Skip to content
New issue

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

Sign up for GitHub

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

Already on GitHub? Sign in to your account

add swin transformer moe as backbone #209

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

add swin transformer moe as backbone #209

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
36 changes: 35 additions & 1 deletion README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -133,9 +133,43 @@ optimizer_config = dict(

> **Image Classification**: See [Swin Transformer for Image Classification](https://github.com/microsoft/Swin-Transformer).

> **Semantic Segmentation**: See [Swin Transformer for Semantic Segmentation](https://github.com/SwinTransformer/Swin-Transformer-Semantic-Segmentation).
> **Semantic Segmentation**: See [Swin Transformer for Semantic Segmentation](https://github.com/Transformer/Swin-Transformer-Semantic-Segmentation).

> **Self-Supervised Learning**: See [MoBY with Swin Transformer](https://github.com/SwinTransformer/Transformer-SSL).

> **Video Recognition**, See [Video Swin Transformer](https://github.com/SwinTransformer/Video-Swin-Transformer).

## swin-T moe

I added Swin Transformer MoE (referred to as Swin-T MoE hereafter) to the backbone network. MoE is a method that expands the model parameters and improves the model performance. The implementation of Swin Transformer MoE used Microsoft's Tutel framework.

### Install [Tutel](https://github.com/microsoft/tutel)

```
python3 -m pip uninstall tutel -y
python3 -m pip install --user --upgrade git+https://github.com/microsoft/tutel@main
```

You can check out Swin-T MoE at .

```
.\mmdet\models\backbones\swin_transformer_moe.py.
```

I provided the relevant configuration files for reference:

contains the parameters for the Swin-T MoE backbone network:

```
.\configs\swin\cascade_mask_rcnn_swin_moe_tiny_patch4_window7_mstrain_480-800_giou_4conv1f_adamw_3x_coco.py
```

contains the modified configuration for the backbone network:

```
.\configs\swin\cascade_mask_rcnn_swin_moe_tiny_patch4_window7_mstrain_480-800_giou_4conv1f_adamw_3x_coco.py
```

As the output of Swin-T MoE is different from Swin-T, I modified the `extract_feat` function in `.\mmdet\models\detectors\two_stage.py`.

You can change the config according to your needs.
216 changes: 216 additions & 0 deletions configs/_base_/models/cascade_mask_rcnn_swin_moe_fpn.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,216 @@
# model settings
model = dict(
type='CascadeRCNN',
pretrained=None,
backbone=dict(
type='SwinTransformerMoE',
embed_dim=96,
depths=[2, 2, 6, 2],
num_heads=[3, 6, 12, 24],
window_size=7,
mlp_ratio=4.,
qkv_bias=True,
qk_scale=None,
drop_rate=0.,
attn_drop_rate=0.,
drop_path_rate=0.2,
ape=False,
patch_norm=True,
out_indices=(0, 1, 2, 3),
use_checkpoint=False,
moe_blocks=[ [ -1 ], [ -1 ], [ 1, 3, 5 ], [ 1 ] ],
# moe_blocks=[ [ -1 ], [ -1 ], [ -1 ], [ -1 ] ],
num_local_experts=1,
top_value= 1,
capacity_factor= 1.25,
is_gshard_loss= False,
moe_drop= 0.1,
aux_loss_weight= 0.01
),
neck=dict(
type='FPN',
in_channels=[96, 192, 384, 768],
out_channels=256,
num_outs=5),
rpn_head=dict(
type='RPNHead',
in_channels=256,
feat_channels=256,
anchor_generator=dict(
type='AnchorGenerator',
scales=[8],
ratios=[0.5, 1.0, 2.0],
strides=[4, 8, 16, 32, 64]),
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[.0, .0, .0, .0],
target_stds=[1.0, 1.0, 1.0, 1.0]),
loss_cls=dict(
type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0 / 9.0, loss_weight=1.0)),
roi_head=dict(
type='CascadeRoIHead',
num_stages=3,
stage_loss_weights=[1, 0.5, 0.25],
bbox_roi_extractor=dict(
type='SingleRoIExtractor',
roi_layer=dict(type='RoIAlign', output_size=7, sampling_ratio=0),
out_channels=256,
featmap_strides=[4, 8, 16, 32]),
bbox_head=[
dict(
type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
num_classes=80,
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[0., 0., 0., 0.],
target_stds=[0.1, 0.1, 0.2, 0.2]),
reg_class_agnostic=True,
loss_cls=dict(
type='CrossEntropyLoss',
use_sigmoid=False,
loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0,
loss_weight=1.0)),
dict(
type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
num_classes=80,
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[0., 0., 0., 0.],
target_stds=[0.05, 0.05, 0.1, 0.1]),
reg_class_agnostic=True,
loss_cls=dict(
type='CrossEntropyLoss',
use_sigmoid=False,
loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0,
loss_weight=1.0)),
dict(
type='Shared2FCBBoxHead',
in_channels=256,
fc_out_channels=1024,
roi_feat_size=7,
num_classes=80,
bbox_coder=dict(
type='DeltaXYWHBBoxCoder',
target_means=[0., 0., 0., 0.],
target_stds=[0.033, 0.033, 0.067, 0.067]),
reg_class_agnostic=True,
loss_cls=dict(
type='CrossEntropyLoss',
use_sigmoid=False,
loss_weight=1.0),
loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0))
],
mask_roi_extractor=dict(
type='SingleRoIExtractor',
roi_layer=dict(type='RoIAlign', output_size=14, sampling_ratio=0),
out_channels=256,
featmap_strides=[4, 8, 16, 32]),
mask_head=dict(
type='FCNMaskHead',
num_convs=4,
in_channels=256,
conv_out_channels=256,
num_classes=80,
loss_mask=dict(
type='CrossEntropyLoss', use_mask=True, loss_weight=1.0))),
# model training and testing settings
train_cfg = dict(
rpn=dict(
assigner=dict(
type='MaxIoUAssigner',
pos_iou_thr=0.7,
neg_iou_thr=0.3,
min_pos_iou=0.3,
match_low_quality=True,
ignore_iof_thr=-1),
sampler=dict(
type='RandomSampler',
num=256,
pos_fraction=0.5,
neg_pos_ub=-1,
add_gt_as_proposals=False),
allowed_border=0,
pos_weight=-1,
debug=False),
rpn_proposal=dict(
nms_across_levels=False,
nms_pre=2000,
nms_post=2000,
max_per_img=2000,
nms=dict(type='nms', iou_threshold=0.7),
min_bbox_size=0),
rcnn=[
dict(
assigner=dict(
type='MaxIoUAssigner',
pos_iou_thr=0.5,
neg_iou_thr=0.5,
min_pos_iou=0.5,
match_low_quality=False,
ignore_iof_thr=-1),
sampler=dict(
type='RandomSampler',
num=512,
pos_fraction=0.25,
neg_pos_ub=-1,
add_gt_as_proposals=True),
mask_size=28,
pos_weight=-1,
debug=False),
dict(
assigner=dict(
type='MaxIoUAssigner',
pos_iou_thr=0.6,
neg_iou_thr=0.6,
min_pos_iou=0.6,
match_low_quality=False,
ignore_iof_thr=-1),
sampler=dict(
type='RandomSampler',
num=512,
pos_fraction=0.25,
neg_pos_ub=-1,
add_gt_as_proposals=True),
mask_size=28,
pos_weight=-1,
debug=False),
dict(
assigner=dict(
type='MaxIoUAssigner',
pos_iou_thr=0.7,
neg_iou_thr=0.7,
min_pos_iou=0.7,
match_low_quality=False,
ignore_iof_thr=-1),
sampler=dict(
type='RandomSampler',
num=512,
pos_fraction=0.25,
neg_pos_ub=-1,
add_gt_as_proposals=True),
mask_size=28,
pos_weight=-1,
debug=False)
]),
test_cfg = dict(
rpn=dict(
nms_across_levels=False,
nms_pre=1000,
nms_post=1000,
max_per_img=1000,
nms=dict(type='nms', iou_threshold=0.7),
min_bbox_size=0),
rcnn=dict(
score_thr=0.05,
nms=dict(type='nms', iou_threshold=0.5),
max_per_img=100,
mask_thr_binary=0.5)))
Loading