[Refactor] refactor DGCNN configs for different areas (#1587)

* update parf of configs

* update pp-3class configs

configs/_base_/datasets/kitti-3d-3class.py

@@ -4,6 +4,20 @@
 class_names = ['Pedestrian', 'Cyclist', 'Car']
 point_cloud_range = [0, -40, -3, 70.4, 40, 1]
 input_modality = dict(use_lidar=True, use_camera=False)
+
+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/kitti/':
+#         's3://openmmlab/datasets/detection3d/kitti/',
+#         'data/kitti/':
+#         's3://openmmlab/datasets/detection3d/kitti/'
+#     }))
+
 db_sampler = dict(
     data_root=data_root,
     info_path=data_root + 'kitti_dbinfos_train.pkl',
@@ -12,14 +26,14 @@
         filter_by_difficulty=[-1],
         filter_by_min_points=dict(Car=5, Pedestrian=10, Cyclist=10)),
     classes=class_names,
-    sample_groups=dict(Car=12, Pedestrian=6, Cyclist=6))
-
-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://kitti_data/'))
+    sample_groups=dict(Car=12, Pedestrian=6, Cyclist=6),
+    points_loader=dict(
+        type='LoadPointsFromFile',
+        coord_type='LIDAR',
+        load_dim=4,
+        use_dim=4,
+        file_client_args=file_client_args),
+    file_client_args=file_client_args)
 
 train_pipeline = [
     dict(
@@ -113,7 +127,8 @@
             test_mode=False,
             # 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',
+            file_client_args=file_client_args)),
     val=dict(
         type=dataset_type,
         data_root=data_root,
@@ -124,7 +139,8 @@
         modality=input_modality,
         classes=class_names,
         test_mode=True,
-        box_type_3d='LiDAR'),
+        box_type_3d='LiDAR',
+        file_client_args=file_client_args),
     test=dict(
         type=dataset_type,
         data_root=data_root,
@@ -135,6 +151,7 @@
         modality=input_modality,
         classes=class_names,
         test_mode=True,
-        box_type_3d='LiDAR'))
+        box_type_3d='LiDAR',
+        file_client_args=file_client_args))
 
 evaluation = dict(interval=1, pipeline=eval_pipeline)

configs/_base_/datasets/s3dis_seg-3d-13class.py

@@ -3,6 +3,7 @@
 data_root = './data/s3dis/'
 class_names = ('ceiling', 'floor', 'wall', 'beam', 'column', 'window', 'door',
                'table', 'chair', 'sofa', 'bookcase', 'board', 'clutter')
+
 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
@@ -15,6 +16,7 @@
 #         'data/s3dis/':
 #         's3://openmmlab/datasets/detection3d/s3dis_processed/'
 #     }))
+
 num_points = 4096
 train_area = [1, 2, 3, 4, 6]
 test_area = 5

configs/_base_/datasets/scannet-3d-18class.py

@@ -5,15 +5,31 @@
                'bookshelf', 'picture', 'counter', 'desk', 'curtain',
                'refrigerator', 'showercurtrain', 'toilet', 'sink', 'bathtub',
                'garbagebin')
+
+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/scannet/':
+#         's3://openmmlab/datasets/detection3d/scannet_processed/',
+#         'data/scannet/':
+#         's3://openmmlab/datasets/detection3d/scannet_processed/'
+#     }))
+
 train_pipeline = [
     dict(
         type='LoadPointsFromFile',
+        file_client_args=file_client_args,
         coord_type='DEPTH',
         shift_height=True,
         load_dim=6,
         use_dim=[0, 1, 2]),
     dict(
         type='LoadAnnotations3D',
+        file_client_args=file_client_args,
         with_bbox_3d=True,
         with_label_3d=True,
         with_mask_3d=True,
@@ -46,6 +62,7 @@
 test_pipeline = [
     dict(
         type='LoadPointsFromFile',
+        file_client_args=file_client_args,
         coord_type='DEPTH',
         shift_height=True,
         load_dim=6,
@@ -80,6 +97,7 @@
 eval_pipeline = [
     dict(
         type='LoadPointsFromFile',
+        file_client_args=file_client_args,
         coord_type='DEPTH',
         shift_height=False,
         load_dim=6,
@@ -107,22 +125,25 @@
             classes=class_names,
             # we use box_type_3d='LiDAR' in kitti and nuscenes dataset
             # and box_type_3d='Depth' in sunrgbd and scannet dataset.
-            box_type_3d='Depth')),
+            box_type_3d='Depth',
+            file_client_args=file_client_args)),
     val=dict(
         type=dataset_type,
         data_root=data_root,
         ann_file=data_root + 'scannet_infos_val.pkl',
         pipeline=test_pipeline,
         classes=class_names,
         test_mode=True,
-        box_type_3d='Depth'),
+        box_type_3d='Depth',
+        file_client_args=file_client_args),
     test=dict(
         type=dataset_type,
         data_root=data_root,
         ann_file=data_root + 'scannet_infos_val.pkl',
         pipeline=test_pipeline,
         classes=class_names,
         test_mode=True,
-        box_type_3d='Depth'))
+        box_type_3d='Depth',
+        file_client_args=file_client_args))
 
 evaluation = dict(pipeline=eval_pipeline)

configs/_base_/datasets/scannet_seg-3d-20class.py

@@ -6,20 +6,36 @@
                'curtain', 'refrigerator', 'showercurtrain', 'toilet', 'sink',
                'bathtub', 'otherfurniture')
 num_points = 8192
+
+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/scannet/':
+#         's3://openmmlab/datasets/detection3d/scannet_processed/',
+#         'data/scannet/':
+#         's3://openmmlab/datasets/detection3d/scannet_processed/'
+#     }))
+
 train_pipeline = [
     dict(
         type='LoadPointsFromFile',
         coord_type='DEPTH',
         shift_height=False,
         use_color=True,
         load_dim=6,
-        use_dim=[0, 1, 2, 3, 4, 5]),
+        use_dim=[0, 1, 2, 3, 4, 5],
+        file_client_args=file_client_args),
     dict(
         type='LoadAnnotations3D',
         with_bbox_3d=False,
         with_label_3d=False,
         with_mask_3d=False,
-        with_seg_3d=True),
+        with_seg_3d=True,
+        file_client_args=file_client_args),
     dict(
         type='PointSegClassMapping',
         valid_cat_ids=(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, 16, 24, 28,
@@ -44,7 +60,8 @@
         shift_height=False,
         use_color=True,
         load_dim=6,
-        use_dim=[0, 1, 2, 3, 4, 5]),
+        use_dim=[0, 1, 2, 3, 4, 5],
+        file_client_args=file_client_args),
     dict(type='NormalizePointsColor', color_mean=None),
     dict(
         # a wrapper in order to successfully call test function
@@ -81,13 +98,15 @@
         shift_height=False,
         use_color=True,
         load_dim=6,
-        use_dim=[0, 1, 2, 3, 4, 5]),
+        use_dim=[0, 1, 2, 3, 4, 5],
+        file_client_args=file_client_args),
     dict(
         type='LoadAnnotations3D',
         with_bbox_3d=False,
         with_label_3d=False,
         with_mask_3d=False,
-        with_seg_3d=True),
+        with_seg_3d=True,
+        file_client_args=file_client_args),
     dict(
         type='PointSegClassMapping',
         valid_cat_ids=(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 14, 16, 24, 28,
@@ -111,22 +130,25 @@
         classes=class_names,
         test_mode=False,
         ignore_index=len(class_names),
-        scene_idxs=data_root + 'seg_info/train_resampled_scene_idxs.npy'),
+        scene_idxs=data_root + 'seg_info/train_resampled_scene_idxs.npy',
+        file_client_args=file_client_args),
     val=dict(
         type=dataset_type,
         data_root=data_root,
         ann_file=data_root + 'scannet_infos_val.pkl',
         pipeline=test_pipeline,
         classes=class_names,
         test_mode=True,
-        ignore_index=len(class_names)),
+        ignore_index=len(class_names),
+        file_client_args=file_client_args),
     test=dict(
         type=dataset_type,
         data_root=data_root,
         ann_file=data_root + 'scannet_infos_val.pkl',
         pipeline=test_pipeline,
         classes=class_names,
         test_mode=True,
-        ignore_index=len(class_names)))
+        ignore_index=len(class_names),
+        file_client_args=file_client_args))
 
 evaluation = dict(pipeline=eval_pipeline)

configs/_base_/datasets/sunrgbd-3d-10class.py

@@ -2,14 +2,29 @@
 data_root = 'data/sunrgbd/'
 class_names = ('bed', 'table', 'sofa', 'chair', 'toilet', 'desk', 'dresser',
                'night_stand', 'bookshelf', 'bathtub')
+
+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/sunrgbd/':
+#         's3://openmmlab/datasets/detection3d/sunrgbd_processed/',
+#         'data/sunrgbd/':
+#         's3://openmmlab/datasets/detection3d/sunrgbd_processed/'
+#     }))
+
 train_pipeline = [
     dict(
         type='LoadPointsFromFile',
         coord_type='DEPTH',
         shift_height=True,
         load_dim=6,
-        use_dim=[0, 1, 2]),
-    dict(type='LoadAnnotations3D'),
+        use_dim=[0, 1, 2],
+        file_client_args=file_client_args),
+    dict(type='LoadAnnotations3D', file_client_args=file_client_args),
     dict(
         type='RandomFlip3D',
         sync_2d=False,
@@ -30,7 +45,8 @@
         coord_type='DEPTH',
         shift_height=True,
         load_dim=6,
-        use_dim=[0, 1, 2]),
+        use_dim=[0, 1, 2],
+        file_client_args=file_client_args),
     dict(
         type='MultiScaleFlipAug3D',
         img_scale=(1333, 800),
@@ -63,7 +79,8 @@
         coord_type='DEPTH',
         shift_height=False,
         load_dim=6,
-        use_dim=[0, 1, 2]),
+        use_dim=[0, 1, 2],
+        file_client_args=file_client_args),
     dict(
         type='DefaultFormatBundle3D',
         class_names=class_names,
@@ -86,22 +103,25 @@
             filter_empty_gt=False,
             # we use box_type_3d='LiDAR' in kitti and nuscenes dataset
             # and box_type_3d='Depth' in sunrgbd and scannet dataset.
-            box_type_3d='Depth')),
+            box_type_3d='Depth',
+            file_client_args=file_client_args)),
     val=dict(
         type=dataset_type,
         data_root=data_root,
         ann_file=data_root + 'sunrgbd_infos_val.pkl',
         pipeline=test_pipeline,
         classes=class_names,
         test_mode=True,
-        box_type_3d='Depth'),
+        box_type_3d='Depth',
+        file_client_args=file_client_args),
     test=dict(
         type=dataset_type,
         data_root=data_root,
         ann_file=data_root + 'sunrgbd_infos_val.pkl',
         pipeline=test_pipeline,
         classes=class_names,
         test_mode=True,
-        box_type_3d='Depth'))
+        box_type_3d='Depth',
+        file_client_args=file_client_args))
 
 evaluation = dict(pipeline=eval_pipeline)

configs/dgcnn/dgcnn_32x4_cosine_100e_s3dis_seg-3d-13class-area1.py

@@ -0,0 +1,42 @@
+_base_ = [
+    '../_base_/datasets/s3dis_seg-3d-13class.py', '../_base_/models/dgcnn.py',
+    '../_base_/schedules/seg_cosine_100e.py', '../_base_/default_runtime.py'
+]
+
+# data settings
+train_area = [2, 3, 4, 5, 6]
+test_area = 1
+data_root = './data/s3dis/'
+data = dict(
+    samples_per_gpu=32,
+    train=dict(
+        ann_files=[
+            data_root + f's3dis_infos_Area_{i}.pkl' for i in train_area
+        ],
+        scene_idxs=[
+            data_root + f'seg_info/Area_{i}_resampled_scene_idxs.npy'
+            for i in train_area
+        ]),
+    val=dict(
+        ann_files=data_root + f's3dis_infos_Area_{test_area}.pkl',
+        scene_idxs=data_root +
+        f'seg_info/Area_{test_area}_resampled_scene_idxs.npy'),
+    test=dict(ann_files=data_root + f's3dis_infos_Area_{test_area}.pkl'))
+
+evaluation = dict(interval=2)
+
+# model settings
+model = dict(
+    backbone=dict(in_channels=9),  # [xyz, rgb, normalized_xyz]
+    decode_head=dict(
+        num_classes=13, ignore_index=13,
+        loss_decode=dict(class_weight=None)),  # S3DIS doesn't use class_weight
+    test_cfg=dict(
+        num_points=4096,
+        block_size=1.0,
+        sample_rate=0.5,
+        use_normalized_coord=True,
+        batch_size=24))
+
+# runtime settings
+checkpoint_config = dict(interval=2)