fix update of issue #25

README.md

@@ -41,7 +41,7 @@ The main branch works with **PyTorch 1.8** (required by some self-supervised met
 
 ## What's New
 
-[2022-10-12] Update new features and documents of `OpenMixup` v0.2.6 (issue [#24](https://github.com/Westlake-AI/openmixup/issues/24) and issue [#25](https://github.com/Westlake-AI/openmixup/issues/25)).
+[2022-10-15] Update new features and documents of `OpenMixup` v0.2.6 (issue [#24](https://github.com/Westlake-AI/openmixup/issues/24) and issue [#25](https://github.com/Westlake-AI/openmixup/issues/25)).
 
 [2022-09-14] `OpenMixup` v0.2.6 is released (issue [#20](https://github.com/Westlake-AI/openmixup/issues/20)).
 
@@ -113,6 +113,7 @@ Please refer to [Model Zoos](docs/en/model_zoos) for various backbones, mixup me
     - [x] [MViTV2](https://arxiv.org/abs/2112.01526) (CVPR'2022) [[config](https://github.com/Westlake-AI/openmixup/tree/main/configs/classification/imagenet/mvit/)]
     - [x] [RepMLP](https://arxiv.org/abs/2105.01883) (CVPR'2022) [[config](https://github.com/Westlake-AI/openmixup/tree/main/configs/classification/imagenet/repmlp/)]
     - [x] [VAN](https://arxiv.org/abs/2202.09741) (ArXiv'2022) [[config](https://github.com/Westlake-AI/openmixup/tree/main/configs/classification/imagenet/van/)]
+    - [x] [DeiT-3](https://arxiv.org/abs/2204.07118) (ECCV'2022) [[config](https://github.com/Westlake-AI/openmixup/tree/main/configs/classification/imagenet/deit3/)]
     - [x] [LITv2](https://arxiv.org/abs/2205.13213) (NIPS'2022) [[config](https://github.com/Westlake-AI/openmixup/tree/main/configs/classification/imagenet/lit_v2/)]
     - [x] [HorNet](https://arxiv.org/abs/2207.14284) (NIPS'2022) [[config](https://github.com/Westlake-AI/openmixup/tree/main/configs/classification/imagenet/hornet/)]
     - [x] [EdgeNeXt](https://arxiv.org/abs/2206.10589) (ECCVW'2022) [[config](https://github.com/Westlake-AI/openmixup/tree/main/configs/classification/imagenet/edgenext/)]

configs/classification/_base_/datasets/imagenet/basic_tta_sz224_4xbs64.py

@@ -0,0 +1,64 @@
+# dataset settings
+data_source_cfg = dict(type='ImageNet')
+# ImageNet dataset
+data_train_list = 'data/meta/ImageNet/train_labeled_full.txt'
+data_train_root = 'data/ImageNet/train'
+data_test_list = 'data/meta/ImageNet/val_labeled.txt'
+data_test_root = 'data/ImageNet/val/'
+
+dataset_type = 'ClassificationDataset'
+img_norm_cfg = dict(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+train_pipeline = [
+    dict(type='RandomResizedCrop', size=224, interpolation=3),  # bicubic
+    dict(type='RandomHorizontalFlip'),
+]
+test_pipeline_1 = [
+    dict(type='Resize', size=256, interpolation=3),  # 0.85
+    dict(type='RandomHorizontalFlip', p=0.5),
+    dict(type='CenterCrop', size=224),
+    dict(type='ToTensor'),
+    dict(type='Normalize', **img_norm_cfg),
+]
+test_pipeline_2 = [
+    dict(type='Resize', size=256, interpolation=3),  # 0.85
+    dict(type='RandomVerticalFlip', p=0.5),
+    dict(type='PlaceCrop', size=224, start=[0, 5, 10, 15,]),
+    dict(type='ToTensor'),
+    dict(type='Normalize', **img_norm_cfg),
+]
+
+# prefetch
+prefetch = True
+if not prefetch:
+    train_pipeline.extend([dict(type='ToTensor'), dict(type='Normalize', **img_norm_cfg)])
+
+data = dict(
+    imgs_per_gpu=64,
+    workers_per_gpu=8,
+    train=dict(
+        type=dataset_type,
+        data_source=dict(
+            list_file=data_train_list, root=data_train_root,
+            **data_source_cfg),
+        pipeline=train_pipeline,
+        prefetch=prefetch,
+    ),
+    val=dict(
+        type="MultiViewDataset",  # use multi-view for test time augmentations
+        data_source=dict(
+            list_file=data_test_list, root=data_test_root, **data_source_cfg),
+        num_views=[2, 4],
+        pipelines=[test_pipeline_1, test_pipeline_2],
+        prefetch=False,
+    ))
+
+# validation hook
+evaluation = dict(
+    initial=False,
+    interval=1,
+    imgs_per_gpu=128,
+    workers_per_gpu=4,
+    eval_param=dict(topk=(1, 5)))
+
+# checkpoint
+checkpoint_config = dict(interval=1, max_keep_ckpts=1)

configs/classification/_base_/datasets/imagenet/deit3_ft_sz224_8xbs128.py

@@ -0,0 +1,83 @@
+# Refers to `_RAND_INCREASING_TRANSFORMS` in pytorch-image-models
+rand_increasing_policies = [
+    dict(type='AutoContrast'),
+    dict(type='Equalize'),
+    dict(type='Invert'),
+    dict(type='Rotate', magnitude_key='angle', magnitude_range=(0, 30)),
+    dict(type='Posterize', magnitude_key='bits', magnitude_range=(4, 0)),
+    dict(type='Solarize', magnitude_key='thr', magnitude_range=(256, 0)),
+    dict(type='SolarizeAdd', magnitude_key='magnitude', magnitude_range=(0, 110)),
+    dict(type='ColorTransform', magnitude_key='magnitude', magnitude_range=(0, 0.9)),
+    dict(type='Contrast', magnitude_key='magnitude', magnitude_range=(0, 0.9)),
+    dict(type='Brightness', magnitude_key='magnitude', magnitude_range=(0, 0.9)),
+    dict(type='Sharpness', magnitude_key='magnitude', magnitude_range=(0, 0.9)),
+    dict(type='Shear',
+        magnitude_key='magnitude', magnitude_range=(0, 0.3), direction='horizontal'),
+    dict(type='Shear',
+        magnitude_key='magnitude', magnitude_range=(0, 0.3), direction='vertical'),
+    dict(type='Translate',
+        magnitude_key='magnitude', magnitude_range=(0, 0.45), direction='horizontal'),
+    dict(type='Translate',
+        magnitude_key='magnitude', magnitude_range=(0, 0.45), direction='vertical'),
+]
+
+# dataset settings
+data_source_cfg = dict(type='ImageNet')
+# ImageNet dataset
+data_train_list = 'data/meta/ImageNet/train_labeled_full.txt'
+data_train_root = 'data/ImageNet/train'
+data_test_list = 'data/meta/ImageNet/val_labeled.txt'
+data_test_root = 'data/ImageNet/val/'
+
+dataset_type = 'ClassificationDataset'
+img_norm_cfg = dict(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+train_pipeline = [
+    dict(type='RandomResizedCrop', size=224, interpolation=3),  # bicubic
+    dict(type='RandomHorizontalFlip'),
+    dict(type='RandAugment',
+        policies=rand_increasing_policies,
+        num_policies=2, total_level=10,
+        magnitude_level=9, magnitude_std=0.5,
+        hparams=dict(
+            pad_val=[104, 116, 124], interpolation='bicubic')),
+]
+test_pipeline = [
+    dict(type='Resize', size=224, interpolation=3),  # crop-ratio = 1.0
+    dict(type='CenterCrop', size=224),
+    dict(type='ToTensor'),
+    dict(type='Normalize', **img_norm_cfg),
+]
+# prefetch
+prefetch = True
+if not prefetch:
+    train_pipeline.extend([dict(type='ToTensor'), dict(type='Normalize', **img_norm_cfg)])
+
+data = dict(
+    imgs_per_gpu=128,
+    workers_per_gpu=10,
+    train=dict(
+        type=dataset_type,
+        data_source=dict(
+            list_file=data_train_list, root=data_train_root,
+            **data_source_cfg),
+        pipeline=train_pipeline,
+        prefetch=prefetch,
+    ),
+    val=dict(
+        type=dataset_type,
+        data_source=dict(
+            list_file=data_test_list, root=data_test_root, **data_source_cfg),
+        pipeline=test_pipeline,
+        prefetch=False,
+    ))
+
+# validation hook
+evaluation = dict(
+    initial=False,
+    interval=1,
+    imgs_per_gpu=128,
+    workers_per_gpu=4,
+    eval_param=dict(topk=(1, 5)))
+
+# checkpoint
+checkpoint_config = dict(interval=1, max_keep_ckpts=1)

configs/classification/_base_/datasets/imagenet/deit3_sz160_8xbs128.py

@@ -0,0 +1,64 @@
+# dataset settings
+data_source_cfg = dict(type='ImageNet')
+# ImageNet dataset
+data_train_list = 'data/meta/ImageNet/train_labeled_full.txt'
+data_train_root = 'data/ImageNet/train'
+data_test_list = 'data/meta/ImageNet/val_labeled.txt'
+data_test_root = 'data/ImageNet/val/'
+
+dataset_type = 'ClassificationDataset'
+img_norm_cfg = dict(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+train_pipeline = [
+    dict(type='RandomResizedCrop', size=160, interpolation=3),  # bicubic
+    dict(type='RandomHorizontalFlip'),
+    dict(type='RandomAppliedTrans',  # 3-Augment in DeiT III
+        transforms=[
+            dict(type='RandomGrayscale', p=1.),
+            dict(type='Solarization', p=1.),
+            dict(type='GaussianBlur', sigma_min=0.1, sigma_max=2.0, p=1.),
+        ],
+        p=1.0),
+    dict(type='ColorJitter',
+         brightness=0.3, contrast=0.3, saturation=0.3),
+]
+test_pipeline = [
+    dict(type='Resize', size=160, interpolation=3),  # crop-ratio = 1.0
+    dict(type='CenterCrop', size=160),
+    dict(type='ToTensor'),
+    dict(type='Normalize', **img_norm_cfg),
+]
+
+# prefetch
+prefetch = True
+if not prefetch:
+    train_pipeline.extend([dict(type='ToTensor'), dict(type='Normalize', **img_norm_cfg)])
+
+data = dict(
+    imgs_per_gpu=128,
+    workers_per_gpu=8,
+    train=dict(
+        type=dataset_type,
+        data_source=dict(
+            list_file=data_train_list, root=data_train_root,
+            **data_source_cfg),
+        pipeline=train_pipeline,
+        prefetch=prefetch,
+    ),
+    val=dict(
+        type=dataset_type,
+        data_source=dict(
+            list_file=data_test_list, root=data_test_root, **data_source_cfg),
+        pipeline=test_pipeline,
+        prefetch=False,
+    ))
+
+# validation hook
+evaluation = dict(
+    initial=False,
+    interval=1,
+    imgs_per_gpu=128,
+    workers_per_gpu=4,
+    eval_param=dict(topk=(1, 5)))
+
+# checkpoint
+checkpoint_config = dict(interval=1, max_keep_ckpts=1)

configs/classification/_base_/datasets/imagenet/deit3_sz192_8xbs128.py

@@ -0,0 +1,64 @@
+# dataset settings
+data_source_cfg = dict(type='ImageNet')
+# ImageNet dataset
+data_train_list = 'data/meta/ImageNet/train_labeled_full.txt'
+data_train_root = 'data/ImageNet/train'
+data_test_list = 'data/meta/ImageNet/val_labeled.txt'
+data_test_root = 'data/ImageNet/val/'
+
+dataset_type = 'ClassificationDataset'
+img_norm_cfg = dict(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+train_pipeline = [
+    dict(type='RandomResizedCrop', size=192, interpolation=3),  # bicubic
+    dict(type='RandomHorizontalFlip'),
+    dict(type='RandomAppliedTrans',  # 3-Augment in DeiT III
+        transforms=[
+            dict(type='RandomGrayscale', p=1.),
+            dict(type='Solarization', p=1.),
+            dict(type='GaussianBlur', sigma_min=0.1, sigma_max=2.0, p=1.),
+        ],
+        p=1.0),
+    dict(type='ColorJitter',
+         brightness=0.3, contrast=0.3, saturation=0.3),
+]
+test_pipeline = [
+    dict(type='Resize', size=192, interpolation=3),  # crop-ratio = 1.0
+    dict(type='CenterCrop', size=192),
+    dict(type='ToTensor'),
+    dict(type='Normalize', **img_norm_cfg),
+]
+
+# prefetch
+prefetch = True
+if not prefetch:
+    train_pipeline.extend([dict(type='ToTensor'), dict(type='Normalize', **img_norm_cfg)])
+
+data = dict(
+    imgs_per_gpu=128,
+    workers_per_gpu=8,
+    train=dict(
+        type=dataset_type,
+        data_source=dict(
+            list_file=data_train_list, root=data_train_root,
+            **data_source_cfg),
+        pipeline=train_pipeline,
+        prefetch=prefetch,
+    ),
+    val=dict(
+        type=dataset_type,
+        data_source=dict(
+            list_file=data_test_list, root=data_test_root, **data_source_cfg),
+        pipeline=test_pipeline,
+        prefetch=False,
+    ))
+
+# validation hook
+evaluation = dict(
+    initial=False,
+    interval=1,
+    imgs_per_gpu=128,
+    workers_per_gpu=4,
+    eval_param=dict(topk=(1, 5)))
+
+# checkpoint
+checkpoint_config = dict(interval=1, max_keep_ckpts=1)

configs/classification/_base_/datasets/imagenet/deit3_sz224_8xbs128.py

@@ -0,0 +1,64 @@
+# dataset settings
+data_source_cfg = dict(type='ImageNet')
+# ImageNet dataset
+data_train_list = 'data/meta/ImageNet/train_labeled_full.txt'
+data_train_root = 'data/ImageNet/train'
+data_test_list = 'data/meta/ImageNet/val_labeled.txt'
+data_test_root = 'data/ImageNet/val/'
+
+dataset_type = 'ClassificationDataset'
+img_norm_cfg = dict(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])
+train_pipeline = [
+    dict(type='RandomResizedCrop', size=224, interpolation=3),  # bicubic
+    dict(type='RandomHorizontalFlip'),
+    dict(type='RandomAppliedTrans',  # 3-Augment in DeiT III
+        transforms=[
+            dict(type='RandomGrayscale', p=1.),
+            dict(type='Solarization', p=1.),
+            dict(type='GaussianBlur', sigma_min=0.1, sigma_max=2.0, p=1.),
+        ],
+        p=1.0),
+    dict(type='ColorJitter',
+         brightness=0.3, contrast=0.3, saturation=0.3),
+]
+test_pipeline = [
+    dict(type='Resize', size=224, interpolation=3),  # crop-ratio = 1.0
+    dict(type='CenterCrop', size=224),
+    dict(type='ToTensor'),
+    dict(type='Normalize', **img_norm_cfg),
+]
+
+# prefetch
+prefetch = True
+if not prefetch:
+    train_pipeline.extend([dict(type='ToTensor'), dict(type='Normalize', **img_norm_cfg)])
+
+data = dict(
+    imgs_per_gpu=128,
+    workers_per_gpu=8,
+    train=dict(
+        type=dataset_type,
+        data_source=dict(
+            list_file=data_train_list, root=data_train_root,
+            **data_source_cfg),
+        pipeline=train_pipeline,
+        prefetch=prefetch,
+    ),
+    val=dict(
+        type=dataset_type,
+        data_source=dict(
+            list_file=data_test_list, root=data_test_root, **data_source_cfg),
+        pipeline=test_pipeline,
+        prefetch=False,
+    ))
+
+# validation hook
+evaluation = dict(
+    initial=False,
+    interval=1,
+    imgs_per_gpu=128,
+    workers_per_gpu=4,
+    eval_param=dict(topk=(1, 5)))
+
+# checkpoint
+checkpoint_config = dict(interval=1, max_keep_ckpts=1)

configs/classification/_base_/models/deit3/deit3_base_p16_sz192.py

@@ -0,0 +1,24 @@
+# model settings
+model = dict(
+    type='MixUpClassification',
+    pretrained=None,
+    alpha=[0.8, 1.0,],
+    mix_mode=["mixup", "cutmix",],
+    mix_args=dict(),
+    backbone=dict(
+        type='DeiT3',
+        arch='base',
+        img_size=192,
+        patch_size=16,
+        drop_path_rate=0.2),
+    head=dict(
+        type='VisionTransformerClsHead',
+        loss=dict(type='CrossEntropyLoss',  # mixup BCE loss (one-hot encoding)
+            use_soft=False, use_sigmoid=True, loss_weight=1.0),
+        multi_label=True,
+        in_channels=768, num_classes=1000),
+    init_cfg=[
+        dict(type='TruncNormal', layer='Linear', std=.02),
+        dict(type='Constant', layer='LayerNorm', val=1., bias=0.),
+    ],
+)