new file: docs/getting_started.md

	new file:   docs/install.md
	new file:   docs/prepare_dataset.md
	new file:   projects/__init__.py
	new file:   projects/configs/Actformer/Actformer_base.py
	new file:   projects/configs/Actformer/Actformer_small.py

docs/getting_started.md

@@ -0,0 +1,21 @@
+# Prerequisites
+
+**Please ensure you have prepared the environment and the nuScenes dataset.**
+
+# Train and Test
+
+Train ActFormer with 8 GPUs
+
+```
+./tools/dist_train.sh ./projects/configs/Actformer/Actformer_base.py 8
+```
+
+Eval ActFormer with 8 GPUs
+
+```
+./tools/dist_test.sh ./projects/configs/Actformer/Actformer_base.py ./path/to/ckpts.pth 8
+```
+
+# Visualization
+
+see [visual.py](../tools/analysis_tools/visual.py)

docs/install.md

@@ -0,0 +1,53 @@
+# Step-by-step installation instructions
+
+Following <https://mmdetection3d.readthedocs.io/en/latest/getting_started.html#installation>
+
+**a. Create a conda virtual environment and activate it.**
+
+```shell
+conda create -n open-mmlab python=3.8 -y
+conda activate open-mmlab
+```
+
+**b. Install PyTorch and torchvision following the [official instructions](https://pytorch.org/).**
+
+```shell
+pip install torch==1.9.1+cu111 torchvision==0.10.1+cu111 torchaudio==0.9.1 -f https://download.pytorch.org/whl/torch_stable.html
+# Recommended torch>=1.9
+
+```
+
+**c. Install gcc>=5 in conda env (optional).**
+
+```shell
+conda install -c omgarcia gcc-6 # gcc-6.2
+```
+
+**c. Install mmcv-full.**
+
+```shell
+pip install mmcv-full==1.4.0
+#  pip install mmcv-full==1.4.0 -f https://download.openmmlab.com/mmcv/dist/cu111/torch1.9.0/index.html
+```
+
+**d. Install mmdet and mmseg.**
+
+```shell
+pip install mmdet==2.14.0
+pip install mmsegmentation==0.14.1
+```
+
+**e. Install mmdet3d from source code.**
+
+```shell
+git clone https://github.com/open-mmlab/mmdetection3d.git
+cd mmdetection3d
+git checkout v0.17.1 # Other versions may not be compatible.
+python setup.py install
+```
+
+**f. Install timm.**
+
+```shell
+pip install timm
+```

docs/prepare_dataset.md

@@ -0,0 +1,13 @@
+## V2X-Sim
+
+Download v2X-Sim V2.0 full dataset data
+
+**Prepare data**
+
+*We genetate custom annotation files which are different from mmdet3d's*
+
+```
+python tools/create_data.py v2x_sim --root-path ./data/v2x_sim --out-dir ./data/v2x_sim --extra-tag v2x_sim --version v2.0-mini --canbus ./data
+```
+
+Using the above code will generate `v2xsim_infos_temporal_{train,val}.pkl`.

projects/configs/Actformer/Actformer_base.py

@@ -0,0 +1,258 @@
+_base_ = [
+    '../datasets/custom_nus-3d.py',
+    '../_base_/default_runtime.py'
+]
+#
+plugin = True
+plugin_dir = 'projects/mmdet3d_plugin/'
+
+# If point cloud range is changed, the models should also change their point
+# cloud range accordingly
+point_cloud_range = [-51.2, -51.2, -5.0, 51.2, 51.2, 3.0]
+voxel_size = [0.2, 0.2, 8]
+
+
+
+img_norm_cfg = dict(
+    mean=[103.530, 116.280, 123.675], std=[1.0, 1.0, 1.0], to_rgb=False)
+# For nuScenes we usually do 10-class detection
+class_names = [
+    'car', 'truck', 'construction_vehicle', 'bus', 'trailer', 'barrier',
+    'motorcycle', 'bicycle', 'pedestrian', 'traffic_cone'
+]
+
+input_modality = dict(
+    use_lidar=False,
+    use_camera=True,
+    use_radar=False,
+    use_map=False,
+    use_external=True)
+
+_dim_ = 256
+_pos_dim_ = _dim_//2
+_ffn_dim_ = _dim_*2
+_num_levels_ = 4
+bev_h_ = 200
+bev_w_ = 200
+queue_length = 4 # each sequence contains `queue_length` frames.
+
+model = dict(
+    type='BEVFormer',
+    use_grid_mask=True,
+    video_test_mode=True,
+
+    img_backbone=dict(
+        type='ResNet',
+        depth=101,
+        num_stages=4,
+        out_indices=(1, 2, 3),
+        frozen_stages=1,
+        norm_cfg=dict(type='BN2d', requires_grad=False),
+        norm_eval=True,
+        style='caffe',
+        dcn=dict(type='DCNv2', deform_groups=1, fallback_on_stride=False), # original DCNv2 will print log when perform load_state_dict
+        stage_with_dcn=(False, False, True, True)),
+    img_neck=dict(
+        type='FPN',
+        in_channels=[512, 1024, 2048],
+        out_channels=_dim_,
+        start_level=0,
+        add_extra_convs='on_output',
+        num_outs=4,
+        relu_before_extra_convs=True),
+    pts_bbox_head=dict(
+        type='BEVFormerHead',
+        bev_h=bev_h_,
+        bev_w=bev_w_,
+        num_query=900,
+        num_classes=10,
+        in_channels=_dim_,
+        sync_cls_avg_factor=True,
+        with_box_refine=True,
+        as_two_stage=False,
+        transformer=dict(
+            type='PerceptionTransformer',
+            rotate_prev_bev=True,
+            use_shift=True,
+            use_can_bus=True,
+            embed_dims=_dim_,
+            encoder=dict(
+                type='BEVFormerEncoder',
+                num_layers=6,
+                pc_range=point_cloud_range,
+                num_points_in_pillar=4,
+                return_intermediate=False,
+                transformerlayers=dict(
+                    type='BEVFormerLayer',
+                    attn_cfgs=[
+                        dict(
+                            type='TemporalSelfAttention',
+                            embed_dims=_dim_,
+                            num_levels=1),
+                        dict(
+                            type='PoseSelectiveAttention',
+                            pc_range=point_cloud_range,
+                            deformable_attention=dict(
+                                type='MSDeformableAttention3D',
+                                embed_dims=_dim_,
+                                num_points=8,
+                                num_levels=_num_levels_),
+                            embed_dims=_dim_,
+                        )
+                    ],
+                    feedforward_channels=_ffn_dim_,
+                    ffn_dropout=0.1,
+                    operation_order=('self_attn', 'norm', 'cross_attn', 'norm',
+                                     'ffn', 'norm'))),
+            decoder=dict(
+                type='DetectionTransformerDecoder',
+                num_layers=6,
+                return_intermediate=True,
+                transformerlayers=dict(
+                    type='DetrTransformerDecoderLayer',
+                    attn_cfgs=[
+                        dict(
+                            type='MultiheadAttention',
+                            embed_dims=_dim_,
+                            num_heads=8,
+                            dropout=0.1),
+                         dict(
+                            type='CustomMSDeformableAttention',
+                            embed_dims=_dim_,
+                            num_levels=1),
+                    ],
+
+                    feedforward_channels=_ffn_dim_,
+                    ffn_dropout=0.1,
+                    operation_order=('self_attn', 'norm', 'cross_attn', 'norm',
+                                     'ffn', 'norm')))),
+        bbox_coder=dict(
+            type='NMSFreeCoder',
+            post_center_range=[-61.2, -61.2, -10.0, 61.2, 61.2, 10.0],
+            pc_range=point_cloud_range,
+            max_num=300,
+            voxel_size=voxel_size,
+            num_classes=10),
+        positional_encoding=dict(
+            type='LearnedPositionalEncoding',
+            num_feats=_pos_dim_,
+            row_num_embed=bev_h_,
+            col_num_embed=bev_w_,
+            ),
+        loss_cls=dict(
+            type='FocalLoss',
+            use_sigmoid=True,
+            gamma=2.0,
+            alpha=0.25,
+            loss_weight=2.0),
+        loss_bbox=dict(type='L1Loss', loss_weight=0.25),
+        loss_iou=dict(type='GIoULoss', loss_weight=0.0)),
+    # model training and testing settings
+    train_cfg=dict(pts=dict(
+        grid_size=[512, 512, 1],
+        voxel_size=voxel_size,
+        point_cloud_range=point_cloud_range,
+        out_size_factor=4,
+        assigner=dict(
+            type='HungarianAssigner3D',
+            cls_cost=dict(type='FocalLossCost', weight=2.0),
+            reg_cost=dict(type='BBox3DL1Cost', weight=0.25),
+            iou_cost=dict(type='IoUCost', weight=0.0), # Fake cost. This is just to make it compatible with DETR head.
+            pc_range=point_cloud_range))))
+
+dataset_type = 'CustomV2XSIMDataset'
+data_root = 'data/V2X-Sim-2.0/'
+file_client_args = dict(backend='disk')
+
+
+train_pipeline = [
+    dict(type='LoadMultiViewImageFromFiles', to_float32=True),
+    dict(type='PhotoMetricDistortionMultiViewImage'),
+    dict(type='LoadAnnotations3D', with_bbox_3d=True, with_label_3d=True, with_attr_label=False),
+    dict(type='ObjectRangeFilter', point_cloud_range=point_cloud_range),
+    dict(type='ObjectNameFilter', classes=class_names),
+    dict(type='NormalizeMultiviewImage', **img_norm_cfg),
+    dict(type='PadMultiViewImage', size_divisor=32),
+    dict(type='DefaultFormatBundle3D', class_names=class_names),
+    dict(type='CustomCollect3D', keys=['gt_bboxes_3d', 'gt_labels_3d', 'img'])
+]
+
+test_pipeline = [
+    dict(type='LoadMultiViewImageFromFiles', to_float32=True),
+    dict(type='NormalizeMultiviewImage', **img_norm_cfg),
+    dict(type='PadMultiViewImage', size_divisor=32),
+    dict(
+        type='MultiScaleFlipAug3D',
+        img_scale=(1600, 900),
+        pts_scale_ratio=1,
+        flip=False,
+        transforms=[
+            dict(
+                type='DefaultFormatBundle3D',
+                class_names=class_names,
+                with_label=False),
+            dict(type='CustomCollect3D', keys=['img'])
+        ])
+]
+
+data = dict(
+    samples_per_gpu=1,
+    workers_per_gpu=4,
+    train=dict(
+        type=dataset_type,
+        data_root=data_root,
+        ann_file=data_root + 'v2x_sim_infos_temporal_train.pkl',
+        pipeline=train_pipeline,
+        classes=class_names,
+        modality=input_modality,
+        test_mode=False,
+        use_valid_flag=True,
+        bev_size=(bev_h_, bev_w_),
+        queue_length=queue_length,
+        # 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'),
+    val=dict(type=dataset_type,
+             data_root=data_root,
+             ann_file=data_root + 'v2x_sim_infos_temporal_val.pkl',
+             pipeline=test_pipeline,  bev_size=(bev_h_, bev_w_),
+             classes=class_names, modality=input_modality, samples_per_gpu=1),
+    test=dict(type=dataset_type,
+              data_root=data_root,
+              ann_file=data_root + 'v2x_sim_infos_temporal_test.pkl',
+              pipeline=test_pipeline, bev_size=(bev_h_, bev_w_),
+              classes=class_names, modality=input_modality),
+    shuffler_sampler=dict(type='DistributedGroupSampler'),
+    nonshuffler_sampler=dict(type='DistributedSampler')
+)
+
+optimizer = dict(
+    type='AdamW',
+    lr=2e-4,
+    paramwise_cfg=dict(
+        custom_keys={
+            'img_backbone': dict(lr_mult=0.1),
+        }),
+    weight_decay=0.01)
+
+optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))
+# learning policy
+lr_config = dict(
+    policy='CosineAnnealing',
+    warmup='linear',
+    warmup_iters=500,
+    warmup_ratio=1.0 / 3,
+    min_lr_ratio=1e-3)
+total_epochs = 24
+evaluation = dict(interval=1, pipeline=test_pipeline)
+
+runner = dict(type='EpochBasedRunner', max_epochs=total_epochs)
+load_from = 'ckpts/r101_dcn_fcos3d_pretrain.pth'
+log_config = dict(
+    interval=50,
+    hooks=[
+        dict(type='TextLoggerHook'),
+        dict(type='TensorboardLoggerHook')
+    ])
+
+checkpoint_config = dict(interval=1)