Merge a884e49 into 273c714

open-mmlab · Mar 8, 2023 · 5d55ddf · 5d55ddf
2 parents 273c714 + a884e49
commit 5d55ddf
Show file tree

Hide file tree

Showing 95 changed files with 1,386 additions and 803 deletions.
diff --git a/configs/3dssd/3dssd_4xb4_kitti-3d-car.py b/configs/3dssd/3dssd_4xb4_kitti-3d-car.py
@@ -9,6 +9,8 @@
 class_names = ['Car']
 point_cloud_range = [0, -40, -5, 70, 40, 3]
 input_modality = dict(use_lidar=True, use_camera=False)
+backend_args = None
+
 db_sampler = dict(
  data_root=data_root,
  info_path=data_root + 'kitti_dbinfos_train.pkl',
@@ -17,17 +19,20 @@
  classes=class_names,
  sample_groups=dict(Car=15),
  points_loader=dict(
- type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4))
-
-file_client_args = dict(backend='disk')
-# Uncomment the following if use ceph or other file clients.
-# See https://mmcv.readthedocs.io/en/latest/api.html#mmcv.fileio.FileClient
-# for more details.
-# file_client_args = dict(
-# backend='petrel', path_mapping=dict(data='s3:https://kitti_data/'))
+ type='LoadPointsFromFile',
+ coord_type='LIDAR',
+ load_dim=4,
+ use_dim=4,
+ backend_args=backend_args),
+ backend_args=backend_args)
 
 train_pipeline = [
- dict(type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4),
+ dict(
+ type='LoadPointsFromFile',
+ coord_type='LIDAR',
+ load_dim=4,
+ use_dim=4,
+ backend_args=backend_args),
  dict(type='LoadAnnotations3D', with_bbox_3d=True, with_label_3d=True),
  dict(type='PointsRangeFilter', point_cloud_range=point_cloud_range),
  dict(type='ObjectRangeFilter', point_cloud_range=point_cloud_range),
@@ -52,7 +57,12 @@
 ]
 
 test_pipeline = [
- dict(type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4),
+ dict(
+ type='LoadPointsFromFile',
+ coord_type='LIDAR',
+ load_dim=4,
+ use_dim=4,
+ backend_args=backend_args),
  dict(
  type='MultiScaleFlipAug3D',
  img_scale=(1333, 800),

diff --git a/configs/_base_/datasets/kitti-3d-3class.py b/configs/_base_/datasets/kitti-3d-3class.py
@@ -6,6 +6,21 @@
 input_modality = dict(use_lidar=True, use_camera=False)
 metainfo = dict(classes=class_names)
 
+# Example to use different file client
+# Method 1: simply set the data root and let the file I/O module
+# automatically infer from prefix (not support LMDB and Memcache yet)
+
+# data_root = 's3:https://openmmlab/datasets/detection3d/kitti/'
+
+# Method 2: Use backend_args, file_client_args in versions before 1.1.0rc4
+# backend_args = dict(
+# backend='petrel',
+# path_mapping=dict({
+# './data/': 's3:https://openmmlab/datasets/detection3d/',
+# 'data/': 's3:https://openmmlab/datasets/detection3d/'
+# }))
+backend_args = None
+
 db_sampler = dict(
  data_root=data_root,
  info_path=data_root + 'kitti_dbinfos_train.pkl',
@@ -16,14 +31,20 @@
  classes=class_names,
  sample_groups=dict(Car=12, Pedestrian=6, Cyclist=6),
  points_loader=dict(
- type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4))
+ type='LoadPointsFromFile',
+ coord_type='LIDAR',
+ load_dim=4,
+ use_dim=4,
+ backend_args=backend_args),
+ backend_args=backend_args)
 
 train_pipeline = [
  dict(
  type='LoadPointsFromFile',
  coord_type='LIDAR',
  load_dim=4, # x, y, z, intensity
- use_dim=4),
+ use_dim=4,
+ backend_args=backend_args),
  dict(type='LoadAnnotations3D', with_bbox_3d=True, with_label_3d=True),
  dict(type='ObjectSample', db_sampler=db_sampler),
  dict(
@@ -45,7 +66,12 @@
  keys=['points', 'gt_bboxes_3d', 'gt_labels_3d'])
 ]
 test_pipeline = [
- dict(type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4),
+ dict(
+ type='LoadPointsFromFile',
+ coord_type='LIDAR',
+ load_dim=4,
+ use_dim=4,
+ backend_args=backend_args),
  dict(
  type='MultiScaleFlipAug3D',
  img_scale=(1333, 800),
@@ -66,7 +92,12 @@
 # construct a pipeline for data and gt loading in show function
 # please keep its loading function consistent with test_pipeline (e.g. client)
 eval_pipeline = [
- dict(type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4),
+ dict(
+ type='LoadPointsFromFile',
+ coord_type='LIDAR',
+ load_dim=4,
+ use_dim=4,
+ backend_args=backend_args),
  dict(type='Pack3DDetInputs', keys=['points'])
 ]
 train_dataloader = dict(
@@ -88,7 +119,8 @@
  metainfo=metainfo,
  # we use box_type_3d='LiDAR' in kitti and nuscenes dataset
  # and box_type_3d='Depth' in sunrgbd and scannet dataset.
- box_type_3d='LiDAR')))
+ box_type_3d='LiDAR',
+ backend_args=backend_args)))
 val_dataloader = dict(
  batch_size=1,
  num_workers=1,
@@ -104,7 +136,8 @@
  modality=input_modality,
  test_mode=True,
  metainfo=metainfo,
- box_type_3d='LiDAR'))
+ box_type_3d='LiDAR',
+ backend_args=backend_args))
 test_dataloader = dict(
  batch_size=1,
  num_workers=1,
@@ -120,11 +153,13 @@
  modality=input_modality,
  test_mode=True,
  metainfo=metainfo,
- box_type_3d='LiDAR'))
+ box_type_3d='LiDAR',
+ backend_args=backend_args))
 val_evaluator = dict(
  type='KittiMetric',
  ann_file=data_root + 'kitti_infos_val.pkl',
- metric='bbox')
+ metric='bbox',
+ backend_args=backend_args)
 test_evaluator = val_evaluator
 
 vis_backends = [dict(type='LocalVisBackend')]

diff --git a/configs/_base_/datasets/kitti-3d-car.py b/configs/_base_/datasets/kitti-3d-car.py
@@ -6,6 +6,21 @@
 input_modality = dict(use_lidar=True, use_camera=False)
 metainfo = dict(classes=class_names)
 
+# Example to use different file client
+# Method 1: simply set the data root and let the file I/O module
+# automatically infer from prefix (not support LMDB and Memcache yet)
+
+# data_root = 's3:https://openmmlab/datasets/detection3d/kitti/'
+
+# Method 2: Use backend_args, file_client_args in versions before 1.1.0rc4
+# backend_args = dict(
+# backend='petrel',
+# path_mapping=dict({
+# './data/': 's3:https://openmmlab/datasets/detection3d/',
+# 'data/': 's3:https://openmmlab/datasets/detection3d/'
+# }))
+backend_args = None
+
 db_sampler = dict(
  data_root=data_root,
  info_path=data_root + 'kitti_dbinfos_train.pkl',
@@ -14,14 +29,20 @@
  classes=class_names,
  sample_groups=dict(Car=15),
  points_loader=dict(
- type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4))
+ type='LoadPointsFromFile',
+ coord_type='LIDAR',
+ load_dim=4,
+ use_dim=4,
+ backend_args=backend_args),
+ backend_args=backend_args)
 
 train_pipeline = [
  dict(
  type='LoadPointsFromFile',
  coord_type='LIDAR',
  load_dim=4, # x, y, z, intensity
- use_dim=4),
+ use_dim=4,
+ backend_args=backend_args),
  dict(type='LoadAnnotations3D', with_bbox_3d=True, with_label_3d=True),
  dict(type='ObjectSample', db_sampler=db_sampler),
  dict(
@@ -43,7 +64,12 @@
  keys=['points', 'gt_bboxes_3d', 'gt_labels_3d'])
 ]
 test_pipeline = [
- dict(type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4),
+ dict(
+ type='LoadPointsFromFile',
+ coord_type='LIDAR',
+ load_dim=4,
+ use_dim=4,
+ backend_args=backend_args),
  dict(
  type='MultiScaleFlipAug3D',
  img_scale=(1333, 800),
@@ -64,7 +90,12 @@
 # construct a pipeline for data and gt loading in show function
 # please keep its loading function consistent with test_pipeline (e.g. client)
 eval_pipeline = [
- dict(type='LoadPointsFromFile', coord_type='LIDAR', load_dim=4, use_dim=4),
+ dict(
+ type='LoadPointsFromFile',
+ coord_type='LIDAR',
+ load_dim=4,
+ use_dim=4,
+ backend_args=backend_args),
  dict(type='Pack3DDetInputs', keys=['points'])
 ]
 train_dataloader = dict(
@@ -86,7 +117,8 @@
  metainfo=metainfo,
  # we use box_type_3d='LiDAR' in kitti and nuscenes dataset
  # and box_type_3d='Depth' in sunrgbd and scannet dataset.
- box_type_3d='LiDAR')))
+ box_type_3d='LiDAR',
+ backend_args=backend_args)))
 val_dataloader = dict(
  batch_size=1,
  num_workers=1,
@@ -102,7 +134,8 @@
  modality=input_modality,
  test_mode=True,
  metainfo=metainfo,
- box_type_3d='LiDAR'))
+ box_type_3d='LiDAR',
+ backend_args=backend_args))
 test_dataloader = dict(
  batch_size=1,
  num_workers=1,
@@ -118,11 +151,13 @@
  modality=input_modality,
  test_mode=True,
  metainfo=metainfo,
- box_type_3d='LiDAR'))
+ box_type_3d='LiDAR',
+ backend_args=backend_args))
 val_evaluator = dict(
  type='KittiMetric',
  ann_file=data_root + 'kitti_infos_val.pkl',
- metric='bbox')
+ metric='bbox',
+ backend_args=backend_args)
 test_evaluator = val_evaluator
 
 vis_backends = [dict(type='LocalVisBackend')]

diff --git a/configs/_base_/datasets/kitti-mono3d.py b/configs/_base_/datasets/kitti-mono3d.py
@@ -4,15 +4,23 @@
 input_modality = dict(use_lidar=False, use_camera=True)
 metainfo = dict(classes=class_names)
 
-file_client_args = dict(backend='disk')
-# Uncomment the following if use ceph or other file clients.
-# See https://mmcv.readthedocs.io/en/latest/api.html#mmcv.fileio.FileClient
-# for more details.
-# file_client_args = dict(
-# backend='petrel', path_mapping=dict(data='s3:https://kitti_data/'))
+# Example to use different file client
+# Method 1: simply set the data root and let the file I/O module
+# automatically infer from prefix (not support LMDB and Memcache yet)
+
+# data_root = 's3:https://openmmlab/datasets/detection3d/kitti/'
+
+# Method 2: Use backend_args, file_client_args in versions before 1.1.0rc4
+# backend_args = dict(
+# backend='petrel',
+# path_mapping=dict({
+# './data/': 's3:https://openmmlab/datasets/detection3d/',
+# 'data/': 's3:https://openmmlab/datasets/detection3d/'
+# }))
+backend_args = None
 
 train_pipeline = [
- dict(type='LoadImageFromFileMono3D'),
+ dict(type='LoadImageFromFileMono3D', backend_args=backend_args),
  dict(
  type='LoadAnnotations3D',
  with_bbox=True,
@@ -31,12 +39,12 @@
  ]),
 ]
 test_pipeline = [
- dict(type='LoadImageFromFileMono3D'),
+ dict(type='LoadImageFromFileMono3D', backend_args=backend_args),
  dict(type='Resize', scale=(1242, 375), keep_ratio=True),
  dict(type='Pack3DDetInputs', keys=['img'])
 ]
 eval_pipeline = [
- dict(type='LoadImageFromFileMono3D'),
+ dict(type='LoadImageFromFileMono3D', backend_args=backend_args),
  dict(type='Pack3DDetInputs', keys=['img'])
 ]
 
@@ -57,7 +65,8 @@
  metainfo=metainfo,
  # we use box_type_3d='Camera' in monocular 3d
  # detection task
- box_type_3d='Camera'))
+ box_type_3d='Camera',
+ backend_args=backend_args))
 val_dataloader = dict(
  batch_size=1,
  num_workers=2,
@@ -74,13 +83,15 @@
  load_type='fov_image_based',
  metainfo=metainfo,
  test_mode=True,
- box_type_3d='Camera'))
+ box_type_3d='Camera',
+ backend_args=backend_args))
 test_dataloader = val_dataloader
 
 val_evaluator = dict(
  type='KittiMetric',
  ann_file=data_root + 'kitti_infos_val.pkl',
- metric='bbox')
+ metric='bbox',
+ backend_args=backend_args)
 
 test_evaluator = val_evaluator