Skip to content
New issue

Have a question about this project? # for a free GitHub account to open an issue and contact its maintainers and the community.

#

By clicking “#”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? # to your account

Fix dcff related issue [464] & [469]，add url for model & log #473

Open
wants to merge 6 commits into
base: main
Choose a base branch
from
Open

Fix dcff related issue [464] & [469]，add url for model & log #473

Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
6c3ff54
fix issue 464&469
Lxtccc Mar 7, 2023
107334c
add data url
Lxtccc Mar 7, 2023
814d24d
update
Lxtccc Mar 7, 2023
48038d2
fix ci
Lxtccc Mar 7, 2023
0fe347f
fix conflict
Lxtccc Mar 16, 2023
ec87694
update readme
Lxtccc Mar 16, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
32 changes: 14 additions & 18 deletions configs/pruning/mmcls/dcff/README.md
View file
Original file line number Diff line number Diff line change
@@ -10,27 +10,27 @@ The mainstream approach for filter pruning is usually either to force a hard-cod

### 1. Classification

| Dataset | Backbone | Params(M) | FLOPs(M) | lr_type | Top-1 (%) | Top-5 (%) | CPrate | Config | Download |
| :------: | :----------: | :-------: | :------: | :-----: | :-------: | :-------: | :---------------------------------------------: | :--------------------------------------------------: | :--------------------------: |
| ImageNet | DCFFResNet50 | 15.16 | 2260 | step | 73.96 | 91.66 | \[0.0\]+\[0.35,0.4,0.1\]\*10+\[0.3,0.3,0.1\]\*6 | [config](../../mmcls/dcff/dcff_resnet_8xb32_in1k.py) | [model](<>) \| \[log\] (\<>) |
| Dataset | Backbone | Params(M) | FLOPs(M) | lr_type | Top-1 (%) | Top-5 (%) | CPrate | Config | Download |
| :------: | :----------: | :-------: | :------: | :-----: | :-------: | :-------: | :---------------------------------------------: | :--------------------------------------------------: | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
| ImageNet | DCFFResNet50 | 13.32 | 2255 | step | 74.46 | 91.99 | \[0.0\]+\[0.35,0.4,0.1\]\*10+\[0.3,0.3,0.1\]\*6 | [config](../../mmcls/dcff/dcff_resnet_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmcls/dcff_cls_fix_subnet_weight.pth) \\ [log](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmcls/dcff_mmcls_sup_20220906_131949.log) |

### 2. Detection

| Dataset | Method | Backbone | Style | Lr schd | Params(M) | FLOPs(M) | bbox AP | CPrate | Config | Download |
| :-----: | :---------: | :----------: | :-----: | :-----: | :-------: | :------: | :-----: | :---------------------------------------------: | :---------------------------------------------------------------: | :--------------------------: |
| COCO | Faster_RCNN | DCFFResNet50 | pytorch | step | 33.31 | 168320 | 35.8 | \[0.0\]+\[0.35,0.4,0.1\]\*10+\[0.3,0.3,0.1\]\*6 | [config](../../mmdet/dcff/dcff_faster_rcnn_resnet50_8xb4_coco.py) | [model](<>) \| \[log\] (\<>) |
| Dataset | Method | Backbone | Style | Lr schd | Params(M) | FLOPs(M) | bbox AP | CPrate | Config | Download |
| :-----: | :---------: | :----------: | :-----: | :-----: | :-------: | :------: | :-----: | :---------------------------------------------: | :---------------------------------------------------------------: | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
| COCO | Faster_RCNN | DCFFResNet50 | pytorch | step | 33.31 | 168320 | 35.8 | \[0.0\]+\[0.35,0.4,0.1\]\*10+\[0.3,0.3,0.1\]\*6 | [config](../../mmdet/dcff/dcff_faster_rcnn_resnet50_8xb4_coco.py) | [model](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmdet/dcff_mmdet.pth) \\ [log](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmdet/dcff_mmdet_sup_20220909_103653.log) |

### 3. Segmentation

| Dataset | Method | Backbone | crop size | Lr schd | Params(M) | FLOPs(M) | mIoU | CPrate | Config | Download |
| :--------: | :-------: | :-------------: | :-------: | :-----: | :-------: | :------: | :---: | :-----------------------------------------------------------------: | :-------------------------------------------------------------------: | :--------------------------: |
| Cityscapes | PointRend | DCFFResNetV1c50 | 512x1024 | 160k | 18.43 | 74410 | 76.75 | \[0.0, 0.0, 0.0\] + \[0.35, 0.4, 0.1\] * 10 + \[0.3, 0.3, 0.1\] * 6 | [config](../../mmseg/dcff/dcff_pointrend_resnet50_8xb2_cityscapes.py) | [model](<>) \| \[log\] (\<>) |
| Dataset | Method | Backbone | crop size | Lr schd | Params(M) | FLOPs(M) | mIoU | CPrate | Config | Download |
| :--------: | :-------: | :-------------: | :-------: | :-----: | :-------: | :------: | :---: | :-----------------------------------------------------------------: | :-------------------------------------------------------------------: | :---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
| Cityscapes | PointRend | DCFFResNetV1c50 | 512x1024 | 160k | 18.43 | 74410 | 76.75 | \[0.0, 0.0, 0.0\] + \[0.35, 0.4, 0.1\] * 10 + \[0.3, 0.3, 0.1\] * 6 | [config](../../mmseg/dcff/dcff_pointrend_resnet50_8xb2_cityscapes.py) | [model](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmseg/dcff_mmseg.pth) \\ [log](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmseg/dcff_mmpose_sup_20220908_172111.log) |

### 4. Pose

| Dataset | Method | Backbone | crop size | total epochs | Params(M) | FLOPs(M) | AP | CPrate | Config | Download |
| :-----: | :-------------: | :----------: | :-------: | :----------: | :-------: | :------: | :--: | :--------------------------------------------------------: | :---------------------------------------------------------------: | :--------------------------: |
| COCO | TopDown HeatMap | DCFFResNet50 | 256x192 | 300 | 26.95 | 4290 | 68.3 | \[0.0\] + \[0.2, 0.2, 0.1\] * 10 + \[0.15, 0.15, 0.1\] * 6 | [config](../../mmpose/dcff/dcff_topdown_heatmap_resnet50_coco.py) | [model](<>) \| \[log\] (\<>) |
| Dataset | Method | Backbone | crop size | total epochs | Params(M) | FLOPs(M) | AP | CPrate | Config | Download |
| :-----: | :-------------: | :----------: | :-------: | :----------: | :-------: | :------: | :--: | :--------------------------------------------------------: | :---------------------------------------------------------------: | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
| COCO | TopDown HeatMap | DCFFResNet50 | 256x192 | 300 | 26.95 | 4290 | 68.3 | \[0.0\] + \[0.2, 0.2, 0.1\] * 10 + \[0.15, 0.15, 0.1\] * 6 | [config](../../mmpose/dcff/dcff_topdown_heatmap_resnet50_coco.py) | [model](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmpose/dcff_pose.pth) \\ [log](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmpose/dcff_mmpose_20220908_140331.log) |

## Citation

@@ -59,24 +59,20 @@ Then set layers' pruning rates `target_pruning_ratio` by `resnet_cls.json`.

### Train DCFF

#### Classification

##### ImageNet

```bash
CUDA_VISIBLE_DEVICES=0,1,2,3 PORT=29500 ./tools/dist_train.sh \
configs/pruning/mmcls/dcff/dcff_resnet50_8xb32_in1k.py 4 \
configs/pruning/mmcls/dcff/dcff_resnet_8xb32_in1k.py 4 \
--work-dir $WORK_DIR
```

### Test DCFF

#### Classification

##### ImageNet

```bash
CUDA_VISIBLE_DEVICES=0,1,2,3 PORT=29500 ./tools/dist_test.sh \
configs/pruning/mmcls/dcff/dcff_compact_resnet50_8xb32_in1k.py \
$CKPT 1 --work-dir $WORK_DIR
$WORK_DIR/fix_subnet_weight.pth 1 --work-dir $WORK_DIR
```
6 changes: 4 additions & 2 deletions configs/pruning/mmcls/dcff/dcff_compact_resnet_8xb32_in1k.py
View file
Original file line number Diff line number Diff line change
@@ -4,10 +4,12 @@
_base_.model = dict(
_scope_='mmrazor',
type='sub_model',
cfg=dict(
cfg_path='mmcls::resnet/resnet50_8xb32_in1k.py', pretrained=False),
cfg=_base_.architecture,
fix_subnet='configs/pruning/mmcls/dcff/fix_subnet.json',
mode='mutator',
init_cfg=dict(
type='Pretrained',
prefix='architecture',
checkpoint='configs/pruning/mmcls/dcff/fix_subnet_weight.pth'))

_base_.val_cfg = dict()
30 changes: 13 additions & 17 deletions configs/pruning/mmdet/dcff/README.md
View file
Original file line number Diff line number Diff line change
@@ -10,27 +10,27 @@ The mainstream approach for filter pruning is usually either to force a hard-cod

### 1. Classification

| Dataset | Backbone | Params(M) | FLOPs(M) | lr_type | Top-1 (%) | Top-5 (%) | CPrate | Config | Download |
| :------: | :----------: | :-------: | :------: | :-----: | :-------: | :-------: | :---------------------------------------------: | :--------------------------------------------------: | :--------------------------: |
| ImageNet | DCFFResNet50 | 15.16 | 2260 | step | 73.96 | 91.66 | \[0.0\]+\[0.35,0.4,0.1\]\*10+\[0.3,0.3,0.1\]\*6 | [config](../../mmcls/dcff/dcff_resnet_8xb32_in1k.py) | [model](<>) \| \[log\] (\<>) |
| Dataset | Backbone | Params(M) | FLOPs(M) | lr_type | Top-1 (%) | Top-5 (%) | CPrate | Config | Download |
| :------: | :----------: | :-------: | :------: | :-----: | :-------: | :-------: | :---------------------------------------------: | :--------------------------------------------------: | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
| ImageNet | DCFFResNet50 | 13.32 | 2255 | step | 74.46 | 91.99 | \[0.0\]+\[0.35,0.4,0.1\]\*10+\[0.3,0.3,0.1\]\*6 | [config](../../mmcls/dcff/dcff_resnet_8xb32_in1k.py) | [model](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmcls/dcff_cls_fix_subnet_weight.pth) \\ [log](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmcls/dcff_mmcls_sup_20220906_131949.log) |

### 2. Detection

| Dataset | Method | Backbone | Style | Lr schd | Params(M) | FLOPs(M) | bbox AP | CPrate | Config | Download |
| :-----: | :---------: | :----------: | :-----: | :-----: | :-------: | :------: | :-----: | :---------------------------------------------: | :---------------------------------------------------------------: | :--------------------------: |
| COCO | Faster_RCNN | DCFFResNet50 | pytorch | step | 33.31 | 168320 | 35.8 | \[0.0\]+\[0.35,0.4,0.1\]\*10+\[0.3,0.3,0.1\]\*6 | [config](../../mmdet/dcff/dcff_faster_rcnn_resnet50_8xb4_coco.py) | [model](<>) \| \[log\] (\<>) |
| Dataset | Method | Backbone | Style | Lr schd | Params(M) | FLOPs(M) | bbox AP | CPrate | Config | Download |
| :-----: | :---------: | :----------: | :-----: | :-----: | :-------: | :------: | :-----: | :---------------------------------------------: | :---------------------------------------------------------------: | :--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
| COCO | Faster_RCNN | DCFFResNet50 | pytorch | step | 33.31 | 168320 | 35.8 | \[0.0\]+\[0.35,0.4,0.1\]\*10+\[0.3,0.3,0.1\]\*6 | [config](../../mmdet/dcff/dcff_faster_rcnn_resnet50_8xb4_coco.py) | [model](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmdet/dcff_mmdet.pth) \\ [log](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmdet/dcff_mmdet_sup_20220909_103653.log) |

### 3. Segmentation

| Dataset | Method | Backbone | crop size | Lr schd | Params(M) | FLOPs(M) | mIoU | CPrate | Config | Download |
| :--------: | :-------: | :-------------: | :-------: | :-----: | :-------: | :------: | :---: | :-----------------------------------------------------------------: | :-------------------------------------------------------------------: | :--------------------------: |
| Cityscapes | PointRend | DCFFResNetV1c50 | 512x1024 | 160k | 18.43 | 74410 | 76.75 | \[0.0, 0.0, 0.0\] + \[0.35, 0.4, 0.1\] * 10 + \[0.3, 0.3, 0.1\] * 6 | [config](../../mmseg/dcff/dcff_pointrend_resnet50_8xb2_cityscapes.py) | [model](<>) \| \[log\] (\<>) |
| Dataset | Method | Backbone | crop size | Lr schd | Params(M) | FLOPs(M) | mIoU | CPrate | Config | Download |
| :--------: | :-------: | :-------------: | :-------: | :-----: | :-------: | :------: | :---: | :-----------------------------------------------------------------: | :-------------------------------------------------------------------: | :---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
| Cityscapes | PointRend | DCFFResNetV1c50 | 512x1024 | 160k | 18.43 | 74410 | 76.75 | \[0.0, 0.0, 0.0\] + \[0.35, 0.4, 0.1\] * 10 + \[0.3, 0.3, 0.1\] * 6 | [config](../../mmseg/dcff/dcff_pointrend_resnet50_8xb2_cityscapes.py) | [model](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmseg/dcff_mmseg.pth) \\ [log](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmseg/dcff_mmpose_sup_20220908_172111.log) |

### 4. Pose

| Dataset | Method | Backbone | crop size | total epochs | Params(M) | FLOPs(M) | AP | CPrate | Config | Download |
| :-----: | :-------------: | :----------: | :-------: | :----------: | :-------: | :------: | :--: | :--------------------------------------------------------: | :---------------------------------------------------------------: | :--------------------------: |
| COCO | TopDown HeatMap | DCFFResNet50 | 256x192 | 300 | 26.95 | 4290 | 68.3 | \[0.0\] + \[0.2, 0.2, 0.1\] * 10 + \[0.15, 0.15, 0.1\] * 6 | [config](../../mmpose/dcff/dcff_topdown_heatmap_resnet50_coco.py) | [model](<>) \| \[log\] (\<>) |
| Dataset | Method | Backbone | crop size | total epochs | Params(M) | FLOPs(M) | AP | CPrate | Config | Download |
| :-----: | :-------------: | :----------: | :-------: | :----------: | :-------: | :------: | :--: | :--------------------------------------------------------: | :---------------------------------------------------------------: | :------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: |
| COCO | TopDown HeatMap | DCFFResNet50 | 256x192 | 300 | 26.95 | 4290 | 68.3 | \[0.0\] + \[0.2, 0.2, 0.1\] * 10 + \[0.15, 0.15, 0.1\] * 6 | [config](../../mmpose/dcff/dcff_topdown_heatmap_resnet50_coco.py) | [model](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmpose/dcff_pose.pth) \\ [log](https://download.openmmlab.com/mmrazor/v1/pruning/dcff/mmpose/dcff_mmpose_20220908_140331.log) |

## Citation

@@ -59,8 +59,6 @@ Then set layers' pruning rates `target_pruning_ratio` by `resnet_det.json`.

### Train DCFF

#### Detection

##### COCO

```bash
@@ -71,12 +69,10 @@ CUDA_VISIBLE_DEVICES=0,1,2,3 PORT=29500 ./tools/dist_train.sh \

### Test DCFF

#### Detection

##### COCO

```bash
CUDA_VISIBLE_DEVICES=0 PORT=29500 ./tools/dist_test.sh \
configs/pruning/mmdet/dcff/dcff_compact_faster_rcnn_resnet50_8xb4_coco.py \
$CKPT 1 --work-dir $WORK_DIR
$WORK_DIR/fix_subnet_weight.pth 1 --work-dir $WORK_DIR
```
3 changes: 3 additions & 0 deletions configs/pruning/mmdet/dcff/dcff_compact_faster_rcnn_resnet50_8xb4_coco.py
View file
Original file line number Diff line number Diff line change
@@ -9,4 +9,7 @@
mode='mutator',
init_cfg=dict(
type='Pretrained',
prefix='architecture',
checkpoint='configs/pruning/mmdet/dcff/fix_subnet_weight.pth'))

_base_.val_cfg = dict()
Loading