add PP-HGNetV2, PP-HGNet models (#1882)

README.md

@@ -63,7 +63,7 @@ PaddleX 3.0 覆盖了 16 条产业级模型产线，其中 9 条基础产线可
     <td>通用图像分类</td>
     <td>图像分类</td>
     <td>CLIP_vit_base_patch16_224<br/>CLIP_vit_large_patch14_224<details>
-    <summary><b>more</b></summary><br/>ConvNeXt_tiny<br/>MobileNetV2_x0_25<br/>MobileNetV2_x0_5<br/>MobileNetV2_x1_0<br/>MobileNetV2_x1_5<br/>MobileNetV2_x2_0<br/>MobileNetV3_large_x0_35<br/>MobileNetV3_large_x0_5<br/>MobileNetV3_large_x0_75<br/>MobileNetV3_large_x1_0<br/>MobileNetV3_large_x1_25<br/>MobileNetV3_small_x0_35<br/>MobileNetV3_small_x0_5<br/>MobileNetV3_small_x0_75<br/>MobileNetV3_small_x1_0<br/>MobileNetV3_small_x1_25<br/>PP-HGNet_small<br/>PP-HGNetV2-B0<br/>PP-HGNetV2-B4<br/>PP-HGNetV2-B6<br/>PP-LCNet_x0_25<br/>PP-LCNet_x0_35<br/>PP-LCNet_x0_5<br/>PP-LCNet_x0_75<br/>PP-LCNet_x1_0<br/>PP-LCNet_x1_5<br/>PP-LCNet_x2_0<br/>PP-LCNet_x2_5<br/>ResNet18<br/>ResNet34<br/>ResNet50<br/>ResNet101<br/>ResNet152<br/>SwinTransformer_base_patch4_window7_224</details></td>
+    <summary><b>more</b></summary><br/>ConvNeXt_tiny<br/>ConvNeXt_small<br/>ConvNeXt_base_224<br/>ConvNeXt_base_384<br/>ConvNeXt_large_224<br/>ConvNeXt_large_384<br/>MobileNetV1_x0_25<br/>MobileNetV1_x0_5<br/>MobileNetV1_x0_75<br/>MobileNetV1_x1_0<br/>MobileNetV2_x0_25<br/>MobileNetV2_x0_5<br/>MobileNetV2_x1_0<br/>MobileNetV2_x1_5<br/>MobileNetV2_x2_0<br/>MobileNetV3_large_x0_35<br/>MobileNetV3_large_x0_5<br/>MobileNetV3_large_x0_75<br/>MobileNetV3_large_x1_0<br/>MobileNetV3_large_x1_25<br/>MobileNetV3_small_x0_35<br/>MobileNetV3_small_x0_5<br/>MobileNetV3_small_x0_75<br/>MobileNetV3_small_x1_0<br/>MobileNetV3_small_x1_25<br/>PP-HGNet_tiny<br/>PP-HGNet_small<br/>PP-HGNet_base<br/>PP-HGNetV2-B0<br/>PP-HGNetV2-B1<br/>PP-HGNetV2-B2<br/>PP-HGNetV2-B3<br/>PP-HGNetV2-B4<br/>PP-HGNetV2-B5<br/>PP-HGNetV2-B6<br/>PP-LCNet_x0_25<br/>PP-LCNet_x0_35<br/>PP-LCNet_x0_5<br/>PP-LCNet_x0_75<br/>PP-LCNet_x1_0<br/>PP-LCNet_x1_5<br/>PP-LCNet_x2_0<br/>PP-LCNet_x2_5<br/>PP-LCNetV2_small<br/>PP-LCNetV2_base<br/>PP-LCNetV2_large<br/>ResNet18<br/>ResNet18_vd<br/>ResNet34<br/>ResNet34_vd<br/>ResNet50<br/>ResNet50_vd<br/>ResNet101<br/>ResNet101_vd<br/>ResNet152<br/>ResNet152_vd<br/>ResNet200_vd<br/>SwinTransformer_tiny_patch4_window7_224<br/>SwinTransformer_small_patch4_window7_224<br/>SwinTransformer_base_patch4_window7_224<br/>SwinTransformer_base_patch4_window12_384<br/>SwinTransformer_large_patch4_window7_224<br/>SwinTransformer_large_patch4_window12_384</details></td>
   </tr>
   <tr>
     <td>基础产线</td>

docs/tutorials/models/support_model_list.md

@@ -60,12 +60,18 @@
 ### 6.PP-HGNet 系列
 | 模型名称 | config |
 | :--- | :---: |
+| PP-HGNet_tiny | [PP-HGNet_tiny.yaml](../../../paddlex/configs/image_classification/PP-HGNet_tiny.yaml)|
 | PP-HGNet_small | [PP-HGNet_small.yaml](../../../paddlex/configs/image_classification/PP-HGNet_small.yaml)|
+| PP-HGNet_base | [PP-HGNet_base.yaml](../../../paddlex/configs/image_classification/PP-HGNet_base.yaml)|
 ### 7.PP-HGNetV2 系列
 | 模型名称 | config |
 | :--- | :---: |
 | PP-HGNetV2-B0 | [PP-HGNetV2-B0.yaml](../../../paddlex/configs/image_classification/PP-HGNetV2-B0.yaml)|
+| PP-HGNetV2-B1 | [PP-HGNetV2-B1.yaml](../../../paddlex/configs/image_classification/PP-HGNetV2-B1.yaml)|
+| PP-HGNetV2-B2 | [PP-HGNetV2-B2.yaml](../../../paddlex/configs/image_classification/PP-HGNetV2-B2.yaml)|
+| PP-HGNetV2-B3 | [PP-HGNetV2-B3.yaml](../../../paddlex/configs/image_classification/PP-HGNetV2-B3.yaml)|
 | PP-HGNetV2-B4 | [PP-HGNetV2-B4.yaml](../../../paddlex/configs/image_classification/PP-HGNetV2-B4.yaml)|
+| PP-HGNetV2-B5 | [PP-HGNetV2-B5.yaml](../../../paddlex/configs/image_classification/PP-HGNetV2-B5.yaml)|
 | PP-HGNetV2-B6 | [PP-HGNetV2-B6.yaml](../../../paddlex/configs/image_classification/PP-HGNetV2-B6.yaml)|
 ### 8.CLIP 系列
 | 模型名称 | config |

paddlex/configs/image_classification/PP-HGNetV2-B1.yaml

@@ -0,0 +1,38 @@
+Global:
+  model: PP-HGNetV2-B1
+  mode: check_dataset # check_dataset/train/evaluate/predict
+  dataset_dir: "/paddle/dataset/paddlex/cls/cls_flowers_examples"
+  device: gpu:0,1,2,3
+  output: "output"
+
+CheckDataset:
+  convert: 
+    enable: False
+    src_dataset_type: null
+  split: 
+    enable: False
+    train_percent: null
+    val_percent: null
+
+Train:
+  num_classes: 102
+  epochs_iters: 20
+  batch_size: 64
+  learning_rate: 0.05
+  pretrain_weight_path: null
+  warmup_steps: 5
+  resume_path: null
+  log_interval: 1
+  eval_interval: 1
+  save_interval: 1
+
+Evaluate:
+  weight_path: "output/best_model.pdparams"
+  log_interval: 1
+
+Predict:
+  model_dir: "output/best_model"
+  input_path: "https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_image_classification_001.jpg"
+  kernel_option:
+    run_mode: paddle
+    batch_size: 1

paddlex/configs/image_classification/PP-HGNetV2-B2.yaml

@@ -0,0 +1,38 @@
+Global:
+  model: PP-HGNetV2-B2
+  mode: check_dataset # check_dataset/train/evaluate/predict
+  dataset_dir: "/paddle/dataset/paddlex/cls/cls_flowers_examples"
+  device: gpu:0,1,2,3
+  output: "output"
+
+CheckDataset:
+  convert: 
+    enable: False
+    src_dataset_type: null
+  split: 
+    enable: False
+    train_percent: null
+    val_percent: null
+
+Train:
+  num_classes: 102
+  epochs_iters: 20
+  batch_size: 64
+  learning_rate: 0.05
+  pretrain_weight_path: null
+  warmup_steps: 5
+  resume_path: null
+  log_interval: 1
+  eval_interval: 1
+  save_interval: 1
+
+Evaluate:
+  weight_path: "output/best_model.pdparams"
+  log_interval: 1
+
+Predict:
+  model_dir: "output/best_model"
+  input_path: "https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_image_classification_001.jpg"
+  kernel_option:
+    run_mode: paddle
+    batch_size: 1

paddlex/configs/image_classification/PP-HGNetV2-B3.yaml

@@ -0,0 +1,38 @@
+Global:
+  model: PP-HGNetV2-B3
+  mode: check_dataset # check_dataset/train/evaluate/predict
+  dataset_dir: "/paddle/dataset/paddlex/cls/cls_flowers_examples"
+  device: gpu:0,1,2,3
+  output: "output"
+
+CheckDataset:
+  convert: 
+    enable: False
+    src_dataset_type: null
+  split: 
+    enable: False
+    train_percent: null
+    val_percent: null
+
+Train:
+  num_classes: 102
+  epochs_iters: 20
+  batch_size: 64
+  learning_rate: 0.05
+  pretrain_weight_path: null
+  warmup_steps: 5
+  resume_path: null
+  log_interval: 1
+  eval_interval: 1
+  save_interval: 1
+
+Evaluate:
+  weight_path: "output/best_model.pdparams"
+  log_interval: 1
+
+Predict:
+  model_dir: "output/best_model"
+  input_path: "https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_image_classification_001.jpg"
+  kernel_option:
+    run_mode: paddle
+    batch_size: 1

paddlex/configs/image_classification/PP-HGNetV2-B5.yaml

@@ -0,0 +1,38 @@
+Global:
+  model: PP-HGNetV2-B5
+  mode: check_dataset # check_dataset/train/evaluate/predict
+  dataset_dir: "/paddle/dataset/paddlex/cls/cls_flowers_examples"
+  device: gpu:0,1,2,3
+  output: "output"
+
+CheckDataset:
+  convert: 
+    enable: False
+    src_dataset_type: null
+  split: 
+    enable: False
+    train_percent: null
+    val_percent: null
+
+Train:
+  num_classes: 102
+  epochs_iters: 20
+  batch_size: 64
+  learning_rate: 0.05
+  pretrain_weight_path: null
+  warmup_steps: 5
+  resume_path: null
+  log_interval: 1
+  eval_interval: 1
+  save_interval: 1
+
+Evaluate:
+  weight_path: "output/best_model.pdparams"
+  log_interval: 1
+
+Predict:
+  model_dir: "output/best_model"
+  input_path: "https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_image_classification_001.jpg"
+  kernel_option:
+    run_mode: paddle
+    batch_size: 1

paddlex/configs/image_classification/PP-HGNet_base.yaml

@@ -0,0 +1,38 @@
+Global:
+  model: PP-HGNet_base
+  mode: check_dataset # check_dataset/train/evaluate/predict
+  dataset_dir: "/paddle/dataset/paddlex/cls/cls_flowers_examples"
+  device: gpu:0,1,2,3
+  output: "output"
+
+CheckDataset:
+  convert: 
+    enable: False
+    src_dataset_type: null
+  split: 
+    enable: False
+    train_percent: null
+    val_percent: null
+
+Train:
+  num_classes: 102
+  epochs_iters: 20
+  batch_size: 64
+  learning_rate: 0.25
+  pretrain_weight_path: null
+  warmup_steps: 5
+  resume_path: null
+  log_interval: 1
+  eval_interval: 1
+  save_interval: 1
+
+Evaluate:
+  weight_path: "output/best_model.pdparams"
+  log_interval: 1
+
+Predict:
+  model_dir: "output/best_model"
+  input_path: "https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_image_classification_001.jpg"
+  kernel_option:
+    run_mode: paddle
+    batch_size: 1

paddlex/configs/image_classification/PP-HGNet_tiny.yaml

@@ -0,0 +1,38 @@
+Global:
+  model: PP-HGNet_tiny
+  mode: check_dataset # check_dataset/train/evaluate/predict
+  dataset_dir: "/paddle/dataset/paddlex/cls/cls_flowers_examples"
+  device: gpu:0,1,2,3
+  output: "output"
+
+CheckDataset:
+  convert: 
+    enable: False
+    src_dataset_type: null
+  split: 
+    enable: False
+    train_percent: null
+    val_percent: null
+
+Train:
+  num_classes: 102
+  epochs_iters: 20
+  batch_size: 64
+  learning_rate: 0.25
+  pretrain_weight_path: null
+  warmup_steps: 5
+  resume_path: null
+  log_interval: 1
+  eval_interval: 1
+  save_interval: 1
+
+Evaluate:
+  weight_path: "output/best_model.pdparams"
+  log_interval: 1
+
+Predict:
+  model_dir: "output/best_model"
+  input_path: "https://paddle-model-ecology.bj.bcebos.com/paddlex/imgs/demo_image/general_image_classification_001.jpg"
+  kernel_option:
+    run_mode: paddle
+    batch_size: 1

paddlex/modules/base/predictor/utils/official_models.py

@@ -145,12 +145,24 @@ SwinTransformer_large_patch4_window7_224_infer.tar",
     "SwinTransformer_large_patch4_window12_384":
     "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0/\
 SwinTransformer_large_patch4_window12_384_infer.tar",
+    "PP-HGNet_tiny":
+    "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0/PP-HGNet_tiny_infer.tar",
     "PP-HGNet_small":
     "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0/PP-HGNet_small_infer.tar",
+    "PP-HGNet_base":
+    "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0/PP-HGNet_base_infer.tar",
     "PP-HGNetV2-B0":
     "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0/PP-HGNetV2-B0_infer.tar",
+    "PP-HGNetV2-B1":
+    "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0/PP-HGNetV2-B1_infer.tar",
+    "PP-HGNetV2-B2":
+    "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0/PP-HGNetV2-B2_infer.tar",
+    "PP-HGNetV2-B3":
+    "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0/PP-HGNetV2-B3_infer.tar",
     "PP-HGNetV2-B4":
     "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0/PP-HGNetV2-B4_infer.tar",
+    "PP-HGNetV2-B5":
+    "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0/PP-HGNetV2-B5_infer.tar",
     "PP-HGNetV2-B6":
     "https://paddle-model-ecology.bj.bcebos.com/paddlex/official_inference_model/paddle3.0/PP-HGNetV2-B6_infer.tar",
     "CLIP_vit_base_patch16_224":

paddlex/modules/image_classification/model_list.py

@@ -40,9 +40,15 @@ MODELS = [
     'MobileNetV3_small_x0_75',
     'MobileNetV3_small_x1_0',
     'MobileNetV3_small_x1_25',
+    'PP-HGNet_tiny',
     'PP-HGNet_small',
+    'PP-HGNet_base',
     'PP-HGNetV2-B0',
+    'PP-HGNetV2-B1',
+    'PP-HGNetV2-B2',
+    'PP-HGNetV2-B3',
     'PP-HGNetV2-B4',
+    'PP-HGNetV2-B5',
     'PP-HGNetV2-B6',
     'PP-LCNet_x0_25',
     'PP-LCNet_x0_35',

paddlex/repo_apis/PaddleClas_api/cls/register.py

@@ -177,6 +177,13 @@ register_model_info({
 })
 
 register_model_info({
+    'model_name': 'PP-HGNet_tiny',
+    'suite': 'Cls',
+    'config_path': osp.join(PDX_CONFIG_DIR, 'PP-HGNet_tiny.yaml'),
+    'supported_apis': ['train', 'evaluate', 'predict', 'export'],
+})
+
+register_model_info({
     'model_name': 'PP-HGNet_small',
     'suite': 'Cls',
     'config_path': osp.join(PDX_CONFIG_DIR, 'PP-HGNet_small.yaml'),
@@ -184,6 +191,13 @@ register_model_info({
 })
 
 register_model_info({
+    'model_name': 'PP-HGNet_base',
+    'suite': 'Cls',
+    'config_path': osp.join(PDX_CONFIG_DIR, 'PP-HGNet_base.yaml'),
+    'supported_apis': ['train', 'evaluate', 'predict', 'export'],
+})
+
+register_model_info({
     'model_name': 'PP-HGNetV2-B0',
     'suite': 'Cls',
     'config_path': osp.join(PDX_CONFIG_DIR, 'PP-HGNetV2-B0.yaml'),
@@ -191,6 +205,27 @@ register_model_info({
 })
 
 register_model_info({
+    'model_name': 'PP-HGNetV2-B1',
+    'suite': 'Cls',
+    'config_path': osp.join(PDX_CONFIG_DIR, 'PP-HGNetV2-B1.yaml'),
+    'supported_apis': ['train', 'evaluate', 'predict', 'export'],
+})
+
+register_model_info({
+    'model_name': 'PP-HGNetV2-B2',
+    'suite': 'Cls',
+    'config_path': osp.join(PDX_CONFIG_DIR, 'PP-HGNetV2-B2.yaml'),
+    'supported_apis': ['train', 'evaluate', 'predict', 'export'],
+})
+
+register_model_info({
+    'model_name': 'PP-HGNetV2-B3',
+    'suite': 'Cls',
+    'config_path': osp.join(PDX_CONFIG_DIR, 'PP-HGNetV2-B3.yaml'),
+    'supported_apis': ['train', 'evaluate', 'predict', 'export'],
+})
+
+register_model_info({
     'model_name': 'PP-HGNetV2-B4',
     'suite': 'Cls',
     'config_path': osp.join(PDX_CONFIG_DIR, 'PP-HGNetV2-B4.yaml'),
@@ -198,6 +233,13 @@ register_model_info({
 })
 
 register_model_info({
+    'model_name': 'PP-HGNetV2-B5',
+    'suite': 'Cls',
+    'config_path': osp.join(PDX_CONFIG_DIR, 'PP-HGNetV2-B5.yaml'),
+    'supported_apis': ['train', 'evaluate', 'predict', 'export'],
+})
+
+register_model_info({
     'model_name': 'PP-HGNetV2-B6',
     'suite': 'Cls',
     'config_path': osp.join(PDX_CONFIG_DIR, 'PP-HGNetV2-B6.yaml'),

paddlex/repo_apis/PaddleClas_api/configs/PP-HGNetV2-B1.yaml

@@ -0,0 +1,164 @@
+## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
+# global configs
+Global:
+  checkpoints: null
+  pretrained_model: null
+  output_dir: ./output/
+  device: gpu
+  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
+  # training model under @to_static
+  to_static: False
+  use_dali: False
+
+# mixed precision training
+AMP:
+  use_amp: True
+  use_fp16_test: False
+  scale_loss: 128.0
+  use_dynamic_loss_scaling: True
+  use_promote: False
+  # O1: mixed fp16, O2: pure fp16
+  level: O1
+
+# model architecture
+Arch:
+  name: PPHGNetV2_B1
+  class_num: 102
+  pretrained: True # ssld pretrained
+ 
+# 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
+    # for global bs 1024, when finetune training, you need to reduce learning_rate manually
+    learning_rate: 0.5
+    warmup_epoch: 5
+  regularizer:
+    name: 'L2'
+    coeff: 0.00002
+
+
+# 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
+            interpolation: bicubic
+            backend: pil
+        - RandFlipImage:
+            flip_code: 1
+        - TimmAutoAugment:
+            config_str: rand-m7-mstd0.5-inc1
+            interpolation: bicubic
+            img_size: 224
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+            order: ''
+        - RandomErasing:
+            EPSILON: 0.25
+            sl: 0.02
+            sh: 1.0/3.0
+            r1: 0.3
+            attempt: 10
+            use_log_aspect: True
+            mode: pixel
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 128
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 16
+      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: 232
+            interpolation: bicubic
+            backend: pil
+        - 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: 128
+      drop_last: False
+      shuffle: False
+    loader:
+      num_workers: 16
+      use_shared_memory: True
+
+Infer:
+  infer_imgs: docs/images/inference_deployment/whl_demo.jpg
+  batch_size: 10
+  transforms:
+    - DecodeImage:
+        to_rgb: True
+        channel_first: False
+    - ResizeImage:
+        resize_short: 232
+        interpolation: bicubic
+        backend: pil
+    - 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]

paddlex/repo_apis/PaddleClas_api/configs/PP-HGNetV2-B2.yaml

@@ -0,0 +1,164 @@
+## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
+# global configs
+Global:
+  checkpoints: null
+  pretrained_model: null
+  output_dir: ./output/
+  device: gpu
+  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
+  # training model under @to_static
+  to_static: False
+  use_dali: False
+
+# mixed precision training
+AMP:
+  use_amp: True
+  use_fp16_test: False
+  scale_loss: 128.0
+  use_dynamic_loss_scaling: True
+  use_promote: False
+  # O1: mixed fp16, O2: pure fp16
+  level: O1
+
+# model architecture
+Arch:
+  name: PPHGNetV2_B2
+  class_num: 102
+  pretrained: True # ssld pretrained
+ 
+# 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
+    # for global bs 1024, when finetune training, you need to reduce learning_rate manually
+    learning_rate: 0.5
+    warmup_epoch: 5
+  regularizer:
+    name: 'L2'
+    coeff: 0.00002
+
+
+# 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
+            interpolation: bicubic
+            backend: pil
+        - RandFlipImage:
+            flip_code: 1
+        - TimmAutoAugment:
+            config_str: rand-m7-mstd0.5-inc1
+            interpolation: bicubic
+            img_size: 224
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+            order: ''
+        - RandomErasing:
+            EPSILON: 0.25
+            sl: 0.02
+            sh: 1.0/3.0
+            r1: 0.3
+            attempt: 10
+            use_log_aspect: True
+            mode: pixel
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 128
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 16
+      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: 232
+            interpolation: bicubic
+            backend: pil
+        - 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: 128
+      drop_last: False
+      shuffle: False
+    loader:
+      num_workers: 16
+      use_shared_memory: True
+
+Infer:
+  infer_imgs: docs/images/inference_deployment/whl_demo.jpg
+  batch_size: 10
+  transforms:
+    - DecodeImage:
+        to_rgb: True
+        channel_first: False
+    - ResizeImage:
+        resize_short: 232
+        interpolation: bicubic
+        backend: pil
+    - 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]

paddlex/repo_apis/PaddleClas_api/configs/PP-HGNetV2-B3.yaml

@@ -0,0 +1,164 @@
+## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
+# global configs
+Global:
+  checkpoints: null
+  pretrained_model: null
+  output_dir: ./output/
+  device: gpu
+  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
+  # training model under @to_static
+  to_static: False
+  use_dali: False
+
+# mixed precision training
+AMP:
+  use_amp: True
+  use_fp16_test: False
+  scale_loss: 128.0
+  use_dynamic_loss_scaling: True
+  use_promote: False
+  # O1: mixed fp16, O2: pure fp16
+  level: O1
+
+# model architecture
+Arch:
+  name: PPHGNetV2_B3
+  class_num: 102
+  pretrained: True # ssld pretrained
+ 
+# 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
+    # for global bs 1024, when finetune training, you need to reduce learning_rate manually
+    learning_rate: 0.5
+    warmup_epoch: 5
+  regularizer:
+    name: 'L2'
+    coeff: 0.00002
+
+
+# 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
+            interpolation: bicubic
+            backend: pil
+        - RandFlipImage:
+            flip_code: 1
+        - TimmAutoAugment:
+            config_str: rand-m7-mstd0.5-inc1
+            interpolation: bicubic
+            img_size: 224
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+            order: ''
+        - RandomErasing:
+            EPSILON: 0.25
+            sl: 0.02
+            sh: 1.0/3.0
+            r1: 0.3
+            attempt: 10
+            use_log_aspect: True
+            mode: pixel
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 128
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 16
+      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: 232
+            interpolation: bicubic
+            backend: pil
+        - 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: 128
+      drop_last: False
+      shuffle: False
+    loader:
+      num_workers: 16
+      use_shared_memory: True
+
+Infer:
+  infer_imgs: docs/images/inference_deployment/whl_demo.jpg
+  batch_size: 10
+  transforms:
+    - DecodeImage:
+        to_rgb: True
+        channel_first: False
+    - ResizeImage:
+        resize_short: 232
+        interpolation: bicubic
+        backend: pil
+    - 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]

paddlex/repo_apis/PaddleClas_api/configs/PP-HGNetV2-B5.yaml

@@ -0,0 +1,164 @@
+## Note: This config is only used for finetune training. The ImageNet metrics in PaddleClas are not trained through this config.
+# global configs
+Global:
+  checkpoints: null
+  pretrained_model: null
+  output_dir: ./output/
+  device: gpu
+  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
+  # training model under @to_static
+  to_static: False
+  use_dali: False
+
+# mixed precision training
+AMP:
+  use_amp: True
+  use_fp16_test: False
+  scale_loss: 128.0
+  use_dynamic_loss_scaling: True
+  use_promote: False
+  # O1: mixed fp16, O2: pure fp16
+  level: O1
+
+# model architecture
+Arch:
+  name: PPHGNetV2_B5
+  class_num: 102
+  pretrained: True # ssld pretrained
+ 
+# 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
+    # for global bs 1024, when finetune training, you need to reduce learning_rate manually
+    learning_rate: 0.5
+    warmup_epoch: 5
+  regularizer:
+    name: 'L2'
+    coeff: 0.00002
+
+
+# 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
+            interpolation: bicubic
+            backend: pil
+        - RandFlipImage:
+            flip_code: 1
+        - TimmAutoAugment:
+            config_str: rand-m7-mstd0.5-inc1
+            interpolation: bicubic
+            img_size: 224
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+            order: ''
+        - RandomErasing:
+            EPSILON: 0.25
+            sl: 0.02
+            sh: 1.0/3.0
+            r1: 0.3
+            attempt: 10
+            use_log_aspect: True
+            mode: pixel
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 128
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 16
+      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: 232
+            interpolation: bicubic
+            backend: pil
+        - 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: 128
+      drop_last: False
+      shuffle: False
+    loader:
+      num_workers: 16
+      use_shared_memory: True
+
+Infer:
+  infer_imgs: docs/images/inference_deployment/whl_demo.jpg
+  batch_size: 10
+  transforms:
+    - DecodeImage:
+        to_rgb: True
+        channel_first: False
+    - ResizeImage:
+        resize_short: 232
+        interpolation: bicubic
+        backend: pil
+    - 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]

paddlex/repo_apis/PaddleClas_api/configs/PP-HGNet_base.yaml

@@ -0,0 +1,169 @@
+# global configs
+Global:
+  checkpoints: null
+  pretrained_model: null
+  output_dir: ./output/
+  device: gpu
+  save_interval: 1
+  eval_during_train: True
+  eval_interval: 1
+  epochs: 600
+  print_batch_step: 10
+  use_visualdl: False
+  # used for static mode and model export
+  image_shape: [3, 224, 224]
+  save_inference_dir: ./inference
+  # training model under @to_static
+  to_static: False
+  use_dali: False
+
+# mixed precision training
+AMP:
+  use_amp: True
+  use_fp16_test: False
+  scale_loss: 128.0
+  use_dynamic_loss_scaling: True
+  use_promote: False
+  # O1: mixed fp16, O2: pure fp16
+  level: O1
+
+# model architecture
+Arch:
+  name: PPHGNet_base
+  class_num: 102
+ 
+# 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: 0.5
+    warmup_epoch: 5
+  regularizer:
+    name: 'L2'
+    coeff: 0.00004
+
+
+# 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
+            interpolation: bicubic
+            backend: pil
+        - RandFlipImage:
+            flip_code: 1
+        - TimmAutoAugment:
+            config_str: rand-m15-mstd0.5-inc1
+            interpolation: bicubic
+            img_size: 224
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+            order: ''
+        - RandomErasing:
+            EPSILON: 0.4
+            sl: 0.02
+            sh: 1.0/3.0
+            r1: 0.3
+            attempt: 10
+            use_log_aspect: True
+            mode: pixel
+      batch_transform_ops:
+        - OpSampler:
+            MixupOperator:
+              alpha: 0.4
+              prob: 0.5
+            CutmixOperator:
+              alpha: 1.0
+              prob: 0.5
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 128
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 16
+      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: 236
+            interpolation: bicubic
+            backend: pil
+        - 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: 128
+      drop_last: False
+      shuffle: False
+    loader:
+      num_workers: 16
+      use_shared_memory: True
+
+Infer:
+  infer_imgs: docs/images/inference_deployment/whl_demo.jpg
+  batch_size: 10
+  transforms:
+    - DecodeImage:
+        to_rgb: True
+        channel_first: False
+    - ResizeImage:
+        resize_short: 236
+        interpolation: bicubic
+        backend: pil
+    - 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]

paddlex/repo_apis/PaddleClas_api/configs/PP-HGNet_tiny.yaml

@@ -0,0 +1,170 @@
+# global configs
+Global:
+  checkpoints: null
+  pretrained_model: null
+  output_dir: ./output/
+  device: gpu
+  save_interval: 1
+  eval_during_train: True
+  eval_interval: 1
+  epochs: 600
+  print_batch_step: 10
+  use_visualdl: False
+  # used for static mode and model export
+  image_shape: [3, 224, 224]
+  save_inference_dir: ./inference
+  # training model under @to_static
+  to_static: False
+  use_dali: False
+
+# mixed precision training
+AMP:
+  use_amp: True
+  use_fp16_test: False
+  scale_loss: 128.0
+  use_dynamic_loss_scaling: True
+  use_promote: False
+  # O1: mixed fp16, O2: pure fp16
+  level: O1
+
+# model architecture
+Arch:
+  name: PPHGNet_tiny
+  class_num: 102
+  pretrained: True
+ 
+# 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: 0.5
+    warmup_epoch: 5
+  regularizer:
+    name: 'L2'
+    coeff: 0.00004
+
+
+# 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
+            interpolation: bicubic
+            backend: pil
+        - RandFlipImage:
+            flip_code: 1
+        - TimmAutoAugment:
+            config_str: rand-m7-mstd0.5-inc1
+            interpolation: bicubic
+            img_size: 224
+        - NormalizeImage:
+            scale: 1.0/255.0
+            mean: [0.485, 0.456, 0.406]
+            std: [0.229, 0.224, 0.225]
+            order: ''
+        - RandomErasing:
+            EPSILON: 0.25
+            sl: 0.02
+            sh: 1.0/3.0
+            r1: 0.3
+            attempt: 10
+            use_log_aspect: True
+            mode: pixel
+      batch_transform_ops:
+        - OpSampler:
+            MixupOperator:
+              alpha: 0.2
+              prob: 0.5
+            CutmixOperator:
+              alpha: 1.0
+              prob: 0.5
+
+    sampler:
+      name: DistributedBatchSampler
+      batch_size: 128
+      drop_last: False
+      shuffle: True
+    loader:
+      num_workers: 16
+      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: 232
+            interpolation: bicubic
+            backend: pil
+        - 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: 128
+      drop_last: False
+      shuffle: False
+    loader:
+      num_workers: 16
+      use_shared_memory: True
+
+Infer:
+  infer_imgs: docs/images/inference_deployment/whl_demo.jpg
+  batch_size: 10
+  transforms:
+    - DecodeImage:
+        to_rgb: True
+        channel_first: False
+    - ResizeImage:
+        resize_short: 232
+        interpolation: bicubic
+        backend: pil
+    - 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]