mmdetection3d
mmdetection3d copied to clipboard
open-mmlab
KeyError: 'TransFusionDetector is not in the mmengine::model registry. Please check whether the value of `TransFusionDetector` is correct or it was registered as expected.
Prerequisite
- [x] I have searched Issues and Discussions but cannot get the expected help.
- [x] I have read the FAQ documentation but cannot get the expected help.
- [x] The bug has not been fixed in the latest version (dev-1.x) or latest version (dev-1.0).
Task
I have modified the scripts/configs, or I'm working on my own tasks/models/datasets.
Branch
main branch https://github.com/open-mmlab/mmdetection3d
Environment
I'm using the official dockerfile
Reproduces the problem - code sample
transfusion.py inside mmdet3d/models/detectors
import mmcv
import torch
#from mmcv.runner import force_fp32
from os import path as osp
from torch import nn as nn
from torch.nn import functional as F
from mmdet3d.structures import Box3DMode, Coord3DMode, bbox3d2result
from mmdet3d.models.test_time_augs import merge_aug_bboxes_3d
from mmdet.models.utils import multi_apply
from mmdet3d.registry import MODELS
from .mvx_two_stage import MVXTwoStageDetector
@MODELS.register_module()
class TransFusionDetector(MVXTwoStageDetector):
"""Base class of Multi-modality VoxelNet."""
def __init__(self, **kwargs):
super(TransFusionDetector, self).__init__(**kwargs)
self.freeze_img = kwargs.get('freeze_img', True)
self.init_weights(pretrained=kwargs.get('pretrained', None))
def init_weights(self, pretrained=None):
"""Initialize model weights."""
super(TransFusionDetector, self).init_weights(pretrained)
if self.freeze_img:
if self.with_img_backbone:
for param in self.img_backbone.parameters():
param.requires_grad = False
if self.with_img_neck:
for param in self.img_neck.parameters():
param.requires_grad = False
def extract_img_feat(self, img, img_metas):
"""Extract features of images."""
if self.with_img_backbone and img is not None:
input_shape = img.shape[-2:]
# update real input shape of each single img
for img_meta in img_metas:
img_meta.update(input_shape=input_shape)
if img.dim() == 5 and img.size(0) == 1:
img.squeeze_(0)
elif img.dim() == 5 and img.size(0) > 1:
B, N, C, H, W = img.size()
img = img.view(B * N, C, H, W)
img_feats = self.img_backbone(img.float())
else:
return None
if self.with_img_neck:
img_feats = self.img_neck(img_feats)
return img_feats
def extract_pts_feat(self, pts, img_feats, img_metas):
"""Extract features of points."""
if not self.with_pts_bbox:
return None
voxels, num_points, coors = self.voxelize(pts)
voxel_features = self.pts_voxel_encoder(voxels, num_points, coors,
)
batch_size = coors[-1, 0] + 1
x = self.pts_middle_encoder(voxel_features, coors, batch_size)
x = self.pts_backbone(x)
if self.with_pts_neck:
x = self.pts_neck(x)
return x
@torch.no_grad()
#@force_fp32()
def voxelize(self, points):
"""Apply dynamic voxelization to points.
Args:
points (list[torch.Tensor]): Points of each sample.
Returns:
tuple[torch.Tensor]: Concatenated points, number of points
per voxel, and coordinates.
"""
voxels, coors, num_points = [], [], []
for res in points:
res_voxels, res_coors, res_num_points = self.pts_voxel_layer(res)
voxels.append(res_voxels)
coors.append(res_coors)
num_points.append(res_num_points)
voxels = torch.cat(voxels, dim=0)
num_points = torch.cat(num_points, dim=0)
coors_batch = []
for i, coor in enumerate(coors):
coor_pad = F.pad(coor, (1, 0), mode='constant', value=i)
coors_batch.append(coor_pad)
coors_batch = torch.cat(coors_batch, dim=0)
return voxels, num_points, coors_batch
def forward_train(self,
points=None,
img_metas=None,
gt_bboxes_3d=None,
gt_labels_3d=None,
gt_labels=None,
gt_bboxes=None,
img=None,
proposals=None,
gt_bboxes_ignore=None):
"""Forward training function.
Args:
points (list[torch.Tensor], optional): Points of each sample.
Defaults to None.
img_metas (list[dict], optional): Meta information of each sample.
Defaults to None.
gt_bboxes_3d (list[:obj:`BaseInstance3DBoxes`], optional):
Ground truth 3D boxes. Defaults to None.
gt_labels_3d (list[torch.Tensor], optional): Ground truth labels
of 3D boxes. Defaults to None.
gt_labels (list[torch.Tensor], optional): Ground truth labels
of 2D boxes in images. Defaults to None.
gt_bboxes (list[torch.Tensor], optional): Ground truth 2D boxes in
images. Defaults to None.
img (torch.Tensor optional): Images of each sample with shape
(N, C, H, W). Defaults to None.
proposals ([list[torch.Tensor], optional): Predicted proposals
used for training Fast RCNN. Defaults to None.
gt_bboxes_ignore (list[torch.Tensor], optional): Ground truth
2D boxes in images to be ignored. Defaults to None.
Returns:
dict: Losses of different branches.
"""
img_feats, pts_feats = self.extract_feat(
points, img=img, img_metas=img_metas)
losses = dict()
if pts_feats:
losses_pts = self.forward_pts_train(pts_feats, img_feats, gt_bboxes_3d,
gt_labels_3d, img_metas,
gt_bboxes_ignore)
losses.update(losses_pts)
if img_feats:
losses_img = self.forward_img_train(
img_feats,
img_metas=img_metas,
gt_bboxes=gt_bboxes,
gt_labels=gt_labels,
gt_bboxes_ignore=gt_bboxes_ignore,
proposals=proposals)
losses.update(losses_img)
return losses
def forward_pts_train(self,
pts_feats,
img_feats,
gt_bboxes_3d,
gt_labels_3d,
img_metas,
gt_bboxes_ignore=None):
"""Forward function for point cloud branch.
Args:
pts_feats (list[torch.Tensor]): Features of point cloud branch
gt_bboxes_3d (list[:obj:`BaseInstance3DBoxes`]): Ground truth
boxes for each sample.
gt_labels_3d (list[torch.Tensor]): Ground truth labels for
boxes of each sampole
img_metas (list[dict]): Meta information of samples.
gt_bboxes_ignore (list[torch.Tensor], optional): Ground truth
boxes to be ignored. Defaults to None.
Returns:
dict: Losses of each branch.
"""
outs = self.pts_bbox_head(pts_feats, img_feats, img_metas)
loss_inputs = [gt_bboxes_3d, gt_labels_3d, outs]
losses = self.pts_bbox_head.loss(*loss_inputs)
return losses
def simple_test_pts(self, x, x_img, img_metas, rescale=False):
"""Test function of point cloud branch."""
outs = self.pts_bbox_head(x, x_img, img_metas)
bbox_list = self.pts_bbox_head.get_bboxes(
outs, img_metas, rescale=rescale)
bbox_results = [
bbox3d2result(bboxes, scores, labels)
for bboxes, scores, labels in bbox_list
]
return bbox_results
def simple_test(self, points, img_metas, img=None, rescale=False):
"""Test function without augmentaiton."""
img_feats, pts_feats = self.extract_feat(
points, img=img, img_metas=img_metas)
bbox_list = [dict() for i in range(len(img_metas))]
if pts_feats and self.with_pts_bbox:
bbox_pts = self.simple_test_pts(
pts_feats, img_feats, img_metas, rescale=rescale)
for result_dict, pts_bbox in zip(bbox_list, bbox_pts):
result_dict['pts_bbox'] = pts_bbox
if img_feats and self.with_img_bbox:
bbox_img = self.simple_test_img(
img_feats, img_metas, rescale=rescale)
for result_dict, img_bbox in zip(bbox_list, bbox_img):
result_dict['img_bbox'] = img_bbox
return bbox_list
__init__.py inside mmdet3d/models/detectors
# Copyright (c) OpenMMLab. All rights reserved.
from .base import Base3DDetector
from .centerpoint import CenterPoint
from .dfm import DfM
from .dynamic_voxelnet import DynamicVoxelNet
from .fcos_mono3d import FCOSMono3D
from .groupfree3dnet import GroupFree3DNet
from .h3dnet import H3DNet
from .imvotenet import ImVoteNet
from .imvoxelnet import ImVoxelNet
from .mink_single_stage import MinkSingleStage3DDetector
from .multiview_dfm import MultiViewDfM
from .mvx_faster_rcnn import DynamicMVXFasterRCNN, MVXFasterRCNN
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
from .ssd3dnet import SSD3DNet
from .votenet import VoteNet
from .voxelnet import VoxelNet
from .transfusion import TransFusionDetector
__all__ = [
'Base3DDetector', 'VoxelNet', 'DynamicVoxelNet', 'MVXTwoStageDetector',
'DynamicMVXFasterRCNN', 'MVXFasterRCNN', 'PartA2', 'VoteNet', 'H3DNet',
'CenterPoint', 'SSD3DNet', 'ImVoteNet', 'SingleStageMono3DDetector',
'FCOSMono3D', 'ImVoxelNet', 'GroupFree3DNet', 'PointRCNN', 'SMOKEMono3D',
'SASSD', 'MinkSingleStage3DDetector', 'MultiViewDfM', 'DfM',
'PointVoxelRCNN',
'TransFusionDetector'
]
transfusion config file: /models/tf/transfusion_nusc_voxel_L.py
point_cloud_range = [-54.0, -54.0, -5.0, 54.0, 54.0, 3.0]
class_names = [
'car', 'truck', 'construction_vehicle', 'bus', 'trailer', 'barrier',
'motorcycle', 'bicycle', 'pedestrian', 'traffic_cone'
]
voxel_size = [0.075, 0.075, 0.2]
out_size_factor = 8
evaluation = dict(interval=1)
dataset_type = 'NuScenesDataset'
data_root = '/data/nuscenes/'
input_modality = dict(
use_lidar=True,
use_camera=False,
use_radar=False,
use_map=False,
use_external=False)
train_pipeline = [
dict(
type='LoadPointsFromFile',
coord_type='LIDAR',
load_dim=5,
use_dim=[0, 1, 2, 3, 4],
),
dict(
type='LoadPointsFromMultiSweeps',
sweeps_num=10,
use_dim=[0, 1, 2, 3, 4],
),
dict(type='LoadAnnotations3D', with_bbox_3d=True, with_label_3d=True),
dict(
type='ObjectSample',
db_sampler=dict(
data_root=None,
info_path=data_root + 'nuscenes_dbinfos_train.pkl',
rate=1.0,
prepare=dict(
filter_by_difficulty=[-1],
filter_by_min_points=dict(
car=5,
truck=5,
bus=5,
trailer=5,
construction_vehicle=5,
traffic_cone=5,
barrier=5,
motorcycle=5,
bicycle=5,
pedestrian=5)),
classes=class_names,
sample_groups=dict(
car=2,
truck=3,
construction_vehicle=7,
bus=4,
trailer=6,
barrier=2,
motorcycle=6,
bicycle=6,
pedestrian=2,
traffic_cone=2),
points_loader=dict(
type='LoadPointsFromFile',
coord_type='LIDAR',
load_dim=5,
use_dim=[0, 1, 2, 3, 4],
))),
dict(
type='GlobalRotScaleTrans',
rot_range=[-0.3925 * 2, 0.3925 * 2],
scale_ratio_range=[0.9, 1.1],
translation_std=[0.5, 0.5, 0.5]),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(type='PointsRangeFilter', point_cloud_range=point_cloud_range),
dict(type='ObjectRangeFilter', point_cloud_range=point_cloud_range),
dict(type='ObjectNameFilter', classes=class_names),
dict(type='PointShuffle'),
dict(type='DefaultFormatBundle3D', class_names=class_names),
dict(type='Collect3D', keys=['points', 'gt_bboxes_3d', 'gt_labels_3d'])
]
test_pipeline = [
dict(
type='LoadPointsFromFile',
coord_type='LIDAR',
load_dim=5,
use_dim=[0, 1, 2, 3, 4],
),
dict(
type='LoadPointsFromMultiSweeps',
sweeps_num=10,
use_dim=[0, 1, 2, 3, 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=[1.0, 1.0],
translation_std=[0, 0, 0]),
dict(type='RandomFlip3D'),
dict(
type='DefaultFormatBundle3D',
class_names=class_names,
with_label=False),
dict(type='Collect3D', keys=['points'])
])
]
data = dict(
samples_per_gpu=2,
workers_per_gpu=6,
train=dict(
type='CBGSDataset',
dataset=dict(
type=dataset_type,
data_root=data_root,
ann_file=data_root + '/nuscenes_infos_train.pkl',
load_interval=1,
pipeline=train_pipeline,
classes=class_names,
modality=input_modality,
test_mode=False,
box_type_3d='LiDAR')),
val=dict(
type=dataset_type,
data_root=data_root,
ann_file=data_root + '/nuscenes_infos_val.pkl',
load_interval=1,
pipeline=test_pipeline,
classes=class_names,
modality=input_modality,
test_mode=True,
box_type_3d='LiDAR'),
test=dict(
type=dataset_type,
data_root=data_root,
ann_file=data_root + '/nuscenes_infos_val.pkl',
load_interval=1,
pipeline=test_pipeline,
classes=class_names,
modality=input_modality,
test_mode=True,
box_type_3d='LiDAR'))
model = dict(
type='TransFusionDetector',
pts_voxel_layer=dict(
max_num_points=10,
voxel_size=voxel_size,
max_voxels=(120000, 160000),
point_cloud_range=point_cloud_range),
pts_voxel_encoder=dict(
type='HardSimpleVFE',
num_features=5,
),
pts_middle_encoder=dict(
type='SparseEncoder',
in_channels=5,
sparse_shape=[41, 1440, 1440],
output_channels=128,
order=('conv', 'norm', 'act'),
encoder_channels=((16, 16, 32), (32, 32, 64), (64, 64, 128), (128, 128)),
encoder_paddings=((0, 0, 1), (0, 0, 1), (0, 0, [0, 1, 1]), (0, 0)),
block_type='basicblock'),
pts_backbone=dict(
type='SECOND',
in_channels=256,
out_channels=[128, 256],
layer_nums=[5, 5],
layer_strides=[1, 2],
norm_cfg=dict(type='BN', eps=0.001, momentum=0.01),
conv_cfg=dict(type='Conv2d', bias=False)),
pts_neck=dict(
type='SECONDFPN',
in_channels=[128, 256],
out_channels=[256, 256],
upsample_strides=[1, 2],
norm_cfg=dict(type='BN', eps=0.001, momentum=0.01),
upsample_cfg=dict(type='deconv', bias=False),
use_conv_for_no_stride=True),
pts_bbox_head=dict(
type='TransFusionHead',
num_proposals=200,
auxiliary=True,
in_channels=256 * 2,
hidden_channel=128,
num_classes=len(class_names),
num_decoder_layers=1,
num_heads=8,
learnable_query_pos=False,
initialize_by_heatmap=True,
nms_kernel_size=3,
ffn_channel=256,
dropout=0.1,
bn_momentum=0.1,
activation='relu',
common_heads=dict(center=(2, 2), height=(1, 2), dim=(3, 2), rot=(2, 2), vel=(2, 2)),
bbox_coder=dict(
type='TransFusionBBoxCoder',
pc_range=point_cloud_range[:2],
voxel_size=voxel_size[:2],
out_size_factor=out_size_factor,
post_center_range=[-61.2, -61.2, -10.0, 61.2, 61.2, 10.0],
score_threshold=0.0,
code_size=10,
),
loss_cls=dict(type='FocalLoss', use_sigmoid=True, gamma=2, alpha=0.25, reduction='mean', loss_weight=1.0),
# loss_iou=dict(type='CrossEntropyLoss', use_sigmoid=True, reduction='mean', loss_weight=0.0),
loss_bbox=dict(type='L1Loss', reduction='mean', loss_weight=0.25),
loss_heatmap=dict(type='GaussianFocalLoss', reduction='mean', loss_weight=1.0),
),
train_cfg=dict(
pts=dict(
dataset='nuScenes',
assigner=dict(
type='HungarianAssigner3D',
iou_calculator=dict(type='BboxOverlaps3D', coordinate='lidar'),
cls_cost=dict(type='FocalLossCost', gamma=2, alpha=0.25, weight=0.15),
reg_cost=dict(type='BBoxBEVL1Cost', weight=0.25),
iou_cost=dict(type='IoU3DCost', weight=0.25)
),
pos_weight=-1,
gaussian_overlap=0.1,
min_radius=2,
grid_size=[1440, 1440, 40], # [x_len, y_len, 1]
voxel_size=voxel_size,
out_size_factor=out_size_factor,
code_weights=[1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.2, 0.2],
point_cloud_range=point_cloud_range)),
test_cfg=dict(
pts=dict(
dataset='nuScenes',
grid_size=[1440, 1440, 40],
out_size_factor=out_size_factor,
pc_range=point_cloud_range[0:2],
voxel_size=voxel_size[:2],
nms_type=None,
)))
optimizer = dict(type='AdamW', lr=0.0001, weight_decay=0.01) # for 8gpu * 2sample_per_gpu
optimizer_config = dict(grad_clip=dict(max_norm=0.1, norm_type=2))
lr_config = dict(
policy='cyclic',
target_ratio=(10, 0.0001),
cyclic_times=1,
step_ratio_up=0.4)
momentum_config = dict(
policy='cyclic',
target_ratio=(0.8947368421052632, 1),
cyclic_times=1,
step_ratio_up=0.4)
total_epochs = 20
checkpoint_config = dict(interval=1)
log_config = dict(
interval=50,
hooks=[dict(type='TextLoggerHook'),
dict(type='TensorboardLoggerHook')])
dist_params = dict(backend='nccl')
log_level = 'INFO'
work_dir = None
load_from = None
resume_from = None
workflow = [('train', 1)]
gpu_ids = range(0, 8)
Reproduces the problem - command or script
python tools/train.py /models/tf/transfusion_nusc_voxel_L.py
Reproduces the problem - error message
Traceback (most recent call last):
File "tools/train.py", line 147, in <module>
main()
File "tools/train.py", line 135, in main
runner = Runner.from_cfg(cfg)
File "/opt/conda/lib/python3.7/site-packages/mmengine/runner/runner.py", line 489, in from_cfg
cfg=cfg,
File "/opt/conda/lib/python3.7/site-packages/mmengine/runner/runner.py", line 429, in __init__
self.model = self.build_model(model)
File "/opt/conda/lib/python3.7/site-packages/mmengine/runner/runner.py", line 836, in build_model
model = MODELS.build(model)
File "/opt/conda/lib/python3.7/site-packages/mmengine/registry/registry.py", line 570, in build
return self.build_func(cfg, *args, **kwargs, registry=self)
File "/opt/conda/lib/python3.7/site-packages/mmengine/registry/build_functions.py", line 232, in build_model_from_cfg
return build_from_cfg(cfg, registry, default_args)
File "/opt/conda/lib/python3.7/site-packages/mmengine/registry/build_functions.py", line 101, in build_from_cfg
f'{obj_type} is not in the {registry.scope}::{registry.name} registry. ' # noqa: E501
KeyError: 'TransFusionDetector is not in the mmengine::model registry. Please check whether the value of `TransFusionDetector` is correct or it was registered as expected. More details can be found at https://mmengine.readthedocs.io/en/latest/advanced_tutorials/config.html#import-the-custom-module'
Additional information
I'm trying to integrate the TransFusion model into the most recent version of MMDetection3D. To do so, I'm adding the required Python files into mmdet3d/models, following the instructions in customize_models.md.
As can be seen in the code I’ve shared, I import MODELS from mmdet3d.registry and add the line:
@MODELS.register_module()
before defining the class. I’ve also modified the __init __.pyfile to include this new detector.
However, when I run the training script, I get an error saying that the detector is not in the registry. It’s not clear to me what I’m doing wrong. Is there any step I might have missed? Could someone explain how to fix this?
Thanks in advance for your help.
