first commit
This commit is contained in:
admin
73
Seg_All_In_One_MMSeg/configs/stdc/README.md
Normal file
73
Seg_All_In_One_MMSeg/configs/stdc/README.md
Normal file
@@ -0,0 +1,73 @@
|
||||
# STDC
|
||||
|
||||
> [Rethinking BiSeNet For Real-time Semantic Segmentation](https://arxiv.org/abs/2104.13188)
|
||||
|
||||
## Introduction
|
||||
|
||||
<!-- [ALGORITHM] -->
|
||||
|
||||
<a href="https://github.com/MichaelFan01/STDC-Seg">Official Repo</a>
|
||||
|
||||
<a href="https://github.com/open-mmlab/mmsegmentation/blob/v0.20.0/mmseg/models/backbones/stdc.py#L394">Code Snippet</a>
|
||||
|
||||
## Abstract
|
||||
|
||||
<!-- [ABSTRACT] -->
|
||||
|
||||
BiSeNet has been proved to be a popular two-stream network for real-time segmentation. However, its principle of adding an extra path to encode spatial information is time-consuming, and the backbones borrowed from pretrained tasks, e.g., image classification, may be inefficient for image segmentation due to the deficiency of task-specific design. To handle these problems, we propose a novel and efficient structure named Short-Term Dense Concatenate network (STDC network) by removing structure redundancy. Specifically, we gradually reduce the dimension of feature maps and use the aggregation of them for image representation, which forms the basic module of STDC network. In the decoder, we propose a Detail Aggregation module by integrating the learning of spatial information into low-level layers in single-stream manner. Finally, the low-level features and deep features are fused to predict the final segmentation results. Extensive experiments on Cityscapes and CamVid dataset demonstrate the effectiveness of our method by achieving promising trade-off between segmentation accuracy and inference speed. On Cityscapes, we achieve 71.9% mIoU on the test set with a speed of 250.4 FPS on NVIDIA GTX 1080Ti, which is 45.2% faster than the latest methods, and achieve 76.8% mIoU with 97.0 FPS while inferring on higher resolution images.
|
||||
|
||||
<!-- [IMAGE] -->
|
||||
|
||||
<div align=center>
|
||||
<img src="https://user-images.githubusercontent.com/24582831/143640374-d0709587-edb2-4821-bb60-340035f6ad8f.png" width="60%"/>
|
||||
</div>
|
||||
|
||||
## Usage
|
||||
|
||||
We have provided [ImageNet Pretrained STDCNet Weights](https://drive.google.com/drive/folders/1wROFwRt8qWHD4jSo8Zu1gp1d6oYJ3ns1) models converted from [official repo](https://github.com/MichaelFan01/STDC-Seg).
|
||||
|
||||
If you want to convert keys on your own to use official repositories' pre-trained models, we also provide a script [`stdc2mmseg.py`](../../tools/model_converters/stdc2mmseg.py) in the tools directory to convert the key of models from [the official repo](https://github.com/MichaelFan01/STDC-Seg) to MMSegmentation style.
|
||||
|
||||
```shell
|
||||
python tools/model_converters/stdc2mmseg.py ${PRETRAIN_PATH} ${STORE_PATH} ${STDC_TYPE}
|
||||
```
|
||||
|
||||
E.g.
|
||||
|
||||
```shell
|
||||
python tools/model_converters/stdc2mmseg.py ./STDCNet813M_73.91.tar ./pretrained/stdc1.pth STDC1
|
||||
|
||||
python tools/model_converters/stdc2mmseg.py ./STDCNet1446_76.47.tar ./pretrained/stdc2.pth STDC2
|
||||
```
|
||||
|
||||
This script convert model from `PRETRAIN_PATH` and store the converted model in `STORE_PATH`.
|
||||
|
||||
## Results and models
|
||||
|
||||
### Cityscapes
|
||||
|
||||
| Method | Backbone | Crop Size | Lr schd | Mem (GB) | Inf time (fps) | Device | mIoU | mIoU(ms+flip) | config | download |
|
||||
| ------ | -------------------- | --------- | ------: | -------- | -------------- | ------ | ----: | ------------- | ----------------------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
||||
| STDC | STDC1 (No Pretrain) | 512x1024 | 80000 | 7.15 | 23.06 | V100 | 71.82 | 73.89 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/stdc/stdc1_4xb12-80k_cityscapes-512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc1_512x1024_80k_cityscapes/stdc1_512x1024_80k_cityscapes_20220224_073048-74e6920a.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc1_512x1024_80k_cityscapes/stdc1_512x1024_80k_cityscapes_20220224_073048.log.json) |
|
||||
| STDC | STDC1 | 512x1024 | 80000 | - | - | V100 | 74.94 | 76.97 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/stdc/stdc1_in1k-pre_4xb12-80k_cityscapes-512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc1_in1k-pre_512x1024_80k_cityscapes/stdc1_in1k-pre_512x1024_80k_cityscapes_20220224_141648-3d4c2981.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc1_in1k-pre_512x1024_80k_cityscapes/stdc1_in1k-pre_512x1024_80k_cityscapes_20220224_141648.log.json) |
|
||||
| STDC | STDC2 (No Pretrain) | 512x1024 | 80000 | 8.27 | 23.71 | V100 | 73.15 | 76.13 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/stdc/stdc2_4xb12-80k_cityscapes-512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc2_512x1024_80k_cityscapes/stdc2_512x1024_80k_cityscapes_20220222_132015-fb1e3a1a.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc2_512x1024_80k_cityscapes/stdc2_512x1024_80k_cityscapes_20220222_132015.log.json) |
|
||||
| STDC | STDC2 | 512x1024 | 80000 | - | - | V100 | 76.67 | 78.67 | [config](https://github.com/open-mmlab/mmsegmentation/blob/main/configs/stdc/stdc2_in1k-pre_4xb12-80k_cityscapes-512x1024.py) | [model](https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc2_in1k-pre_512x1024_80k_cityscapes/stdc2_in1k-pre_512x1024_80k_cityscapes_20220224_073048-1f8f0f6c.pth) \| [log](https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc2_in1k-pre_512x1024_80k_cityscapes/stdc2_in1k-pre_512x1024_80k_cityscapes_20220224_073048.log.json) |
|
||||
|
||||
Note:
|
||||
|
||||
- For STDC on Cityscapes dataset, default setting is 4 GPUs with 12 samples per GPU in training.
|
||||
- `No Pretrain` means the model is trained from scratch.
|
||||
- The FPS is for reference only. The environment is also different from paper setting, whose input size is `512x1024` and `768x1536`, i.e., 50% and 75% of our input size, respectively and using TensorRT.
|
||||
- The parameter `fusion_kernel` in `STDCHead` is not learnable. In official repo, `find_unused_parameters=True` is set [here](https://github.com/MichaelFan01/STDC-Seg/blob/59ff37fbd693b99972c76fcefe97caa14aeb619f/train.py#L220). You may check it by printing model parameters of original repo on your own.
|
||||
|
||||
## Citation
|
||||
|
||||
```bibtex
|
||||
@inproceedings{fan2021rethinking,
|
||||
title={Rethinking BiSeNet For Real-time Semantic Segmentation},
|
||||
author={Fan, Mingyuan and Lai, Shenqi and Huang, Junshi and Wei, Xiaoming and Chai, Zhenhua and Luo, Junfeng and Wei, Xiaolin},
|
||||
booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition},
|
||||
pages={9716--9725},
|
||||
year={2021}
|
||||
}
|
||||
```
|
||||
107
Seg_All_In_One_MMSeg/configs/stdc/metafile.yaml
Normal file
107
Seg_All_In_One_MMSeg/configs/stdc/metafile.yaml
Normal file
@@ -0,0 +1,107 @@
|
||||
Collections:
|
||||
- Name: STDC
|
||||
License: Apache License 2.0
|
||||
Metadata:
|
||||
Training Data:
|
||||
- Cityscapes
|
||||
Paper:
|
||||
Title: Rethinking BiSeNet For Real-time Semantic Segmentation
|
||||
URL: https://arxiv.org/abs/2104.13188
|
||||
README: configs/stdc/README.md
|
||||
Frameworks:
|
||||
- PyTorch
|
||||
Models:
|
||||
- Name: stdc1_4xb12-80k_cityscapes-512x1024
|
||||
In Collection: STDC
|
||||
Results:
|
||||
Task: Semantic Segmentation
|
||||
Dataset: Cityscapes
|
||||
Metrics:
|
||||
mIoU: 71.82
|
||||
mIoU(ms+flip): 73.89
|
||||
Config: configs/stdc/stdc1_4xb12-80k_cityscapes-512x1024.py
|
||||
Metadata:
|
||||
Training Data: Cityscapes
|
||||
Batch Size: 48
|
||||
Architecture:
|
||||
- STDC1
|
||||
- STDC
|
||||
Training Resources: 4x V100 GPUS
|
||||
Memory (GB): 7.15
|
||||
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc1_512x1024_80k_cityscapes/stdc1_512x1024_80k_cityscapes_20220224_073048-74e6920a.pth
|
||||
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc1_512x1024_80k_cityscapes/stdc1_512x1024_80k_cityscapes_20220224_073048.log.json
|
||||
Paper:
|
||||
Title: Rethinking BiSeNet For Real-time Semantic Segmentation
|
||||
URL: https://arxiv.org/abs/2104.13188
|
||||
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.20.0/mmseg/models/backbones/stdc.py#L394
|
||||
Framework: PyTorch
|
||||
- Name: stdc1_in1k-pre_4xb12-80k_cityscapes-512x1024
|
||||
In Collection: STDC
|
||||
Results:
|
||||
Task: Semantic Segmentation
|
||||
Dataset: Cityscapes
|
||||
Metrics:
|
||||
mIoU: 74.94
|
||||
mIoU(ms+flip): 76.97
|
||||
Config: configs/stdc/stdc1_in1k-pre_4xb12-80k_cityscapes-512x1024.py
|
||||
Metadata:
|
||||
Training Data: Cityscapes
|
||||
Batch Size: 48
|
||||
Architecture:
|
||||
- STDC1
|
||||
- STDC
|
||||
Training Resources: 4x V100 GPUS
|
||||
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc1_in1k-pre_512x1024_80k_cityscapes/stdc1_in1k-pre_512x1024_80k_cityscapes_20220224_141648-3d4c2981.pth
|
||||
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc1_in1k-pre_512x1024_80k_cityscapes/stdc1_in1k-pre_512x1024_80k_cityscapes_20220224_141648.log.json
|
||||
Paper:
|
||||
Title: Rethinking BiSeNet For Real-time Semantic Segmentation
|
||||
URL: https://arxiv.org/abs/2104.13188
|
||||
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.20.0/mmseg/models/backbones/stdc.py#L394
|
||||
Framework: PyTorch
|
||||
- Name: stdc2_4xb12-80k_cityscapes-512x1024
|
||||
In Collection: STDC
|
||||
Results:
|
||||
Task: Semantic Segmentation
|
||||
Dataset: Cityscapes
|
||||
Metrics:
|
||||
mIoU: 73.15
|
||||
mIoU(ms+flip): 76.13
|
||||
Config: configs/stdc/stdc2_4xb12-80k_cityscapes-512x1024.py
|
||||
Metadata:
|
||||
Training Data: Cityscapes
|
||||
Batch Size: 48
|
||||
Architecture:
|
||||
- STDC2
|
||||
- STDC
|
||||
Training Resources: 4x V100 GPUS
|
||||
Memory (GB): 8.27
|
||||
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc2_512x1024_80k_cityscapes/stdc2_512x1024_80k_cityscapes_20220222_132015-fb1e3a1a.pth
|
||||
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc2_512x1024_80k_cityscapes/stdc2_512x1024_80k_cityscapes_20220222_132015.log.json
|
||||
Paper:
|
||||
Title: Rethinking BiSeNet For Real-time Semantic Segmentation
|
||||
URL: https://arxiv.org/abs/2104.13188
|
||||
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.20.0/mmseg/models/backbones/stdc.py#L394
|
||||
Framework: PyTorch
|
||||
- Name: stdc2_in1k-pre_4xb12-80k_cityscapes-512x1024
|
||||
In Collection: STDC
|
||||
Results:
|
||||
Task: Semantic Segmentation
|
||||
Dataset: Cityscapes
|
||||
Metrics:
|
||||
mIoU: 76.67
|
||||
mIoU(ms+flip): 78.67
|
||||
Config: configs/stdc/stdc2_in1k-pre_4xb12-80k_cityscapes-512x1024.py
|
||||
Metadata:
|
||||
Training Data: Cityscapes
|
||||
Batch Size: 48
|
||||
Architecture:
|
||||
- STDC2
|
||||
- STDC
|
||||
Training Resources: 4x V100 GPUS
|
||||
Weights: https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc2_in1k-pre_512x1024_80k_cityscapes/stdc2_in1k-pre_512x1024_80k_cityscapes_20220224_073048-1f8f0f6c.pth
|
||||
Training log: https://download.openmmlab.com/mmsegmentation/v0.5/stdc/stdc2_in1k-pre_512x1024_80k_cityscapes/stdc2_in1k-pre_512x1024_80k_cityscapes_20220224_073048.log.json
|
||||
Paper:
|
||||
Title: Rethinking BiSeNet For Real-time Semantic Segmentation
|
||||
URL: https://arxiv.org/abs/2104.13188
|
||||
Code: https://github.com/open-mmlab/mmsegmentation/blob/v0.20.0/mmseg/models/backbones/stdc.py#L394
|
||||
Framework: PyTorch
|
||||
@@ -0,0 +1,94 @@
|
||||
_base_ = [
|
||||
'../_base_/models/stdc.py',
|
||||
'../_base_/datasets/publicdataset_autolaparo.py',
|
||||
'../_base_/default_runtime.py',
|
||||
'../_base_/schedules/schedule_300e_val1_check10.py',
|
||||
]
|
||||
|
||||
norm_cfg = dict(
|
||||
type='BN',
|
||||
)
|
||||
|
||||
crop_size = (512, 512)
|
||||
|
||||
data_preprocessor = dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
123.62464353460942,
|
||||
85.34836259209033,
|
||||
82.31539425671558,
|
||||
],
|
||||
std=[
|
||||
47.172211618459315,
|
||||
47.08256715323592,
|
||||
48.135121265163605,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
)
|
||||
|
||||
model = dict(
|
||||
data_preprocessor=dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
123.62464353460942,
|
||||
85.34836259209033,
|
||||
82.31539425671558,
|
||||
],
|
||||
std=[
|
||||
47.172211618459315,
|
||||
47.08256715323592,
|
||||
48.135121265163605,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
),
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
stdc_type='STDCNet1',
|
||||
),
|
||||
init_cfg=dict(
|
||||
type='Pretrained',
|
||||
checkpoint='./My_Local_Model/open_mmlab/stdc1.pth',
|
||||
),
|
||||
),
|
||||
decode_head=dict(
|
||||
num_classes=10,
|
||||
loss_decode=dict(
|
||||
type='DiceLoss',
|
||||
use_sigmoid=False,
|
||||
loss_weight=1.0,
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
optim_wrapper = dict(
|
||||
type='OptimWrapper',
|
||||
_delete_=True,
|
||||
optimizer=dict(
|
||||
type='AdamW',
|
||||
lr=0.0001,
|
||||
weight_decay=0.0005,
|
||||
),
|
||||
clip_grad=dict(
|
||||
max_norm=1,
|
||||
norm_type=2,
|
||||
),
|
||||
)
|
||||
|
||||
param_scheduler = [
|
||||
dict(
|
||||
type='LinearLR',
|
||||
start_factor=1e-06,
|
||||
by_epoch=True,
|
||||
begin=0,
|
||||
end=10,
|
||||
),
|
||||
dict(
|
||||
type='PolyLR',
|
||||
power=0.9,
|
||||
begin=10,
|
||||
end=300,
|
||||
eta_min=1e-05,
|
||||
by_epoch=True,
|
||||
),
|
||||
]
|
||||
|
||||
@@ -0,0 +1,94 @@
|
||||
_base_ = [
|
||||
'../_base_/models/stdc.py',
|
||||
'../_base_/datasets/publicdataset_cholecseg8k.py',
|
||||
'../_base_/default_runtime.py',
|
||||
'../_base_/schedules/schedule_300e_val1_check10.py',
|
||||
]
|
||||
|
||||
norm_cfg = dict(
|
||||
type='BN',
|
||||
)
|
||||
|
||||
crop_size = (512, 512)
|
||||
|
||||
data_preprocessor = dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
85.65740418979115,
|
||||
53.99282220050495,
|
||||
46.074045888534535,
|
||||
],
|
||||
std=[
|
||||
72.24589167201978,
|
||||
56.76979155397199,
|
||||
49.056637115061775,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
)
|
||||
|
||||
model = dict(
|
||||
data_preprocessor=dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
85.65740418979115,
|
||||
53.99282220050495,
|
||||
46.074045888534535,
|
||||
],
|
||||
std=[
|
||||
72.24589167201978,
|
||||
56.76979155397199,
|
||||
49.056637115061775,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
),
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
stdc_type='STDCNet1',
|
||||
),
|
||||
init_cfg=dict(
|
||||
type='Pretrained',
|
||||
checkpoint='./My_Local_Model/open_mmlab/stdc1.pth',
|
||||
),
|
||||
),
|
||||
decode_head=dict(
|
||||
num_classes=13,
|
||||
loss_decode=dict(
|
||||
type='DiceLoss',
|
||||
use_sigmoid=False,
|
||||
loss_weight=1.0,
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
optim_wrapper = dict(
|
||||
type='OptimWrapper',
|
||||
_delete_=True,
|
||||
optimizer=dict(
|
||||
type='AdamW',
|
||||
lr=0.0001,
|
||||
weight_decay=0.0005,
|
||||
),
|
||||
clip_grad=dict(
|
||||
max_norm=1,
|
||||
norm_type=2,
|
||||
),
|
||||
)
|
||||
|
||||
param_scheduler = [
|
||||
dict(
|
||||
type='LinearLR',
|
||||
start_factor=1e-06,
|
||||
by_epoch=True,
|
||||
begin=0,
|
||||
end=10,
|
||||
),
|
||||
dict(
|
||||
type='PolyLR',
|
||||
power=0.9,
|
||||
begin=10,
|
||||
end=300,
|
||||
eta_min=1e-05,
|
||||
by_epoch=True,
|
||||
),
|
||||
]
|
||||
|
||||
@@ -0,0 +1,94 @@
|
||||
_base_ = [
|
||||
'../_base_/models/stdc.py',
|
||||
'../_base_/datasets/publicdataset_dresden.py',
|
||||
'../_base_/default_runtime.py',
|
||||
'../_base_/schedules/schedule_300e_val1_check10.py',
|
||||
]
|
||||
|
||||
norm_cfg = dict(
|
||||
type='BN',
|
||||
)
|
||||
|
||||
crop_size = (512, 512)
|
||||
|
||||
data_preprocessor = dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
103.172638338208,
|
||||
61.44762740851152,
|
||||
51.407770213021976,
|
||||
],
|
||||
std=[
|
||||
75.77031253622098,
|
||||
54.63616729031377,
|
||||
49.45572239497569,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
)
|
||||
|
||||
model = dict(
|
||||
data_preprocessor=dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
103.172638338208,
|
||||
61.44762740851152,
|
||||
51.407770213021976,
|
||||
],
|
||||
std=[
|
||||
75.77031253622098,
|
||||
54.63616729031377,
|
||||
49.45572239497569,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
),
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
stdc_type='STDCNet1',
|
||||
),
|
||||
init_cfg=dict(
|
||||
type='Pretrained',
|
||||
checkpoint='./My_Local_Model/open_mmlab/stdc1.pth',
|
||||
),
|
||||
),
|
||||
decode_head=dict(
|
||||
num_classes=11,
|
||||
loss_decode=dict(
|
||||
type='DiceLoss',
|
||||
use_sigmoid=False,
|
||||
loss_weight=1.0,
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
optim_wrapper = dict(
|
||||
type='OptimWrapper',
|
||||
_delete_=True,
|
||||
optimizer=dict(
|
||||
type='AdamW',
|
||||
lr=0.0001,
|
||||
weight_decay=0.0005,
|
||||
),
|
||||
clip_grad=dict(
|
||||
max_norm=1,
|
||||
norm_type=2,
|
||||
),
|
||||
)
|
||||
|
||||
param_scheduler = [
|
||||
dict(
|
||||
type='LinearLR',
|
||||
start_factor=1e-06,
|
||||
by_epoch=True,
|
||||
begin=0,
|
||||
end=10,
|
||||
),
|
||||
dict(
|
||||
type='PolyLR',
|
||||
power=0.9,
|
||||
begin=10,
|
||||
end=300,
|
||||
eta_min=1e-05,
|
||||
by_epoch=True,
|
||||
),
|
||||
]
|
||||
|
||||
@@ -0,0 +1,94 @@
|
||||
_base_ = [
|
||||
'../_base_/models/stdc.py',
|
||||
'../_base_/datasets/publicdataset_endovis_2017.py',
|
||||
'../_base_/default_runtime.py',
|
||||
'../_base_/schedules/schedule_300e_val1_check10.py',
|
||||
]
|
||||
|
||||
norm_cfg = dict(
|
||||
type='BN',
|
||||
)
|
||||
|
||||
crop_size = (512, 512)
|
||||
|
||||
data_preprocessor = dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
122.21429912990676,
|
||||
77.0821859677977,
|
||||
87.03836664626716,
|
||||
],
|
||||
std=[
|
||||
50.53335800365262,
|
||||
42.895340354037465,
|
||||
47.739426483390446,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
)
|
||||
|
||||
model = dict(
|
||||
data_preprocessor=dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
122.21429912990676,
|
||||
77.0821859677977,
|
||||
87.03836664626716,
|
||||
],
|
||||
std=[
|
||||
50.53335800365262,
|
||||
42.895340354037465,
|
||||
47.739426483390446,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
),
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
stdc_type='STDCNet1',
|
||||
),
|
||||
init_cfg=dict(
|
||||
type='Pretrained',
|
||||
checkpoint='./My_Local_Model/open_mmlab/stdc1.pth',
|
||||
),
|
||||
),
|
||||
decode_head=dict(
|
||||
num_classes=8,
|
||||
loss_decode=dict(
|
||||
type='DiceLoss',
|
||||
use_sigmoid=False,
|
||||
loss_weight=1.0,
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
optim_wrapper = dict(
|
||||
type='OptimWrapper',
|
||||
_delete_=True,
|
||||
optimizer=dict(
|
||||
type='AdamW',
|
||||
lr=0.0001,
|
||||
weight_decay=0.0005,
|
||||
),
|
||||
clip_grad=dict(
|
||||
max_norm=1,
|
||||
norm_type=2,
|
||||
),
|
||||
)
|
||||
|
||||
param_scheduler = [
|
||||
dict(
|
||||
type='LinearLR',
|
||||
start_factor=1e-06,
|
||||
by_epoch=True,
|
||||
begin=0,
|
||||
end=10,
|
||||
),
|
||||
dict(
|
||||
type='PolyLR',
|
||||
power=0.9,
|
||||
begin=10,
|
||||
end=300,
|
||||
eta_min=1e-05,
|
||||
by_epoch=True,
|
||||
),
|
||||
]
|
||||
|
||||
@@ -0,0 +1,94 @@
|
||||
_base_ = [
|
||||
'../_base_/models/stdc.py',
|
||||
'../_base_/datasets/publicdataset_endovis_2018.py',
|
||||
'../_base_/default_runtime.py',
|
||||
'../_base_/schedules/schedule_300e_val1_check10.py',
|
||||
]
|
||||
|
||||
norm_cfg = dict(
|
||||
type='BN',
|
||||
)
|
||||
|
||||
crop_size = (512, 512)
|
||||
|
||||
data_preprocessor = dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
122.21429912990676,
|
||||
77.0821859677977,
|
||||
87.03836664626716,
|
||||
],
|
||||
std=[
|
||||
50.53335800365262,
|
||||
42.895340354037465,
|
||||
47.739426483390446,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
)
|
||||
|
||||
model = dict(
|
||||
data_preprocessor=dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
122.21429912990676,
|
||||
77.0821859677977,
|
||||
87.03836664626716,
|
||||
],
|
||||
std=[
|
||||
50.53335800365262,
|
||||
42.895340354037465,
|
||||
47.739426483390446,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
),
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
stdc_type='STDCNet1',
|
||||
),
|
||||
init_cfg=dict(
|
||||
type='Pretrained',
|
||||
checkpoint='./My_Local_Model/open_mmlab/stdc1.pth',
|
||||
),
|
||||
),
|
||||
decode_head=dict(
|
||||
num_classes=8,
|
||||
loss_decode=dict(
|
||||
type='DiceLoss',
|
||||
use_sigmoid=False,
|
||||
loss_weight=1.0,
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
optim_wrapper = dict(
|
||||
type='OptimWrapper',
|
||||
_delete_=True,
|
||||
optimizer=dict(
|
||||
type='AdamW',
|
||||
lr=0.0001,
|
||||
weight_decay=0.0005,
|
||||
),
|
||||
clip_grad=dict(
|
||||
max_norm=1,
|
||||
norm_type=2,
|
||||
),
|
||||
)
|
||||
|
||||
param_scheduler = [
|
||||
dict(
|
||||
type='LinearLR',
|
||||
start_factor=1e-06,
|
||||
by_epoch=True,
|
||||
begin=0,
|
||||
end=10,
|
||||
),
|
||||
dict(
|
||||
type='PolyLR',
|
||||
power=0.9,
|
||||
begin=10,
|
||||
end=300,
|
||||
eta_min=1e-05,
|
||||
by_epoch=True,
|
||||
),
|
||||
]
|
||||
|
||||
@@ -0,0 +1,94 @@
|
||||
_base_ = [
|
||||
'../_base_/models/stdc.py',
|
||||
'../_base_/datasets/publicdataset_autolaparo.py',
|
||||
'../_base_/default_runtime.py',
|
||||
'../_base_/schedules/schedule_300e_val1_check10.py',
|
||||
]
|
||||
|
||||
norm_cfg = dict(
|
||||
type='BN',
|
||||
)
|
||||
|
||||
crop_size = (512, 512)
|
||||
|
||||
data_preprocessor = dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
123.62464353460942,
|
||||
85.34836259209033,
|
||||
82.31539425671558,
|
||||
],
|
||||
std=[
|
||||
47.172211618459315,
|
||||
47.08256715323592,
|
||||
48.135121265163605,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
)
|
||||
|
||||
model = dict(
|
||||
data_preprocessor=dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
123.62464353460942,
|
||||
85.34836259209033,
|
||||
82.31539425671558,
|
||||
],
|
||||
std=[
|
||||
47.172211618459315,
|
||||
47.08256715323592,
|
||||
48.135121265163605,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
),
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
stdc_type='STDCNet2',
|
||||
),
|
||||
init_cfg=dict(
|
||||
type='Pretrained',
|
||||
checkpoint='./My_Local_Model/open_mmlab/stdc2.pth',
|
||||
),
|
||||
),
|
||||
decode_head=dict(
|
||||
num_classes=10,
|
||||
loss_decode=dict(
|
||||
type='DiceLoss',
|
||||
use_sigmoid=False,
|
||||
loss_weight=1.0,
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
optim_wrapper = dict(
|
||||
type='OptimWrapper',
|
||||
_delete_=True,
|
||||
optimizer=dict(
|
||||
type='AdamW',
|
||||
lr=0.0001,
|
||||
weight_decay=0.0005,
|
||||
),
|
||||
clip_grad=dict(
|
||||
max_norm=1,
|
||||
norm_type=2,
|
||||
),
|
||||
)
|
||||
|
||||
param_scheduler = [
|
||||
dict(
|
||||
type='LinearLR',
|
||||
start_factor=1e-06,
|
||||
by_epoch=True,
|
||||
begin=0,
|
||||
end=10,
|
||||
),
|
||||
dict(
|
||||
type='PolyLR',
|
||||
power=0.9,
|
||||
begin=10,
|
||||
end=300,
|
||||
eta_min=1e-05,
|
||||
by_epoch=True,
|
||||
),
|
||||
]
|
||||
|
||||
@@ -0,0 +1,94 @@
|
||||
_base_ = [
|
||||
'../_base_/models/stdc.py',
|
||||
'../_base_/datasets/publicdataset_cholecseg8k.py',
|
||||
'../_base_/default_runtime.py',
|
||||
'../_base_/schedules/schedule_300e_val1_check10.py',
|
||||
]
|
||||
|
||||
norm_cfg = dict(
|
||||
type='BN',
|
||||
)
|
||||
|
||||
crop_size = (512, 512)
|
||||
|
||||
data_preprocessor = dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
85.65740418979115,
|
||||
53.99282220050495,
|
||||
46.074045888534535,
|
||||
],
|
||||
std=[
|
||||
72.24589167201978,
|
||||
56.76979155397199,
|
||||
49.056637115061775,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
)
|
||||
|
||||
model = dict(
|
||||
data_preprocessor=dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
85.65740418979115,
|
||||
53.99282220050495,
|
||||
46.074045888534535,
|
||||
],
|
||||
std=[
|
||||
72.24589167201978,
|
||||
56.76979155397199,
|
||||
49.056637115061775,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
),
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
stdc_type='STDCNet2',
|
||||
),
|
||||
init_cfg=dict(
|
||||
type='Pretrained',
|
||||
checkpoint='./My_Local_Model/open_mmlab/stdc2.pth',
|
||||
),
|
||||
),
|
||||
decode_head=dict(
|
||||
num_classes=13,
|
||||
loss_decode=dict(
|
||||
type='DiceLoss',
|
||||
use_sigmoid=False,
|
||||
loss_weight=1.0,
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
optim_wrapper = dict(
|
||||
type='OptimWrapper',
|
||||
_delete_=True,
|
||||
optimizer=dict(
|
||||
type='AdamW',
|
||||
lr=0.0001,
|
||||
weight_decay=0.0005,
|
||||
),
|
||||
clip_grad=dict(
|
||||
max_norm=1,
|
||||
norm_type=2,
|
||||
),
|
||||
)
|
||||
|
||||
param_scheduler = [
|
||||
dict(
|
||||
type='LinearLR',
|
||||
start_factor=1e-06,
|
||||
by_epoch=True,
|
||||
begin=0,
|
||||
end=10,
|
||||
),
|
||||
dict(
|
||||
type='PolyLR',
|
||||
power=0.9,
|
||||
begin=10,
|
||||
end=300,
|
||||
eta_min=1e-05,
|
||||
by_epoch=True,
|
||||
),
|
||||
]
|
||||
|
||||
@@ -0,0 +1,94 @@
|
||||
_base_ = [
|
||||
'../_base_/models/stdc.py',
|
||||
'../_base_/datasets/publicdataset_dresden.py',
|
||||
'../_base_/default_runtime.py',
|
||||
'../_base_/schedules/schedule_300e_val1_check10.py',
|
||||
]
|
||||
|
||||
norm_cfg = dict(
|
||||
type='BN',
|
||||
)
|
||||
|
||||
crop_size = (512, 512)
|
||||
|
||||
data_preprocessor = dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
103.172638338208,
|
||||
61.44762740851152,
|
||||
51.407770213021976,
|
||||
],
|
||||
std=[
|
||||
75.77031253622098,
|
||||
54.63616729031377,
|
||||
49.45572239497569,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
)
|
||||
|
||||
model = dict(
|
||||
data_preprocessor=dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
103.172638338208,
|
||||
61.44762740851152,
|
||||
51.407770213021976,
|
||||
],
|
||||
std=[
|
||||
75.77031253622098,
|
||||
54.63616729031377,
|
||||
49.45572239497569,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
),
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
stdc_type='STDCNet2',
|
||||
),
|
||||
init_cfg=dict(
|
||||
type='Pretrained',
|
||||
checkpoint='./My_Local_Model/open_mmlab/stdc2.pth',
|
||||
),
|
||||
),
|
||||
decode_head=dict(
|
||||
num_classes=11,
|
||||
loss_decode=dict(
|
||||
type='DiceLoss',
|
||||
use_sigmoid=False,
|
||||
loss_weight=1.0,
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
optim_wrapper = dict(
|
||||
type='OptimWrapper',
|
||||
_delete_=True,
|
||||
optimizer=dict(
|
||||
type='AdamW',
|
||||
lr=0.0001,
|
||||
weight_decay=0.0005,
|
||||
),
|
||||
clip_grad=dict(
|
||||
max_norm=1,
|
||||
norm_type=2,
|
||||
),
|
||||
)
|
||||
|
||||
param_scheduler = [
|
||||
dict(
|
||||
type='LinearLR',
|
||||
start_factor=1e-06,
|
||||
by_epoch=True,
|
||||
begin=0,
|
||||
end=10,
|
||||
),
|
||||
dict(
|
||||
type='PolyLR',
|
||||
power=0.9,
|
||||
begin=10,
|
||||
end=300,
|
||||
eta_min=1e-05,
|
||||
by_epoch=True,
|
||||
),
|
||||
]
|
||||
|
||||
@@ -0,0 +1,94 @@
|
||||
_base_ = [
|
||||
'../_base_/models/stdc.py',
|
||||
'../_base_/datasets/publicdataset_endovis_2017.py',
|
||||
'../_base_/default_runtime.py',
|
||||
'../_base_/schedules/schedule_300e_val1_check10.py',
|
||||
]
|
||||
|
||||
norm_cfg = dict(
|
||||
type='BN',
|
||||
)
|
||||
|
||||
crop_size = (512, 512)
|
||||
|
||||
data_preprocessor = dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
122.21429912990676,
|
||||
77.0821859677977,
|
||||
87.03836664626716,
|
||||
],
|
||||
std=[
|
||||
50.53335800365262,
|
||||
42.895340354037465,
|
||||
47.739426483390446,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
)
|
||||
|
||||
model = dict(
|
||||
data_preprocessor=dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
122.21429912990676,
|
||||
77.0821859677977,
|
||||
87.03836664626716,
|
||||
],
|
||||
std=[
|
||||
50.53335800365262,
|
||||
42.895340354037465,
|
||||
47.739426483390446,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
),
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
stdc_type='STDCNet2',
|
||||
),
|
||||
init_cfg=dict(
|
||||
type='Pretrained',
|
||||
checkpoint='./My_Local_Model/open_mmlab/stdc2.pth',
|
||||
),
|
||||
),
|
||||
decode_head=dict(
|
||||
num_classes=8,
|
||||
loss_decode=dict(
|
||||
type='DiceLoss',
|
||||
use_sigmoid=False,
|
||||
loss_weight=1.0,
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
optim_wrapper = dict(
|
||||
type='OptimWrapper',
|
||||
_delete_=True,
|
||||
optimizer=dict(
|
||||
type='AdamW',
|
||||
lr=0.0001,
|
||||
weight_decay=0.0005,
|
||||
),
|
||||
clip_grad=dict(
|
||||
max_norm=1,
|
||||
norm_type=2,
|
||||
),
|
||||
)
|
||||
|
||||
param_scheduler = [
|
||||
dict(
|
||||
type='LinearLR',
|
||||
start_factor=1e-06,
|
||||
by_epoch=True,
|
||||
begin=0,
|
||||
end=10,
|
||||
),
|
||||
dict(
|
||||
type='PolyLR',
|
||||
power=0.9,
|
||||
begin=10,
|
||||
end=300,
|
||||
eta_min=1e-05,
|
||||
by_epoch=True,
|
||||
),
|
||||
]
|
||||
|
||||
@@ -0,0 +1,94 @@
|
||||
_base_ = [
|
||||
'../_base_/models/stdc.py',
|
||||
'../_base_/datasets/publicdataset_endovis_2018.py',
|
||||
'../_base_/default_runtime.py',
|
||||
'../_base_/schedules/schedule_300e_val1_check10.py',
|
||||
]
|
||||
|
||||
norm_cfg = dict(
|
||||
type='BN',
|
||||
)
|
||||
|
||||
crop_size = (512, 512)
|
||||
|
||||
data_preprocessor = dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
122.21429912990676,
|
||||
77.0821859677977,
|
||||
87.03836664626716,
|
||||
],
|
||||
std=[
|
||||
50.53335800365262,
|
||||
42.895340354037465,
|
||||
47.739426483390446,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
)
|
||||
|
||||
model = dict(
|
||||
data_preprocessor=dict(
|
||||
size=(512, 512),
|
||||
mean=[
|
||||
122.21429912990676,
|
||||
77.0821859677977,
|
||||
87.03836664626716,
|
||||
],
|
||||
std=[
|
||||
50.53335800365262,
|
||||
42.895340354037465,
|
||||
47.739426483390446,
|
||||
],
|
||||
bgr_to_rgb=False,
|
||||
),
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
stdc_type='STDCNet2',
|
||||
),
|
||||
init_cfg=dict(
|
||||
type='Pretrained',
|
||||
checkpoint='./My_Local_Model/open_mmlab/stdc2.pth',
|
||||
),
|
||||
),
|
||||
decode_head=dict(
|
||||
num_classes=8,
|
||||
loss_decode=dict(
|
||||
type='DiceLoss',
|
||||
use_sigmoid=False,
|
||||
loss_weight=1.0,
|
||||
),
|
||||
),
|
||||
)
|
||||
|
||||
optim_wrapper = dict(
|
||||
type='OptimWrapper',
|
||||
_delete_=True,
|
||||
optimizer=dict(
|
||||
type='AdamW',
|
||||
lr=0.0001,
|
||||
weight_decay=0.0005,
|
||||
),
|
||||
clip_grad=dict(
|
||||
max_norm=1,
|
||||
norm_type=2,
|
||||
),
|
||||
)
|
||||
|
||||
param_scheduler = [
|
||||
dict(
|
||||
type='LinearLR',
|
||||
start_factor=1e-06,
|
||||
by_epoch=True,
|
||||
begin=0,
|
||||
end=10,
|
||||
),
|
||||
dict(
|
||||
type='PolyLR',
|
||||
power=0.9,
|
||||
begin=10,
|
||||
end=300,
|
||||
eta_min=1e-05,
|
||||
by_epoch=True,
|
||||
),
|
||||
]
|
||||
|
||||
@@ -0,0 +1,21 @@
|
||||
_base_ = [
|
||||
'../_base_/models/stdc.py', '../_base_/datasets/cityscapes.py',
|
||||
'../_base_/default_runtime.py', '../_base_/schedules/schedule_80k.py'
|
||||
]
|
||||
crop_size = (512, 1024)
|
||||
data_preprocessor = dict(size=crop_size)
|
||||
model = dict(data_preprocessor=data_preprocessor)
|
||||
param_scheduler = [
|
||||
dict(type='LinearLR', by_epoch=False, start_factor=0.1, begin=0, end=1000),
|
||||
dict(
|
||||
type='PolyLR',
|
||||
eta_min=1e-4,
|
||||
power=0.9,
|
||||
begin=1000,
|
||||
end=80000,
|
||||
by_epoch=False,
|
||||
)
|
||||
]
|
||||
train_dataloader = dict(batch_size=12, num_workers=4)
|
||||
val_dataloader = dict(batch_size=1, num_workers=4)
|
||||
test_dataloader = val_dataloader
|
||||
@@ -0,0 +1,6 @@
|
||||
checkpoint = 'https://download.openmmlab.com/mmsegmentation/v0.5/pretrain/stdc/stdc1_20220308-5368626c.pth' # noqa
|
||||
_base_ = './stdc1_4xb12-80k_cityscapes-512x1024.py'
|
||||
model = dict(
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
init_cfg=dict(type='Pretrained', checkpoint=checkpoint))))
|
||||
@@ -0,0 +1,2 @@
|
||||
_base_ = './stdc1_4xb12-80k_cityscapes-512x1024.py'
|
||||
model = dict(backbone=dict(backbone_cfg=dict(stdc_type='STDCNet2')))
|
||||
@@ -0,0 +1,6 @@
|
||||
checkpoint = 'https://download.openmmlab.com/mmsegmentation/v0.5/pretrain/stdc/stdc2_20220308-7dbd9127.pth' # noqa
|
||||
_base_ = './stdc2_4xb12-80k_cityscapes-512x1024.py'
|
||||
model = dict(
|
||||
backbone=dict(
|
||||
backbone_cfg=dict(
|
||||
init_cfg=dict(type='Pretrained', checkpoint=checkpoint))))
|
||||
Reference in New Issue
Block a user