Merge pull request #944 from cuicheng01/develop

Update LeViT configs

README_ch.md

@@ -9,7 +9,7 @@
 **近期更新**
 
 - 2021.06.22,23,24 PaddleClas官方研发团队带来技术深入解读三日直播课，6月22日、23日、24日晚上20:30，[直播地址](https://live.bilibili.com/21689802)
-- 🔥🔥🔥: 2021.06.16 PaddleClas v2.2版本升级，集成Metric learning，向量检索等组件。新增商品识别、动漫人物识别、车辆识别和logo识别等4个图像识别应用。新增LeViT、Twins、TNT、DLA、HarDNet、RedNet系列30个预训练模型。
+- 🔥🔥🔥: 2021.06.16 PaddleClas v2.2版本升级，集成Metric learning，向量检索等组件。新增商品识别、动漫人物识别、车辆识别和logo识别等4个图像识别应用。新增LeViT、TNT、DLA、HarDNet、RedNet系列24个预训练模型。
 - [more](./docs/zh_CN/update_history.md)
 
 ## 特性
@@ -17,7 +17,7 @@
 - 实用的图像识别系统：集成了目标检测、特征学习、图像检索等模块，广泛适用于各类图像识别任务。
 提供商品识别、车辆识别、logo识别和动漫人物识别等4个场景应用示例。
 
-- 丰富的预训练模型库：提供了35个系列共164个ImageNet预训练模型，其中6个精选系列模型支持结构快速修改。
+- 丰富的预训练模型库：提供了33个系列共150个ImageNet预训练模型，其中6个精选系列模型支持结构快速修改。
 
 - 全面易用的特征学习组件：集成arcmargin, triplet loss等12度量学习方法，通过配置文件即可随意组合切换。
 

README_en.md

@@ -8,7 +8,7 @@ PaddleClas is an image recognition toolset for industry and academia, helping us
 
 **Recent updates**
 
-- 🔥🔥🔥: 2021.06.16 PaddleClas release/2.2. Add metric learning and vector search modules. Add product recognition, animation character recognition, vehicle recognition and logo recognition. Added 30 pretrained models of LeViT, Twins, TNT, DLA, HarDNet, and RedNet, and the accuracy is roughly the same as that of the paper.
+- 🔥🔥🔥: 2021.06.16 PaddleClas release/2.2. Add metric learning and vector search modules. Add product recognition, animation character recognition, vehicle recognition and logo recognition. Added 24 pretrained models of LeViT, TNT, DLA, HarDNet, and RedNet, and the accuracy is roughly the same as that of the paper.
 - [more](./docs/en/update_history_en.md)
 
 ## Features
@@ -16,7 +16,7 @@ PaddleClas is an image recognition toolset for industry and academia, helping us
 - A practical image recognition system consist of detection, feature learning and retrieval modules, widely applicable to all types of image recognition tasks.
 Four sample solutions are provided, including product recognition, vehicle recognition, logo recognition and animation character recognition.
 
-- Rich library of pre-trained models: Provide a total of 164 ImageNet pre-trained models in 34 series, among which 6 selected series of models support fast structural modification.
+- Rich library of pre-trained models: Provide a total of 150 ImageNet pre-trained models in 33 series, among which 6 selected series of models support fast structural modification.
 
 - Comprehensive and easy-to-use feature learning components: 12 metric learning methods are integrated and can be combined and switched at will through configuration files.
 

docs/en/update_history_en.md

 # Release Notes
 
-- 2021.05.14
-   - Add `SwinTransformer` series pretrained models, whose Top-1 Acc on ImageNet-1k dataset reaches 87.19%.
-
 - 2021.04.15
    - Add `MixNet` and `ReXNet` pretrained models, `MixNet_L`'s Top-1 Acc on ImageNet-1k reaches 78.6% and `ReXNet_3_0` reaches 82.09%.
 - 2021.01.27

docs/zh_CN/update_history.md

 # 更新日志
 
 
-- 2021.05.14
-   - 添加`SwinTransformer` 系列模型，在ImageNet-1k上，Top1 Acc可达87.19%
 - 2021.04.15
    - 添加`MixNet_L`和`ReXNet_3_0`系列模型，在ImageNet-1k上`MixNet` 模型Top1 Acc可达78.6%，`ReXNet`模型可达82.09%
 - 2021.01.27

ppcls/configs/ImageNet/Twins/pcpvt_base.yaml → ppcls/configs/ImageNet/DeiT/DeiT_base_distilled_patch16_224.yaml

@@ -17,7 +17,7 @@ Global:
 
 # model architecture
 Arch:
-  name: pcpvt_base
+  name: DeiT_base_distilled_patch16_224
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/DeiT/DeiT_base_distilled_patch16_384.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, 384, 384]
+  save_inference_dir: ./inference
+
+# model architecture
+Arch:
+  name: DeiT_base_distilled_patch16_384
+ 
+# 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: 0.1
+    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: 384
+        - 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:
+    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: 426
+        - CropImage:
+            size: 384
+        - 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: 426
+    - CropImage:
+        size: 384
+    - 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/Twins/alt_gvt_base.yaml → ppcls/configs/ImageNet/DeiT/DeiT_base_patch16_224.yaml

@@ -17,7 +17,7 @@ Global:
 
 # model architecture
 Arch:
-  name: alt_gvt_base
+  name: DeiT_base_patch16_224
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/DeiT/DeiT_base_patch16_384.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, 384, 384]
+  save_inference_dir: ./inference
+
+# model architecture
+Arch:
+  name: DeiT_base_patch16_384
+ 
+# 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: 0.1
+    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: 384
+        - 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:
+    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: 426
+        - CropImage:
+            size: 384
+        - 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: 426
+    - CropImage:
+        size: 384
+    - 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/DeiT/DeiT_small_distilled_patch16_224.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: DeiT_small_distilled_patch16_224
+ 
+# 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: 0.1
+    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:
+    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/Twins/alt_gvt_small.yaml → ppcls/configs/ImageNet/DeiT/DeiT_small_patch16_224.yaml

@@ -17,7 +17,7 @@ Global:
 
 # model architecture
 Arch:
-  name: alt_gvt_small
+  name: DeiT_small_patch16_224
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/DeiT/DeiT_tiny_distilled_patch16_224.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: DeiT_tiny_distilled_patch16_224
+ 
+# 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: 0.1
+    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:
+    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/Twins/alt_gvt_large.yaml → ppcls/configs/ImageNet/DeiT/DeiT_tiny_patch16_224.yaml

@@ -17,7 +17,7 @@ Global:
 
 # model architecture
 Arch:
-  name: alt_gvt_large
+  name: DeiT_tiny_patch16_224
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/Inception/GoogLeNet.yaml

@@ -22,11 +22,10 @@ Arch:
 # loss function config for traing/eval process
 Loss:
   Train:
-    - CELoss:
+    - GoogLeNetLoss:
         weight: 1.0
-        epsilon: 0.1
   Eval:
-    - CELoss:
+    - GoogLeNetLoss:
         weight: 1.0
 
 
@@ -36,6 +35,7 @@ Optimizer:
   lr:
     name: Cosine
     learning_rate: 0.01
+    warmup_epoch: 5
   regularizer:
     name: 'L2'
     coeff: 0.0001
@@ -77,6 +77,9 @@ 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:
@@ -119,8 +122,8 @@ Infer:
 
 Metric:
   Train:
-    - TopkAcc:
+    - GoogLeNetTopkAcc:
         topk: [1, 5]
   Eval:
-    - TopkAcc:
+    - GoogLeNetTopkAcc:
         topk: [1, 5]

ppcls/configs/ImageNet/LeViT/LeViT_128.yaml

@@ -17,7 +17,7 @@ Global:
 
 # model architecture
 Arch:
-  name: LeViT128
+  name: LeViT_128
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/LeViT/LeViT_128S.yaml

@@ -17,7 +17,7 @@ Global:
 
 # model architecture
 Arch:
-  name: LeViT128S
+  name: LeViT_128S
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/LeViT/LeViT_192.yaml

@@ -17,7 +17,7 @@ Global:
 
 # model architecture
 Arch:
-  name: LeViT192
+  name: LeViT_192
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/LeViT/LeViT_256.yaml

@@ -17,7 +17,7 @@ Global:
 
 # model architecture
 Arch:
-  name: LeViT256
+  name: LeViT_256
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/LeViT/LeViT_384.yaml

@@ -17,7 +17,7 @@ Global:
 
 # model architecture
 Arch:
-  name: LeViT384
+  name: LeViT_384
  
 # loss function config for traing/eval process
 Loss:

ppcls/configs/ImageNet/Twins/pcpvt_large.yaml → ppcls/configs/ImageNet/VisionTransformer/ViT_base_patch16_224.yaml

@@ -17,7 +17,7 @@ Global:
 
 # model architecture
 Arch:
-  name: pcpvt_large
+  name: ViT_base_patch16_224
  
 # loss function config for traing/eval process
 Loss:
@@ -59,8 +59,8 @@ DataLoader:
             flip_code: 1
         - NormalizeImage:
             scale: 1.0/255.0
-            mean: [0.485, 0.456, 0.406]
-            std: [0.229, 0.224, 0.225]
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
             order: ''
 
     sampler:
@@ -87,8 +87,8 @@ DataLoader:
             size: 224
         - NormalizeImage:
             scale: 1.0/255.0
-            mean: [0.485, 0.456, 0.406]
-            std: [0.229, 0.224, 0.225]
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
             order: ''
     sampler:
       name: DistributedBatchSampler
@@ -112,8 +112,8 @@ Infer:
         size: 224
     - NormalizeImage:
         scale: 1.0/255.0
-        mean: [0.485, 0.456, 0.406]
-        std: [0.229, 0.224, 0.225]
+        mean: [0.5, 0.5, 0.5]
+        std: [0.5, 0.5, 0.5]
         order: ''
     - ToCHWImage:
   PostProcess:

ppcls/configs/ImageNet/VisionTransformer/ViT_base_patch16_384.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, 384, 384]
+  save_inference_dir: ./inference
+
+# model architecture
+Arch:
+  name: ViT_base_patch16_224
+ 
+# 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: 0.1
+    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: 384
+        - RandFlipImage:
+            flip_code: 1
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            order: ''
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 64
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+  Eval:
+    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: 384
+        - CropImage:
+            size: 384
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            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: 384
+    - CropImage:
+        size: 384
+    - NormalizeImage:
+        scale: 1.0/255.0
+        mean: [0.5, 0.5, 0.5]
+        std: [0.5, 0.5, 0.5]
+        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/VisionTransformer/ViT_base_patch32_384.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, 384, 384]
+  save_inference_dir: ./inference
+
+# model architecture
+Arch:
+  name: ViT_base_patch32_384
+ 
+# 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: 0.1
+    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: 384
+        - RandFlipImage:
+            flip_code: 1
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            order: ''
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 64
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+  Eval:
+    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: 384
+        - CropImage:
+            size: 384
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            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: 384
+    - CropImage:
+        size: 384
+    - NormalizeImage:
+        scale: 1.0/255.0
+        mean: [0.5, 0.5, 0.5]
+        std: [0.5, 0.5, 0.5]
+        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/Twins/pcpvt_small.yaml → ppcls/configs/ImageNet/VisionTransformer/ViT_huge_patch16_224.yaml

@@ -17,7 +17,7 @@ Global:
 
 # model architecture
 Arch:
-  name: pcpvt_small
+  name:  ViT_huge_patch16_224
  
 # loss function config for traing/eval process
 Loss:
@@ -59,8 +59,8 @@ DataLoader:
             flip_code: 1
         - NormalizeImage:
             scale: 1.0/255.0
-            mean: [0.485, 0.456, 0.406]
-            std: [0.229, 0.224, 0.225]
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
             order: ''
 
     sampler:
@@ -87,8 +87,8 @@ DataLoader:
             size: 224
         - NormalizeImage:
             scale: 1.0/255.0
-            mean: [0.485, 0.456, 0.406]
-            std: [0.229, 0.224, 0.225]
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
             order: ''
     sampler:
       name: DistributedBatchSampler
@@ -112,8 +112,8 @@ Infer:
         size: 224
     - NormalizeImage:
         scale: 1.0/255.0
-        mean: [0.485, 0.456, 0.406]
-        std: [0.229, 0.224, 0.225]
+        mean: [0.5, 0.5, 0.5]
+        std: [0.5, 0.5, 0.5]
         order: ''
     - ToCHWImage:
   PostProcess:

ppcls/configs/ImageNet/VisionTransformer/ViT_huge_patch32_384.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, 384, 384]
+  save_inference_dir: ./inference
+
+# model architecture
+Arch:
+  name: ViT_huge_patch32_384
+ 
+# 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: 0.1
+    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: 384
+        - RandFlipImage:
+            flip_code: 1
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            order: ''
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 64
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+  Eval:
+    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: 384
+        - CropImage:
+            size: 384
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            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: 384
+    - CropImage:
+        size: 384
+    - NormalizeImage:
+        scale: 1.0/255.0
+        mean: [0.5, 0.5, 0.5]
+        std: [0.5, 0.5, 0.5]
+        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/VisionTransformer/ViT_large_patch16_224.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: ViT_large_patch16_224
+ 
+# 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: 0.1
+    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: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            order: ''
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 64
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+  Eval:
+    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: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            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: [0.5, 0.5, 0.5]
+        std: [0.5, 0.5, 0.5]
+        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/VisionTransformer/ViT_large_patch16_384.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, 384, 384]
+  save_inference_dir: ./inference
+
+# model architecture
+Arch:
+  name: ViT_large_patch16_384
+ 
+# 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: 0.1
+    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: 384
+        - RandFlipImage:
+            flip_code: 1
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            order: ''
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 64
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+  Eval:
+    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: 384
+        - CropImage:
+            size: 384
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            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: 384
+    - CropImage:
+        size: 384
+    - NormalizeImage:
+        scale: 1.0/255.0
+        mean: [0.5, 0.5, 0.5]
+        std: [0.5, 0.5, 0.5]
+        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/VisionTransformer/ViT_large_patch32_384.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, 384, 384]
+  save_inference_dir: ./inference
+
+# model architecture
+Arch:
+  name: ViT_large_patch32_384
+ 
+# 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: 0.1
+    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: 384
+        - RandFlipImage:
+            flip_code: 1
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            order: ''
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 64
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+  Eval:
+    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: 384
+        - CropImage:
+            size: 384
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            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: 384
+    - CropImage:
+        size: 384
+    - NormalizeImage:
+        scale: 1.0/255.0
+        mean: [0.5, 0.5, 0.5]
+        std: [0.5, 0.5, 0.5]
+        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/VisionTransformer/ViT_small_patch16_224.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: ViT_small_patch16_224
+ 
+# 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: 0.1
+    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: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            order: ''
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 64
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 4
+      use_shared_memory: True
+
+  Eval:
+    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: [0.5, 0.5, 0.5]
+            std: [0.5, 0.5, 0.5]
+            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: [0.5, 0.5, 0.5]
+        std: [0.5, 0.5, 0.5]
+        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/loss/__init__.py

@@ -5,6 +5,7 @@ import paddle.nn as nn
 from ppcls.utils import logger
 
 from .celoss import CELoss
+from .googlenetloss import GoogLeNetLoss
 from .centerloss import CenterLoss
 from .emlloss import EmlLoss
 from .msmloss import MSMLoss

ppcls/loss/googlenetloss.py

+# 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
+import paddle.nn as nn
+import paddle.nn.functional as F
+
+
+class GoogLeNetLoss(nn.Layer):
+    """
+    Cross entropy loss used after googlenet
+    """
+    def __init__(self, epsilon=None):
+        super().__init__()
+        assert (epsilon is None or epsilon <= 0 or epsilon >= 1), "googlenet is not support label_smooth"
+            
+
+    def forward(self, inputs, label):
+        input0, input1, input2 = inputs
+        if isinstance(input0, dict):
+            input0 = input0["logits"]
+        if isinstance(input1, dict):
+            input1 = input1["logits"]
+        if isinstance(input2, dict):
+            input2 = input2["logits"]
+
+        loss0 = F.cross_entropy(input0, label=label, soft_label=False)
+        loss1 = F.cross_entropy(input1, label=label, soft_label=False)
+        loss2 = F.cross_entropy(input2, label=label, soft_label=False)
+        loss = loss0 + 0.3 * loss1 + 0.3 * loss2
+        loss = loss.mean()
+        return {"GooleNetLoss": loss}

ppcls/metric/__init__.py

@@ -18,6 +18,7 @@ from collections import OrderedDict
 
 from .metrics import TopkAcc, mAP, mINP, Recallk
 from .metrics import DistillationTopkAcc
+from .metrics import GoogLeNetTopkAcc
 
 class CombinedMetrics(nn.Layer):
     def __init__(self, config_list):

ppcls/metric/metrics.py

@@ -120,3 +120,16 @@ class DistillationTopkAcc(TopkAcc):
         if self.feature_key is not None:
             x = x[self.feature_key]
         return super().forward(x, label)
+
+
+class GoogLeNetTopkAcc(TopkAcc):
+    def __init__(self, topk=(1, 5)):
+        super().__init__()
+        assert isinstance(topk, (int, list, tuple))
+        if isinstance(topk, int):
+            topk = [topk]
+        self.topk = topk
+
+    def forward(self, x, label):
+        return super().forward(x[0], label)
+