-
Notifications
You must be signed in to change notification settings - Fork 9.5k
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
1 parent
af816d3
commit 82d2a6e
Showing
20 changed files
with
864 additions
and
97 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
14 changes: 14 additions & 0 deletions
14
configs/glip/glip_atss_swin-l_fpn_dyhead_16xb2_ms-2x_funtune_coco.py
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,14 @@ | ||
| _base_ = './glip_atss_swin-t_b_fpn_dyhead_16xb2_ms-2x_funtune_coco.py' | ||
|
|
||
| model = dict( | ||
| backbone=dict( | ||
| embed_dims=192, | ||
| depths=[2, 2, 18, 2], | ||
| num_heads=[6, 12, 24, 48], | ||
| window_size=12, | ||
| drop_path_rate=0.4, | ||
| ), | ||
| neck=dict(in_channels=[384, 768, 1536]), | ||
| bbox_head=dict(early_fuse=True, num_dyhead_blocks=8, use_checkpoint=True)) | ||
|
|
||
| load_from = 'https://download.openmmlab.com/mmdetection/v3.0/glip/glip_l_mmdet-abfe026b.pth' # noqa |
155 changes: 155 additions & 0 deletions
155
configs/glip/glip_atss_swin-t_a_fpn_dyhead_16xb2_ms-2x_funtune_coco.py
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,155 @@ | ||
| _base_ = [ | ||
| '../_base_/datasets/coco_detection.py', | ||
| '../_base_/schedules/schedule_1x.py', '../_base_/default_runtime.py' | ||
| ] | ||
| load_from = 'https://download.openmmlab.com/mmdetection/v3.0/glip/glip_tiny_a_mmdet-b3654169.pth' # noqa | ||
| lang_model_name = 'bert-base-uncased' | ||
|
|
||
| model = dict( | ||
| type='GLIP', | ||
| data_preprocessor=dict( | ||
| type='DetDataPreprocessor', | ||
| mean=[103.53, 116.28, 123.675], | ||
| std=[57.375, 57.12, 58.395], | ||
| bgr_to_rgb=False, | ||
| pad_size_divisor=32), | ||
| backbone=dict( | ||
| type='SwinTransformer', | ||
| embed_dims=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, | ||
| patch_norm=True, | ||
| out_indices=(1, 2, 3), | ||
| with_cp=False, | ||
| convert_weights=False), | ||
| neck=dict( | ||
| type='FPN_DropBlock', | ||
| in_channels=[192, 384, 768], | ||
| out_channels=256, | ||
| start_level=0, | ||
| relu_before_extra_convs=True, | ||
| add_extra_convs='on_output', | ||
| num_outs=5), | ||
| bbox_head=dict( | ||
| type='ATSSVLFusionHead', | ||
| lang_model_name=lang_model_name, | ||
| num_classes=80, | ||
| in_channels=256, | ||
| feat_channels=256, | ||
| anchor_generator=dict( | ||
| type='AnchorGenerator', | ||
| ratios=[1.0], | ||
| octave_base_scale=8, | ||
| scales_per_octave=1, | ||
| strides=[8, 16, 32, 64, 128], | ||
| center_offset=0.5), | ||
| bbox_coder=dict( | ||
| type='DeltaXYWHBBoxCoderForGLIP', | ||
| target_means=[.0, .0, .0, .0], | ||
| target_stds=[0.1, 0.1, 0.2, 0.2]), | ||
| loss_cls=dict( | ||
| type='FocalLoss', | ||
| use_sigmoid=True, | ||
| gamma=2.0, | ||
| alpha=0.25, | ||
| loss_weight=1.0), | ||
| loss_bbox=dict(type='GIoULoss', loss_weight=2.0), | ||
| loss_centerness=dict( | ||
| type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0)), | ||
| language_model=dict(type='BertModel', name=lang_model_name), | ||
| train_cfg=dict( | ||
| assigner=dict( | ||
| type='ATSSAssigner', | ||
| topk=9, | ||
| iou_calculator=dict(type='BboxOverlaps2D_GLIP')), | ||
| allowed_border=-1, | ||
| pos_weight=-1, | ||
| debug=False), | ||
| test_cfg=dict( | ||
| nms_pre=1000, | ||
| min_bbox_size=0, | ||
| score_thr=0.05, | ||
| nms=dict(type='nms', iou_threshold=0.6), | ||
| max_per_img=100)) | ||
|
|
||
| # dataset settings | ||
| train_pipeline = [ | ||
| dict( | ||
| type='LoadImageFromFile', | ||
| imdecode_backend='pillow', | ||
| backend_args=_base_.backend_args), | ||
| dict(type='LoadAnnotations', with_bbox=True), | ||
| dict(type='GTBoxSubOne_GLIP'), | ||
| dict( | ||
| type='RandomChoiceResize', | ||
| scales=[(1333, 480), (1333, 560), (1333, 640), (1333, 720), | ||
| (1333, 800)], | ||
| keep_ratio=True, | ||
| resize_type='FixScaleResize', | ||
| backend='pillow'), | ||
| dict(type='RandomFlip_GLIP', prob=0.5), | ||
| dict(type='FilterAnnotations', min_gt_bbox_wh=(1, 1)), | ||
| dict( | ||
| type='PackDetInputs', | ||
| meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape', | ||
| 'scale_factor', 'flip', 'flip_direction', 'text', | ||
| 'custom_entities')) | ||
| ] | ||
|
|
||
| test_pipeline = [ | ||
| dict( | ||
| type='LoadImageFromFile', | ||
| backend_args=_base_.backend_args, | ||
| imdecode_backend='pillow'), | ||
| dict( | ||
| type='FixScaleResize', | ||
| scale=(800, 1333), | ||
| keep_ratio=True, | ||
| backend='pillow'), | ||
| dict(type='LoadAnnotations', with_bbox=True), | ||
| dict( | ||
| type='PackDetInputs', | ||
| meta_keys=('img_id', 'img_path', 'ori_shape', 'img_shape', | ||
| 'scale_factor', 'text', 'custom_entities')) | ||
| ] | ||
|
|
||
| train_dataloader = dict( | ||
| dataset=dict( | ||
| _delete_=True, | ||
| type='RepeatDataset', | ||
| times=2, | ||
| dataset=dict( | ||
| type=_base_.dataset_type, | ||
| data_root=_base_.data_root, | ||
| ann_file='annotations/instances_train2017.json', | ||
| data_prefix=dict(img='train2017/'), | ||
| filter_cfg=dict(filter_empty_gt=True, min_size=32), | ||
| pipeline=train_pipeline, | ||
| return_classes=True, | ||
| backend_args=_base_.backend_args))) | ||
|
|
||
| val_dataloader = dict( | ||
| dataset=dict(pipeline=test_pipeline, return_classes=True)) | ||
| test_dataloader = val_dataloader | ||
|
|
||
| # We did not adopt the official 24e optimizer strategy | ||
| # because the results indicate that the current strategy is superior. | ||
| optim_wrapper = dict( | ||
| _delete_=True, | ||
| type='OptimWrapper', | ||
| optimizer=dict( | ||
| type='AdamW', lr=0.00002, betas=(0.9, 0.999), weight_decay=0.05), | ||
| paramwise_cfg=dict( | ||
| custom_keys={ | ||
| 'absolute_pos_embed': dict(decay_mult=0.), | ||
| 'relative_position_bias_table': dict(decay_mult=0.), | ||
| 'norm': dict(decay_mult=0.) | ||
| }), | ||
| clip_grad=None) |
9 changes: 9 additions & 0 deletions
9
configs/glip/glip_atss_swin-t_b_fpn_dyhead_16xb2_ms-2x_funtune_coco.py
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,9 @@ | ||
| _base_ = './glip_atss_swin-t_a_fpn_dyhead_16xb2_ms-2x_funtune_coco.py' | ||
|
|
||
| model = dict(bbox_head=dict(early_fuse=True, use_checkpoint=True)) | ||
|
|
||
| load_from = 'https://download.openmmlab.com/mmdetection/v3.0/glip/glip_tiny_b_mmdet-6dfbd102.pth' # noqa | ||
|
|
||
| optim_wrapper = dict( | ||
| optimizer=dict(lr=0.00001), | ||
| clip_grad=dict(_delete_=True, max_norm=1, norm_type=2)) |
3 changes: 3 additions & 0 deletions
3
configs/glip/glip_atss_swin-t_c_fpn_dyhead_16xb2_ms-2x_funtune_coco.py
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,3 @@ | ||
| _base_ = './glip_atss_swin-t_b_fpn_dyhead_16xb2_ms-2x_funtune_coco.py' | ||
|
|
||
| load_from = 'https://download.openmmlab.com/mmdetection/v3.0/glip/glip_tiny_c_mmdet-2fc427dd.pth' # noqa |
3 changes: 3 additions & 0 deletions
3
configs/glip/glip_atss_swin-t_fpn_dyhead_16xb2_ms-2x_funtune_coco.py
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,3 @@ | ||
| _base_ = './glip_atss_swin-t_b_fpn_dyhead_16xb2_ms-2x_funtune_coco.py' | ||
|
|
||
| load_from = 'https://download.openmmlab.com/mmdetection/v3.0/glip/glip_tiny_mmdet-c24ce662.pth' # noqa |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.