diff --git a/ppcls/arch/backbone/__init__.py b/ppcls/arch/backbone/__init__.py
index 7260a3442..b4f60f36b 100644
--- a/ppcls/arch/backbone/__init__.py
+++ b/ppcls/arch/backbone/__init__.py
@@ -12,9 +12,14 @@
# See the License for the specific language governing permissions and
# limitations under the License.
-from ppcls.arch.backbone.model_zoo.resnet import ResNet18, ResNet34, ResNet50, ResNet101, ResNet152
+from ppcls.arch.backbone.legendary_models.mobilenet_v1 import MobileNetV1_x0_25, MobileNetV1_x0_5, MobileNetV1_x0_75, MobileNetV1
+from ppcls.arch.backbone.legendary_models.mobilenet_v3 import 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.arch.backbone.legendary_models.resnet import ResNet18, ResNet18_vd, ResNet34, ResNet34_vd, ResNet50, ResNet50_vd, ResNet101, ResNet101_vd, ResNet152, ResNet152_vd, ResNet200_vd
+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_vd import ResNet18_vd, ResNet34_vd, ResNet50_vd, ResNet101_vd, ResNet152_vd, ResNet200_vd
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
@@ -23,22 +28,17 @@ from ppcls.arch.backbone.model_zoo.se_resnext_vd import SE_ResNeXt50_vd_32x4d, S
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
from ppcls.arch.backbone.model_zoo.densenet import DenseNet121, DenseNet161, DenseNet169, DenseNet201, DenseNet264
-from ppcls.arch.backbone.model_zoo.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_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.arch.backbone.model_zoo.efficientnet import EfficientNetB0, EfficientNetB1, EfficientNetB2, EfficientNetB3, EfficientNetB4, EfficientNetB5, EfficientNetB6, EfficientNetB7, EfficientNetB0_small
from ppcls.arch.backbone.model_zoo.resnest import ResNeSt50_fast_1s1x64d, ResNeSt50, ResNeSt101
from ppcls.arch.backbone.model_zoo.googlenet import GoogLeNet
-from ppcls.arch.backbone.model_zoo.mobilenet_v1 import MobileNetV1_x0_25, MobileNetV1_x0_5, MobileNetV1_x0_75, MobileNetV1
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.mobilenet_v3 import 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.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.alexnet import AlexNet
-from ppcls.arch.backbone.model_zoo.inception_v3 import InceptionV3
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.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.vgg import VGG11, VGG13, VGG16, VGG19
from ppcls.arch.backbone.model_zoo.darknet import DarkNet53
from ppcls.arch.backbone.model_zoo.regnet import RegNetX_200MF, RegNetX_4GF, RegNetX_32GF, RegNetY_200MF, RegNetY_4GF, RegNetY_32GF
from ppcls.arch.backbone.model_zoo.vision_transformer import 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
diff --git a/ppcls/configs/ImageNet/HRNet/HRNet_W18_C.yaml b/ppcls/configs/ImageNet/HRNet/HRNet_W18_C.yaml
new file mode 100644
index 000000000..a5e677411
--- /dev/null
+++ b/ppcls/configs/ImageNet/HRNet/HRNet_W18_C.yaml
@@ -0,0 +1,124 @@
+# 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: "HRNet_W18_C"
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/HRNet/HRNet_W30_C.yaml b/ppcls/configs/ImageNet/HRNet/HRNet_W30_C.yaml
new file mode 100644
index 000000000..50bd7144d
--- /dev/null
+++ b/ppcls/configs/ImageNet/HRNet/HRNet_W30_C.yaml
@@ -0,0 +1,124 @@
+# 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: "HRNet_W30_C"
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/HRNet/HRNet_W32_C.yaml b/ppcls/configs/ImageNet/HRNet/HRNet_W32_C.yaml
new file mode 100644
index 000000000..62129ae8c
--- /dev/null
+++ b/ppcls/configs/ImageNet/HRNet/HRNet_W32_C.yaml
@@ -0,0 +1,124 @@
+# 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: "HRNet_W32_C"
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/HRNet/HRNet_W40_C.yaml b/ppcls/configs/ImageNet/HRNet/HRNet_W40_C.yaml
new file mode 100644
index 000000000..f9af0c422
--- /dev/null
+++ b/ppcls/configs/ImageNet/HRNet/HRNet_W40_C.yaml
@@ -0,0 +1,124 @@
+# 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: "HRNet_W40_C"
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/HRNet/HRNet_W44_C.yaml b/ppcls/configs/ImageNet/HRNet/HRNet_W44_C.yaml
new file mode 100644
index 000000000..40cdf5a40
--- /dev/null
+++ b/ppcls/configs/ImageNet/HRNet/HRNet_W44_C.yaml
@@ -0,0 +1,124 @@
+# 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: "HRNet_W44_C"
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/HRNet/HRNet_W48_C.yaml b/ppcls/configs/ImageNet/HRNet/HRNet_W48_C.yaml
new file mode 100644
index 000000000..51c796cf9
--- /dev/null
+++ b/ppcls/configs/ImageNet/HRNet/HRNet_W48_C.yaml
@@ -0,0 +1,124 @@
+# 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: "HRNet_W48_C"
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/HRNet/HRNet_W64_C.yaml b/ppcls/configs/ImageNet/HRNet/HRNet_W64_C.yaml
new file mode 100644
index 000000000..00c9fff91
--- /dev/null
+++ b/ppcls/configs/ImageNet/HRNet/HRNet_W64_C.yaml
@@ -0,0 +1,124 @@
+# 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: "HRNet_W64_C"
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/Ineption/InceptionV3.yaml b/ppcls/configs/ImageNet/Ineption/InceptionV3.yaml
new file mode 100644
index 000000000..66607f764
--- /dev/null
+++ b/ppcls/configs/ImageNet/Ineption/InceptionV3.yaml
@@ -0,0 +1,124 @@
+# 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: "InceptionV3"
+
+# 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: Cosine
+ learning_rate: 0.045
+ 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:
+ - RandCropImage:
+ size: 299
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+Infer:
+ infer_imgs: "docs/images/whl/demo.jpg"
+ batch_size: 10
+ transforms:
+ - DecodeImage:
+ to_rgb: True
+ channel_first: False
+ - ResizeImage:
+ resize_short: 320
+ - CropImage:
+ size: 299
+ - 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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1.yaml b/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1.yaml
new file mode 100644
index 000000000..13184416d
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1.yaml
@@ -0,0 +1,124 @@
+# 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: "MobileNetV1"
+
+# 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: 0.00003
+
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_25.yaml b/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_25.yaml
new file mode 100644
index 000000000..75428a6c2
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_25.yaml
@@ -0,0 +1,124 @@
+# 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: "MobileNetV1_x0_25"
+
+# 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: 0.00003
+
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_5.yaml b/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_5.yaml
new file mode 100644
index 000000000..b516f5540
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_5.yaml
@@ -0,0 +1,124 @@
+# 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: "MobileNetV1_x0_5"
+
+# 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: 0.00003
+
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_75.yaml b/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_75.yaml
new file mode 100644
index 000000000..700f5c651
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV1/MobileNetV1_x0_75.yaml
@@ -0,0 +1,124 @@
+# 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: "MobileNetV1_x0_75"
+
+# 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: 0.00003
+
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_35.yaml b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_35.yaml
new file mode 100644
index 000000000..d18c211ae
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_35.yaml
@@ -0,0 +1,122 @@
+# 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: 360
+ 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: "MobileNetV3_large_x0_35"
+
+# 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: Cosine
+ learning_rate: 1.3
+ 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ mean: [0.485, 0.456, 0.406]
+ std: [0.229, 0.224, 0.225]
+ order: ''
+
+ sampler:
+ name: DistributedBatchSampler
+ batch_size: 512
+ drop_last: False
+ shuffle: True
+ loader:
+ num_workers: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_5.yaml b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_5.yaml
new file mode 100644
index 000000000..c1ff15e2b
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_5.yaml
@@ -0,0 +1,122 @@
+# 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: 360
+ 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: "MobileNetV3_large_x0_5"
+
+# 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: Cosine
+ learning_rate: 1.3
+ 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ mean: [0.485, 0.456, 0.406]
+ std: [0.229, 0.224, 0.225]
+ order: ''
+
+ sampler:
+ name: DistributedBatchSampler
+ batch_size: 512
+ drop_last: False
+ shuffle: True
+ loader:
+ num_workers: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_75.yaml b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_75.yaml
new file mode 100644
index 000000000..fd387f001
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x0_75.yaml
@@ -0,0 +1,122 @@
+# 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: 360
+ 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: "MobileNetV3_large_x0_75"
+
+# 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: Cosine
+ learning_rate: 1.3
+ 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ mean: [0.485, 0.456, 0.406]
+ std: [0.229, 0.224, 0.225]
+ order: ''
+
+ sampler:
+ name: DistributedBatchSampler
+ batch_size: 512
+ drop_last: False
+ shuffle: True
+ loader:
+ num_workers: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_0.yaml b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_0.yaml
new file mode 100644
index 000000000..b9072ed21
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_0.yaml
@@ -0,0 +1,122 @@
+# 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: 360
+ 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: "MobileNetV3_large_x1_0"
+
+# 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: Cosine
+ learning_rate: 1.3
+ 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ mean: [0.485, 0.456, 0.406]
+ std: [0.229, 0.224, 0.225]
+ order: ''
+
+ sampler:
+ name: DistributedBatchSampler
+ batch_size: 512
+ drop_last: False
+ shuffle: True
+ loader:
+ num_workers: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_25.yaml b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_25.yaml
new file mode 100644
index 000000000..472942a7e
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_large_x1_25.yaml
@@ -0,0 +1,122 @@
+# 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: 360
+ 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: "MobileNetV3_large_x1_25"
+
+# 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: Cosine
+ learning_rate: 1.3
+ 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ mean: [0.485, 0.456, 0.406]
+ std: [0.229, 0.224, 0.225]
+ order: ''
+
+ sampler:
+ name: DistributedBatchSampler
+ batch_size: 512
+ drop_last: False
+ shuffle: True
+ loader:
+ num_workers: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_35.yaml b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_35.yaml
new file mode 100644
index 000000000..4312730f2
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_35.yaml
@@ -0,0 +1,122 @@
+# 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: 360
+ 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: "MobileNetV3_small_x0_35"
+
+# 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: Cosine
+ learning_rate: 1.3
+ 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ mean: [0.485, 0.456, 0.406]
+ std: [0.229, 0.224, 0.225]
+ order: ''
+
+ sampler:
+ name: DistributedBatchSampler
+ batch_size: 512
+ drop_last: False
+ shuffle: True
+ loader:
+ num_workers: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_5.yaml b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_5.yaml
new file mode 100644
index 000000000..427a2793e
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_5.yaml
@@ -0,0 +1,122 @@
+# 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: 360
+ 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: "MobileNetV3_small_x0_5"
+
+# 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: Cosine
+ learning_rate: 1.3
+ 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ mean: [0.485, 0.456, 0.406]
+ std: [0.229, 0.224, 0.225]
+ order: ''
+
+ sampler:
+ name: DistributedBatchSampler
+ batch_size: 512
+ drop_last: False
+ shuffle: True
+ loader:
+ num_workers: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_75.yaml b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_75.yaml
new file mode 100644
index 000000000..82246d4e4
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x0_75.yaml
@@ -0,0 +1,122 @@
+# 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: 360
+ 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: "MobileNetV3_small_x0_75"
+
+# 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: Cosine
+ learning_rate: 1.3
+ 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ mean: [0.485, 0.456, 0.406]
+ std: [0.229, 0.224, 0.225]
+ order: ''
+
+ sampler:
+ name: DistributedBatchSampler
+ batch_size: 512
+ drop_last: False
+ shuffle: True
+ loader:
+ num_workers: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_0.yaml b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_0.yaml
new file mode 100644
index 000000000..f2810c678
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_0.yaml
@@ -0,0 +1,122 @@
+# 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: 360
+ 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: "MobileNetV3_small_x1_0"
+
+# 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: Cosine
+ learning_rate: 1.3
+ 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ mean: [0.485, 0.456, 0.406]
+ std: [0.229, 0.224, 0.225]
+ order: ''
+
+ sampler:
+ name: DistributedBatchSampler
+ batch_size: 512
+ drop_last: False
+ shuffle: True
+ loader:
+ num_workers: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_25.yaml b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_25.yaml
new file mode 100644
index 000000000..d42bddc0b
--- /dev/null
+++ b/ppcls/configs/ImageNet/MobileNetV3/MobileNetV3_small_x1_25.yaml
@@ -0,0 +1,122 @@
+# 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: 360
+ 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: "MobileNetV3_small_x1_25"
+
+# 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: Cosine
+ learning_rate: 1.3
+ 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ mean: [0.485, 0.456, 0.406]
+ std: [0.229, 0.224, 0.225]
+ order: ''
+
+ sampler:
+ name: DistributedBatchSampler
+ batch_size: 512
+ drop_last: False
+ shuffle: True
+ loader:
+ num_workers: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/ResNet/ResNet101.yaml b/ppcls/configs/ImageNet/ResNet/ResNet101.yaml
new file mode 100644
index 000000000..ed1405e71
--- /dev/null
+++ b/ppcls/configs/ImageNet/ResNet/ResNet101.yaml
@@ -0,0 +1,124 @@
+# 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: "ResNet101"
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/ResNet/ResNet101_vd.yaml b/ppcls/configs/ImageNet/ResNet/ResNet101_vd.yaml
new file mode 100644
index 000000000..ab25cdcce
--- /dev/null
+++ b/ppcls/configs/ImageNet/ResNet/ResNet101_vd.yaml
@@ -0,0 +1,122 @@
+# 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: "ResNet101_vd"
+
+# 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: 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/ResNet/ResNet152.yaml b/ppcls/configs/ImageNet/ResNet/ResNet152.yaml
new file mode 100644
index 000000000..bc49c6bcc
--- /dev/null
+++ b/ppcls/configs/ImageNet/ResNet/ResNet152.yaml
@@ -0,0 +1,124 @@
+# 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: "ResNet152"
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/ResNet/ResNet152_vd.yaml b/ppcls/configs/ImageNet/ResNet/ResNet152_vd.yaml
new file mode 100644
index 000000000..e190e68f5
--- /dev/null
+++ b/ppcls/configs/ImageNet/ResNet/ResNet152_vd.yaml
@@ -0,0 +1,122 @@
+# 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: "ResNet152_vd"
+
+# 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: 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/ResNet/ResNet18.yaml b/ppcls/configs/ImageNet/ResNet/ResNet18.yaml
new file mode 100644
index 000000000..382a1a386
--- /dev/null
+++ b/ppcls/configs/ImageNet/ResNet/ResNet18.yaml
@@ -0,0 +1,124 @@
+# 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: "ResNet18"
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/ResNet/ResNet18_vd.yaml b/ppcls/configs/ImageNet/ResNet/ResNet18_vd.yaml
new file mode 100644
index 000000000..dcc1508fd
--- /dev/null
+++ b/ppcls/configs/ImageNet/ResNet/ResNet18_vd.yaml
@@ -0,0 +1,122 @@
+# 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: "ResNet18_vd"
+
+# 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: Cosine
+ learning_rate: 0.1
+ regularizer:
+ name: 'L2'
+ coeff: 0.00007
+
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/ResNet/ResNet200_vd.yaml b/ppcls/configs/ImageNet/ResNet/ResNet200_vd.yaml
new file mode 100644
index 000000000..84b8acf03
--- /dev/null
+++ b/ppcls/configs/ImageNet/ResNet/ResNet200_vd.yaml
@@ -0,0 +1,122 @@
+# 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: "ResNet200_vd"
+
+# 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: 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/ResNet/ResNet34.yaml b/ppcls/configs/ImageNet/ResNet/ResNet34.yaml
new file mode 100644
index 000000000..07f7036ea
--- /dev/null
+++ b/ppcls/configs/ImageNet/ResNet/ResNet34.yaml
@@ -0,0 +1,124 @@
+# 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: "ResNet34"
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/ResNet/ResNet34_vd.yaml b/ppcls/configs/ImageNet/ResNet/ResNet34_vd.yaml
new file mode 100644
index 000000000..92f8bf3d7
--- /dev/null
+++ b/ppcls/configs/ImageNet/ResNet/ResNet34_vd.yaml
@@ -0,0 +1,122 @@
+# 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: "ResNet101_vd"
+
+# 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: Cosine
+ learning_rate: 0.1
+ regularizer:
+ name: 'L2'
+ coeff: 0.00007
+
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/ResNet/ResNet50.yaml b/ppcls/configs/ImageNet/ResNet/ResNet50.yaml
index 654498e6b..a55aeb4e8 100644
--- a/ppcls/configs/ImageNet/ResNet/ResNet50.yaml
+++ b/ppcls/configs/ImageNet/ResNet/ResNet50.yaml
@@ -45,52 +45,52 @@ Optimizer:
# data loader for train and eval
DataLoader:
Train:
- # Dataset:
- # Sampler:
- # Loader:
- batch_size: 256
- num_workers: 4
- file_list: "./dataset/ILSVRC2012/train_list.txt"
- data_dir: "./dataset/ILSVRC2012/"
- shuffle_seed: 0
- transforms:
- - DecodeImage:
- to_rgb: True
- channel_first: False
- - RandCropImage:
- size: 224
- - RandFlipImage:
- flip_code: 1
- - NormalizeImage:
- scale: 1./255.
- mean: [0.485, 0.456, 0.406]
- std: [0.229, 0.224, 0.225]
- order: ''
- - ToCHWImage:
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/train_list.txt"
+ transform_ops:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
Eval:
# TOTO: modify to the latest trainer
- # Dataset:
- # Sampler:
- # Loader:
- batch_size: 128
- num_workers: 4
- file_list: "./dataset/ILSVRC2012/val_list.txt"
- data_dir: "./dataset/ILSVRC2012/"
- shuffle_seed: 0
- 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:
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
Infer:
infer_imgs: "docs/images/whl/demo.jpg"
diff --git a/ppcls/configs/ImageNet/ResNet/ResNet50_vd.yaml b/ppcls/configs/ImageNet/ResNet/ResNet50_vd.yaml
new file mode 100644
index 000000000..8d4305fa1
--- /dev/null
+++ b/ppcls/configs/ImageNet/ResNet/ResNet50_vd.yaml
@@ -0,0 +1,122 @@
+# 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: "ResNet50_vd"
+
+# 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: Cosine
+ learning_rate: 0.1
+ regularizer:
+ name: 'L2'
+ coeff: 0.00007
+
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/VGG/VGG11.yaml b/ppcls/configs/ImageNet/VGG/VGG11.yaml
new file mode 100644
index 000000000..34648346b
--- /dev/null
+++ b/ppcls/configs/ImageNet/VGG/VGG11.yaml
@@ -0,0 +1,122 @@
+# 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: 90
+ 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: "VGG11"
+
+# 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: Cosine
+ learning_rate: 0.1
+ regularizer:
+ name: 'L2'
+ coeff: 0.0002
+
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/VGG/VGG13.yaml b/ppcls/configs/ImageNet/VGG/VGG13.yaml
new file mode 100644
index 000000000..c57081bb7
--- /dev/null
+++ b/ppcls/configs/ImageNet/VGG/VGG13.yaml
@@ -0,0 +1,122 @@
+# 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: 90
+ 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: "VGG13"
+
+# 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: Cosine
+ learning_rate: 0.01
+ regularizer:
+ name: 'L2'
+ coeff: 0.0003
+
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/VGG/VGG16.yaml b/ppcls/configs/ImageNet/VGG/VGG16.yaml
new file mode 100644
index 000000000..6e1c4045d
--- /dev/null
+++ b/ppcls/configs/ImageNet/VGG/VGG16.yaml
@@ -0,0 +1,122 @@
+# 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: 90
+ 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: "VGG16"
+
+# 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: Cosine
+ learning_rate: 0.01
+ regularizer:
+ name: 'L2'
+ coeff: 0.0004
+
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/configs/ImageNet/VGG/VGG19.yaml b/ppcls/configs/ImageNet/VGG/VGG19.yaml
new file mode 100644
index 000000000..28b5a101b
--- /dev/null
+++ b/ppcls/configs/ImageNet/VGG/VGG19.yaml
@@ -0,0 +1,122 @@
+# 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: 150
+ 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: "VGG19"
+
+# 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: Cosine
+ learning_rate: 0.01
+ regularizer:
+ name: 'L2'
+ coeff: 0.0004
+
+
+# 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:
+ - RandCropImage:
+ size: 224
+ - RandFlipImage:
+ flip_code: 1
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+ Eval:
+ # TOTO: modify to the latest trainer
+ dataset:
+ name: ImageNetDataset
+ image_root: "./dataset/ILSVRC2012/"
+ cls_label_path: "./dataset/ILSVRC2012/val_list.txt"
+ transform_ops:
+ - ResizeImage:
+ size: 224
+ - NormalizeImage:
+ scale: 0.00392157
+ 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: 6
+ use_shared_memory: False
+
+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:
+ - Topk:
+ k: [1, 5]
+ Eval:
+ - Topk:
+ k: [1, 5]
+
diff --git a/ppcls/data/__init__.py b/ppcls/data/__init__.py
index 8abed57ef..7ffa6fd53 100644
--- a/ppcls/data/__init__.py
+++ b/ppcls/data/__init__.py
@@ -18,13 +18,13 @@ from paddle.io import DistributedBatchSampler, BatchSampler, DataLoader
from ppcls.utils import logger
-from . import datasets
+from . import dataset
from . import imaug
from . import samplers
# dataset
-from .datasets.imagenet_dataset import ImageNetDataset
+from .dataset.imagenet_dataset import ImageNetDataset
from .dataset.multilabel_dataset import MultiLabelDataset
-from .datasets.common_dataset import create_operators
+from .dataset.common_dataset import create_operators
# sampler
from .samplers import DistributedRandomIdentitySampler
@@ -35,6 +35,7 @@ def build_dataloader(config, mode, device, seed=None):
assert mode in ['Train', 'Eval', 'Test'], "Mode should be Train, Eval or Test."
# build dataset
config_dataset = config[mode]['dataset']
+ config_dataset = copy.deepcopy(config_dataset)
dataset_name = config_dataset.pop('name')
if 'batch_transform_ops' in config_dataset:
batch_transform = config_dataset.pop('batch_transform_ops')
@@ -105,7 +106,7 @@ def build_dataloader(config, mode, device, seed=None):
logger.info("build data_loader({}) success...".format(data_loader))
- return dataloader
+ return data_loader
'''
# TODO: fix the format
diff --git a/ppcls/data/dataset/common_dataset.py b/ppcls/data/dataset/common_dataset.py
index bb9359ca1..bcbfc92e3 100644
--- a/ppcls/data/dataset/common_dataset.py
+++ b/ppcls/data/dataset/common_dataset.py
@@ -40,7 +40,6 @@ def create_operators(params):
assert isinstance(params, list), ('operator config should be a list')
ops = []
for operator in params:
- print(operator)
assert isinstance(operator,
dict) and len(operator) == 1, "yaml format error"
op_name = list(operator)[0]
diff --git a/ppcls/data/postprocess/__init__.py b/ppcls/data/postprocess/__init__.py
index 8d78ec814..6ef0ea819 100644
--- a/ppcls/data/postprocess/__init__.py
+++ b/ppcls/data/postprocess/__init__.py
@@ -14,9 +14,9 @@
import copy
import importlib
-from . import topk_process
+from . import topk
-from .topk_process import Topk
+from .topk import Topk
def build_postprocess(config):
diff --git a/ppcls/data/preprocess/__init__.py b/ppcls/data/preprocess/__init__.py
index ce4813523..ebd0df420 100644
--- a/ppcls/data/preprocess/__init__.py
+++ b/ppcls/data/preprocess/__init__.py
@@ -40,7 +40,6 @@ def transform(data, ops=[]):
""" transform """
for op in ops:
data = op(data)
- #print(data.shape, op)
return data
diff --git a/ppcls/engine/trainer.py b/ppcls/engine/trainer.py
index 458ddb35f..c56e83763 100644
--- a/ppcls/engine/trainer.py
+++ b/ppcls/engine/trainer.py
@@ -104,7 +104,7 @@ class Trainer(object):
metric_func = self._build_metric_info(self.config["Metric"])
- train_dataloader = build_dataloader(self.config["DataLoader"], "train",
+ train_dataloader = build_dataloader(self.config["DataLoader"], "Train",
self.device)
step_each_epoch = len(train_dataloader)
@@ -217,7 +217,7 @@ class Trainer(object):
def eval(self, epoch_id=0):
output_info = dict()
- eval_dataloader = build_dataloader(self.config["DataLoader"], "eval",
+ eval_dataloader = build_dataloader(self.config["DataLoader"], "Eval",
self.device)
self.model.eval()
diff --git a/ppcls/losses/triplet.py b/ppcls/losses/triplet.py
index beef5def0..bba222adc 100644
--- a/ppcls/losses/triplet.py
+++ b/ppcls/losses/triplet.py
@@ -48,7 +48,6 @@ class TripletLossV2(nn.Layer):
# `dist_ap` means distance(anchor, positive)
## both `dist_ap` and `relative_p_inds` with shape [N, 1]
- #print(is_pos.shape, dist.shape, type(is_pos), type(dist), paddle.reshape(paddle.masked_select(dist, is_pos),(bs, -1)))
'''
dist_ap, relative_p_inds = paddle.max(
paddle.reshape(dist[is_pos], (bs, -1)), axis=1, keepdim=True)
@@ -98,7 +97,6 @@ class TripletLoss(nn.Layer):
"""
inputs = input["features"]
- #print(inputs.shape, targets.shape)
bs = inputs.shape[0]
# Compute pairwise distance, replace by the official when merged
dist = paddle.pow(inputs, 2).sum(axis=1, keepdim=True).expand([bs, bs])
diff --git a/tools/run.sh b/tools/run.sh
index be89623f3..345b62758 100755
--- a/tools/run.sh
+++ b/tools/run.sh
@@ -3,5 +3,5 @@
python3.7 -m paddle.distributed.launch \
--gpus="0,1,2,3" \
tools/train.py \
- -c ./configs/ResNet/ResNet50.yaml \
+ -c ./ppcls/configs/ImageNet/ResNet/ResNet50.yaml \
-o print_interval=10