my CenterPoint config for KITTI dataset

This config file is modified from both official code released for KITTI and OpenPCDet version for WAYMO. And if anyone wants to get a pre-training model, please let me know. Update:

• 2022-06-30: I’ve uploaded the pre-trained model centerpoint@KITTI to the Google Drive. It is available at Performance and Models section in the repo

CLASS_NAMES: ['Car', 'Pedestrian', 'Cyclist']

DATA_CONFIG:
    _BASE_CONFIG_: cfgs/dataset_configs/kitti_dataset.yaml

MODEL:
    NAME: CenterPoint

    VFE:
        NAME: MeanVFE

    BACKBONE_3D:
        NAME: VoxelResBackBone8x

    MAP_TO_BEV:
        NAME: HeightCompression
        NUM_BEV_FEATURES: 256

    BACKBONE_2D:
        NAME: BaseBEVBackbone

        LAYER_NUMS: [5]
        LAYER_STRIDES: [1]
        NUM_FILTERS: [128]
        UPSAMPLE_STRIDES: [2]
        NUM_UPSAMPLE_FILTERS: [256]

    DENSE_HEAD:
        NAME: CenterHead
        CLASS_AGNOSTIC: False

        CLASS_NAMES_EACH_HEAD: [
            [ 'Car', 'Pedestrian', 'Cyclist' ]
        ]

        SHARED_CONV_CHANNEL: 64
        USE_BIAS_BEFORE_NORM: True
        NUM_HM_CONV: 2 #  heatmap
        SEPARATE_HEAD_CFG:
            HEAD_ORDER: ['center', 'center_z', 'dim', 'rot']
            HEAD_DICT: {
                'center': {'out_channels': 2, 'num_conv': 2}, # offset
                'center_z': {'out_channels': 1, 'num_conv': 2},
                'dim': {'out_channels': 3, 'num_conv': 2},
                'rot': {'out_channels': 2, 'num_conv': 2},
            }

        TARGET_ASSIGNER_CONFIG:
            FEATURE_MAP_STRIDE: 4
            NUM_MAX_OBJS: 500
            GAUSSIAN_OVERLAP: 0.1
            MIN_RADIUS: 2

        LOSS_CONFIG:
            LOSS_WEIGHTS: {
                'cls_weight': 1.0,
                'loc_weight': 2.0,
                'code_weights': [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0]
            }
        POST_PROCESSING:
            SCORE_THRESH: 0.1
            POST_CENTER_LIMIT_RANGE: [-75.2, -75.2, -2, 75.2, 75.2, 4]
            MAX_OBJ_PER_SAMPLE: 500
            NMS_CONFIG:
                NMS_TYPE: nms_gpu
                NMS_THRESH: 0.7
                NMS_PRE_MAXSIZE: 4096
                NMS_POST_MAXSIZE: 500

    POST_PROCESSING:
        RECALL_THRESH_LIST: [0.3, 0.5, 0.7]
        SCORE_THRESH: 0.1
        OUTPUT_RAW_SCORE: False

        EVAL_METRIC: kitti

        NMS_CONFIG:
            MULTI_CLASSES_NMS: False
            NMS_TYPE: nms_gpu
            NMS_THRESH: 0.01
            NMS_PRE_MAXSIZE: 4096
            NMS_POST_MAXSIZE: 500


OPTIMIZATION:
    BATCH_SIZE_PER_GPU: 4
    NUM_EPOCHS: 80

    OPTIMIZER: adam_onecycle
    LR: 0.003
    WEIGHT_DECAY: 0.01
    MOMENTUM: 0.9

    MOMS: [0.95, 0.85]
    PCT_START: 0.4
    DIV_FACTOR: 10
    DECAY_STEP_LIST: [35, 45]
    LR_DECAY: 0.1
    LR_CLIP: 0.0000001

    LR_WARMUP: False
    WARMUP_EPOCH: 1

    GRAD_NORM_CLIP: 10

[OuyangJunyuan]

Hi, I have a question about ‘FEATURE_MAP_STRIDE: 4’. It seems to be 8, since the backbone downsamples the feature map to 1/8. I have read the code of centerhead and config file of that on Waymo, I think 8 should be the right value. However, stride=4 gets better results than 8. And I have also conducted experiments on stride=1 and stride=2, they get nearly 0 mAP. Hope to get your reply.

Indeed, what you thought are approximately right about the Backbone3D. But there is one thing you has ignored, i.e., BEV_Backbone. Please kindly refer to configuration file that you can find UPSAMPLE_STRIDES: [2] in it, resulting in 1/8 * 2 = 1/4 here.

[Rapisurazurite]

I need a pre-training mode, can you give me?

Hi, can you share your pretrained model? Thanks [e-mail]

has been sent~

Hi, I used your pretrain-model to test kitti testing velodyne bin, but got error result. python demo.py --cfg_file cfgs/kitti_models/centerpoints.yaml --ckpt ../checkpoints/centerpoint_kitti_80.pth --data_path ../testing/velodyne/000003.bin

you can set NMS_THRESH to 0.01 in MODEL.DENSE_HEAD.POST_PROCESSING config. Centerpoint model do nms in centerhead, so the config of MODEL.POST_PROCESSING is useless.