open-mmlab · mzr1996 · Apr 28, 2022 · Mar 29, 2022 · Apr 11, 2022 · Apr 11, 2022
diff --git a/configs/_base_/models/densenet/densenet121.py b/configs/_base_/models/densenet/densenet121.py
@@ -0,0 +1,11 @@
+# Model settings
+model = dict(
+    type='ImageClassifier',
+    backbone=dict(type='DenseNet', arch='121'),
+    neck=dict(type='GlobalAveragePooling'),
+    head=dict(
+        type='LinearClsHead',
+        num_classes=1000,
+        in_channels=1024,
+        loss=dict(type='CrossEntropyLoss', loss_weight=1.0),
+    ))
diff --git a/configs/_base_/models/densenet/densenet161.py b/configs/_base_/models/densenet/densenet161.py
@@ -0,0 +1,11 @@
+# Model settings
+model = dict(
+    type='ImageClassifier',
+    backbone=dict(type='DenseNet', arch='161'),
+    neck=dict(type='GlobalAveragePooling'),
+    head=dict(
+        type='LinearClsHead',
+        num_classes=1000,
+        in_channels=2208,
+        loss=dict(type='CrossEntropyLoss', loss_weight=1.0),
+    ))
diff --git a/configs/_base_/models/densenet/densenet169.py b/configs/_base_/models/densenet/densenet169.py
@@ -0,0 +1,11 @@
+# Model settings
+model = dict(
+    type='ImageClassifier',
+    backbone=dict(type='DenseNet', arch='169'),
+    neck=dict(type='GlobalAveragePooling'),
+    head=dict(
+        type='LinearClsHead',
+        num_classes=1000,
+        in_channels=1664,
+        loss=dict(type='CrossEntropyLoss', loss_weight=1.0),
+    ))
diff --git a/configs/_base_/models/densenet/densenet201.py b/configs/_base_/models/densenet/densenet201.py
@@ -0,0 +1,11 @@
+# Model settings
+model = dict(
+    type='ImageClassifier',
+    backbone=dict(type='DenseNet', arch='201'),
+    neck=dict(type='GlobalAveragePooling'),
+    head=dict(
+        type='LinearClsHead',
+        num_classes=1000,
+        in_channels=1920,
+        loss=dict(type='CrossEntropyLoss', loss_weight=1.0),
+    ))
diff --git a/configs/densenet/README.md b/configs/densenet/README.md
@@ -0,0 +1,41 @@
+# DenseNet
+
+> [Densely Connected Convolutional Networks](https://arxiv.org/abs/1608.06993)
+<!-- [ALGORITHM] -->
+
+## Abstract
+
+Recent work has shown that convolutional networks can be substantially deeper, more accurate, and efficient to train if they contain shorter connections between layers close to the input and those close to the output. In this paper, we embrace this observation and introduce the Dense Convolutional Network (DenseNet), which connects each layer to every other layer in a feed-forward fashion. Whereas traditional convolutional networks with L layers have L connections - one between each layer and its subsequent layer - our network has L(L+1)/2 direct connections. For each layer, the feature-maps of all preceding layers are used as inputs, and its own feature-maps are used as inputs into all subsequent layers. DenseNets have several compelling advantages: they alleviate the vanishing-gradient problem, strengthen feature propagation, encourage feature reuse, and substantially reduce the number of parameters. We evaluate our proposed architecture on four highly competitive object recognition benchmark tasks (CIFAR-10, CIFAR-100, SVHN, and ImageNet). DenseNets obtain significant improvements over the state-of-the-art on most of them, whilst requiring less computation to achieve high performance.
+
+<div align=center>
+<img src="https://user-images.githubusercontent.com/42952108/162675098-9a670883-b13a-4a5a-a9c9-06c39c616a0a.png" width="100%"/>
+</div>
+
+## Results and models
+
+### ImageNet-1k
+
+|      Model      | Params(M) | Flops(G) | Top-1 (%) | Top-5 (%) | Config | Download |
+|:---------------:|:---------:|:--------:|:---------:|:---------:|:------:|:--------:|
+| DenseNet121\*   | 7.98      | 2.88     | 74.96     | 92.21     | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/densenet/densenet121_4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/densenet/densenet121_4xb256_in1k_20220426-07450f99.pth) |
+| DenseNet169\*   | 14.15     | 3.42     | 76.08     | 93.11     | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/densenet/densenet169_4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/densenet/densenet169_4xb256_in1k_20220426-a2889902.pth) |
+| DenseNet201\*   | 20.01     | 4.37     | 77.32     | 93.64     | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/densenet/densenet201_4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/densenet/densenet201_4xb256_in1k_20220426-05cae4ef.pth) |
+| DenseNet161\*   | 28.68     | 7.82     | 77.61     | 93.83     | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/densenet/densenet161_4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/densenet/densenet161_4xb256_in1k_20220426-ee6a80a9.pth) |
+
+*Models with \* are converted from [pytorch](https://pytorch.org/vision/stable/models.html), guided by [original repo](https://github.com/liuzhuang13/DenseNet). The config files of these models are only for inference. We don't ensure these config files' training accuracy and welcome you to contribute your reproduction results.*
+
+
+## Citation
+
+```bibtex
+@misc{https://doi.org/10.48550/arxiv.1608.06993,
+      doi = {10.48550/ARXIV.1608.06993},
+      url = {https://arxiv.org/abs/1608.06993},
+      author = {Huang, Gao and Liu, Zhuang and van der Maaten, Laurens and Weinberger, Kilian Q.},
+      keywords = {Computer Vision and Pattern Recognition (cs.CV), Machine Learning (cs.LG), FOS: Computer and information sciences, FOS: Computer and information sciences},
+      title = {Densely Connected Convolutional Networks},
+      publisher = {arXiv},
+      year = {2016},
+      copyright = {arXiv.org perpetual, non-exclusive license}
+}
+```
diff --git a/configs/densenet/densenet121_4xb256_in1k.py b/configs/densenet/densenet121_4xb256_in1k.py
@@ -0,0 +1,10 @@
+_base_ = [
+    '../_base_/models/densenet/densenet121.py',
+    '../_base_/datasets/imagenet_bs64.py',
+    '../_base_/schedules/imagenet_bs256.py',
+    '../_base_/default_runtime.py',
+]
+
+data = dict(samples_per_gpu=256)
+
+runner = dict(type='EpochBasedRunner', max_epochs=90)
diff --git a/configs/densenet/densenet161_4xb256_in1k.py b/configs/densenet/densenet161_4xb256_in1k.py
@@ -0,0 +1,10 @@
+_base_ = [
+    '../_base_/models/densenet/densenet161.py',
+    '../_base_/datasets/imagenet_bs64.py',
+    '../_base_/schedules/imagenet_bs256.py',
+    '../_base_/default_runtime.py',
+]
+
+data = dict(samples_per_gpu=256)
+
+runner = dict(type='EpochBasedRunner', max_epochs=90)
diff --git a/configs/densenet/densenet169_4xb256_in1k.py b/configs/densenet/densenet169_4xb256_in1k.py
@@ -0,0 +1,10 @@
+_base_ = [
+    '../_base_/models/densenet/densenet169.py',
+    '../_base_/datasets/imagenet_bs64.py',
+    '../_base_/schedules/imagenet_bs256.py',
+    '../_base_/default_runtime.py',
+]
+
+data = dict(samples_per_gpu=256)
+
+runner = dict(type='EpochBasedRunner', max_epochs=90)
diff --git a/configs/densenet/densenet201_4xb256_in1k.py b/configs/densenet/densenet201_4xb256_in1k.py
@@ -0,0 +1,10 @@
+_base_ = [
+    '../_base_/models/densenet/densenet201.py',
+    '../_base_/datasets/imagenet_bs64.py',
+    '../_base_/schedules/imagenet_bs256.py',
+    '../_base_/default_runtime.py',
+]
+
+data = dict(samples_per_gpu=256)
+
+runner = dict(type='EpochBasedRunner', max_epochs=90)
diff --git a/configs/densenet/metafile.yml b/configs/densenet/metafile.yml
@@ -0,0 +1,76 @@
+Collections:
+  - Name: DenseNet
+    Metadata:
+      Training Data: ImageNet-1k
+      Architecture:
+        - DenseBlock
+    Paper:
+      URL: https://arxiv.org/abs/1608.06993
+      Title: Densely Connected Convolutional Networks
+    README: configs/densenet/README.md
+
+Models:
+  - Name: densenet121_4xb256_in1k
+    Metadata:
+      FLOPs: 2881695488
+      Parameters: 7978856
+    In Collections: DenseNet
+    Results:
+      - Dataset: ImageNet-1k
+        Metrics:
+          Top 1 Accuracy: 74.96
+          Top 5 Accuracy: 92.21
+        Task: Image Classification
+    Weights: https://download.openmmlab.com/mmclassification/v0/densenet/densenet121_4xb256_in1k_20220426-07450f99.pth
+    Config: configs/densenet/densenet121_4xb256_in1k.py
+    Converted From:
+      Weights: https://download.pytorch.org/models/densenet121-a639ec97.pth
+      Code: https://github.com/pytorch/vision/blob/main/torchvision/models/densenet.py
+  - Name: densenet169_4xb256_in1k
+    Metadata:
+      FLOPs: 3416860160
+      Parameters: 14149480
+    In Collections: DenseNet
+    Results:
+      - Dataset: ImageNet-1k
+        Metrics:
+          Top 1 Accuracy: 76.08
+          Top 5 Accuracy: 93.11
+        Task: Image Classification
+    Weights: https://download.openmmlab.com/mmclassification/v0/densenet/densenet169_4xb256_in1k_20220426-a2889902.pth
+    Config: configs/densenet/densenet169_4xb256_in1k.py
+    Converted From:
+      Weights: https://download.pytorch.org/models/densenet169-b2777c0a.pth
+      Code: https://github.com/pytorch/vision/blob/main/torchvision/models/densenet.py
+  - Name: densenet201_4xb256_in1k
+    Metadata:
+      FLOPs: 4365236736
+      Parameters: 20013928
+    In Collections: DenseNet
+    Results:
+      - Dataset: ImageNet-1k
+        Metrics:
+          Top 1 Accuracy: 77.32
+          Top 5 Accuracy: 93.64
+        Task: Image Classification
+    Weights: https://download.openmmlab.com/mmclassification/v0/densenet/densenet201_4xb256_in1k_20220426-05cae4ef.pth
+    Config: configs/densenet/densenet201_4xb256_in1k.py
+    Converted From:
+      Weights: https://download.pytorch.org/models/densenet201-c1103571.pth
+      Code: https://github.com/pytorch/vision/blob/main/torchvision/models/densenet.py
+  - Name: densenet161_4xb256_in1k
+    Metadata:
+      FLOPs: 7816363968
+      Parameters: 28681000
+    In Collections: DenseNet
+    Results:
+      - Dataset: ImageNet-1k
+        Metrics:
+          Top 1 Accuracy: 77.61
+          Top 5 Accuracy: 93.83
+        Task: Image Classification
+    Weights: https://download.openmmlab.com/mmclassification/v0/densenet/densenet161_4xb256_in1k_20220426-ee6a80a9.pth
+    Config: configs/densenet/densenet161_4xb256_in1k.py
+    Converted From:
+      Weights: https://download.pytorch.org/models/densenet161-8d451a50.pth
+      Code: https://github.com/pytorch/vision/blob/main/torchvision/models/densenet.py
diff --git a/docs/en/api/models.rst b/docs/en/api/models.rst
@@ -55,6 +55,7 @@ Backbones
    Conformer
    ConvMixer
    ConvNeXt
+   DenseNet
    DistilledVisionTransformer
    EfficientNet
    HRNet

diff --git a/docs/en/model_zoo.md b/docs/en/model_zoo.md
@@ -133,6 +133,10 @@ The ResNet family models below are trained by standard data augmentations, i.e.,
 | CSPDarkNet50\*  |  27.64 | 5.04 | 80.05 | 95.07  | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/cspnet/cspdarknet50_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/cspnet/cspdarknet50_3rdparty_8xb32_in1k_20220329-bd275287.pth) |
 |  CSPResNet50\*  |  21.62 | 3.48 | 79.55 | 94.68  | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/cspnet/cspresnet50_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/cspnet/cspresnet50_3rdparty_8xb32_in1k_20220329-dd6dddfb.pth) |
 |  CSPResNeXt50\* | 20.57 | 3.11 | 79.96 | 94.96 | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/cspnet/cspresnext50_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/cspnet/cspresnext50_3rdparty_8xb32_in1k_20220329-2cc84d21.pth) |
+|  DenseNet121\*   | 7.98      | 2.88     | 74.96     | 92.21     | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/densenet/densenet121_4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/densenet/densenet121_4xb256_in1k_20220426-07450f99.pth) |
+|  DenseNet169\*   | 14.15     | 3.42     | 76.08     | 93.11     | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/densenet/densenet169_4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/densenet/densenet169_4xb256_in1k_20220426-a2889902.pth) |
+|  DenseNet201\*   | 20.01     | 4.37     | 77.32     | 93.64     | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/densenet/densenet201_4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/densenet/densenet201_4xb256_in1k_20220426-05cae4ef.pth) |
+|  DenseNet161\*   | 28.68     | 7.82     | 77.61     | 93.83     | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/densenet/densenet161_4xb256_in1k.py) | [model](https://download.openmmlab.com/mmclassification/v0/densenet/densenet161_4xb256_in1k_20220426-ee6a80a9.pth) |
 |  VAN-T\*   |   4.11   |    0.88   |   75.41   |   93.02   |  [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/van/van-tiny_8xb128_in1k.py)  | [model](https://download.openmmlab.com/mmclassification/v0/van/van-tiny_8xb128_in1k_20220427-8ac0feec.pth)  |
 |  VAN-S\*   |   13.86   |    2.52   |   81.01   |    95.63   | [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/van/van-small_8xb128_in1k.py)  | [model](https://download.openmmlab.com/mmclassification/v0/van/van-small_8xb128_in1k_20220427-bd6a9edd.pth)  |
 |  VAN-B\*   |   26.58   |   5.03   |   82.80   |    96.21   |   [config](https://github.com/open-mmlab/mmclassification/blob/master/configs/van/van-base_8xb128_in1k.py)  | [model](https://download.openmmlab.com/mmclassification/v0/van/van-base_8xb128_in1k_20220427-5275471d.pth)  |

diff --git a/docs/zh_CN/api/models.rst b/docs/zh_CN/api/models.rst
@@ -55,6 +55,7 @@ Backbones
    Conformer
    ConvMixer
    ConvNeXt
+   DenseNet
    DistilledVisionTransformer
    EfficientNet
    HRNet

diff --git a/mmcls/models/backbones/__init__.py b/mmcls/models/backbones/__init__.py
@@ -5,6 +5,7 @@
 from .convnext import ConvNeXt
 from .cspnet import CSPDarkNet, CSPNet, CSPResNet, CSPResNeXt
 from .deit import DistilledVisionTransformer
+from .densenet import DenseNet
 from .efficientnet import EfficientNet
 from .hrnet import HRNet
 from .lenet import LeNet5
@@ -41,5 +42,5 @@
     'Conformer', 'MlpMixer', 'DistilledVisionTransformer', 'PCPVT', 'SVT',
     'EfficientNet', 'ConvNeXt', 'HRNet', 'ResNetV1c', 'ConvMixer',
     'CSPDarkNet', 'CSPResNet', 'CSPResNeXt', 'CSPNet', 'RepMLPNet',
-    'PoolFormer', 'VAN'
+    'PoolFormer', 'DenseNet', 'VAN'
 ]