Skip to content

GitLab

Commit ca8a762a authored by chenzk's avatar chenzk
Browse files

v1.0

parents
Hide whitespace changes
Inline Side-by-side
Showing with 2062 additions and 0 deletions
configs/animal_2d_keypoint/topdown_heatmap/ak/td-hm_hrnet-w32_8xb32-300e_animalkingdom_P2-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
train_cfg = dict(max_epochs=300, val_interval=10)
# optimizer
optim_wrapper = dict(optimizer=dict(
type='AdamW',
lr=5e-4,
))
# learning policy
param_scheduler = [
dict(
type='LinearLR', begin=0, end=500, start_factor=0.001,
by_epoch=False), # warm-up
dict(
type='MultiStepLR',
begin=0,
end=210,
milestones=[170, 200],
gamma=0.1,
by_epoch=True)
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=512)
# hooks
default_hooks = dict(checkpoint=dict(save_best='PCK', rule='greater'))
# codec settings
codec = dict(
type='MSRAHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
type='HRNet',
in_channels=3,
extra=dict(
stage1=dict(
num_modules=1,
num_branches=1,
block='BOTTLENECK',
num_blocks=(4, ),
num_channels=(64, )),
stage2=dict(
num_modules=1,
num_branches=2,
block='BASIC',
num_blocks=(4, 4),
num_channels=(32, 64)),
stage3=dict(
num_modules=4,
num_branches=3,
block='BASIC',
num_blocks=(4, 4, 4),
num_channels=(32, 64, 128)),
stage4=dict(
num_modules=3,
num_branches=4,
block='BASIC',
num_blocks=(4, 4, 4, 4),
num_channels=(32, 64, 128, 256))),
init_cfg=dict(
type='Pretrained',
checkpoint='https://download.openmmlab.com/mmpose/'
'pretrain_models/hrnet_w32-36af842e.pth'),
),
head=dict(
type='HeatmapHead',
in_channels=32,
out_channels=23,
deconv_out_channels=None,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=True,
))
# base dataset settings
dataset_type = 'AnimalKingdomDataset'
data_mode = 'topdown'
data_root = 'data/ak/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(type='RandomBBoxTransform'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P2/train.json',
data_prefix=dict(img='images/'),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=24,
num_workers=2,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P2/test.json',
data_prefix=dict(img='images/'),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = val_dataloader
# evaluators
val_evaluator = [dict(type='PCKAccuracy', thr=0.05), dict(type='AUC')]
test_evaluator = val_evaluator
configs/animal_2d_keypoint/topdown_heatmap/ak/td-hm_hrnet-w32_8xb32-300e_animalkingdom_P3_amphibian-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
train_cfg = dict(max_epochs=300, val_interval=10)
# optimizer
optim_wrapper = dict(optimizer=dict(
type='AdamW',
lr=5e-4,
))
# learning policy
param_scheduler = [
dict(
type='LinearLR', begin=0, end=500, start_factor=0.001,
by_epoch=False), # warm-up
dict(
type='MultiStepLR',
begin=0,
end=210,
milestones=[170, 200],
gamma=0.1,
by_epoch=True)
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=512)
# hooks
default_hooks = dict(checkpoint=dict(save_best='PCK', rule='greater'))
# codec settings
codec = dict(
type='MSRAHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
type='HRNet',
in_channels=3,
extra=dict(
stage1=dict(
num_modules=1,
num_branches=1,
block='BOTTLENECK',
num_blocks=(4, ),
num_channels=(64, )),
stage2=dict(
num_modules=1,
num_branches=2,
block='BASIC',
num_blocks=(4, 4),
num_channels=(32, 64)),
stage3=dict(
num_modules=4,
num_branches=3,
block='BASIC',
num_blocks=(4, 4, 4),
num_channels=(32, 64, 128)),
stage4=dict(
num_modules=3,
num_branches=4,
block='BASIC',
num_blocks=(4, 4, 4, 4),
num_channels=(32, 64, 128, 256))),
init_cfg=dict(
type='Pretrained',
checkpoint='https://download.openmmlab.com/mmpose/'
'pretrain_models/hrnet_w32-36af842e.pth'),
),
head=dict(
type='HeatmapHead',
in_channels=32,
out_channels=23,
deconv_out_channels=None,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=True,
))
# base dataset settings
dataset_type = 'AnimalKingdomDataset'
data_mode = 'topdown'
data_root = 'data/ak/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(type='RandomBBoxTransform'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P3_amphibian/train.json',
data_prefix=dict(img='images/'),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=24,
num_workers=2,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P3_amphibian/test.json',
data_prefix=dict(img='images/'),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = val_dataloader
# evaluators
val_evaluator = [dict(type='PCKAccuracy', thr=0.05), dict(type='AUC')]
test_evaluator = val_evaluator
configs/animal_2d_keypoint/topdown_heatmap/ak/td-hm_hrnet-w32_8xb32-300e_animalkingdom_P3_bird-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
train_cfg = dict(max_epochs=300, val_interval=10)
# optimizer
optim_wrapper = dict(optimizer=dict(
type='AdamW',
lr=5e-4,
))
# learning policy
param_scheduler = [
dict(
type='LinearLR', begin=0, end=500, start_factor=0.001,
by_epoch=False), # warm-up
dict(
type='MultiStepLR',
begin=0,
end=210,
milestones=[170, 200],
gamma=0.1,
by_epoch=True)
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=512)
# hooks
default_hooks = dict(checkpoint=dict(save_best='PCK', rule='greater'))
# codec settings
codec = dict(
type='MSRAHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
type='HRNet',
in_channels=3,
extra=dict(
stage1=dict(
num_modules=1,
num_branches=1,
block='BOTTLENECK',
num_blocks=(4, ),
num_channels=(64, )),
stage2=dict(
num_modules=1,
num_branches=2,
block='BASIC',
num_blocks=(4, 4),
num_channels=(32, 64)),
stage3=dict(
num_modules=4,
num_branches=3,
block='BASIC',
num_blocks=(4, 4, 4),
num_channels=(32, 64, 128)),
stage4=dict(
num_modules=3,
num_branches=4,
block='BASIC',
num_blocks=(4, 4, 4, 4),
num_channels=(32, 64, 128, 256))),
init_cfg=dict(
type='Pretrained',
checkpoint='https://download.openmmlab.com/mmpose/'
'pretrain_models/hrnet_w32-36af842e.pth'),
),
head=dict(
type='HeatmapHead',
in_channels=32,
out_channels=23,
deconv_out_channels=None,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=True,
))
# base dataset settings
dataset_type = 'AnimalKingdomDataset'
data_mode = 'topdown'
data_root = 'data/ak/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(type='RandomBBoxTransform'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P3_bird/train.json',
data_prefix=dict(img='images/'),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=24,
num_workers=2,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P3_bird/test.json',
data_prefix=dict(img='images/'),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = val_dataloader
# evaluators
val_evaluator = [dict(type='PCKAccuracy', thr=0.05), dict(type='AUC')]
test_evaluator = val_evaluator
configs/animal_2d_keypoint/topdown_heatmap/ak/td-hm_hrnet-w32_8xb32-300e_animalkingdom_P3_fish-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
train_cfg = dict(max_epochs=300, val_interval=10)
# optimizer
optim_wrapper = dict(optimizer=dict(
type='AdamW',
lr=5e-4,
))
# learning policy
param_scheduler = [
dict(
type='LinearLR', begin=0, end=500, start_factor=0.001,
by_epoch=False), # warm-up
dict(
type='MultiStepLR',
begin=0,
end=210,
milestones=[170, 200],
gamma=0.1,
by_epoch=True)
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=512)
# hooks
default_hooks = dict(checkpoint=dict(save_best='PCK', rule='greater'))
# codec settings
codec = dict(
type='MSRAHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
type='HRNet',
in_channels=3,
extra=dict(
stage1=dict(
num_modules=1,
num_branches=1,
block='BOTTLENECK',
num_blocks=(4, ),
num_channels=(64, )),
stage2=dict(
num_modules=1,
num_branches=2,
block='BASIC',
num_blocks=(4, 4),
num_channels=(32, 64)),
stage3=dict(
num_modules=4,
num_branches=3,
block='BASIC',
num_blocks=(4, 4, 4),
num_channels=(32, 64, 128)),
stage4=dict(
num_modules=3,
num_branches=4,
block='BASIC',
num_blocks=(4, 4, 4, 4),
num_channels=(32, 64, 128, 256))),
init_cfg=dict(
type='Pretrained',
checkpoint='https://download.openmmlab.com/mmpose/'
'pretrain_models/hrnet_w32-36af842e.pth'),
),
head=dict(
type='HeatmapHead',
in_channels=32,
out_channels=23,
deconv_out_channels=None,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=True,
))
# base dataset settings
dataset_type = 'AnimalKingdomDataset'
data_mode = 'topdown'
data_root = 'data/ak/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(type='RandomBBoxTransform'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P3_fish/train.json',
data_prefix=dict(img='images/'),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=24,
num_workers=2,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P3_fish/test.json',
data_prefix=dict(img='images/'),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = val_dataloader
# evaluators
val_evaluator = [dict(type='PCKAccuracy', thr=0.05), dict(type='AUC')]
test_evaluator = val_evaluator
configs/animal_2d_keypoint/topdown_heatmap/ak/td-hm_hrnet-w32_8xb32-300e_animalkingdom_P3_mammal-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
train_cfg = dict(max_epochs=300, val_interval=10)
# optimizer
optim_wrapper = dict(optimizer=dict(
type='AdamW',
lr=5e-4,
))
# learning policy
param_scheduler = [
dict(
type='LinearLR', begin=0, end=500, start_factor=0.001,
by_epoch=False), # warm-up
dict(
type='MultiStepLR',
begin=0,
end=210,
milestones=[170, 200],
gamma=0.1,
by_epoch=True)
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=512)
# hooks
default_hooks = dict(checkpoint=dict(save_best='PCK', rule='greater'))
# codec settings
codec = dict(
type='MSRAHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
type='HRNet',
in_channels=3,
extra=dict(
stage1=dict(
num_modules=1,
num_branches=1,
block='BOTTLENECK',
num_blocks=(4, ),
num_channels=(64, )),
stage2=dict(
num_modules=1,
num_branches=2,
block='BASIC',
num_blocks=(4, 4),
num_channels=(32, 64)),
stage3=dict(
num_modules=4,
num_branches=3,
block='BASIC',
num_blocks=(4, 4, 4),
num_channels=(32, 64, 128)),
stage4=dict(
num_modules=3,
num_branches=4,
block='BASIC',
num_blocks=(4, 4, 4, 4),
num_channels=(32, 64, 128, 256))),
init_cfg=dict(
type='Pretrained',
checkpoint='https://download.openmmlab.com/mmpose/'
'pretrain_models/hrnet_w32-36af842e.pth'),
),
head=dict(
type='HeatmapHead',
in_channels=32,
out_channels=23,
deconv_out_channels=None,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=True,
))
# base dataset settings
dataset_type = 'AnimalKingdomDataset'
data_mode = 'topdown'
data_root = 'data/ak/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(type='RandomBBoxTransform'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P3_mammal/train.json',
data_prefix=dict(img='images/'),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=24,
num_workers=2,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P3_mammal/test.json',
data_prefix=dict(img='images/'),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = val_dataloader
# evaluators
val_evaluator = [dict(type='PCKAccuracy', thr=0.05), dict(type='AUC')]
test_evaluator = val_evaluator
configs/animal_2d_keypoint/topdown_heatmap/ak/td-hm_hrnet-w32_8xb32-300e_animalkingdom_P3_reptile-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
train_cfg = dict(max_epochs=300, val_interval=10)
# optimizer
optim_wrapper = dict(optimizer=dict(
type='AdamW',
lr=5e-4,
))
# learning policy
param_scheduler = [
dict(
type='LinearLR', begin=0, end=500, start_factor=0.001,
by_epoch=False), # warm-up
dict(
type='MultiStepLR',
begin=0,
end=210,
milestones=[170, 200],
gamma=0.1,
by_epoch=True)
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=512)
# hooks
default_hooks = dict(checkpoint=dict(save_best='PCK', rule='greater'))
# codec settings
codec = dict(
type='MSRAHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
type='HRNet',
in_channels=3,
extra=dict(
stage1=dict(
num_modules=1,
num_branches=1,
block='BOTTLENECK',
num_blocks=(4, ),
num_channels=(64, )),
stage2=dict(
num_modules=1,
num_branches=2,
block='BASIC',
num_blocks=(4, 4),
num_channels=(32, 64)),
stage3=dict(
num_modules=4,
num_branches=3,
block='BASIC',
num_blocks=(4, 4, 4),
num_channels=(32, 64, 128)),
stage4=dict(
num_modules=3,
num_branches=4,
block='BASIC',
num_blocks=(4, 4, 4, 4),
num_channels=(32, 64, 128, 256))),
init_cfg=dict(
type='Pretrained',
checkpoint='https://download.openmmlab.com/mmpose/'
'pretrain_models/hrnet_w32-36af842e.pth'),
),
head=dict(
type='HeatmapHead',
in_channels=32,
out_channels=23,
deconv_out_channels=None,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=True,
))
# base dataset settings
dataset_type = 'AnimalKingdomDataset'
data_mode = 'topdown'
data_root = 'data/ak/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(type='RandomBBoxTransform'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P3_reptile/train.json',
data_prefix=dict(img='images/'),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=24,
num_workers=2,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ak_P3_reptile/test.json',
data_prefix=dict(img='images/'),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = val_dataloader
# evaluators
val_evaluator = [dict(type='PCKAccuracy', thr=0.05), dict(type='AUC')]
test_evaluator = val_evaluator
configs/animal_2d_keypoint/topdown_heatmap/animalpose/hrnet_animalpose.md 0 → 100644
View file @ ca8a762a
<!-- [ALGORITHM] -->
<details>
<summary align="right"><a href="http://openaccess.thecvf.com/content_CVPR_2019/html/Sun_Deep_High-Resolution_Representation_Learning_for_Human_Pose_Estimation_CVPR_2019_paper.html">HRNet (CVPR'2019)</a></summary>
```bibtex
@inproceedings{sun2019deep,
title={Deep high-resolution representation learning for human pose estimation},
author={Sun, Ke and Xiao, Bin and Liu, Dong and Wang, Jingdong},
booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition},
pages={5693--5703},
year={2019}
}
```
</details>
<!-- [DATASET] -->
<details>
<summary align="right"><a href="http://openaccess.thecvf.com/content_ICCV_2019/html/Cao_Cross-Domain_Adaptation_for_Animal_Pose_Estimation_ICCV_2019_paper.html">Animal-Pose (ICCV'2019)</a></summary>
```bibtex
@InProceedings{Cao_2019_ICCV,
author = {Cao, Jinkun and Tang, Hongyang and Fang, Hao-Shu and Shen, Xiaoyong and Lu, Cewu and Tai, Yu-Wing},
title = {Cross-Domain Adaptation for Animal Pose Estimation},
booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
month = {October},
year = {2019}
}
```
</details>
Results on AnimalPose validation set (1117 instances)
| Arch | Input Size | AP | AP<sup>50</sup> | AP<sup>75</sup> | AR | AR<sup>50</sup> | ckpt | log |
| :-------------------------------------------- | :--------: | :---: | :-------------: | :-------------: | :---: | :-------------: | :-------------------------------------------: | :-------------------------------------------: |
| [pose_hrnet_w32](/configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_hrnet-w32_8xb64-210e_animalpose-256x256.py) | 256x256 | 0.740 | 0.959 | 0.833 | 0.780 | 0.965 | [ckpt](https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w32_animalpose_256x256-1aa7f075_20210426.pth) | [log](https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w32_animalpose_256x256_20210426.log.json) |
| [pose_hrnet_w48](/configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_hrnet-w48_8xb64-210e_animalpose-256x256.py) | 256x256 | 0.738 | 0.958 | 0.831 | 0.778 | 0.962 | [ckpt](https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w48_animalpose_256x256-34644726_20210426.pth) | [log](https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w48_animalpose_256x256_20210426.log.json) |
configs/animal_2d_keypoint/topdown_heatmap/animalpose/hrnet_animalpose.yml 0 → 100644
View file @ ca8a762a
Models:
- Config: configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_hrnet-w32_8xb64-210e_animalpose-256x256.py
In Collection: HRNet
Metadata:
Architecture: &id001
- HRNet
Training Data: Animal-Pose
Name: td-hm_hrnet-w32_8xb64-210e_animalpose-256x256
Results:
- Dataset: Animal-Pose
Metrics:
AP: 0.740
AP@0.5: 0.959
AP@0.75: 0.833
AR: 0.780
AR@0.5: 0.965
Task: Animal 2D Keypoint
Weights: https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w32_animalpose_256x256-1aa7f075_20210426.pth
- Config: configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_hrnet-w48_8xb64-210e_animalpose-256x256.py
In Collection: HRNet
Metadata:
Architecture: *id001
Training Data: Animal-Pose
Name: td-hm_hrnet-w48_8xb64-210e_animalpose-256x256
Results:
- Dataset: Animal-Pose
Metrics:
AP: 0.738
AP@0.5: 0.958
AP@0.75: 0.831
AR: 0.778
AR@0.5: 0.962
Task: Animal 2D Keypoint
Weights: https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w48_animalpose_256x256-34644726_20210426.pth
configs/animal_2d_keypoint/topdown_heatmap/animalpose/resnet_animalpose.md 0 → 100644
View file @ ca8a762a
<!-- [ALGORITHM] -->
<details>
<summary align="right"><a href="http://openaccess.thecvf.com/content_ECCV_2018/html/Bin_Xiao_Simple_Baselines_for_ECCV_2018_paper.html">SimpleBaseline2D (ECCV'2018)</a></summary>
```bibtex
@inproceedings{xiao2018simple,
title={Simple baselines for human pose estimation and tracking},
author={Xiao, Bin and Wu, Haiping and Wei, Yichen},
booktitle={Proceedings of the European conference on computer vision (ECCV)},
pages={466--481},
year={2018}
}
```
</details>
<!-- [DATASET] -->
<details>
<summary align="right"><a href="http://openaccess.thecvf.com/content_ICCV_2019/html/Cao_Cross-Domain_Adaptation_for_Animal_Pose_Estimation_ICCV_2019_paper.html">Animal-Pose (ICCV'2019)</a></summary>
```bibtex
@InProceedings{Cao_2019_ICCV,
author = {Cao, Jinkun and Tang, Hongyang and Fang, Hao-Shu and Shen, Xiaoyong and Lu, Cewu and Tai, Yu-Wing},
title = {Cross-Domain Adaptation for Animal Pose Estimation},
booktitle = {The IEEE International Conference on Computer Vision (ICCV)},
month = {October},
year = {2019}
}
```
</details>
Results on AnimalPose validation set (1117 instances)
| Arch | Input Size | AP | AP<sup>50</sup> | AP<sup>75</sup> | AR | AR<sup>50</sup> | ckpt | log |
| :-------------------------------------------- | :--------: | :---: | :-------------: | :-------------: | :---: | :-------------: | :-------------------------------------------: | :-------------------------------------------: |
| [pose_resnet_50](/configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_res50_8xb64-210e_animalpose-256x256.py) | 256x256 | 0.691 | 0.947 | 0.770 | 0.736 | 0.955 | [ckpt](https://download.openmmlab.com/mmpose/animal/resnet/res50_animalpose_256x256-e1f30bff_20210426.pth) | [log](https://download.openmmlab.com/mmpose/animal/resnet/res50_animalpose_256x256_20210426.log.json) |
| [pose_resnet_101](/configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_res101_8xb64-210e_animalpose-256x256.py) | 256x256 | 0.696 | 0.948 | 0.774 | 0.736 | 0.951 | [ckpt](https://download.openmmlab.com/mmpose/animal/resnet/res101_animalpose_256x256-85563f4a_20210426.pth) | [log](https://download.openmmlab.com/mmpose/animal/resnet/res101_animalpose_256x256_20210426.log.json) |
| [pose_resnet_152](/configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_res152_8xb32-210e_animalpose-256x256.py) | 256x256 | 0.704 | 0.938 | 0.786 | 0.748 | 0.946 | [ckpt](https://download.openmmlab.com/mmpose/animal/resnet/res152_animalpose_256x256-a0a7506c_20210426.pth) | [log](https://download.openmmlab.com/mmpose/animal/resnet/res152_animalpose_256x256_20210426.log.json) |
configs/animal_2d_keypoint/topdown_heatmap/animalpose/resnet_animalpose.yml 0 → 100644
View file @ ca8a762a
Models:
- Config: configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_res50_8xb64-210e_animalpose-256x256.py
In Collection: SimpleBaseline2D
Metadata:
Architecture: &id001
- SimpleBaseline2D
- ResNet
Training Data: Animal-Pose
Name: td-hm_res50_8xb64-210e_animalpose-256x256
Results:
- Dataset: Animal-Pose
Metrics:
AP: 0.691
AP@0.5: 0.947
AP@0.75: 0.770
AR: 0.736
AR@0.5: 0.955
Task: Animal 2D Keypoint
Weights: https://download.openmmlab.com/mmpose/animal/resnet/res50_animalpose_256x256-e1f30bff_20210426.pth
- Config: configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_res101_8xb64-210e_animalpose-256x256.py
In Collection: SimpleBaseline2D
Metadata:
Architecture: *id001
Training Data: Animal-Pose
Name: td-hm_res101_8xb64-210e_animalpose-256x256
Results:
- Dataset: Animal-Pose
Metrics:
AP: 0.696
AP@0.5: 0.948
AP@0.75: 0.774
AR: 0.736
AR@0.5: 0.951
Task: Animal 2D Keypoint
Weights: https://download.openmmlab.com/mmpose/animal/resnet/res101_animalpose_256x256-85563f4a_20210426.pth
- Config: configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_res152_8xb32-210e_animalpose-256x256.py
In Collection: SimpleBaseline2D
Metadata:
Architecture: *id001
Training Data: Animal-Pose
Name: td-hm_res152_8xb32-210e_animalpose-256x256
Results:
- Dataset: Animal-Pose
Metrics:
AP: 0.704
AP@0.5: 0.938
AP@0.75: 0.786
AR: 0.748
AR@0.5: 0.946
Task: Animal 2D Keypoint
Weights: https://download.openmmlab.com/mmpose/animal/resnet/res152_animalpose_256x256-a0a7506c_20210426.pth
configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_hrnet-w32_8xb64-210e_animalpose-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
train_cfg = dict(max_epochs=210, val_interval=10)
# optimizer
optim_wrapper = dict(optimizer=dict(
type='Adam',
lr=5e-4,
))
# learning policy
param_scheduler = [
dict(
type='LinearLR', begin=0, end=500, start_factor=0.001,
by_epoch=False), # warm-up
dict(
type='MultiStepLR',
begin=0,
end=210,
milestones=[170, 200],
gamma=0.1,
by_epoch=True)
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=512)
# hooks
default_hooks = dict(checkpoint=dict(save_best='coco/AP', rule='greater'))
# codec settings
codec = dict(
type='MSRAHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
type='HRNet',
in_channels=3,
extra=dict(
stage1=dict(
num_modules=1,
num_branches=1,
block='BOTTLENECK',
num_blocks=(4, ),
num_channels=(64, )),
stage2=dict(
num_modules=1,
num_branches=2,
block='BASIC',
num_blocks=(4, 4),
num_channels=(32, 64)),
stage3=dict(
num_modules=4,
num_branches=3,
block='BASIC',
num_blocks=(4, 4, 4),
num_channels=(32, 64, 128)),
stage4=dict(
num_modules=3,
num_branches=4,
block='BASIC',
num_blocks=(4, 4, 4, 4),
num_channels=(32, 64, 128, 256))),
init_cfg=dict(
type='Pretrained',
checkpoint='https://download.openmmlab.com/mmpose/'
'pretrain_models/hrnet_w32-36af842e.pth'),
),
head=dict(
type='HeatmapHead',
in_channels=32,
out_channels=20,
deconv_out_channels=None,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=True,
))
# base dataset settings
dataset_type = 'AnimalPoseDataset'
data_mode = 'topdown'
data_root = 'data/animalpose/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(type='RandomBBoxTransform'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=64,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/animalpose_train.json',
data_prefix=dict(img=''),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/animalpose_val.json',
data_prefix=dict(img=''),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = val_dataloader
# evaluators
val_evaluator = dict(
type='CocoMetric', ann_file=data_root + 'annotations/animalpose_val.json')
test_evaluator = val_evaluator
configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_hrnet-w48_8xb64-210e_animalpose-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
train_cfg = dict(max_epochs=210, val_interval=10)
# optimizer
optim_wrapper = dict(optimizer=dict(
type='Adam',
lr=5e-4,
))
# learning policy
param_scheduler = [
dict(
type='LinearLR', begin=0, end=500, start_factor=0.001,
by_epoch=False), # warm-up
dict(
type='MultiStepLR',
begin=0,
end=210,
milestones=[170, 200],
gamma=0.1,
by_epoch=True)
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=512)
# hooks
default_hooks = dict(checkpoint=dict(save_best='coco/AP', rule='greater'))
# codec settings
codec = dict(
type='MSRAHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
type='HRNet',
in_channels=3,
extra=dict(
stage1=dict(
num_modules=1,
num_branches=1,
block='BOTTLENECK',
num_blocks=(4, ),
num_channels=(64, )),
stage2=dict(
num_modules=1,
num_branches=2,
block='BASIC',
num_blocks=(4, 4),
num_channels=(48, 96)),
stage3=dict(
num_modules=4,
num_branches=3,
block='BASIC',
num_blocks=(4, 4, 4),
num_channels=(48, 96, 192)),
stage4=dict(
num_modules=3,
num_branches=4,
block='BASIC',
num_blocks=(4, 4, 4, 4),
num_channels=(48, 96, 192, 384))),
init_cfg=dict(
type='Pretrained',
checkpoint='https://download.openmmlab.com/mmpose/'
'pretrain_models/hrnet_w48-8ef0771d.pth'),
),
head=dict(
type='HeatmapHead',
in_channels=48,
out_channels=20,
deconv_out_channels=None,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=True,
))
# base dataset settings
dataset_type = 'AnimalPoseDataset'
data_mode = 'topdown'
data_root = 'data/animalpose/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(type='RandomBBoxTransform'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=64,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/animalpose_train.json',
data_prefix=dict(img=''),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/animalpose_val.json',
data_prefix=dict(img=''),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = val_dataloader
# evaluators
val_evaluator = dict(
type='CocoMetric', ann_file=data_root + 'annotations/animalpose_val.json')
test_evaluator = val_evaluator
configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_res101_8xb64-210e_animalpose-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
train_cfg = dict(max_epochs=210, val_interval=10)
# optimizer
optim_wrapper = dict(optimizer=dict(
type='Adam',
lr=5e-4,
))
# learning policy
param_scheduler = [
dict(
type='LinearLR', begin=0, end=500, start_factor=0.001,
by_epoch=False), # warm-up
dict(
type='MultiStepLR',
begin=0,
end=210,
milestones=[170, 200],
gamma=0.1,
by_epoch=True)
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=512)
# hooks
default_hooks = dict(checkpoint=dict(save_best='coco/AP', rule='greater'))
# codec settings
codec = dict(
type='MSRAHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
type='ResNet',
depth=101,
init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet101'),
),
head=dict(
type='HeatmapHead',
in_channels=2048,
out_channels=20,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=True,
))
# base dataset settings
dataset_type = 'AnimalPoseDataset'
data_mode = 'topdown'
data_root = 'data/animalpose/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(type='RandomBBoxTransform'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=64,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/animalpose_train.json',
data_prefix=dict(img=''),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/animalpose_val.json',
data_prefix=dict(img=''),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = val_dataloader
# evaluators
val_evaluator = dict(
type='CocoMetric', ann_file=data_root + 'annotations/animalpose_val.json')
test_evaluator = val_evaluator
configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_res152_8xb32-210e_animalpose-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
train_cfg = dict(max_epochs=210, val_interval=10)
# optimizer
optim_wrapper = dict(optimizer=dict(
type='Adam',
lr=5e-4,
))
# learning policy
param_scheduler = [
dict(
type='LinearLR', begin=0, end=500, start_factor=0.001,
by_epoch=False), # warm-up
dict(
type='MultiStepLR',
begin=0,
end=210,
milestones=[170, 200],
gamma=0.1,
by_epoch=True)
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=256)
# hooks
default_hooks = dict(checkpoint=dict(save_best='coco/AP', rule='greater'))
# codec settings
codec = dict(
type='MSRAHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
type='ResNet',
depth=152,
init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet152'),
),
head=dict(
type='HeatmapHead',
in_channels=2048,
out_channels=20,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=True,
))
# base dataset settings
dataset_type = 'AnimalPoseDataset'
data_mode = 'topdown'
data_root = 'data/animalpose/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(type='RandomBBoxTransform'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/animalpose_train.json',
data_prefix=dict(img=''),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/animalpose_val.json',
data_prefix=dict(img=''),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = val_dataloader
# evaluators
val_evaluator = dict(
type='CocoMetric', ann_file=data_root + 'annotations/animalpose_val.json')
test_evaluator = val_evaluator
configs/animal_2d_keypoint/topdown_heatmap/animalpose/td-hm_res50_8xb64-210e_animalpose-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
train_cfg = dict(max_epochs=210, val_interval=10)
# optimizer
optim_wrapper = dict(optimizer=dict(
type='Adam',
lr=5e-4,
))
# learning policy
param_scheduler = [
dict(
type='LinearLR', begin=0, end=500, start_factor=0.001,
by_epoch=False), # warm-up
dict(
type='MultiStepLR',
begin=0,
end=210,
milestones=[170, 200],
gamma=0.1,
by_epoch=True)
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=512)
# hooks
default_hooks = dict(checkpoint=dict(save_best='coco/AP', rule='greater'))
# codec settings
codec = dict(
type='MSRAHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
type='ResNet',
depth=50,
init_cfg=dict(type='Pretrained', checkpoint='torchvision://resnet50'),
),
head=dict(
type='HeatmapHead',
in_channels=2048,
out_channels=20,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=True,
))
# base dataset settings
dataset_type = 'AnimalPoseDataset'
data_mode = 'topdown'
data_root = 'data/animalpose/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(type='RandomBBoxTransform'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size']),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=64,
num_workers=2,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/animalpose_train.json',
data_prefix=dict(img=''),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=32,
num_workers=2,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/animalpose_val.json',
data_prefix=dict(img=''),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = val_dataloader
# evaluators
val_evaluator = dict(
type='CocoMetric', ann_file=data_root + 'annotations/animalpose_val.json')
test_evaluator = val_evaluator
configs/animal_2d_keypoint/topdown_heatmap/ap10k/cspnext-m_udp_8xb64-210e_ap10k-256x256.py 0 → 100644
View file @ ca8a762a
_base_ = ['../../../_base_/default_runtime.py']
# runtime
max_epochs = 210
stage2_num_epochs = 30
base_lr = 4e-3
train_cfg = dict(max_epochs=max_epochs, val_interval=10)
randomness = dict(seed=21)
# optimizer
optim_wrapper = dict(
type='OptimWrapper',
optimizer=dict(type='AdamW', lr=base_lr, weight_decay=0.05),
paramwise_cfg=dict(
norm_decay_mult=0, bias_decay_mult=0, bypass_duplicate=True))
# learning policy
param_scheduler = [
dict(
type='LinearLR',
start_factor=1.0e-5,
by_epoch=False,
begin=0,
end=1000),
dict(
# use cosine lr from 105 to 210 epoch
type='CosineAnnealingLR',
eta_min=base_lr * 0.05,
begin=max_epochs // 2,
end=max_epochs,
T_max=max_epochs // 2,
by_epoch=True,
convert_to_iter_based=True),
]
# automatically scaling LR based on the actual training batch size
auto_scale_lr = dict(base_batch_size=1024)
# codec settings
codec = dict(
type='UDPHeatmap', input_size=(256, 256), heatmap_size=(64, 64), sigma=2)
# model settings
model = dict(
type='TopdownPoseEstimator',
data_preprocessor=dict(
type='PoseDataPreprocessor',
mean=[123.675, 116.28, 103.53],
std=[58.395, 57.12, 57.375],
bgr_to_rgb=True),
backbone=dict(
_scope_='mmdet',
type='CSPNeXt',
arch='P5',
expand_ratio=0.5,
deepen_factor=0.67,
widen_factor=0.75,
out_indices=(4, ),
channel_attention=True,
norm_cfg=dict(type='SyncBN'),
act_cfg=dict(type='SiLU'),
init_cfg=dict(
type='Pretrained',
prefix='backbone.',
checkpoint='https://download.openmmlab.com/mmdetection/v3.0/'
'rtmdet/cspnext_rsb_pretrain/'
'cspnext-m_8xb256-rsb-a1-600e_in1k-ecb3bbd9.pth')),
head=dict(
type='HeatmapHead',
in_channels=768,
out_channels=17,
loss=dict(type='KeypointMSELoss', use_target_weight=True),
decoder=codec),
test_cfg=dict(
flip_test=True,
flip_mode='heatmap',
shift_heatmap=False,
))
# base dataset settings
dataset_type = 'AP10KDataset'
data_mode = 'topdown'
data_root = 'data/ap10k/'
# pipelines
train_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(
type='RandomBBoxTransform', scale_factor=[0.6, 1.4], rotate_factor=80),
dict(type='TopdownAffine', input_size=codec['input_size'], use_udp=True),
dict(type='mmdet.YOLOXHSVRandomAug'),
dict(
type='Albumentation',
transforms=[
dict(type='Blur', p=0.1),
dict(type='MedianBlur', p=0.1),
dict(
type='CoarseDropout',
max_holes=1,
max_height=0.4,
max_width=0.4,
min_holes=1,
min_height=0.2,
min_width=0.2,
p=1.),
]),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
val_pipeline = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='TopdownAffine', input_size=codec['input_size'], use_udp=True),
dict(type='PackPoseInputs')
]
train_pipeline_stage2 = [
dict(type='LoadImage'),
dict(type='GetBBoxCenterScale'),
dict(type='RandomFlip', direction='horizontal'),
dict(type='RandomHalfBody'),
dict(
type='RandomBBoxTransform',
shift_factor=0.,
scale_factor=[0.75, 1.25],
rotate_factor=60),
dict(type='TopdownAffine', input_size=codec['input_size'], use_udp=True),
dict(type='mmdet.YOLOXHSVRandomAug'),
dict(
type='Albumentation',
transforms=[
dict(type='Blur', p=0.1),
dict(type='MedianBlur', p=0.1),
dict(
type='CoarseDropout',
max_holes=1,
max_height=0.4,
max_width=0.4,
min_holes=1,
min_height=0.2,
min_width=0.2,
p=0.5),
]),
dict(type='GenerateTarget', encoder=codec),
dict(type='PackPoseInputs')
]
# data loaders
train_dataloader = dict(
batch_size=64,
num_workers=10,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ap10k-train-split1.json',
data_prefix=dict(img='data/'),
pipeline=train_pipeline,
))
val_dataloader = dict(
batch_size=32,
num_workers=10,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ap10k-val-split1.json',
data_prefix=dict(img='data/'),
test_mode=True,
pipeline=val_pipeline,
))
test_dataloader = dict(
batch_size=32,
num_workers=10,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False, round_up=False),
dataset=dict(
type=dataset_type,
data_root=data_root,
data_mode=data_mode,
ann_file='annotations/ap10k-test-split1.json',
data_prefix=dict(img='data/'),
test_mode=True,
pipeline=val_pipeline,
))
# hooks
default_hooks = dict(
checkpoint=dict(save_best='coco/AP', rule='greater', max_keep_ckpts=1))
custom_hooks = [
dict(
type='EMAHook',
ema_type='ExpMomentumEMA',
momentum=0.0002,
update_buffers=True,
priority=49),
dict(
type='mmdet.PipelineSwitchHook',
switch_epoch=max_epochs - stage2_num_epochs,
switch_pipeline=train_pipeline_stage2)
]
# evaluators
val_evaluator = dict(
type='CocoMetric',
ann_file=data_root + 'annotations/ap10k-val-split1.json')
test_evaluator = dict(
type='CocoMetric',
ann_file=data_root + 'annotations/ap10k-test-split1.json')
configs/animal_2d_keypoint/topdown_heatmap/ap10k/cspnext_udp_ap10k.md 0 → 100644
View file @ ca8a762a
<!-- [ALGORITHM] -->
<details>
<summary align="right"><a href="https://arxiv.org/abs/2212.07784">RTMDet (ArXiv 2022)</a></summary>
```bibtex
@misc{lyu2022rtmdet,
title={RTMDet: An Empirical Study of Designing Real-Time Object Detectors},
author={Chengqi Lyu and Wenwei Zhang and Haian Huang and Yue Zhou and Yudong Wang and Yanyi Liu and Shilong Zhang and Kai Chen},
year={2022},
eprint={2212.07784},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
```
</details>
<!-- [ALGORITHM] -->
<details>
<summary align="right"><a href="http://openaccess.thecvf.com/content_CVPR_2020/html/Huang_The_Devil_Is_in_the_Details_Delving_Into_Unbiased_Data_CVPR_2020_paper.html">UDP (CVPR'2020)</a></summary>
```bibtex
@InProceedings{Huang_2020_CVPR,
author = {Huang, Junjie and Zhu, Zheng and Guo, Feng and Huang, Guan},
title = {The Devil Is in the Details: Delving Into Unbiased Data Processing for Human Pose Estimation},
booktitle = {The IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2020}
}
```
</details>
<!-- [DATASET] -->
<details>
<summary align="right"><a href="https://arxiv.org/abs/2108.12617">AP-10K (NeurIPS'2021)</a></summary>
```bibtex
@misc{yu2021ap10k,
title={AP-10K: A Benchmark for Animal Pose Estimation in the Wild},
author={Hang Yu and Yufei Xu and Jing Zhang and Wei Zhao and Ziyu Guan and Dacheng Tao},
year={2021},
eprint={2108.12617},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
```
</details>
Results on AP-10K validation set
| Arch | Input Size | AP | AP<sup>50</sup> | AP<sup>75</sup> | AP<sup>M</sup> | AP<sup>L</sup> | ckpt | log |
| :----------------------------------------- | :--------: | :---: | :-------------: | :-------------: | :------------: | :------------: | :-----------------------------------------: | :----------------------------------------: |
| [pose_cspnext_m](/configs/animal_2d_keypoint/topdown_heatmap/ap10k/cspnext-m_udp_8xb64-210e_ap10k-256x256.py) | 256x256 | 0.703 | 0.944 | 0.776 | 0.513 | 0.710 | [ckpt](https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/cspnext-m_udp-ap10k_pt-in1k_210e-256x256-1f2d947a_20230123.pth) | [log](https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/cspnext-m_udp-ap10k_pt-in1k_210e-256x256-1f2d947a_20230123.json) |
configs/animal_2d_keypoint/topdown_heatmap/ap10k/cspnext_udp_ap10k.yml 0 → 100644
View file @ ca8a762a
Models:
- Config: configs/animal_2d_keypoint/topdown_heatmap/ap10k/cspnext-m_udp_8xb64-210e_ap10k-256x256.py
In Collection: UDP
Metadata:
Architecture: &id001
- UDP
- HRNet
Training Data: AP-10K
Name: cspnext-m_udp_8xb64-210e_ap10k-256x256
Results:
- Dataset: AP-10K
Metrics:
AP: 0.703
AP@0.5: 0.944
AP@0.75: 0.776
AP (L): 0.71
AP (M): 0.513
Task: Animal 2D Keypoint
Weights: https://download.openmmlab.com/mmpose/v1/projects/rtmposev1/cspnext-m_udp-ap10k_pt-in1k_210e-256x256-1f2d947a_20230123.pth
configs/animal_2d_keypoint/topdown_heatmap/ap10k/hrnet_ap10k.md 0 → 100644
View file @ ca8a762a
<!-- [ALGORITHM] -->
<details>
<summary align="right"><a href="http://openaccess.thecvf.com/content_CVPR_2019/html/Sun_Deep_High-Resolution_Representation_Learning_for_Human_Pose_Estimation_CVPR_2019_paper.html">HRNet (CVPR'2019)</a></summary>
```bibtex
@inproceedings{sun2019deep,
title={Deep high-resolution representation learning for human pose estimation},
author={Sun, Ke and Xiao, Bin and Liu, Dong and Wang, Jingdong},
booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition},
pages={5693--5703},
year={2019}
}
```
</details>
<!-- [DATASET] -->
<details>
<summary align="right"><a href="https://arxiv.org/abs/2108.12617">AP-10K (NeurIPS'2021)</a></summary>
```bibtex
@misc{yu2021ap10k,
title={AP-10K: A Benchmark for Animal Pose Estimation in the Wild},
author={Hang Yu and Yufei Xu and Jing Zhang and Wei Zhao and Ziyu Guan and Dacheng Tao},
year={2021},
eprint={2108.12617},
archivePrefix={arXiv},
primaryClass={cs.CV}
}
```
</details>
Results on AP-10K validation set
| Arch | Input Size | AP | AP<sup>50</sup> | AP<sup>75</sup> | AP<sup>M</sup> | AP<sup>L</sup> | ckpt | log |
| :----------------------------------------- | :--------: | :---: | :-------------: | :-------------: | :------------: | :------------: | :-----------------------------------------: | :----------------------------------------: |
| [pose_hrnet_w32](/configs/animal_2d_keypoint/topdown_heatmap/ap10k/td-hm_hrnet-w32_8xb64-210e_ap10k-256x256.py) | 256x256 | 0.722 | 0.935 | 0.789 | 0.557 | 0.729 | [ckpt](https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w32_ap10k_256x256-18aac840_20211029.pth) | [log](https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w32_ap10k_256x256-18aac840_20211029.log.json) |
| [pose_hrnet_w48](/configs/animal_2d_keypoint/topdown_heatmap/ap10k/td-hm_hrnet-w48_8xb64-210e_ap10k-256x256.py) | 256x256 | 0.728 | 0.936 | 0.802 | 0.577 | 0.735 | [ckpt](https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w48_ap10k_256x256-d95ab412_20211029.pth) | [log](https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w48_ap10k_256x256-d95ab412_20211029.log.json) |
configs/animal_2d_keypoint/topdown_heatmap/ap10k/hrnet_ap10k.yml 0 → 100644
View file @ ca8a762a
Models:
- Config: configs/animal_2d_keypoint/topdown_heatmap/ap10k/td-hm_hrnet-w32_8xb64-210e_ap10k-256x256.py
In Collection: HRNet
Metadata:
Architecture: &id001
- HRNet
Training Data: AP-10K
Name: td-hm_hrnet-w32_8xb64-210e_ap10k-256x256
Results:
- Dataset: AP-10K
Metrics:
AP: 0.722
AP@0.5: 0.935
AP@0.75: 0.789
AP (L): 0.729
AP (M): 0.557
Task: Animal 2D Keypoint
Weights: https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w32_ap10k_256x256-18aac840_20211029.pth
- Config: configs/animal_2d_keypoint/topdown_heatmap/ap10k/td-hm_hrnet-w48_8xb64-210e_ap10k-256x256.py
In Collection: HRNet
Metadata:
Architecture: *id001
Training Data: AP-10K
Name: td-hm_hrnet-w48_8xb64-210e_ap10k-256x256
Results:
- Dataset: AP-10K
Metrics:
AP: 0.728
AP@0.5: 0.936
AP@0.75: 0.802
AP (L): 0.735
AP (M): 0.577
Task: Animal 2D Keypoint
Weights: https://download.openmmlab.com/mmpose/animal/hrnet/hrnet_w48_ap10k_256x256-d95ab412_20211029.pth
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment