[Feature] Provide URLs of Swin Transformer pretrained models (open-mmlab#1389)

* [Feature] Provide URLs of Swin Transformer pretrained models

* [Feature] Add multi machine `dist_train`. (open-mmlab#1383)

* Add training startup documentation

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* fix

* delete pretrained=None in all six config files

* [Fix] make arguments effective in tools/confusion_matrix.py (open-mmlab#1401)

* add an argument for customizing `title' of the output figure

* fix `color_theme' arguments not passing to plot function

Signed-off-by: code14 <[email protected]>

* colab notebook: fix outdated link for doc (open-mmlab#1392)

* colab notebook: fix outdated link for doc

Fixed outdated link for how to customize your datasets by reorganizing data.

* fix lint

* fix typo (open-mmlab#1405)

* [Fix] Fix windows-style path in `md2yml.py` in Windows pre-commit.  (open-mmlab#1407)

* test

* avoid windows path

* [Fix] fix the config name style description (open-mmlab#1414)

Co-authored-by: FangjianLin <[email protected]>
Co-authored-by: Cody Wong <[email protected]>
Co-authored-by: Nemo Xiong <[email protected]>
Co-authored-by: Xiangxu-0103 <[email protected]>
Co-authored-by: Rockey <[email protected]>

Author: 6 people

.DS_Store

@@ -34,9 +34,9 @@ This paper presents a new vision Transformer, called Swin Transformer, that capa
 
 ## Usage
 
-To use other repositories' pre-trained models, it is necessary to convert keys.
+We have provided pretrained models converted from [official repo](https://github.com/microsoft/Swin-Transformer)．
 
-We provide a script [`swin2mmseg.py`](../../tools/model_converters/swin2mmseg.py) in the tools directory to convert the key of models from [the official repo](https://github.com/SwinTransformer/Swin-Transformer-Semantic-Segmentation) to MMSegmentation style.
+If you want to convert keys on your own to use official repositories' pre-trained models, we also provide a script [`swin2mmseg.py`](../../tools/model_converters/swin2mmseg.py) in the tools directory to convert the key of models from [the official repo](https://github.com/SwinTransformer/Swin-Transformer-Semantic-Segmentation) to MMSegmentation style.
 
 ```shell
 python tools/model_converters/swin2mmseg.py ${PRETRAIN_PATH} ${STORE_PATH}
@@ -50,6 +50,17 @@ python tools/model_converters/swin2mmseg.py https://github.com/SwinTransformer/s
 
 This script convert model from `PRETRAIN_PATH` and store the converted model in `STORE_PATH`.
 
+In our default setting, pretrained models and their corresponding [original models](https://github.com/microsoft/Swin-Transforme) models could be defined below:
+
+  | pretrained models | original models |
+  | ------ | -------- |
+  |pretrain/swin_tiny_patch4_window7_224.pth | [swin_tiny_patch4_window7_224.pth](https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_tiny_patch4_window7_224.pth) |
+  |pretrain/swin_small_patch4_window7_224.pth | [swin_small_patch4_window7_224.pth](https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_small_patch4_window7_224.pth) |
+  |pretrain/swin_base_patch4_window7_224.pth | [swin_base_patch4_window7_224.pth](https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_base_patch4_window7_224.pth) |
+  |pretrain/swin_base_patch4_window7_224_22k.pth | [swin_base_patch4_window7_224_22k.pth](https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_base_patch4_window7_224_22k.pth) |
+  |pretrain/swin_base_patch4_window12_384.pth | [swin_base_patch4_window12_384.pth](https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_base_patch4_window12_384.pth) |
+  |pretrain/swin_base_patch4_window12_384_22k.pth | [swin_base_patch4_window12_384_22k.pth](https://github.com/SwinTransformer/storage/releases/download/v1.0.0/swin_base_patch4_window12_384_22k.pth) |
+
 ## Results and models
 
 ### ADE20K

configs/swin/README.md

@@ -2,9 +2,10 @@
     'upernet_swin_tiny_patch4_window7_512x512_160k_ade20k_'
     'pretrain_224x224_1K.py'
 ]
+checkpoint_file = 'https://download.openmmlab.com/mmsegmentation/v0.5/pretrain/swin/swin_base_patch4_window12_384_20220317-55b0104a.pth'  # noqa
 model = dict(
-    pretrained='pretrain/swin_base_patch4_window12_384.pth',
     backbone=dict(
+        init_cfg=dict(type='Pretrained', checkpoint=checkpoint_file),
         pretrain_img_size=384,
         embed_dims=128,
         depths=[2, 2, 18, 2],

configs/swin/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_1K.py

@@ -2,4 +2,7 @@
     './upernet_swin_base_patch4_window12_512x512_160k_ade20k_'
     'pretrain_384x384_1K.py'
 ]
-model = dict(pretrained='pretrain/swin_base_patch4_window12_384_22k.pth')
+checkpoint_file = 'https://download.openmmlab.com/mmsegmentation/v0.5/pretrain/swin/swin_base_patch4_window12_384_22k_20220317-e5c09f74.pth'  # noqa
+model = dict(
+    backbone=dict(
+        init_cfg=dict(type='Pretrained', checkpoint=checkpoint_file)))

configs/swin/upernet_swin_base_patch4_window12_512x512_160k_ade20k_pretrain_384x384_22K.py

@@ -2,9 +2,12 @@
     './upernet_swin_tiny_patch4_window7_512x512_160k_ade20k_'
     'pretrain_224x224_1K.py'
 ]
+checkpoint_file = 'https://download.openmmlab.com/mmsegmentation/v0.5/pretrain/swin/swin_base_patch4_window7_224_20220317-e9b98025.pth'  # noqa
 model = dict(
-    pretrained='pretrain/swin_base_patch4_window7_224.pth',
     backbone=dict(
-        embed_dims=128, depths=[2, 2, 18, 2], num_heads=[4, 8, 16, 32]),
+        init_cfg=dict(type='Pretrained', checkpoint=checkpoint_file),
+        embed_dims=128,
+        depths=[2, 2, 18, 2],
+        num_heads=[4, 8, 16, 32]),
     decode_head=dict(in_channels=[128, 256, 512, 1024], num_classes=150),
     auxiliary_head=dict(in_channels=512, num_classes=150))

configs/swin/upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K.py

@@ -2,4 +2,7 @@
     './upernet_swin_base_patch4_window7_512x512_160k_ade20k_'
     'pretrain_224x224_1K.py'
 ]
-model = dict(pretrained='pretrain/swin_base_patch4_window7_224_22k.pth')
+checkpoint_file = 'https://download.openmmlab.com/mmsegmentation/v0.5/pretrain/swin/swin_base_patch4_window7_224_22k_20220317-4f79f7c0.pth'  # noqa
+model = dict(
+    backbone=dict(
+        init_cfg=dict(type='Pretrained', checkpoint=checkpoint_file)))

configs/swin/upernet_swin_base_patch4_window7_512x512_160k_ade20k_pretrain_224x224_22K.py

@@ -2,8 +2,10 @@
     './upernet_swin_tiny_patch4_window7_512x512_160k_ade20k_'
     'pretrain_224x224_1K.py'
 ]
+checkpoint_file = 'https://download.openmmlab.com/mmsegmentation/v0.5/pretrain/swin/swin_small_patch4_window7_224_20220317-7ba6d6dd.pth'  # noqa
 model = dict(
-    pretrained='pretrain/swin_small_patch4_window7_224.pth',
-    backbone=dict(depths=[2, 2, 18, 2]),
+    backbone=dict(
+        init_cfg=dict(type='Pretrained', checkpoint=checkpoint_file),
+        depths=[2, 2, 18, 2]),
     decode_head=dict(in_channels=[96, 192, 384, 768], num_classes=150),
     auxiliary_head=dict(in_channels=384, num_classes=150))

configs/swin/upernet_swin_small_patch4_window7_512x512_160k_ade20k_pretrain_224x224_1K.py

@@ -2,9 +2,10 @@
     '../_base_/models/upernet_swin.py', '../_base_/datasets/ade20k.py',
     '../_base_/default_runtime.py', '../_base_/schedules/schedule_160k.py'
 ]
+checkpoint_file = 'https://download.openmmlab.com/mmsegmentation/v0.5/pretrain/swin/swin_tiny_patch4_window7_224_20220317-1cdeb081.pth'  # noqa
 model = dict(
-    pretrained='pretrain/swin_tiny_patch4_window7_224.pth',
     backbone=dict(
+        init_cfg=dict(type='Pretrained', checkpoint=checkpoint_file),
         embed_dims=96,
         depths=[2, 2, 6, 2],
         num_heads=[3, 6, 12, 24],