s3dis

[Wang-hui-001]

Hello, I would like to train solely on the S3DIS dataset, but my results are extremely poor. Could you please help me understand why? To train on S3DIS only, I should only need to modify the configuration file, right? I’ve made changes to the config as follows:
base = ['mmdet3d::base/default_runtime.py']
custom_imports = dict(imports=['unidet3d'])

classes_scannet = ['cabinet', 'bed', 'chair', 'sofa', 'table', 'door', 'window', 'bookshelf',
'picture', 'counter', 'desk', 'curtain', 'refrigerator', 'showercurtrain',
'toilet', 'sink', 'bathtub', 'otherfurniture']
classes_s3dis = ['table', 'chair', 'sofa', 'bookcase', 'board']

classes_multiscan = ['door', 'table', 'chair', 'cabinet', 'window', 'sofa', 'microwave', 'pillow',
'tv_monitor', 'curtain', 'trash_can', 'suitcase', 'sink', 'backpack', 'bed',
'refrigerator', 'toilet']

classes_3rscan = classes_scannet

classes_scannetpp = ['table', 'door', 'ceiling lamp', 'cabinet', 'blinds', 'curtain', 'chair', 'storage cabinet', 'office chair', 'bookshelf', 'whiteboard', 'window', 'box',
'monitor', 'shelf', 'heater', 'kitchen cabinet', 'sofa', 'bed', 'trash can', 'book', 'plant', 'blanket', 'tv', 'computer tower', 'refrigerator', 'jacket',
'sink', 'bag', 'picture', 'pillow', 'towel', 'suitcase', 'backpack', 'crate', 'keyboard', 'rack', 'toilet', 'printer', 'poster', 'painting', 'microwave', 'shoes',
'socket', 'bottle', 'bucket', 'cushion', 'basket', 'shoe rack', 'telephone', 'file folder', 'laptop', 'plant pot', 'exhaust fan', 'cup', 'coat hanger', 'light switch',
'speaker', 'table lamp', 'kettle', 'smoke detector', 'container', 'power strip', 'slippers', 'paper bag', 'mouse', 'cutting board', 'toilet paper', 'paper towel',
'pot', 'clock', 'pan', 'tap', 'jar', 'soap dispenser', 'binder', 'bowl', 'tissue box', 'whiteboard eraser', 'toilet brush', 'spray bottle', 'headphones', 'stapler', 'marker']

classes_arkitscenes = ['cabinet', 'refrigerator', 'shelf', 'stove', 'bed',
'sink', 'washer', 'toilet', 'bathtub', 'oven',
'dishwasher', 'fireplace', 'stool', 'chair', 'table',
'tv_monitor', 'sofa']

model settings

num_channels=32
voxel_size=0.02

model = dict(
type='UniDet3D',
data_preprocessor=dict(type='Det3DDataPreprocessor_'),
in_channels=6,
num_channels=num_channels,
voxel_size=voxel_size,
min_spatial_shape=128,
query_thr=3000,
bbox_by_mask=[True],
target_by_distance=[False],
use_superpoints=[True],
fast_nms=[True],
backbone=dict(
type='SpConvUNet',
num_planes=[num_channels * (i + 1) for i in range(5)],
return_blocks=True),
decoder=dict(
type='UniDet3DEncoder',
num_layers=6,
datasets_classes=[classes_s3dis],
in_channels=num_channels,
d_model=256,
num_heads=8,
hidden_dim=1024,
dropout=0.0,
activation_fn='gelu',
datasets=['s3dis'],
angles=[False]),
criterion=dict(
type='UniDet3DCriterion',
datasets=['s3dis'],
datasets_weights=[1],
bbox_loss_simple=dict(
type='UniDet3DAxisAlignedIoULoss',
mode='diou',
reduction='none'),
bbox_loss_rotated=dict(
type='UniDet3DRotatedIoU3DLoss',
mode='diou',
reduction='none'),
matcher=dict(
type='UniMatcher',
costs=[
dict(type='QueryClassificationCost', weight=0.5),
dict(type='BboxCostJointTraining',
weight=2.0,
loss_simple=dict(
type='UniDet3DAxisAlignedIoULoss',
mode='diou',
reduction='none'),
loss_rotated=dict(
type='UniDet3DRotatedIoU3DLoss',
mode='diou',
reduction='none'))]),
loss_weight=[0.5, 1.0],
non_object_weight=0.1,
topk=[6],
iter_matcher=True),
train_cfg=dict(topk=6),
test_cfg=dict(
low_sp_thr=0.18,
up_sp_thr=0.81,
topk_insts=1000,
score_thr=0,
iou_thr=[0.5]))

scannet dataset settings

metainfo_scannet = dict(classes=classes_scannet)
data_root_scannet = 'data/scannet/'

max_class_scannet = 20
dataset_type_scannet = 'ScanNetDetDataset'
data_prefix_scannet = dict(
pts='points',
pts_instance_mask='instance_mask',
pts_semantic_mask='semantic_mask',
sp_pts_mask='super_points')

train_pipeline_scannet = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(
type='LoadAnnotations3D_',
with_bbox_3d=False,
with_label_3d=False,
with_mask_3d=True,
with_seg_3d=True,
with_sp_mask_3d=True),
dict(type='GlobalAlignment', rotation_axis=2),
dict(type='PointSegClassMapping'),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[-3.14, 3.14],
scale_ratio_range=[0.8, 1.2],
translation_std=[0.1, 0.1, 0.1],
shift_height=False),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='PointDetClassMappingScanNet',
num_classes=max_class_scannet,
stuff_classes=[0, 1]),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=0.5),
dict(
type='Pack3DDetInputs_',
keys=[
'points', 'gt_labels_3d', 'pts_semantic_mask', 'pts_instance_mask',
'sp_pts_mask', 'gt_sp_masks', 'elastic_coords'
])
]
test_pipeline_scannet = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(
type='LoadAnnotations3D_',
with_bbox_3d=False,
with_label_3d=False,
with_mask_3d=True,
with_seg_3d=True,
with_sp_mask_3d=True),
dict(type='GlobalAlignment', rotation_axis=2),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

s3dis dataset settings

dataset_type_s3dis = 'S3DISSegDetDataset'
data_root_s3dis = 'data/s3dis/'
data_prefix_s3dis = dict(
pts='points',
pts_instance_mask='instance_mask',
pts_semantic_mask='semantic_mask',
sp_pts_mask='super_points')

train_area = [1, 2, 3, 4, 6]
test_area = 5

train_pipeline_s3dis = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(
type='LoadAnnotations3D_',
with_label_3d=False,
with_bbox_3d=False,
with_mask_3d=True,
with_seg_3d=True,
with_sp_mask_3d=True),
dict(
type='PointSample_',
num_points=180000),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[0.0, 0.0],
scale_ratio_range=[0.9, 1.1],
translation_std=[.1, .1, .1],
shift_height=False),
dict(type='PointDetClassMappingS3DIS',
classes=[7, 8, 9, 10, 11]),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=-1),
dict(
type='Pack3DDetInputs_',
keys=[
'points', 'elastic_coords', 'gt_labels_3d',
'sp_pts_mask', 'gt_sp_masks',
'pts_semantic_mask', 'pts_instance_mask'
])
]
test_pipeline_s3dis = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(
type='LoadAnnotations3D_',
with_bbox_3d=False,
with_label_3d=False,
with_mask_3d=True,
with_seg_3d=True,
with_sp_mask_3d=True),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(
type='PointSample_',
num_points=180000),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

multiscan dataset settings

data_root_multiscan = 'data/multiscan/bins'
dataset_type_multiscan = 'MultiScan_'
data_prefix_multiscan = dict(
pts='points',
pts_instance_mask='instance_mask',
pts_semantic_mask='semantic_mask',
sp_pts_mask='super_points')

train_pipeline_multiscan = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=True,
with_bbox_3d=True,
with_sp_mask_3d=True),
dict(type='PointSample_',
num_points=100000),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[0, 0],
scale_ratio_range=[0.9, 1.1],
translation_std=[0.1, 0.1, 0.1],
shift_height=False),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=-1),
dict(
type='Pack3DDetInputs_',
keys=['points', 'elastic_coords', 'gt_bboxes_3d',
'gt_labels_3d', 'sp_pts_mask'])
]
test_pipeline_multiscan = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=False,
with_bbox_3d=False,
with_sp_mask_3d=True),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(type='PointSample_', num_points=100000),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])
]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

3rscan dataset settings

data_root_3rscan = 'data/3rscan'
dataset_type_3rscan = 'ThreeRScan_'
data_prefix_3rscan = dict(
pts='points',
pts_instance_mask='instance_mask',
pts_semantic_mask='semantic_mask',
sp_pts_mask='super_points_spt')

train_pipeline_3rscan = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=True,
with_bbox_3d=True,
with_sp_mask_3d=True),
dict(type='PointSample_',
num_points=100000),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[0, 0],
scale_ratio_range=[0.9, 1.1],
translation_std=[0.1, 0.1, 0.1],
shift_height=False),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=-1),
dict(
type='Pack3DDetInputs_',
keys=['points', 'elastic_coords', 'gt_bboxes_3d',
'gt_labels_3d', 'sp_pts_mask'])
]
test_pipeline_3rscan = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=False,
with_bbox_3d=False,
with_sp_mask_3d=True),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(type='PointSample_', num_points=100000),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])
]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

scannetpp dataset settings

data_root_scannetpp = 'data/scannetpp/bins'
dataset_type_scannetpp = 'Scannetpp_'
data_prefix_scannetpp = dict(
pts='points',
pts_instance_mask='instance_mask',
pts_semantic_mask='semantic_mask',
sp_pts_mask='super_points_spt')

train_pipeline_scannetpp = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=True,
with_bbox_3d=True,
with_sp_mask_3d=True),
dict(type='PointSample_',
num_points=200000),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[0, 0],
scale_ratio_range=[0.9, 1.1],
translation_std=[0.1, 0.1, 0.1],
shift_height=False),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=-1),
dict(
type='Pack3DDetInputs_',
keys=['points', 'elastic_coords', 'gt_bboxes_3d',
'gt_labels_3d', 'sp_pts_mask'])
]
test_pipeline_scannetpp = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=False,
with_bbox_3d=False,
with_sp_mask_3d=True),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(type='PointSample_', num_points=200000),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])
]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

arkitscenes dataset settings

dataset_type_arkitscenes = 'ARKitScenesOfflineDataset'
data_root_arkitscenes = 'data/arkitscenes'
data_prefix_arkitscenes = dict(
pts='offline_prepared_data',
sp_pts_mask='super_points')

train_pipeline_arkitscenes = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=True,
with_bbox_3d=True,
with_sp_mask_3d=True),
dict(type='PointSample_', num_points=100000),
dict(
type='DenormalizePointsColor',
color_mean=[0, 0, 0],
color_std=[255, 255, 255]),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[-0.5, 0.5],
scale_ratio_range=[0.9, 1.1],
translation_std=[0.1, 0.1, 0.1],
shift_height=False),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=-1),
dict(
type='Pack3DDetInputs_',
keys=['points', 'elastic_coords', 'gt_bboxes_3d',
'gt_labels_3d', 'sp_pts_mask'])
]
test_pipeline_arkitscenes = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=False,
with_bbox_3d=False,
with_sp_mask_3d=True),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(type='PointSample_', num_points=100000),
dict(
type='DenormalizePointsColor',
color_mean=[0, 0, 0],
color_std=[255, 255, 255]),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])
]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

run settings

train_dataloader = dict(
batch_size=8,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type='ConcatDataset_',
# datasets=[dict(
# type=dataset_type_scannet,
# ann_file='scannet_infos_train.pkl',
# data_prefix=data_prefix_scannet,
# data_root=data_root_scannet,
# metainfo=metainfo_scannet,
# pipeline=train_pipeline_scannet,
# ignore_index=max_class_scannet,
# scene_idxs=None,
# test_mode=False)] #+
datasets=[dict(
type=dataset_type_s3dis,
data_root=data_root_s3dis,
ann_file=f's3dis_sp_infos_Area_{i}.pkl',
partition=0.33,
pipeline=train_pipeline_s3dis,
filter_empty_gt=True,
data_prefix=data_prefix_s3dis,
box_type_3d='Depth',
backend_args=None) for i in train_area]
# [dict(
# type=dataset_type_multiscan,
# ann_file='multiscan_infos_train.pkl',
# partition=0.25,
# data_prefix=data_prefix_multiscan,
# data_root=data_root_multiscan,
# pipeline=train_pipeline_multiscan,
# test_mode=False)] +
# [dict(
# type=dataset_type_3rscan,
# ann_file='3rscan_infos_train.pkl',
# partition=0.15,
# data_prefix=data_prefix_3rscan,
# data_root=data_root_3rscan,
# pipeline=train_pipeline_3rscan,
# test_mode=False)] +
# [dict(
# type=dataset_type_scannetpp,
# ann_file='scannetpp_infos_train.pkl',
# partition=0.33,
# data_prefix=data_prefix_scannetpp,
# data_root=data_root_scannetpp,
# pipeline=train_pipeline_scannetpp,
# test_mode=False)] +
# [dict(
# type=dataset_type_arkitscenes,
# ann_file='arkitscenes_offline_infos_train.pkl',
# partition=0.08,
# data_prefix=data_prefix_arkitscenes,
# data_root=data_root_arkitscenes,
# pipeline=train_pipeline_arkitscenes,
# test_mode=False)]
))

val_dataloader = dict(
batch_size=1,
num_workers=1,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type='ConcatDataset_',
datasets=
# [dict(
# type=dataset_type_scannet,
# ann_file='scannet_infos_val.pkl',
# data_prefix=data_prefix_scannet,
# data_root=data_root_scannet,
# metainfo=metainfo_scannet,
# pipeline=test_pipeline_scannet,
# ignore_index=max_class_scannet,
# test_mode=True)] #+
[dict(
type=dataset_type_s3dis,
data_root=data_root_s3dis,
ann_file=f's3dis_sp_infos_Area_{test_area}.pkl',
pipeline=test_pipeline_s3dis,
test_mode=True,
data_prefix=data_prefix_s3dis,
box_type_3d='Depth',
backend_args=None)]
# [dict(
# type=dataset_type_multiscan,
# ann_file='multiscan_infos_val.pkl',
# data_prefix=data_prefix_multiscan,
# data_root=data_root_multiscan,
# pipeline=test_pipeline_multiscan,
# test_mode=True)] +
# [dict(
# type=dataset_type_3rscan,
# ann_file='3rscan_infos_val.pkl',
# data_prefix=data_prefix_3rscan,
# data_root=data_root_3rscan,
# pipeline=test_pipeline_3rscan,
# test_mode=True)] +
# [dict(
# type=dataset_type_scannetpp,
# ann_file='scannetpp_infos_val.pkl',
# data_prefix=data_prefix_scannetpp,
# data_root=data_root_scannetpp,
# pipeline=test_pipeline_scannetpp,
# test_mode=True)] +
# [dict(
# type=dataset_type_arkitscenes,
# ann_file='arkitscenes_offline_infos_val.pkl',
# data_prefix=data_prefix_arkitscenes,
# data_root=data_root_arkitscenes,
# pipeline=test_pipeline_arkitscenes,
# test_mode=True)]
))

test_dataloader = val_dataloader

load_from = 'work_dirs/tmp/oneformer3d_1xb4_scannet.pth'

test_evaluator = dict(type='IndoorMetric_',
datasets=['s3dis'],
datasets_classes=[classes_s3dis])

val_evaluator = test_evaluator

optim_wrapper = dict(
type='OptimWrapper',
optimizer=dict(type='AdamW', lr=0.0001 * 2*2, weight_decay=0.05),
clip_grad=dict(max_norm=10, norm_type=2))

param_scheduler = dict(type='PolyLR', begin=0, end=1024, power=0.9)

custom_hooks = [dict(type='EmptyCacheHook', after_iter=True)]
default_hooks = dict(
checkpoint=dict(interval=1, max_keep_ckpts=16))

train_cfg = dict(
type='EpochBasedTrainLoop',
max_epochs=1024,
dynamic_intervals=[(1, 16), (1024 - 16, 1)])
val_cfg = dict(type='ValLoop')
test_cfg = dict(type='TestLoop')

[filaPro]

Is it fine when training on all 6 datasets?
Also I noticed that you modified learning rate. Not sure if it is a good idea.

[Wang-hui-001]

Hello, I have not tried running the experiments on all six datasets, but I can confirm that my setup works correctly on the ScanNet dataset alone.
Now, for the paper I need to add a second dataset, so I would like to report results on S3DIS. I followed the instructions in the config file: I commented out all other datasets and kept only the S3DIS section, using exactly the same data paths and hyper-parameters that you provided.
Your published mAP@0.25 is 63.6, whereas my reproduction only reaches 5.2.
Could you please advise what might cause this huge gap and suggest a solution?

[col14m]

Hi! Could you test the released model trained on 6 datasets only on s3dis?

You only need to change this part of the config:

val_dataloader = dict(
    batch_size=1,
    num_workers=1,
    persistent_workers=True,
    drop_last=False,
    sampler=dict(type='DefaultSampler', shuffle=False),
    dataset=dict(
        type='ConcatDataset_',
        datasets= \
                # [dict(
                #     type=dataset_type_scannet,
                #     ann_file='scannet_infos_val.pkl',
                #     data_prefix=data_prefix_scannet,
                #     data_root=data_root_scannet,
                #     metainfo=metainfo_scannet,
                #     pipeline=test_pipeline_scannet,
                #     ignore_index=max_class_scannet,
                #     test_mode=True)] + \
                [dict(
                    type=dataset_type_s3dis,
                    data_root=data_root_s3dis,
                    ann_file=f's3dis_sp_infos_Area_{test_area}.pkl',
                    pipeline=test_pipeline_s3dis,
                    test_mode=True,
                    data_prefix=data_prefix_s3dis,
                    box_type_3d='Depth',
                    backend_args=None)] #+ \
                # [dict(
                #     type=dataset_type_multiscan,
                #     ann_file='multiscan_infos_val.pkl',
                #     data_prefix=data_prefix_multiscan,
                #     data_root=data_root_multiscan,
                #     pipeline=test_pipeline_multiscan,
                #     test_mode=True)] + \
                # [dict(
                #     type=dataset_type_3rscan,
                #     ann_file='3rscan_infos_val.pkl',
                #     data_prefix=data_prefix_3rscan,
                #     data_root=data_root_3rscan,
                #     pipeline=test_pipeline_3rscan,
                #     test_mode=True)] + \
                # [dict(
                #     type=dataset_type_scannetpp,
                #     ann_file='scannetpp_infos_val.pkl',
                #     data_prefix=data_prefix_scannetpp,
                #     data_root=data_root_scannetpp,
                #     pipeline=test_pipeline_scannetpp,
                #     test_mode=True)] + \
                # [dict(
                #     type=dataset_type_arkitscenes,
                #     ann_file='arkitscenes_offline_infos_val.pkl',
                #     data_prefix=data_prefix_arkitscenes,
                #     data_root=data_root_arkitscenes,
                #     pipeline=test_pipeline_arkitscenes,
                #     test_mode=True)] 
                    ))

[Wang-hui-001]

Hello, thank you very much for providing the solution. I have conducted multiple tests following your suggestions, but the results were not satisfactory. Below is the code I modified according to your advice.
base = ['mmdet3d::base/default_runtime.py']
custom_imports = dict(imports=['unidet3d'])

classes_scannet = ['cabinet', 'bed', 'chair', 'sofa', 'table', 'door', 'window', 'bookshelf',
'picture', 'counter', 'desk', 'curtain', 'refrigerator', 'showercurtrain',
'toilet', 'sink', 'bathtub', 'otherfurniture']
classes_s3dis = ['table', 'chair', 'sofa', 'bookcase', 'board']

classes_multiscan = ['door', 'table', 'chair', 'cabinet', 'window', 'sofa', 'microwave', 'pillow',
'tv_monitor', 'curtain', 'trash_can', 'suitcase', 'sink', 'backpack', 'bed',
'refrigerator', 'toilet']

classes_3rscan = classes_scannet

classes_scannetpp = ['table', 'door', 'ceiling lamp', 'cabinet', 'blinds', 'curtain', 'chair', 'storage cabinet', 'office chair', 'bookshelf', 'whiteboard', 'window', 'box',
'monitor', 'shelf', 'heater', 'kitchen cabinet', 'sofa', 'bed', 'trash can', 'book', 'plant', 'blanket', 'tv', 'computer tower', 'refrigerator', 'jacket',
'sink', 'bag', 'picture', 'pillow', 'towel', 'suitcase', 'backpack', 'crate', 'keyboard', 'rack', 'toilet', 'printer', 'poster', 'painting', 'microwave', 'shoes',
'socket', 'bottle', 'bucket', 'cushion', 'basket', 'shoe rack', 'telephone', 'file folder', 'laptop', 'plant pot', 'exhaust fan', 'cup', 'coat hanger', 'light switch',
'speaker', 'table lamp', 'kettle', 'smoke detector', 'container', 'power strip', 'slippers', 'paper bag', 'mouse', 'cutting board', 'toilet paper', 'paper towel',
'pot', 'clock', 'pan', 'tap', 'jar', 'soap dispenser', 'binder', 'bowl', 'tissue box', 'whiteboard eraser', 'toilet brush', 'spray bottle', 'headphones', 'stapler', 'marker']

classes_arkitscenes = ['cabinet', 'refrigerator', 'shelf', 'stove', 'bed',
'sink', 'washer', 'toilet', 'bathtub', 'oven',
'dishwasher', 'fireplace', 'stool', 'chair', 'table',
'tv_monitor', 'sofa']

model settings

num_channels=32
voxel_size=0.02

model = dict(
type='UniDet3D',
data_preprocessor=dict(type='Det3DDataPreprocessor_'),
in_channels=6,
num_channels=num_channels,
voxel_size=voxel_size,
min_spatial_shape=128,
query_thr=3000,
bbox_by_mask=[True, True, False, False, False, False],
target_by_distance=[False, False, True, True, True, True],
use_superpoints=[True, True, True, False, False, False],
fast_nms=[True, False, True, True, True, None],
backbone=dict(
type='SpConvUNet',
num_planes=[num_channels * (i + 1) for i in range(5)],
return_blocks=True),
decoder=dict(
type='UniDet3DEncoder',
num_layers=6,
datasets_classes=[classes_scannet, classes_s3dis,
classes_multiscan, classes_3rscan,
classes_scannetpp, classes_arkitscenes],
in_channels=num_channels,
d_model=256,
num_heads=8,
hidden_dim=1024,
dropout=0.0,
activation_fn='gelu',
datasets=['scannet', 's3dis', 'multiscan', '3rscan',
'scannetpp', 'arkitscenes'],
angles=[False, False, False, False, False, True]),
criterion=dict(
type='UniDet3DCriterion',
datasets=['scannet', 's3dis', 'multiscan', '3rscan',
'scannetpp', 'arkitscenes'],
datasets_weights=[1, 1, 1, 1, 1, 1],
bbox_loss_simple=dict(
type='UniDet3DAxisAlignedIoULoss',
mode='diou',
reduction='none'),
bbox_loss_rotated=dict(
type='UniDet3DRotatedIoU3DLoss',
mode='diou',
reduction='none'),
matcher=dict(
type='UniMatcher',
costs=[
dict(type='QueryClassificationCost', weight=0.5),
dict(type='BboxCostJointTraining',
weight=2.0,
loss_simple=dict(
type='UniDet3DAxisAlignedIoULoss',
mode='diou',
reduction='none'),
loss_rotated=dict(
type='UniDet3DRotatedIoU3DLoss',
mode='diou',
reduction='none'))]),
loss_weight=[0.5, 1.0],
non_object_weight=0.1,
topk=[6, 6, 3, 3, 3, 3],
iter_matcher=True),
train_cfg=dict(topk=6),
test_cfg=dict(
low_sp_thr=0.18,
up_sp_thr=0.81,
topk_insts=1000,
score_thr=0,
iou_thr=[0.5, 0.55, 0.55, 0.55, 0.55, 0.55]))

scannet dataset settings

metainfo_scannet = dict(classes=classes_scannet)
data_root_scannet = '/home/wanghui/DownLoads/unidet3d-master/data/scannet/'

max_class_scannet = 20
dataset_type_scannet = 'ScanNetDetDataset'
data_prefix_scannet = dict(
pts='points',
pts_instance_mask='instance_mask',
pts_semantic_mask='semantic_mask',
sp_pts_mask='super_points')

train_pipeline_scannet = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(
type='LoadAnnotations3D_',
with_bbox_3d=False,
with_label_3d=False,
with_mask_3d=True,
with_seg_3d=True,
with_sp_mask_3d=True),
dict(type='GlobalAlignment', rotation_axis=2),
dict(type='PointSegClassMapping'),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[-3.14, 3.14],
scale_ratio_range=[0.8, 1.2],
translation_std=[0.1, 0.1, 0.1],
shift_height=False),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='PointDetClassMappingScanNet',
num_classes=max_class_scannet,
stuff_classes=[0, 1]),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=0.5),
dict(
type='Pack3DDetInputs_',
keys=[
'points', 'gt_labels_3d', 'pts_semantic_mask', 'pts_instance_mask',
'sp_pts_mask', 'gt_sp_masks', 'elastic_coords'
])
]
test_pipeline_scannet = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(
type='LoadAnnotations3D_',
with_bbox_3d=False,
with_label_3d=False,
with_mask_3d=True,
with_seg_3d=True,
with_sp_mask_3d=True),
dict(type='GlobalAlignment', rotation_axis=2),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

s3dis dataset settings

dataset_type_s3dis = 'S3DISSegDetDataset'
data_root_s3dis = '/home/wanghui/DownLoads/unidet3d-master/data/s3dis/'
data_prefix_s3dis = dict(
pts='points',
pts_instance_mask='instance_mask',
pts_semantic_mask='semantic_mask',
sp_pts_mask='super_points')

train_area = [1, 2, 3, 4, 6]
test_area = 5

train_pipeline_s3dis = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(
type='LoadAnnotations3D_',
with_label_3d=False,
with_bbox_3d=False,
with_mask_3d=True,
with_seg_3d=True,
with_sp_mask_3d=True),
dict(
type='PointSample_',
num_points=180000),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[0.0, 0.0],
scale_ratio_range=[0.9, 1.1],
translation_std=[.1, .1, .1],
shift_height=False),
dict(type='PointDetClassMappingS3DIS',
classes=[7, 8, 9, 10, 11]),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=-1),
dict(
type='Pack3DDetInputs_',
keys=[
'points', 'elastic_coords', 'gt_labels_3d',
'sp_pts_mask', 'gt_sp_masks',
'pts_semantic_mask', 'pts_instance_mask'
])
]
test_pipeline_s3dis = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(
type='LoadAnnotations3D_',
with_bbox_3d=False,
with_label_3d=False,
with_mask_3d=True,
with_seg_3d=True,
with_sp_mask_3d=True),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(
type='PointSample_',
num_points=180000),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

multiscan dataset settings

data_root_multiscan = 'data/multiscan/bins'
dataset_type_multiscan = 'MultiScan_'
data_prefix_multiscan = dict(
pts='points',
pts_instance_mask='instance_mask',
pts_semantic_mask='semantic_mask',
sp_pts_mask='super_points')

train_pipeline_multiscan = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=True,
with_bbox_3d=True,
with_sp_mask_3d=True),
dict(type='PointSample_',
num_points=100000),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[0, 0],
scale_ratio_range=[0.9, 1.1],
translation_std=[0.1, 0.1, 0.1],
shift_height=False),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=-1),
dict(
type='Pack3DDetInputs_',
keys=['points', 'elastic_coords', 'gt_bboxes_3d',
'gt_labels_3d', 'sp_pts_mask'])
]
test_pipeline_multiscan = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=False,
with_bbox_3d=False,
with_sp_mask_3d=True),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(type='PointSample_', num_points=100000),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])
]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

3rscan dataset settings

data_root_3rscan = 'data/3rscan'
dataset_type_3rscan = 'ThreeRScan_'
data_prefix_3rscan = dict(
pts='points',
pts_instance_mask='instance_mask',
pts_semantic_mask='semantic_mask',
sp_pts_mask='super_points_spt')

train_pipeline_3rscan = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=True,
with_bbox_3d=True,
with_sp_mask_3d=True),
dict(type='PointSample_',
num_points=100000),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[0, 0],
scale_ratio_range=[0.9, 1.1],
translation_std=[0.1, 0.1, 0.1],
shift_height=False),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=-1),
dict(
type='Pack3DDetInputs_',
keys=['points', 'elastic_coords', 'gt_bboxes_3d',
'gt_labels_3d', 'sp_pts_mask'])
]
test_pipeline_3rscan = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=False,
with_bbox_3d=False,
with_sp_mask_3d=True),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(type='PointSample_', num_points=100000),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])
]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

scannetpp dataset settings

data_root_scannetpp = 'data/scannetpp/bins'
dataset_type_scannetpp = 'Scannetpp_'
data_prefix_scannetpp = dict(
pts='points',
pts_instance_mask='instance_mask',
pts_semantic_mask='semantic_mask',
sp_pts_mask='super_points_spt')

train_pipeline_scannetpp = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=True,
with_bbox_3d=True,
with_sp_mask_3d=True),
dict(type='PointSample_',
num_points=200000),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[0, 0],
scale_ratio_range=[0.9, 1.1],
translation_std=[0.1, 0.1, 0.1],
shift_height=False),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=-1),
dict(
type='Pack3DDetInputs_',
keys=['points', 'elastic_coords', 'gt_bboxes_3d',
'gt_labels_3d', 'sp_pts_mask'])
]
test_pipeline_scannetpp = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=False,
with_bbox_3d=False,
with_sp_mask_3d=True),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(type='PointSample_', num_points=200000),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])
]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

arkitscenes dataset settings

dataset_type_arkitscenes = 'ARKitScenesOfflineDataset'
data_root_arkitscenes = 'data/arkitscenes'
data_prefix_arkitscenes = dict(
pts='offline_prepared_data',
sp_pts_mask='super_points')

train_pipeline_arkitscenes = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=True,
with_bbox_3d=True,
with_sp_mask_3d=True),
dict(type='PointSample_', num_points=100000),
dict(
type='DenormalizePointsColor',
color_mean=[0, 0, 0],
color_std=[255, 255, 255]),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5]),
dict(
type='RandomFlip3D',
sync_2d=False,
flip_ratio_bev_horizontal=0.5,
flip_ratio_bev_vertical=0.5),
dict(
type='GlobalRotScaleTrans',
rot_range=[-0.5, 0.5],
scale_ratio_range=[0.9, 1.1],
translation_std=[0.1, 0.1, 0.1],
shift_height=False),
dict(
type='ElasticTransfrom',
gran=[6, 20],
mag=[40, 160],
voxel_size=voxel_size,
p=-1),
dict(
type='Pack3DDetInputs_',
keys=['points', 'elastic_coords', 'gt_bboxes_3d',
'gt_labels_3d', 'sp_pts_mask'])
]
test_pipeline_arkitscenes = [
dict(
type='LoadPointsFromFile',
coord_type='DEPTH',
shift_height=False,
use_color=True,
load_dim=6,
use_dim=[0, 1, 2, 3, 4, 5]),
dict(type='LoadAnnotations3D_',
with_label_3d=False,
with_bbox_3d=False,
with_sp_mask_3d=True),
dict(
type='MultiScaleFlipAug3D',
img_scale=(1333, 800),
pts_scale_ratio=1,
flip=False,
transforms=[
dict(type='PointSample_', num_points=100000),
dict(
type='DenormalizePointsColor',
color_mean=[0, 0, 0],
color_std=[255, 255, 255]),
dict(
type='NormalizePointsColor_',
color_mean=[127.5, 127.5, 127.5])
]),
dict(type='Pack3DDetInputs_', keys=['points', 'sp_pts_mask'])
]

run settings

train_dataloader = dict(
batch_size=8,
num_workers=8,
persistent_workers=True,
sampler=dict(type='DefaultSampler', shuffle=True),
dataset=dict(
type='ConcatDataset_',
datasets=
[dict(
type=dataset_type_scannet,
ann_file='/home/wanghui/DownLoads/unidet3d-master/data/scannet/scannet_infos_train.pkl',
data_prefix=data_prefix_scannet,
data_root=data_root_scannet,
metainfo=metainfo_scannet,
pipeline=train_pipeline_scannet,
ignore_index=max_class_scannet,
scene_idxs=None,
test_mode=False)]# +
# [dict(
# type=dataset_type_s3dis,
# data_root=data_root_s3dis,
# ann_file=f's3dis_sp_infos_Area_{i}.pkl',
# partition=0.33,
# pipeline=train_pipeline_s3dis,
# filter_empty_gt=True,
# data_prefix=data_prefix_s3dis,
# box_type_3d='Depth',
# backend_args=None) for i in train_area]
# [dict(
# type=dataset_type_multiscan,
# ann_file='multiscan_infos_train.pkl',
# partition=0.25,
# data_prefix=data_prefix_multiscan,
# data_root=data_root_multiscan,
# pipeline=train_pipeline_multiscan,
# test_mode=False)] +
# [dict(
# type=dataset_type_3rscan,
# ann_file='3rscan_infos_train.pkl',
# partition=0.15,
# data_prefix=data_prefix_3rscan,
# data_root=data_root_3rscan,
# pipeline=train_pipeline_3rscan,
# test_mode=False)] +
# [dict(
# type=dataset_type_scannetpp,
# ann_file='scannetpp_infos_train.pkl',
# partition=0.33,
# data_prefix=data_prefix_scannetpp,
# data_root=data_root_scannetpp,
# pipeline=train_pipeline_scannetpp,
# test_mode=False)] +
# [dict(
# type=dataset_type_arkitscenes,
# ann_file='arkitscenes_offline_infos_train.pkl',
# partition=0.08,
# data_prefix=data_prefix_arkitscenes,
# data_root=data_root_arkitscenes,
# pipeline=train_pipeline_arkitscenes,
# test_mode=False)]
))

val_dataloader = dict(
batch_size=1,
num_workers=1,
persistent_workers=True,
drop_last=False,
sampler=dict(type='DefaultSampler', shuffle=False),
dataset=dict(
type='ConcatDataset_',
datasets=
# [dict(
# type=dataset_type_scannet,
# ann_file='/home/wanghui/DownLoads/unidet3d-master/data/scannet/scannet_infos_val.pkl',
# data_prefix=data_prefix_scannet,
# data_root=data_root_scannet,
# metainfo=metainfo_scannet,
# pipeline=test_pipeline_scannet,
# ignore_index=max_class_scannet,
# test_mode=True)] +
[dict(
type=dataset_type_s3dis,
data_root=data_root_s3dis,
ann_file=f's3dis_sp_infos_Area_{test_area}.pkl',
pipeline=test_pipeline_s3dis,
test_mode=True,
data_prefix=data_prefix_s3dis,
box_type_3d='Depth',
backend_args=None)]
# [dict(
# type=dataset_type_multiscan,
# ann_file='multiscan_infos_val.pkl',
# data_prefix=data_prefix_multiscan,
# data_root=data_root_multiscan,
# pipeline=test_pipeline_multiscan,
# test_mode=True)] +
# [dict(
# type=dataset_type_3rscan,
# ann_file='3rscan_infos_val.pkl',
# data_prefix=data_prefix_3rscan,
# data_root=data_root_3rscan,
# pipeline=test_pipeline_3rscan,
# test_mode=True)] +
# [dict(
# type=dataset_type_scannetpp,
# ann_file='scannetpp_infos_val.pkl',
# data_prefix=data_prefix_scannetpp,
# data_root=data_root_scannetpp,
# pipeline=test_pipeline_scannetpp,
# test_mode=True)] +
# [dict(
# type=dataset_type_arkitscenes,
# ann_file='arkitscenes_offline_infos_val.pkl',
# data_prefix=data_prefix_arkitscenes,
# data_root=data_root_arkitscenes,
# pipeline=test_pipeline_arkitscenes,
# test_mode=True)]
))

test_dataloader = val_dataloader

load_from = '/home/wanghui/DownLoads/unidet3d-master/work_dirs/tmp/unidet3d.pth'

test_evaluator = dict(type='IndoorMetric_',
datasets=['scannet', 's3dis', 'multiscan', '3rscan', 'scannetpp', 'arkitscenes'],
datasets_classes=[classes_scannet, classes_s3dis,
classes_multiscan, classes_3rscan,
classes_scannetpp, classes_arkitscenes])

val_evaluator = test_evaluator

optim_wrapper = dict(
type='OptimWrapper',
optimizer=dict(type='AdamW', lr=0.0001 * 2, weight_decay=0.05),
clip_grad=dict(max_norm=10, norm_type=2))

param_scheduler = dict(type='PolyLR', begin=0, end=1024, power=0.9)

custom_hooks = [dict(type='EmptyCacheHook', after_iter=True)]
default_hooks = dict(
checkpoint=dict(interval=1, max_keep_ckpts=16))

train_cfg = dict(
type='EpochBasedTrainLoop',
max_epochs=1024,
dynamic_intervals=[(1, 2), (1024 - 16, 1)])
val_cfg = dict(type='ValLoop')
test_cfg = dict(type='TestLoop')

The following are the test results.
+----------+---------+---------+---------+---------+
| classes | AP_0.25 | AR_0.25 | AP_0.50 | AR_0.50 |
+----------+---------+---------+---------+---------+
| table | 0.0554 | 0.7143 | 0.0076 | 0.2597 |
| chair | 0.0688 | 0.6395 | 0.0119 | 0.2519 |
| sofa | 0.0328 | 0.3636 | 0.0145 | 0.2727 |
| bookcase | 0.0217 | 0.5806 | 0.0053 | 0.3088 |
| board | 0.0530 | 0.3333 | 0.0361 | 0.1190 |
+----------+---------+---------+---------+---------+
| Overall | 0.0464 | 0.5263 | 0.0150 | 0.2424 |
+----------+---------+---------+---------+---------+
The reproduced results on the ScanNet dataset are consistent with those reported in the paper, whereas the performance degrades considerably when training exclusively on the S3DIS dataset.
Could you suggest some approaches for resolving the problem of poor training performance on the S3DIS single dataset?

[col14m]

Please download the preprocessed dataset from here and run inference again.

[Wang-hui-001]

Hello, I have already downloaded the preprocessed dataset, and following your advice I re-downloaded it once again and configured it according to the instructions in the README. However, the results still remain consistently low without any improvement. Could you please explain what other possible reasons might cause such results?