image_root: ./dataset/ILSVRC2012/
+      cls_label_path: ./dataset/ILSVRC2012/val_list.txt
+      transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
+        - ResizeImage:
+            resize_short: 256
+        - CropImage:
+            size: 224
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+            order: ''
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 64
+      drop_last: False
+      shuffle: False
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+Infer:
+  infer_imgs: docs/images/whl/demo.jpg
+  batch_size: 10
+  transforms:
+    - DecodeImage:
+        to_rgb: True
+        channel_first: False
+    - ResizeImage:
+        resize_short: 256
+    - CropImage:
+        size: 224
+    - NormalizeImage:
+        scale: 1.0/255.0
+        mean: [0.485, 0.456, 0.406]
+        std: [0.229, 0.224, 0.225]
+        order: ''
+    - ToCHWImage:
+  PostProcess:
+    name: Topk
+    topk: 5
+    class_id_map_file: ppcls/utils/imagenet1k_label_list.txt
+
+Metric:
+  Train:
+    - TopkAcc:
+        topk: [1, 5]
+  Eval:
+    - TopkAcc:
+        topk: [1, 5]

ppcls/configs/ImageNet/DPN/DPN107.yaml

+# global configs
+Global:
+  checkpoints: null
+  pretrained_model: null
+  output_dir: ./output/
+  device: gpu
+  class_num: 1000
+  save_interval: 1
+  eval_during_train: True
+  eval_interval: 1
+  epochs: 200
+  print_batch_step: 10
+  use_visualdl: False
+  # used for static mode and model export
+  image_shape: [3, 224, 224]
+  save_inference_dir: ./inference
+
+# model architecture
+Arch:
+  name: DPN107
+ 
+# loss function config for traing/eval process
+Loss:
+  Train:
+    - CELoss:
+        weight: 1.0
+        epsilon: 0.1
+  Eval:
+    - CELoss:
+        weight: 1.0
+
+
+Optimizer:
+  name: Momentum
+  momentum: 0.9
+  lr:
+    name: Cosine
+    learning_rate: 0.1
+  regularizer:
+    name: 'L2'
+    coeff: 0.0001
+
+
+# data loader for train and eval
+DataLoader:
+  Train:
+    dataset:
+      name: ImageNetDataset
+      image_root: ./dataset/ILSVRC2012/
+      cls_label_path: ./dataset/ILSVRC2012/train_list.txt
+      transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
+        - RandCropImage:
+            size: 224
+        - RandFlipImage:
+            flip_code: 1
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+            order: ''
+      batch_transform_ops:
+        - MixupOperator:
+            alpha: 0.2
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 64
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+  Eval:
+    # TOTO: modify to the latest trainer
+    dataset: 
+      name: ImageNetDataset
+      image_root: ./dataset/ILSVRC2012/
+      cls_label_path: ./dataset/ILSVRC2012/val_list.txt
+      transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
+        - ResizeImage:
+            resize_short: 256
+        - CropImage:
+            size: 224
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+            order: ''
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 64
+      drop_last: False
+      shuffle: False
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+Infer:
+  infer_imgs: docs/images/whl/demo.jpg
+  batch_size: 10
+  transforms:
+    - DecodeImage:
+        to_rgb: True
+        channel_first: False
+    - ResizeImage:
+        resize_short: 256
+    - CropImage:
+        size: 224
+    - NormalizeImage:
+        scale: 1.0/255.0
+        mean: [0.485, 0.456, 0.406]
+        std: [0.229, 0.224, 0.225]
+        order: ''
+    - ToCHWImage:
+  PostProcess:
+    name: Topk
+    topk: 5
+    class_id_map_file: ppcls/utils/imagenet1k_label_list.txt
+
+Metric:
+  Train:
+  Eval:
+    - TopkAcc:
+        topk: [1, 5]

ppcls/configs/ImageNet/HRNet/HRNet_W18_C.yaml

@@ -50,6 +50,9 @@ DataLoader:
       image_root: ./dataset/ILSVRC2012/
       cls_label_path: ./dataset/ILSVRC2012/train_list.txt
       transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
         - RandCropImage:
             size: 224
         - RandFlipImage:
@@ -76,6 +79,9 @@ DataLoader:
       image_root: ./dataset/ILSVRC2012/
       cls_label_path: ./dataset/ILSVRC2012/val_list.txt
       transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
         - ResizeImage:
             resize_short: 256
         - CropImage:

ppcls/configs/ImageNet/HRNet/HRNet_W30_C.yaml

@@ -50,6 +50,9 @@ DataLoader:
       image_root: ./dataset/ILSVRC2012/
       cls_label_path: ./dataset/ILSVRC2012/train_list.txt
       transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
         - RandCropImage:
             size: 224
         - RandFlipImage:
@@ -76,6 +79,9 @@ DataLoader:
       image_root: ./dataset/ILSVRC2012/
       cls_label_path: ./dataset/ILSVRC2012/val_list.txt
       transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
         - ResizeImage:
             resize_short: 256
         - CropImage:

ppcls/configs/ImageNet/HRNet/HRNet_W32_C.yaml

@@ -50,6 +50,9 @@ DataLoader:
       image_root: ./dataset/ILSVRC2012/
       cls_label_path: ./dataset/ILSVRC2012/train_list.txt
       transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
         - RandCropImage:
             size: 224
         - RandFlipImage:
@@ -76,6 +79,9 @@ DataLoader:
       image_root: ./dataset/ILSVRC2012/
       cls_label_path: ./dataset/ILSVRC2012/val_list.txt
       transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
         - ResizeImage:
             resize_short: 256
         - CropImage:

ppcls/configs/ImageNet/HRNet/HRNet_W40_C.yaml

@@ -50,6 +50,9 @@ DataLoader:
       image_root: ./dataset/ILSVRC2012/
       cls_label_path: ./dataset/ILSVRC2012/train_list.txt
       transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
         - RandCropImage:
             size: 224
         - RandFlipImage:
@@ -76,6 +79,9 @@ DataLoader:
       image_root: ./dataset/ILSVRC2012/
       cls_label_path: ./dataset/ILSVRC2012/val_list.txt
       transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
         - ResizeImage:
             resize_short: 256
         - CropImage:

ppcls/configs/ImageNet/HRNet/HRNet_W44_C.yaml

@@ -50,6 +50,9 @@ DataLoader:
       image_root: ./dataset/ILSVRC2012/
       cls_label_path: ./dataset/ILSVRC2012/train_list.txt
       transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
         - RandCropImage:
             size: 224
         - RandFlipImage:
@@ -76,6 +79,9 @@ DataLoader:
       image_root: ./dataset/ILSVRC2012/
       cls_label_path: ./dataset/ILSVRC2012/val_list.txt
       transform_ops:
+        - DecodeImage:
+            to_rgb: True
+            channel_first: False
         - ResizeImage:
             resize_short: 256
         - CropImage: