[Fix] update pp_mobileseg ckpt links (#3254)

open-mmlab · Aug 10, 2023 · 9801c62 · 9801c62
1 parent e458a46
commit 9801c62
Show file tree

Hide file tree

Showing 3 changed files with 28 additions and 9 deletions.
diff --git a/projects/pp_mobileseg/README.md b/projects/pp_mobileseg/README.md
@@ -30,10 +30,18 @@ Extensive experiments show that PP-MobileSeg achieves a superior params-accuracy
 
 ### ADE20K
 
-| Model             | Backbone          | Training Iters | Batchsize | Train Resolution | mIoU(%) | latency(ms)\* | params(M) | config                                                                                                                    | Links                                                                                                                                                                                                                          |
-| ----------------- | ----------------- | -------------- | --------- | ---------------- | ------- | ------------- | --------- | ------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
-| PP-MobileSeg-Base | StrideFormer-Base | 80000          | 32        | 512x512          | 41.57%  | 265.5         | 5.62      | [config](https://github.com/Yang-Changhui/mmsegmentation/tree/add_ppmobileseg/projects/pp_mobileseg/configs/pp_mobileseg) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/pp_mobileseg/pp_mobileseg_mobilenetv3_2x16_80k_ade20k_512x512_base-ed0be681.pth)\|[log](https://bj.bcebos.com/paddleseg/dygraph/ade20k/pp_mobileseg_base/train.log) |
-| PP-MobileSeg-Tiny | StrideFormer-Tiny | 80000          | 32        | 512x512          | 36.39%  | 215.3         | 1.61      | [config](https://github.com/Yang-Changhui/mmsegmentation/tree/add_ppmobileseg/projects/pp_mobileseg/configs/pp_mobileseg) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/pp_mobileseg/pp_mobileseg_mobilenetv3_2x16_80k_ade20k_512x512_tiny-e4b35e96.pth)\|[log](https://bj.bcebos.com/paddleseg/dygraph/ade20k/pp_mobileseg_tiny/train.log) |
+| Model             | Backbone          | Training Iters | Batchsize | Train Resolution | mIoU(%) | latency(ms)\* | params(M) | config                                                                                                                    | Links                                                                                                                                                                                                                                |
+| ----------------- | ----------------- | -------------- | --------- | ---------------- | ------- | ------------- | --------- | ------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
+| PP-MobileSeg-Base | StrideFormer-Base | 80000          | 32        | 512x512          | 41.57%  | 265.5         | 5.62      | [config](https://github.com/Yang-Changhui/mmsegmentation/tree/add_ppmobileseg/projects/pp_mobileseg/configs/pp_mobileseg) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/pp_mobileseg/pp_mobileseg_mobilenetv3_2xb16_3rdparty-base_512x512-ade20k-f12b44f3.pth)\|[log](https://bj.bcebos.com/paddleseg/dygraph/ade20k/pp_mobileseg_base/train.log) |
+| PP-MobileSeg-Tiny | StrideFormer-Tiny | 80000          | 32        | 512x512          | 36.39%  | 215.3         | 1.61      | [config](https://github.com/Yang-Changhui/mmsegmentation/tree/add_ppmobileseg/projects/pp_mobileseg/configs/pp_mobileseg) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/pp_mobileseg/pp_mobileseg_mobilenetv3_2xb16_3rdparty-tiny_512x512-ade20k-a351ebf5.pth)\|[log](https://bj.bcebos.com/paddleseg/dygraph/ade20k/pp_mobileseg_tiny/train.log) |
+
+## Usage
+
+Same as other models in MMsegmentation, you can run the following command to test the model at ${MMSEG_ROOT}:
+
+```shell
+./tools/dist_test.sh projects/pp_mobileseg/configs/pp_mobileseg/pp_mobileseg_mobilenetv3_2x16_80k_ade20k_512x512_base.py checkpoints/pp_mobileseg_mobilenetv3_2xb16_3rdparty-base_512x512-ade20k-f12b44f3.pth 8
+```
 
 ## Citation
 

diff --git a/...p_mobileseg/configs/pp_mobileseg/pp_mobileseg_mobilenetv3_2x16_80k_ade20k_512x512_base.py b/...p_mobileseg/configs/pp_mobileseg/pp_mobileseg_mobilenetv3_2x16_80k_ade20k_512x512_base.py
@@ -2,12 +2,17 @@
     '../_base_/models/pp_mobile.py', '../_base_/datasets/ade20k.py',
     '../_base_/default_runtime.py', '../_base_/schedules/schedule_80k.py'
 ]
-checkpoint = './models/pp_mobile_base.pth'
+# the custom import path is determined by your workspace path (i.e., where you run the command from) # noqa
+custom_imports = dict(
+    imports=[
+        'projects.pp_mobileseg.backbones', 'projects.pp_mobileseg.decode_head'
+    ],
+    allow_failed_imports=False)
+checkpoint = 'https://download.openmmlab.com/mmsegmentation/v0.5/pp_mobileseg/pp_mobileseg_mobilenetv3_3rdparty-base-ed0be681.pth'  # noqa
 crop_size = (512, 512)
 data_preprocessor = dict(size=crop_size, test_cfg=dict(size_divisor=32))
 norm_cfg = dict(type='SyncBN', requires_grad=True)
 model = dict(
     data_preprocessor=data_preprocessor,
-    backbone=dict(init_cfg=dict(type='Pretrained', checkpoint=checkpoint), ),
-    decode_head=dict(num_classes=150, upsample='intepolate'),
-)
+    backbone=dict(init_cfg=dict(type='Pretrained', checkpoint=checkpoint)),
+    decode_head=dict(num_classes=150, upsample='intepolate'))
diff --git a/...p_mobileseg/configs/pp_mobileseg/pp_mobileseg_mobilenetv3_2x16_80k_ade20k_512x512_tiny.py b/...p_mobileseg/configs/pp_mobileseg/pp_mobileseg_mobilenetv3_2x16_80k_ade20k_512x512_tiny.py
@@ -2,7 +2,13 @@
     '../_base_/models/pp_mobile.py', '../_base_/datasets/ade20k.py',
     '../_base_/default_runtime.py', '../_base_/schedules/schedule_80k.py'
 ]
-checkpoint = './models/pp_mobile_tiny.pth'
+# the custom import path is determined by your workspace path (i.e., where you run the command from) # noqa
+custom_imports = dict(
+    imports=[
+        'projects.pp_mobileseg.backbones', 'projects.pp_mobileseg.decode_head'
+    ],
+    allow_failed_imports=False)
+checkpoint = 'https://download.openmmlab.com/mmsegmentation/v0.5/pp_mobileseg/pp_mobileseg_mobilenetv3_3rdparty-tiny-e4b35e96.pth'  # noqa
 crop_size = (512, 512)
 data_preprocessor = dict(size=crop_size, test_cfg=dict(size_divisor=32))
 norm_cfg = dict(type='SyncBN', requires_grad=True)