Skip to content

Commit

Permalink
[Features] Support PV_RCNN modules (open-mmlab#1957)
Browse files Browse the repository at this point in the history 
* add pvrcnn module code

* add voxelsa

* fix

* fix comments

* fix comments

* fix comments

* add stack sa

* fix

* fix comments

* fix comments

* fix

* add ut

* fix comments
  • Loading branch information
@VVsssssk @ZwwWayne
VVsssssk authored and ZwwWayne committed Dec 3, 2022
1 parent 1fd7153 commit aa544d4
Show file tree
Hide file tree
Showing 18 changed files with 2,310 additions and 11 deletions.
353 changes: 353 additions & 0 deletions configs/pvrcnn/pvrcnn_8xb2-80e_kitti-3d-3class.py
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,353 @@
_base_ = [
'../_base_/datasets/kitti-3d-3class.py',
'../_base_/schedules/cyclic-40e.py', '../_base_/default_runtime.py'
]

voxel_size = [0.05, 0.05, 0.1]
point_cloud_range = [0, -40, -3, 70.4, 40, 1]

data_root = 'data/kitti/'
class_names = ['Pedestrian', 'Cyclist', 'Car']
metainfo = dict(CLASSES=class_names)
db_sampler = dict(
data_root=data_root,
info_path=data_root + 'kitti_dbinfos_train.pkl',
rate=1.0,
prepare=dict(
filter_by_difficulty=[-1],
filter_by_min_points=dict(Car=5, Pedestrian=5, Cyclist=5)),
classes=class_names,
sample_groups=dict(Car=15, Pedestrian=10, Cyclist=10),
points_loader=dict(
type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4))

train_pipeline = [
dict(type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4),
dict(type='LoadAnnotations3D', with_bbox_3d=True, with_label_3d=True),
dict(type='ObjectSample', db_sampler=db_sampler, use_ground_plane=True),
dict(type='RandomFlip3D', flip_ratio_bev_horizontal=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[-0.78539816, 0.78539816],
scale_ratio_range=[0.95, 1.05]),
dict(type='PointsRangeFilter', point_cloud_range=point_cloud_range),
dict(type='ObjectRangeFilter', point_cloud_range=point_cloud_range),
dict(type='PointShuffle'),
dict(
type='Pack3DDetInputs',
keys=['points', 'gt_bboxes_3d', 'gt_labels_3d'])
]
test_pipeline = [
dict(type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(
type='GlobalRotScaleTrans',
rot_range=[0, 0],
scale_ratio_range=[1., 1.],
translation_std=[0, 0, 0]),
dict(type='RandomFlip3D'),
dict(
type='PointsRangeFilter', point_cloud_range=point_cloud_range)
]),
dict(type='Pack3DDetInputs', keys=['points'])
]

model = dict(
type='PointVoxelRCNN',
data_preprocessor=dict(
type='Det3DDataPreprocessor',
voxel=True,
voxel_layer=dict(
max_num_points=5, # max_points_per_voxel
point_cloud_range=point_cloud_range,
voxel_size=voxel_size,
max_voxels=(16000, 40000))),
voxel_encoder=dict(type='HardSimpleVFE'),
middle_encoder=dict(
type='SparseEncoder',
in_channels=4,
sparse_shape=[41, 1600, 1408],
order=('conv', 'norm', 'act'),
encoder_paddings=((0, 0, 0), ((1, 1, 1), 0, 0), ((1, 1, 1), 0, 0),
((0, 1, 1), 0, 0)),
return_middle_feats=True),
points_encoder=dict(
type='VoxelSetAbstraction',
num_keypoints=2048,
fused_out_channel=128,
voxel_size=voxel_size,
point_cloud_range=point_cloud_range,
voxel_sa_cfgs_list=[
dict(
type='StackedSAModuleMSG',
in_channels=16,
scale_factor=1,
radius=(0.4, 0.8),
sample_nums=(16, 16),
mlp_channels=((16, 16), (16, 16)),
use_xyz=True),
dict(
type='StackedSAModuleMSG',
in_channels=32,
scale_factor=2,
radius=(0.8, 1.2),
sample_nums=(16, 32),
mlp_channels=((32, 32), (32, 32)),
use_xyz=True),
dict(
type='StackedSAModuleMSG',
in_channels=64,
scale_factor=4,
radius=(1.2, 2.4),
sample_nums=(16, 32),
mlp_channels=((64, 64), (64, 64)),
use_xyz=True),
dict(
type='StackedSAModuleMSG',
in_channels=64,
scale_factor=8,
radius=(2.4, 4.8),
sample_nums=(16, 32),
mlp_channels=((64, 64), (64, 64)),
use_xyz=True)
],
rawpoints_sa_cfgs=dict(
type='StackedSAModuleMSG',
in_channels=1,
radius=(0.4, 0.8),
sample_nums=(16, 16),
mlp_channels=((16, 16), (16, 16)),
use_xyz=True),
bev_feat_channel=256,
bev_scale_factor=8),
backbone=dict(
type='SECOND',
in_channels=256,
layer_nums=[5, 5],
layer_strides=[1, 2],
out_channels=[128, 256]),
neck=dict(
type='SECONDFPN',
in_channels=[128, 256],
upsample_strides=[1, 2],
out_channels=[256, 256]),
rpn_head=dict(
type='PartA2RPNHead',
num_classes=3,
in_channels=512,
feat_channels=512,
use_direction_classifier=True,
dir_offset=0.78539,
anchor_generator=dict(
type='Anchor3DRangeGenerator',
ranges=[[0, -40.0, -0.6, 70.4, 40.0, -0.6],
[0, -40.0, -0.6, 70.4, 40.0, -0.6],
[0, -40.0, -1.78, 70.4, 40.0, -1.78]],
sizes=[[0.8, 0.6, 1.73], [1.76, 0.6, 1.73], [3.9, 1.6, 1.56]],
rotations=[0, 1.57],
reshape_out=False),
diff_rad_by_sin=True,
assigner_per_size=True,
assign_per_class=True,
bbox_coder=dict(type='DeltaXYZWLHRBBoxCoder'),
loss_cls=dict(
type='mmdet.FocalLoss',
use_sigmoid=True,
gamma=2.0,
alpha=0.25,
loss_weight=1.0),
loss_bbox=dict(
type='mmdet.SmoothL1Loss', beta=1.0 / 9.0, loss_weight=2.0),
loss_dir=dict(
type='mmdet.CrossEntropyLoss', use_sigmoid=False,
loss_weight=0.2)),
roi_head=dict(
type='PVRCNNRoiHead',
num_classes=3,
semantic_head=dict(
type='ForegroundSegmentationHead',
in_channels=640,
extra_width=0.1,
loss_seg=dict(
type='mmdet.FocalLoss',
use_sigmoid=True,
reduction='sum',
gamma=2.0,
alpha=0.25,
activated=True,
loss_weight=1.0)),
bbox_roi_extractor=dict(
type='Batch3DRoIGridExtractor',
grid_size=6,
roi_layer=dict(
type='StackedSAModuleMSG',
in_channels=128,
radius=(0.8, 1.6),
sample_nums=(16, 16),
mlp_channels=((64, 64), (64, 64)),
use_xyz=True,
pool_mod='max'),
),
bbox_head=dict(
type='PVRCNNBBoxHead',
in_channels=128,
grid_size=6,
num_classes=3,
class_agnostic=True,
shared_fc_channels=(256, 256),
reg_channels=(256, 256),
cls_channels=(256, 256),
dropout_ratio=0.3,
with_corner_loss=True,
bbox_coder=dict(type='DeltaXYZWLHRBBoxCoder'),
loss_bbox=dict(
type='mmdet.SmoothL1Loss',
beta=1.0 / 9.0,
reduction='sum',
loss_weight=1.0),
loss_cls=dict(
type='mmdet.CrossEntropyLoss',
use_sigmoid=True,
reduction='sum',
loss_weight=1.0))),
# model training and testing settings
train_cfg=dict(
rpn=dict(
assigner=[
dict( # for Pedestrian
type='Max3DIoUAssigner',
iou_calculator=dict(type='BboxOverlapsNearest3D'),
pos_iou_thr=0.5,
neg_iou_thr=0.35,
min_pos_iou=0.35,
ignore_iof_thr=-1),
dict( # for Cyclist
type='Max3DIoUAssigner',
iou_calculator=dict(type='BboxOverlapsNearest3D'),
pos_iou_thr=0.5,
neg_iou_thr=0.35,
min_pos_iou=0.35,
ignore_iof_thr=-1),
dict( # for Car
type='Max3DIoUAssigner',
iou_calculator=dict(type='BboxOverlapsNearest3D'),
pos_iou_thr=0.6,
neg_iou_thr=0.45,
min_pos_iou=0.45,
ignore_iof_thr=-1)
],
allowed_border=0,
pos_weight=-1,
debug=False),
rpn_proposal=dict(
nms_pre=9000,
nms_post=512,
max_num=512,
nms_thr=0.8,
score_thr=0,
use_rotate_nms=True),
rcnn=dict(
assigner=[
dict( # for Pedestrian
type='Max3DIoUAssigner',
iou_calculator=dict(
type='BboxOverlaps3D', coordinate='lidar'),
pos_iou_thr=0.55,
neg_iou_thr=0.55,
min_pos_iou=0.55,
ignore_iof_thr=-1),
dict( # for Cyclist
type='Max3DIoUAssigner',
iou_calculator=dict(
type='BboxOverlaps3D', coordinate='lidar'),
pos_iou_thr=0.55,
neg_iou_thr=0.55,
min_pos_iou=0.55,
ignore_iof_thr=-1),
dict( # for Car
type='Max3DIoUAssigner',
iou_calculator=dict(
type='BboxOverlaps3D', coordinate='lidar'),
pos_iou_thr=0.55,
neg_iou_thr=0.55,
min_pos_iou=0.55,
ignore_iof_thr=-1)
],
sampler=dict(
type='IoUNegPiecewiseSampler',
num=128,
pos_fraction=0.5,
neg_piece_fractions=[0.8, 0.2],
neg_iou_piece_thrs=[0.55, 0.1],
neg_pos_ub=-1,
add_gt_as_proposals=False,
return_iou=True),
cls_pos_thr=0.75,
cls_neg_thr=0.25)),
test_cfg=dict(
rpn=dict(
nms_pre=1024,
nms_post=100,
max_num=100,
nms_thr=0.7,
score_thr=0,
use_rotate_nms=True),
rcnn=dict(
use_rotate_nms=True,
use_raw_score=True,
nms_thr=0.1,
score_thr=0.1)))
train_dataloader = dict(
batch_size=2,
num_workers=2,
dataset=dict(dataset=dict(pipeline=train_pipeline, metainfo=metainfo)))
test_dataloader = dict(dataset=dict(pipeline=test_pipeline, metainfo=metainfo))
eval_dataloader = dict(dataset=dict(pipeline=test_pipeline, metainfo=metainfo))
lr = 0.001
optim_wrapper = dict(optimizer=dict(lr=lr))
param_scheduler = [
# learning rate scheduler
# During the first 16 epochs, learning rate increases from 0 to lr * 10
# during the next 24 epochs, learning rate decreases from lr * 10 to
# lr * 1e-4
dict(
type='CosineAnnealingLR',
T_max=15,
eta_min=lr * 10,
begin=0,
end=15,
by_epoch=True,
convert_to_iter_based=True),
dict(
type='CosineAnnealingLR',
T_max=25,
eta_min=lr * 1e-4,
begin=15,
end=40,
by_epoch=True,
convert_to_iter_based=True),
# momentum scheduler
# During the first 16 epochs, momentum increases from 0 to 0.85 / 0.95
# during the next 24 epochs, momentum increases from 0.85 / 0.95 to 1
dict(
type='CosineAnnealingMomentum',
T_max=15,
eta_min=0.85 / 0.95,
begin=0,
end=15,
by_epoch=True,
convert_to_iter_based=True),
dict(
type='CosineAnnealingMomentum',
T_max=25,
eta_min=1,
begin=15,
end=40,
by_epoch=True,
convert_to_iter_based=True)
]
1 change: 0 additions & 1 deletion mmdet3d/apis/inference.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -67,7 +67,6 @@ def init_model(config: Union[str, Path, Config],

if checkpoint is not None:
checkpoint = load_checkpoint(model, checkpoint, map_location='cpu')

dataset_meta = checkpoint['meta'].get('dataset_meta', None)
# save the dataset_meta in the model for convenience
if 'dataset_meta' in checkpoint.get('meta', {}):
Expand Down
4 changes: 3 additions & 1 deletion mmdet3d/models/detectors/__init__.py
View file
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@
from .mvx_two_stage import MVXTwoStageDetector
from .parta2 import PartA2
from .point_rcnn import PointRCNN
from .pv_rcnn import PointVoxelRCNN
from .sassd import SASSD
from .single_stage_mono3d import SingleStageMono3DDetector
from .smoke_mono3d import SMOKEMono3D
Expand All @@ -26,5 +27,6 @@
'DynamicMVXFasterRCNN', 'MVXFasterRCNN', 'PartA2', 'VoteNet', 'H3DNet',
'CenterPoint', 'SSD3DNet', 'ImVoteNet', 'SingleStageMono3DDetector',
'FCOSMono3D', 'ImVoxelNet', 'GroupFree3DNet', 'PointRCNN', 'SMOKEMono3D',
'SASSD', 'MinkSingleStage3DDetector', 'MultiViewDfM', 'DfM'
'SASSD', 'MinkSingleStage3DDetector', 'MultiViewDfM', 'DfM',
'PointVoxelRCNN'
]
Loading

0 comments on commit aa544d4

Please # to comment.