Categorical Depth Distribution Network for Monocular 3D Object Detection

Related tags

Deep Learning CaDDN
Overview

CaDDN

CaDDN is a monocular-based 3D object detection method. This repository is based off of [OpenPCDet].

Categorical Depth Distribution Network for Monocular 3D Object Detection
Cody Reading, Ali Harakeh, Julia Chae, and Steven L. Waslander
[Paper]

Overview

Changelog

[2021-03-16] CaDDN v0.3.0 is released.

Introduction

What does CaDDN do?

CaDDN is a general PyTorch-based method for 3D object detection from monocular images. At the time of submission, CaDDN achieved first 1st place among published monocular methods on the Kitti 3D object detection benchmark. We welcome contributions to this project.

CaDDN design pattern

We inherit the design pattern from [OpenPCDet].

  • Data-Model separation with unified point cloud coordinate for easily extending to custom datasets:

  • Unified 3D box definition: (x, y, z, dx, dy, dz, heading).

Model Zoo

KITTI 3D Object Detection Baselines

Selected supported methods are shown in the below table. The results are the 3D detection performance of Car class on the val set of KITTI dataset.

  • All models are trained with 2 Tesla T4 GPUs and are available for download.
  • The training time is measured with 2 Tesla T4 GPUs and PyTorch 1.4.
training time Easy@R40 Moderate@R40 Hard@R40 download
CaDDN ~76 hours 23.77 16.07 13.61 model-774M

Installation

Please refer to INSTALL.md for the installation of CaDDN.

Getting Started

Please refer to GETTING_STARTED.md to learn more usage about this project.

License

CaDDN is released under the Apache 2.0 license.

Acknowledgement

CaDDN is an open source project for monocular-based 3D scene perception. We would like to thank the authors of OpenPCDet for their open-source release of their 3D object detection codebase.

Citation

If you find this project useful in your research, please consider citing:

@article{CaDDN,
    title={Categorical Depth DistributionNetwork for Monocular 3D Object Detection},
    author={Cody Reading and
            Ali Harakeh and
            Julia Chae and
            Steven L. Waslander},
    journal = {CVPR},
    year={2021}
}

Contribution

Welcome to be a member of the CaDDN development team by contributing to this repo, and feel free to contact us for any potential contributions.

Comments
  • Out of memory, can not train on 11G machine

    Out of memory, can not train on 11G machine

    Hi,I set batch_size = 2, and I can not train resnet50 on 2080ti due to out of memory.Do you have any way to reduce the memory usage of the code? I tried to reduce the number of blocks in resnet50,but loss=nan, and error:

    WARNING:root:NaN or Inf found in input tensor.

    Thank you.

    opened by DuZzzs 14
  • Can I train the Nuscene model on OpenPCDet?

    Can I train the Nuscene model on OpenPCDet?

    Hello, I'm trying to use the CaDDN to train the Nuscene model. But it seems that it can not be directly applied. Do you implement the Nuscene model using CaDDN? If not, could you please give me some advice on what to do if I'd like to train the Nuscene model with CaDDN? As I know, the Nuscene dataset has six cameras, but the CaDDN only ran on the front camera. Thank you.

    opened by rockywind 7
  • Cuda Out of Memory with Cam150 image (above 2M image size)

    Cuda Out of Memory with Cam150 image (above 2M image size)

    I train CaDDN on A100 with Kitti Dataset it works fluently. But with our private dataset (captured with lidar and camera 150 degree) (after conversion to Kitti), it produces errors: File "/home/ubuntu/CaDDN/CaDDN/pcdet/models/backbones_3d/ffe/depth_ffe.py", line 93, in create_frustum_features frustum_features = depth_probs * image_features RuntimeError: CUDA out of memory. Tried to allocate 5.53 GiB (GPU 0; 39.59 GiB total capacity; 32.48 GiB already allocated; 4.92 GiB free; 32.99 GiB reserved in total by PyTorch). depth probs: torch.Size([2, 1, 80, 302, 480]) image feature shape: torch.Size([2, 64, 1, 302, 480])

    Pytorch 1.9.0 cudatoolkit 11.1 torchaudio 0.9.0
    torchvision 0.5.0
    It is the same error with cudatoolkit 11.0, pytorch 1.7.1 . Thank you very much !!!

    opened by namnv78 7
  • Can I train the Waymo model on OpenPCDet?

    Can I train the Waymo model on OpenPCDet?

    HELLO, I'm trying to use the OpenPCDet to train the Waymo model. But it seems that it can not be directly applied, do you implement the Waymo model using OpenPCDet? If not, could you please give me some advice on what to do if I'd like to train the Waymo model with OpenPCDet. Thank you.

    opened by Cc-Hy 6
  • How many gpus were used ?

    How many gpus were used ?

    hi, Thank you for sharing your work.

    when training on Kitti benchmark, do you mean by batch size of 4 per GPU? which equals to batch size of 8 when using two GPUs? Am I getting this right..?

    opened by ksh11023 6
  • Run test on pretrained model

    Run test on pretrained model

    i am getting the following error by running this command:

    python ./tools/test.py --cfg_file ./tools/cfgs/kitti_models/CaDDN.yaml --launcher pytorch --eval_all

    error:

    
Traceback (most recent call last):
  File "./tools/test.py", line 199, in <module>
    main()
  File "./tools/test.py", line 190, in main
    model = build_network(model_cfg=cfg.MODEL, num_class=len(cfg.CLASS_NAMES), dataset=test_set)
  File "/home/mm/MasterThesis/CaDDN/pcdet/models/__init__.py", line 14, in build_network
    model = build_detector(
  File "/home/mm/MasterThesis/CaDDN/pcdet/models/detectors/__init__.py", line 23, in build_detector
    model = __all__[model_cfg.NAME](
  File "/home/mm/MasterThesis/CaDDN/pcdet/models/detectors/caddn.py", line 7, in __init__
    self.module_list = self.build_networks()
  File "/home/mm/MasterThesis/CaDDN/pcdet/models/detectors/detector3d_template.py", line 50, in build_networks
    module, model_info_dict = getattr(self, 'build_%s' % module_name)(
  File "/home/mm/MasterThesis/CaDDN/pcdet/models/detectors/detector3d_template.py", line 74, in build_ffe
    ffe_module = ffe.__all__[self.model_cfg.FFE.NAME](
  File "/home/mm/MasterThesis/CaDDN/pcdet/models/backbones_3d/ffe/depth_ffe.py", line 25, in __init__
    self.ddn = ddn.__all__[ddn_cfg.NAME](
  File "/home/mm/MasterThesis/CaDDN/pcdet/models/backbones_3d/ffe/ddn/ddn_deeplabv3.py", line 21, in __init__
    super().__init__(constructor=constructor, **kwargs)
  File "/home/mm/MasterThesis/CaDDN/pcdet/models/backbones_3d/ffe/ddn/ddn_template.py", line 34, in __init__
    self.model = self.get_model(constructor=constructor)
  File "/home/mm/MasterThesis/CaDDN/pcdet/models/backbones_3d/ffe/ddn/ddn_template.py", line 63, in get_model
    pretrained_dict = self.filter_pretrained_dict(model_dict=model_dict, pretrained_dict=pretrained_dict)
  File "/home/mm/MasterThesis/CaDDN/pcdet/models/backbones_3d/ffe/ddn/ddn_template.py", line 88, in filter_pretrained_dict
    pretrained_num_classes = pretrained_dict["classifier.4.weight"].shape[0]
KeyError: 'classifier.4.weight'

    to be more precise: i downloaded the cadnn.pth from the link in README.md and set the path in the config file. also pretrained_dict only contains: pretrained_dict : dict_keys(['epoch', 'it', 'model_state', 'optimizer_state', 'version'])

    installation is based on what you mentioned in INSTALL.md

    opened by EmbeddedAndMore 6
  • How can I transform depth map to real depth ?

    How can I transform depth map to real depth ?

    I notice that you turn the depth map to real depth by dividing 256. But when I change the depth to point cloud, I found this value not correct. Could you tell me why you divide 256 to get the depth?

    opened by kaixinbear 5
  • Rendering 3D bounding boxes onto test images

    Rendering 3D bounding boxes onto test images

    Hello,

    Thanks for your help on the previous issue. I'm running the tools/test.py script with the intent to generate 3D bounding boxes on the test images as a result image set. I'm using the KITTI dataset. Is there a flag I can set for test.py, or use another software package, or follow some other additional steps to be able to render those 3D bounding boxes?

    Also, my next step is to try using a custom generated imageset. I'll be using the pre-trained model even though it's from the perspective of a street light. They would be frames from a video capture. I was going to try to follow the structure of the KITTI dataset and see if I can repurpose the script that generated the datainfos. Do you have any suggestions for the best way to go about using a custom set of images like this?

    Ultimately, it's my goal to see 3D bounding box rendering working on those images.

    Thanks!

    opened by dpwolfe 5
  • Why does it show the loading model's EPOCH is 3?

    Why does it show the loading model's EPOCH is 3?

    Thank you for your help in advance! Why does it show the loading model's EPOCH is 3? 2021-10-22 20:57:44,189 INFO ==> Checkpoint trained from version: pcdet+0.3.0+0000000 2021-10-22 20:57:47,894 INFO ==> Done (loaded 903/903) 2021-10-22 20:57:48,489 INFO *************** EPOCH 3 EVALUATION *****************

    opened by rockywind 4
  • Question about the Voxel Features

    Question about the Voxel Features

    Congratulations on your great work!

    I have read your paper and have several questions that bother me:

    In your work,

    1. Firstly, the voxel grid is first generated.
    2. Secondly, use the gird_to_lidar, lidar_to_cam, cam_to_img transformation to find the correspondence between the grid coordinates and the image coordinates.
    3. Thirdly, use grid_sample to sample features from Frustum to Voxel.
    4. Finally, Voxelcollapse to BEV features

    Since, in my opinion, the BEV features represent the world coordinates. My question is, why not just use BEV features to generate 'BEV grid', which represents the real world (lidar) coordinates? So, the gird_to_lidar step can be omitted. Am I right?

    I am still confused about the 'Voxel Features'. I don'y know what is it used for?

    Thank you very much, looking forward to your replay!

    opened by taylover-pei 4
  • Can I change the lidar coordinate to camera coordinate?

    Can I change the lidar coordinate to camera coordinate?

    Hi, I have eight camera data. The 3D lable is similar to KITTI style, it is based on camera coordinate. I think the network would be confused if converted to lidar coordinate. But, the generated voxel and anchors setting are based on lidar coordinate. What can I do to change the code based on camera coordinate?

    opened by rockywind 4
  • `run.sh` caddn -> codyreading/caddn

    `run.sh` caddn -> codyreading/caddn

    When following our tutorial on using Docker, the pulled Docker Image name did not match the command used to run run.sh, resulting in an error with the docker run command. https://github.com/TRAILab/CaDDN/blob/master/docker/DOCKER.md#get-a-docker-image

    docker pull codyreading/caddn

    CMD="docker run -it \
    --runtime=nvidia \
    --net=host \
    --privileged=true \
    --ipc=host \
    --volume="/tmp/.X11-unix:/tmp/.X11-unix:rw" \
    --volume="$XAUTHORITY:/root/.Xauthority:rw" \
    --env="DISPLAY" \
    --env="QT_X11_NO_MITSHM=1" \
    --hostname="inside-DOCKER" \
    --name="CaDDN" \
    --volume $PROJ_DIR/checkpoints:/CaDDN/checkpoints \
    --volume $PROJ_DIR/data:/CaDDN/data \
    --volume $PROJ_DIR/output:/CaDDN/output \
    --volume $PROJ_DIR/tools:/CaDDN/tools \
    --volume $PROJ_DIR/.git:/CaDDN/.git \
    $DATA_VOLUMES \
    $PCDET_VOLUMES \
    --rm \
    caddn bash <--- codyreading/caddn
    opened by PINTO0309 0
  • Error while training the model

    Error while training the model "ValueError: num_samples should be a positive integer value, but got num_samples=0"

    Code we used for training: !python /content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/train.py --cfg_file /content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/cfgs/kitti_models/CaDDN.yaml

    Log we got: 2022-11-19 21:23:27,419 INFO Start logging 2022-11-19 21:23:27,420 INFO CUDA_VISIBLE_DEVICES=ALL 2022-11-19 21:23:27,420 INFO cfg_file /content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/cfgs/kitti_models/CaDDN.yaml 2022-11-19 21:23:27,420 INFO batch_size 4 2022-11-19 21:23:27,420 INFO epochs 80 2022-11-19 21:23:27,420 INFO workers 8 2022-11-19 21:23:27,420 INFO extra_tag default 2022-11-19 21:23:27,421 INFO ckpt None 2022-11-19 21:23:27,421 INFO pretrained_model None 2022-11-19 21:23:27,421 INFO launcher none 2022-11-19 21:23:27,421 INFO tcp_port 18888 2022-11-19 21:23:27,421 INFO sync_bn False 2022-11-19 21:23:27,421 INFO fix_random_seed False 2022-11-19 21:23:27,421 INFO ckpt_save_interval 1 2022-11-19 21:23:27,421 INFO local_rank 0 2022-11-19 21:23:27,421 INFO max_ckpt_save_num 30 2022-11-19 21:23:27,421 INFO merge_all_iters_to_one_epoch False 2022-11-19 21:23:27,421 INFO set_cfgs None 2022-11-19 21:23:27,421 INFO max_waiting_mins 0 2022-11-19 21:23:27,422 INFO start_epoch 0 2022-11-19 21:23:27,422 INFO save_to_file False 2022-11-19 21:23:27,422 INFO cfg.ROOT_DIR: /content/gdrive/MyDrive/IITH_VCA/CaDDN 2022-11-19 21:23:27,422 INFO cfg.LOCAL_RANK: 0 2022-11-19 21:23:27,422 INFO cfg.CLASS_NAMES: ['Car', 'Pedestrian', 'Cyclist'] 2022-11-19 21:23:27,422 INFO
    cfg.DATA_CONFIG = edict() 2022-11-19 21:23:27,422 INFO cfg.DATA_CONFIG.DATASET: KittiDataset 2022-11-19 21:23:27,422 INFO cfg.DATA_CONFIG.DATA_PATH: /gdrive/MyDrive/3D_Obj/CaDDN/data/kitti 2022-11-19 21:23:27,422 INFO cfg.DATA_CONFIG.POINT_CLOUD_RANGE: [2, -30.08, -3.0, 46.8, 30.08, 1.0] 2022-11-19 21:23:27,422 INFO
    cfg.DATA_CONFIG.DATA_SPLIT = edict() 2022-11-19 21:23:27,422 INFO cfg.DATA_CONFIG.DATA_SPLIT.train: train 2022-11-19 21:23:27,423 INFO cfg.DATA_CONFIG.DATA_SPLIT.test: test 2022-11-19 21:23:27,423 INFO
    cfg.DATA_CONFIG.INFO_PATH = edict() 2022-11-19 21:23:27,423 INFO cfg.DATA_CONFIG.INFO_PATH.train: ['kitti_infos_trainval.pkl'] 2022-11-19 21:23:27,423 INFO cfg.DATA_CONFIG.INFO_PATH.test: ['kitti_infos_test.pkl'] 2022-11-19 21:23:27,423 INFO cfg.DATA_CONFIG.FOV_POINTS_ONLY: True 2022-11-19 21:23:27,423 INFO
    cfg.DATA_CONFIG.DATA_AUGMENTOR = edict() 2022-11-19 21:23:27,424 INFO cfg.DATA_CONFIG.DATA_AUGMENTOR.DISABLE_AUG_LIST: ['placeholder'] 2022-11-19 21:23:27,424 INFO cfg.DATA_CONFIG.DATA_AUGMENTOR.AUG_CONFIG_LIST: [{'NAME': 'random_image_flip', 'ALONG_AXIS_LIST': ['horizontal']}] 2022-11-19 21:23:27,424 INFO
    cfg.DATA_CONFIG.POINT_FEATURE_ENCODING = edict() 2022-11-19 21:23:27,424 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.encoding_type: absolute_coordinates_encoding 2022-11-19 21:23:27,424 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.used_feature_list: ['x', 'y', 'z', 'intensity'] 2022-11-19 21:23:27,424 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.src_feature_list: ['x', 'y', 'z', 'intensity'] 2022-11-19 21:23:27,424 INFO cfg.DATA_CONFIG.DATA_PROCESSOR: [{'NAME': 'mask_points_and_boxes_outside_range', 'REMOVE_OUTSIDE_BOXES': True}, {'NAME': 'calculate_grid_size', 'VOXEL_SIZE': [0.16, 0.16, 0.16]}] 2022-11-19 21:23:27,424 INFO cfg.DATA_CONFIG.BASE_CONFIG: cfgs/dataset_configs/kitti_dataset.yaml 2022-11-19 21:23:27,425 INFO
    cfg.DATA_CONFIG.IMAGE = edict() 2022-11-19 21:23:27,425 INFO cfg.DATA_CONFIG.IMAGE.ENABLED: True 2022-11-19 21:23:27,425 INFO
    cfg.DATA_CONFIG.IMAGE.COLLATE = edict() 2022-11-19 21:23:27,425 INFO cfg.DATA_CONFIG.IMAGE.COLLATE.MODE: Pad 2022-11-19 21:23:27,425 INFO cfg.DATA_CONFIG.IMAGE.COLLATE.CONSTANT_VALUES: [0.485, 0.456, 0.406] 2022-11-19 21:23:27,425 INFO
    cfg.DATA_CONFIG.DEPTH_MAP = edict() 2022-11-19 21:23:27,425 INFO cfg.DATA_CONFIG.DEPTH_MAP.ENABLED: True 2022-11-19 21:23:27,425 INFO cfg.DATA_CONFIG.DEPTH_MAP.DOWNSAMPLE_FACTOR: 4 2022-11-19 21:23:27,426 INFO
    cfg.DATA_CONFIG.CALIB = edict() 2022-11-19 21:23:27,426 INFO cfg.DATA_CONFIG.CALIB.ENABLED: True 2022-11-19 21:23:27,426 INFO
    cfg.MODEL = edict() 2022-11-19 21:23:27,426 INFO cfg.MODEL.NAME: CaDDN 2022-11-19 21:23:27,426 INFO
    cfg.MODEL.FFE = edict() 2022-11-19 21:23:27,426 INFO cfg.MODEL.FFE.NAME: DepthFFE 2022-11-19 21:23:27,426 INFO
    cfg.MODEL.FFE.DDN = edict() 2022-11-19 21:23:27,426 INFO cfg.MODEL.FFE.DDN.NAME: DDNDeepLabV3 2022-11-19 21:23:27,427 INFO cfg.MODEL.FFE.DDN.BACKBONE_NAME: ResNet101 2022-11-19 21:23:27,427 INFO
    cfg.MODEL.FFE.DDN.ARGS = edict() 2022-11-19 21:23:27,427 INFO cfg.MODEL.FFE.DDN.ARGS.feat_extract_layer: layer1 2022-11-19 21:23:27,427 INFO cfg.MODEL.FFE.DDN.ARGS.pretrained_path: ../checkpoints/deeplabv3_resnet101_coco-586e9e4e.pth 2022-11-19 21:23:27,427 INFO
    cfg.MODEL.FFE.CHANNEL_REDUCE = edict() 2022-11-19 21:23:27,427 INFO cfg.MODEL.FFE.CHANNEL_REDUCE.in_channels: 256 2022-11-19 21:23:27,427 INFO cfg.MODEL.FFE.CHANNEL_REDUCE.out_channels: 64 2022-11-19 21:23:27,427 INFO cfg.MODEL.FFE.CHANNEL_REDUCE.kernel_size: 1 2022-11-19 21:23:27,427 INFO cfg.MODEL.FFE.CHANNEL_REDUCE.stride: 1 2022-11-19 21:23:27,428 INFO cfg.MODEL.FFE.CHANNEL_REDUCE.bias: False 2022-11-19 21:23:27,428 INFO
    cfg.MODEL.FFE.DISCRETIZE = edict() 2022-11-19 21:23:27,428 INFO cfg.MODEL.FFE.DISCRETIZE.mode: LID 2022-11-19 21:23:27,428 INFO cfg.MODEL.FFE.DISCRETIZE.num_bins: 80 2022-11-19 21:23:27,428 INFO cfg.MODEL.FFE.DISCRETIZE.depth_min: 2.0 2022-11-19 21:23:27,428 INFO cfg.MODEL.FFE.DISCRETIZE.depth_max: 46.8 2022-11-19 21:23:27,428 INFO
    cfg.MODEL.FFE.DDN_LOSS = edict() 2022-11-19 21:23:27,428 INFO cfg.MODEL.FFE.DDN_LOSS.weight: 3.0 2022-11-19 21:23:27,428 INFO cfg.MODEL.FFE.DDN_LOSS.alpha: 0.25 2022-11-19 21:23:27,429 INFO cfg.MODEL.FFE.DDN_LOSS.gamma: 2.0 2022-11-19 21:23:27,429 INFO cfg.MODEL.FFE.DDN_LOSS.fg_weight: 13 2022-11-19 21:23:27,429 INFO cfg.MODEL.FFE.DDN_LOSS.bg_weight: 1 2022-11-19 21:23:27,429 INFO
    cfg.MODEL.F2V = edict() 2022-11-19 21:23:27,429 INFO cfg.MODEL.F2V.NAME: FrustumToVoxel 2022-11-19 21:23:27,429 INFO
    cfg.MODEL.F2V.SAMPLER = edict() 2022-11-19 21:23:27,429 INFO cfg.MODEL.F2V.SAMPLER.mode: bilinear 2022-11-19 21:23:27,429 INFO cfg.MODEL.F2V.SAMPLER.padding_mode: zeros 2022-11-19 21:23:27,429 INFO
    cfg.MODEL.MAP_TO_BEV = edict() 2022-11-19 21:23:27,430 INFO cfg.MODEL.MAP_TO_BEV.NAME: Conv2DCollapse 2022-11-19 21:23:27,430 INFO cfg.MODEL.MAP_TO_BEV.NUM_BEV_FEATURES: 64 2022-11-19 21:23:27,430 INFO
    cfg.MODEL.MAP_TO_BEV.ARGS = edict() 2022-11-19 21:23:27,430 INFO cfg.MODEL.MAP_TO_BEV.ARGS.kernel_size: 1 2022-11-19 21:23:27,430 INFO cfg.MODEL.MAP_TO_BEV.ARGS.stride: 1 2022-11-19 21:23:27,430 INFO cfg.MODEL.MAP_TO_BEV.ARGS.bias: False 2022-11-19 21:23:27,430 INFO
    cfg.MODEL.BACKBONE_2D = edict() 2022-11-19 21:23:27,430 INFO cfg.MODEL.BACKBONE_2D.NAME: BaseBEVBackbone 2022-11-19 21:23:27,431 INFO cfg.MODEL.BACKBONE_2D.LAYER_NUMS: [10, 10, 10] 2022-11-19 21:23:27,431 INFO cfg.MODEL.BACKBONE_2D.LAYER_STRIDES: [2, 2, 2] 2022-11-19 21:23:27,431 INFO cfg.MODEL.BACKBONE_2D.NUM_FILTERS: [64, 128, 256] 2022-11-19 21:23:27,431 INFO cfg.MODEL.BACKBONE_2D.UPSAMPLE_STRIDES: [1, 2, 4] 2022-11-19 21:23:27,431 INFO cfg.MODEL.BACKBONE_2D.NUM_UPSAMPLE_FILTERS: [128, 128, 128] 2022-11-19 21:23:27,431 INFO
    cfg.MODEL.DENSE_HEAD = edict() 2022-11-19 21:23:27,431 INFO cfg.MODEL.DENSE_HEAD.NAME: AnchorHeadSingle 2022-11-19 21:23:27,431 INFO cfg.MODEL.DENSE_HEAD.CLASS_AGNOSTIC: False 2022-11-19 21:23:27,431 INFO cfg.MODEL.DENSE_HEAD.USE_DIRECTION_CLASSIFIER: True 2022-11-19 21:23:27,432 INFO cfg.MODEL.DENSE_HEAD.DIR_OFFSET: 0.78539 2022-11-19 21:23:27,432 INFO cfg.MODEL.DENSE_HEAD.DIR_LIMIT_OFFSET: 0.0 2022-11-19 21:23:27,432 INFO cfg.MODEL.DENSE_HEAD.NUM_DIR_BINS: 2 2022-11-19 21:23:27,432 INFO cfg.MODEL.DENSE_HEAD.ANCHOR_GENERATOR_CONFIG: [{'class_name': 'Car', 'anchor_sizes': [[3.9, 1.6, 1.56]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-1.78], 'align_center': False, 'feature_map_stride': 2, 'matched_threshold': 0.6, 'unmatched_threshold': 0.45}, {'class_name': 'Pedestrian', 'anchor_sizes': [[0.8, 0.6, 1.73]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-0.6], 'align_center': False, 'feature_map_stride': 2, 'matched_threshold': 0.5, 'unmatched_threshold': 0.35}, {'class_name': 'Cyclist', 'anchor_sizes': [[1.76, 0.6, 1.73]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-0.6], 'align_center': False, 'feature_map_stride': 2, 'matched_threshold': 0.5, 'unmatched_threshold': 0.35}] 2022-11-19 21:23:27,432 INFO
    cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG = edict() 2022-11-19 21:23:27,432 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.NAME: AxisAlignedTargetAssigner 2022-11-19 21:23:27,432 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.POS_FRACTION: -1.0 2022-11-19 21:23:27,432 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.SAMPLE_SIZE: 512 2022-11-19 21:23:27,433 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.NORM_BY_NUM_EXAMPLES: False 2022-11-19 21:23:27,433 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.MATCH_HEIGHT: False 2022-11-19 21:23:27,433 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.BOX_CODER: ResidualCoder 2022-11-19 21:23:27,433 INFO
    cfg.MODEL.DENSE_HEAD.LOSS_CONFIG = edict() 2022-11-19 21:23:27,433 INFO
    cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS = edict() 2022-11-19 21:23:27,433 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.cls_weight: 1.0 2022-11-19 21:23:27,433 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.loc_weight: 2.0 2022-11-19 21:23:27,433 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.dir_weight: 0.2 2022-11-19 21:23:27,433 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.code_weights: [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0] 2022-11-19 21:23:27,434 INFO
    cfg.MODEL.POST_PROCESSING = edict() 2022-11-19 21:23:27,434 INFO cfg.MODEL.POST_PROCESSING.RECALL_THRESH_LIST: [0.3, 0.5, 0.7] 2022-11-19 21:23:27,434 INFO cfg.MODEL.POST_PROCESSING.SCORE_THRESH: 0.1 2022-11-19 21:23:27,434 INFO cfg.MODEL.POST_PROCESSING.OUTPUT_RAW_SCORE: False 2022-11-19 21:23:27,434 INFO cfg.MODEL.POST_PROCESSING.EVAL_METRIC: kitti 2022-11-19 21:23:27,434 INFO
    cfg.MODEL.POST_PROCESSING.NMS_CONFIG = edict() 2022-11-19 21:23:27,434 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.MULTI_CLASSES_NMS: False 2022-11-19 21:23:27,434 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_TYPE: nms_gpu 2022-11-19 21:23:27,434 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_THRESH: 0.01 2022-11-19 21:23:27,435 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_PRE_MAXSIZE: 4096 2022-11-19 21:23:27,435 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_POST_MAXSIZE: 500 2022-11-19 21:23:27,435 INFO
    cfg.OPTIMIZATION = edict() 2022-11-19 21:23:27,435 INFO cfg.OPTIMIZATION.BATCH_SIZE_PER_GPU: 4 2022-11-19 21:23:27,435 INFO cfg.OPTIMIZATION.NUM_EPOCHS: 80 2022-11-19 21:23:27,435 INFO cfg.OPTIMIZATION.OPTIMIZER: adam_onecycle 2022-11-19 21:23:27,435 INFO cfg.OPTIMIZATION.LR: 0.001 2022-11-19 21:23:27,435 INFO cfg.OPTIMIZATION.WEIGHT_DECAY: 0.01 2022-11-19 21:23:27,436 INFO cfg.OPTIMIZATION.MOMENTUM: 0.9 2022-11-19 21:23:27,436 INFO cfg.OPTIMIZATION.MOMS: [0.95, 0.85] 2022-11-19 21:23:27,436 INFO cfg.OPTIMIZATION.PCT_START: 0.4 2022-11-19 21:23:27,436 INFO cfg.OPTIMIZATION.DIV_FACTOR: 10 2022-11-19 21:23:27,436 INFO cfg.OPTIMIZATION.DECAY_STEP_LIST: [35, 45] 2022-11-19 21:23:27,436 INFO cfg.OPTIMIZATION.LR_DECAY: 0.1 2022-11-19 21:23:27,436 INFO cfg.OPTIMIZATION.LR_CLIP: 1e-07 2022-11-19 21:23:27,436 INFO cfg.OPTIMIZATION.LR_WARMUP: False 2022-11-19 21:23:27,436 INFO cfg.OPTIMIZATION.WARMUP_EPOCH: 1 2022-11-19 21:23:27,437 INFO cfg.OPTIMIZATION.GRAD_NORM_CLIP: 10 2022-11-19 21:23:27,437 INFO cfg.TAG: CaDDN 2022-11-19 21:23:27,437 INFO cfg.EXP_GROUP_PATH: content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/cfgs/kitti_models 2022-11-19 21:23:28,525 INFO Loading KITTI dataset 2022-11-19 21:23:28,526 INFO Total samples for KITTI dataset: 0 Traceback (most recent call last): File "/content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/train.py", line 198, in main() File "/content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/train.py", line 112, in main total_epochs=args.epochs File "/content/gdrive/MyDrive/IITH_VCA/CaDDN/pcdet/datasets/init.py", line 69, in build_dataloader drop_last=False, sampler=sampler, timeout=0 File "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py", line 213, in init sampler = RandomSampler(dataset) File "/usr/local/lib/python3.7/dist-packages/torch/utils/data/sampler.py", line 94, in init "value, but got num_samples={}".format(self.num_samples)) ValueError: num_samples should be a positive integer value, but got num_samples=0

    Screen shots of the same: image

    Error screen shot: image

    opened by prabhata07 1
  • KeyError: 'model_state'

    KeyError: 'model_state'

    Facing KeyError: 'model_state' while trying to test with "Pretrained model" Command used:

    !python /content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/test.py --cfg_file /content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/cfgs/kitti_models/CaDDN.yaml --batch_size 1 --ckpt /content/gdrive/MyDrive/IITH_VCA/CaDDN/checkpoints/deeplabv3_resnet101_coco-586e9e4e.pth

    Environemnt: Collab

    Model: as given in the github repo deeplabv3_resnet101_coco-586e9e4e.pth

    Error log:

    2022-11-19 21:16:39,034 INFO Start logging 2022-11-19 21:16:39,035 INFO CUDA_VISIBLE_DEVICES=ALL 2022-11-19 21:16:39,035 INFO cfg_file /content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/cfgs/kitti_models/CaDDN.yaml 2022-11-19 21:16:39,035 INFO batch_size 1 2022-11-19 21:16:39,035 INFO workers 4 2022-11-19 21:16:39,035 INFO extra_tag default 2022-11-19 21:16:39,035 INFO ckpt /content/gdrive/MyDrive/IITH_VCA/CaDDN/checkpoints/deeplabv3_resnet101_coco-586e9e4e.pth 2022-11-19 21:16:39,035 INFO launcher none 2022-11-19 21:16:39,036 INFO tcp_port 18888 2022-11-19 21:16:39,036 INFO local_rank 0 2022-11-19 21:16:39,036 INFO set_cfgs None 2022-11-19 21:16:39,036 INFO max_waiting_mins 30 2022-11-19 21:16:39,036 INFO start_epoch 0 2022-11-19 21:16:39,036 INFO eval_tag default 2022-11-19 21:16:39,036 INFO eval_all False 2022-11-19 21:16:39,036 INFO ckpt_dir None 2022-11-19 21:16:39,036 INFO save_to_file False 2022-11-19 21:16:39,037 INFO cfg.ROOT_DIR: /content/gdrive/MyDrive/IITH_VCA/CaDDN 2022-11-19 21:16:39,037 INFO cfg.LOCAL_RANK: 0 2022-11-19 21:16:39,037 INFO cfg.CLASS_NAMES: ['Car', 'Pedestrian', 'Cyclist'] 2022-11-19 21:16:39,037 INFO
    cfg.DATA_CONFIG = edict() 2022-11-19 21:16:39,037 INFO cfg.DATA_CONFIG.DATASET: KittiDataset 2022-11-19 21:16:39,037 INFO cfg.DATA_CONFIG.DATA_PATH: /gdrive/MyDrive/3D_Obj/CaDDN/data/kitti 2022-11-19 21:16:39,037 INFO cfg.DATA_CONFIG.POINT_CLOUD_RANGE: [2, -30.08, -3.0, 46.8, 30.08, 1.0] 2022-11-19 21:16:39,037 INFO
    cfg.DATA_CONFIG.DATA_SPLIT = edict() 2022-11-19 21:16:39,038 INFO cfg.DATA_CONFIG.DATA_SPLIT.train: train 2022-11-19 21:16:39,038 INFO cfg.DATA_CONFIG.DATA_SPLIT.test: test 2022-11-19 21:16:39,038 INFO
    cfg.DATA_CONFIG.INFO_PATH = edict() 2022-11-19 21:16:39,038 INFO cfg.DATA_CONFIG.INFO_PATH.train: ['kitti_infos_trainval.pkl'] 2022-11-19 21:16:39,038 INFO cfg.DATA_CONFIG.INFO_PATH.test: ['kitti_infos_test.pkl'] 2022-11-19 21:16:39,038 INFO cfg.DATA_CONFIG.FOV_POINTS_ONLY: True 2022-11-19 21:16:39,038 INFO
    cfg.DATA_CONFIG.DATA_AUGMENTOR = edict() 2022-11-19 21:16:39,038 INFO cfg.DATA_CONFIG.DATA_AUGMENTOR.DISABLE_AUG_LIST: ['placeholder'] 2022-11-19 21:16:39,038 INFO cfg.DATA_CONFIG.DATA_AUGMENTOR.AUG_CONFIG_LIST: [{'NAME': 'random_image_flip', 'ALONG_AXIS_LIST': ['horizontal']}] 2022-11-19 21:16:39,038 INFO
    cfg.DATA_CONFIG.POINT_FEATURE_ENCODING = edict() 2022-11-19 21:16:39,039 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.encoding_type: absolute_coordinates_encoding 2022-11-19 21:16:39,039 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.used_feature_list: ['x', 'y', 'z', 'intensity'] 2022-11-19 21:16:39,039 INFO cfg.DATA_CONFIG.POINT_FEATURE_ENCODING.src_feature_list: ['x', 'y', 'z', 'intensity'] 2022-11-19 21:16:39,039 INFO cfg.DATA_CONFIG.DATA_PROCESSOR: [{'NAME': 'mask_points_and_boxes_outside_range', 'REMOVE_OUTSIDE_BOXES': True}, {'NAME': 'calculate_grid_size', 'VOXEL_SIZE': [0.16, 0.16, 0.16]}] 2022-11-19 21:16:39,039 INFO cfg.DATA_CONFIG.BASE_CONFIG: cfgs/dataset_configs/kitti_dataset.yaml 2022-11-19 21:16:39,039 INFO
    cfg.DATA_CONFIG.IMAGE = edict() 2022-11-19 21:16:39,039 INFO cfg.DATA_CONFIG.IMAGE.ENABLED: True 2022-11-19 21:16:39,039 INFO
    cfg.DATA_CONFIG.IMAGE.COLLATE = edict() 2022-11-19 21:16:39,039 INFO cfg.DATA_CONFIG.IMAGE.COLLATE.MODE: Pad 2022-11-19 21:16:39,039 INFO cfg.DATA_CONFIG.IMAGE.COLLATE.CONSTANT_VALUES: [0.485, 0.456, 0.406] 2022-11-19 21:16:39,039 INFO
    cfg.DATA_CONFIG.DEPTH_MAP = edict() 2022-11-19 21:16:39,039 INFO cfg.DATA_CONFIG.DEPTH_MAP.ENABLED: True 2022-11-19 21:16:39,039 INFO cfg.DATA_CONFIG.DEPTH_MAP.DOWNSAMPLE_FACTOR: 4 2022-11-19 21:16:39,040 INFO
    cfg.DATA_CONFIG.CALIB = edict() 2022-11-19 21:16:39,040 INFO cfg.DATA_CONFIG.CALIB.ENABLED: True 2022-11-19 21:16:39,040 INFO
    cfg.MODEL = edict() 2022-11-19 21:16:39,040 INFO cfg.MODEL.NAME: CaDDN 2022-11-19 21:16:39,040 INFO
    cfg.MODEL.FFE = edict() 2022-11-19 21:16:39,040 INFO cfg.MODEL.FFE.NAME: DepthFFE 2022-11-19 21:16:39,040 INFO
    cfg.MODEL.FFE.DDN = edict() 2022-11-19 21:16:39,040 INFO cfg.MODEL.FFE.DDN.NAME: DDNDeepLabV3 2022-11-19 21:16:39,040 INFO cfg.MODEL.FFE.DDN.BACKBONE_NAME: ResNet101 2022-11-19 21:16:39,040 INFO
    cfg.MODEL.FFE.DDN.ARGS = edict() 2022-11-19 21:16:39,040 INFO cfg.MODEL.FFE.DDN.ARGS.feat_extract_layer: layer1 2022-11-19 21:16:39,040 INFO cfg.MODEL.FFE.DDN.ARGS.pretrained_path: ../checkpoints/deeplabv3_resnet101_coco-586e9e4e.pth 2022-11-19 21:16:39,040 INFO
    cfg.MODEL.FFE.CHANNEL_REDUCE = edict() 2022-11-19 21:16:39,041 INFO cfg.MODEL.FFE.CHANNEL_REDUCE.in_channels: 256 2022-11-19 21:16:39,041 INFO cfg.MODEL.FFE.CHANNEL_REDUCE.out_channels: 64 2022-11-19 21:16:39,041 INFO cfg.MODEL.FFE.CHANNEL_REDUCE.kernel_size: 1 2022-11-19 21:16:39,041 INFO cfg.MODEL.FFE.CHANNEL_REDUCE.stride: 1 2022-11-19 21:16:39,041 INFO cfg.MODEL.FFE.CHANNEL_REDUCE.bias: False 2022-11-19 21:16:39,041 INFO
    cfg.MODEL.FFE.DISCRETIZE = edict() 2022-11-19 21:16:39,041 INFO cfg.MODEL.FFE.DISCRETIZE.mode: LID 2022-11-19 21:16:39,041 INFO cfg.MODEL.FFE.DISCRETIZE.num_bins: 80 2022-11-19 21:16:39,041 INFO cfg.MODEL.FFE.DISCRETIZE.depth_min: 2.0 2022-11-19 21:16:39,041 INFO cfg.MODEL.FFE.DISCRETIZE.depth_max: 46.8 2022-11-19 21:16:39,041 INFO
    cfg.MODEL.FFE.DDN_LOSS = edict() 2022-11-19 21:16:39,041 INFO cfg.MODEL.FFE.DDN_LOSS.weight: 3.0 2022-11-19 21:16:39,041 INFO cfg.MODEL.FFE.DDN_LOSS.alpha: 0.25 2022-11-19 21:16:39,042 INFO cfg.MODEL.FFE.DDN_LOSS.gamma: 2.0 2022-11-19 21:16:39,042 INFO cfg.MODEL.FFE.DDN_LOSS.fg_weight: 13 2022-11-19 21:16:39,042 INFO cfg.MODEL.FFE.DDN_LOSS.bg_weight: 1 2022-11-19 21:16:39,042 INFO
    cfg.MODEL.F2V = edict() 2022-11-19 21:16:39,042 INFO cfg.MODEL.F2V.NAME: FrustumToVoxel 2022-11-19 21:16:39,042 INFO
    cfg.MODEL.F2V.SAMPLER = edict() 2022-11-19 21:16:39,042 INFO cfg.MODEL.F2V.SAMPLER.mode: bilinear 2022-11-19 21:16:39,042 INFO cfg.MODEL.F2V.SAMPLER.padding_mode: zeros 2022-11-19 21:16:39,042 INFO
    cfg.MODEL.MAP_TO_BEV = edict() 2022-11-19 21:16:39,042 INFO cfg.MODEL.MAP_TO_BEV.NAME: Conv2DCollapse 2022-11-19 21:16:39,042 INFO cfg.MODEL.MAP_TO_BEV.NUM_BEV_FEATURES: 64 2022-11-19 21:16:39,042 INFO
    cfg.MODEL.MAP_TO_BEV.ARGS = edict() 2022-11-19 21:16:39,042 INFO cfg.MODEL.MAP_TO_BEV.ARGS.kernel_size: 1 2022-11-19 21:16:39,043 INFO cfg.MODEL.MAP_TO_BEV.ARGS.stride: 1 2022-11-19 21:16:39,043 INFO cfg.MODEL.MAP_TO_BEV.ARGS.bias: False 2022-11-19 21:16:39,043 INFO
    cfg.MODEL.BACKBONE_2D = edict() 2022-11-19 21:16:39,043 INFO cfg.MODEL.BACKBONE_2D.NAME: BaseBEVBackbone 2022-11-19 21:16:39,043 INFO cfg.MODEL.BACKBONE_2D.LAYER_NUMS: [10, 10, 10] 2022-11-19 21:16:39,043 INFO cfg.MODEL.BACKBONE_2D.LAYER_STRIDES: [2, 2, 2] 2022-11-19 21:16:39,043 INFO cfg.MODEL.BACKBONE_2D.NUM_FILTERS: [64, 128, 256] 2022-11-19 21:16:39,043 INFO cfg.MODEL.BACKBONE_2D.UPSAMPLE_STRIDES: [1, 2, 4] 2022-11-19 21:16:39,043 INFO cfg.MODEL.BACKBONE_2D.NUM_UPSAMPLE_FILTERS: [128, 128, 128] 2022-11-19 21:16:39,043 INFO
    cfg.MODEL.DENSE_HEAD = edict() 2022-11-19 21:16:39,043 INFO cfg.MODEL.DENSE_HEAD.NAME: AnchorHeadSingle 2022-11-19 21:16:39,043 INFO cfg.MODEL.DENSE_HEAD.CLASS_AGNOSTIC: False 2022-11-19 21:16:39,043 INFO cfg.MODEL.DENSE_HEAD.USE_DIRECTION_CLASSIFIER: True 2022-11-19 21:16:39,044 INFO cfg.MODEL.DENSE_HEAD.DIR_OFFSET: 0.78539 2022-11-19 21:16:39,044 INFO cfg.MODEL.DENSE_HEAD.DIR_LIMIT_OFFSET: 0.0 2022-11-19 21:16:39,044 INFO cfg.MODEL.DENSE_HEAD.NUM_DIR_BINS: 2 2022-11-19 21:16:39,044 INFO cfg.MODEL.DENSE_HEAD.ANCHOR_GENERATOR_CONFIG: [{'class_name': 'Car', 'anchor_sizes': [[3.9, 1.6, 1.56]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-1.78], 'align_center': False, 'feature_map_stride': 2, 'matched_threshold': 0.6, 'unmatched_threshold': 0.45}, {'class_name': 'Pedestrian', 'anchor_sizes': [[0.8, 0.6, 1.73]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-0.6], 'align_center': False, 'feature_map_stride': 2, 'matched_threshold': 0.5, 'unmatched_threshold': 0.35}, {'class_name': 'Cyclist', 'anchor_sizes': [[1.76, 0.6, 1.73]], 'anchor_rotations': [0, 1.57], 'anchor_bottom_heights': [-0.6], 'align_center': False, 'feature_map_stride': 2, 'matched_threshold': 0.5, 'unmatched_threshold': 0.35}] 2022-11-19 21:16:39,044 INFO
    cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG = edict() 2022-11-19 21:16:39,044 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.NAME: AxisAlignedTargetAssigner 2022-11-19 21:16:39,044 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.POS_FRACTION: -1.0 2022-11-19 21:16:39,044 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.SAMPLE_SIZE: 512 2022-11-19 21:16:39,044 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.NORM_BY_NUM_EXAMPLES: False 2022-11-19 21:16:39,044 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.MATCH_HEIGHT: False 2022-11-19 21:16:39,044 INFO cfg.MODEL.DENSE_HEAD.TARGET_ASSIGNER_CONFIG.BOX_CODER: ResidualCoder 2022-11-19 21:16:39,044 INFO
    cfg.MODEL.DENSE_HEAD.LOSS_CONFIG = edict() 2022-11-19 21:16:39,045 INFO
    cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS = edict() 2022-11-19 21:16:39,045 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.cls_weight: 1.0 2022-11-19 21:16:39,045 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.loc_weight: 2.0 2022-11-19 21:16:39,045 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.dir_weight: 0.2 2022-11-19 21:16:39,045 INFO cfg.MODEL.DENSE_HEAD.LOSS_CONFIG.LOSS_WEIGHTS.code_weights: [1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 1.0] 2022-11-19 21:16:39,045 INFO
    cfg.MODEL.POST_PROCESSING = edict() 2022-11-19 21:16:39,045 INFO cfg.MODEL.POST_PROCESSING.RECALL_THRESH_LIST: [0.3, 0.5, 0.7] 2022-11-19 21:16:39,045 INFO cfg.MODEL.POST_PROCESSING.SCORE_THRESH: 0.1 2022-11-19 21:16:39,045 INFO cfg.MODEL.POST_PROCESSING.OUTPUT_RAW_SCORE: False 2022-11-19 21:16:39,045 INFO cfg.MODEL.POST_PROCESSING.EVAL_METRIC: kitti 2022-11-19 21:16:39,045 INFO
    cfg.MODEL.POST_PROCESSING.NMS_CONFIG = edict() 2022-11-19 21:16:39,045 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.MULTI_CLASSES_NMS: False 2022-11-19 21:16:39,045 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_TYPE: nms_gpu 2022-11-19 21:16:39,046 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_THRESH: 0.01 2022-11-19 21:16:39,046 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_PRE_MAXSIZE: 4096 2022-11-19 21:16:39,046 INFO cfg.MODEL.POST_PROCESSING.NMS_CONFIG.NMS_POST_MAXSIZE: 500 2022-11-19 21:16:39,046 INFO
    cfg.OPTIMIZATION = edict() 2022-11-19 21:16:39,046 INFO cfg.OPTIMIZATION.BATCH_SIZE_PER_GPU: 4 2022-11-19 21:16:39,046 INFO cfg.OPTIMIZATION.NUM_EPOCHS: 80 2022-11-19 21:16:39,046 INFO cfg.OPTIMIZATION.OPTIMIZER: adam_onecycle 2022-11-19 21:16:39,046 INFO cfg.OPTIMIZATION.LR: 0.001 2022-11-19 21:16:39,046 INFO cfg.OPTIMIZATION.WEIGHT_DECAY: 0.01 2022-11-19 21:16:39,046 INFO cfg.OPTIMIZATION.MOMENTUM: 0.9 2022-11-19 21:16:39,046 INFO cfg.OPTIMIZATION.MOMS: [0.95, 0.85] 2022-11-19 21:16:39,046 INFO cfg.OPTIMIZATION.PCT_START: 0.4 2022-11-19 21:16:39,047 INFO cfg.OPTIMIZATION.DIV_FACTOR: 10 2022-11-19 21:16:39,047 INFO cfg.OPTIMIZATION.DECAY_STEP_LIST: [35, 45] 2022-11-19 21:16:39,047 INFO cfg.OPTIMIZATION.LR_DECAY: 0.1 2022-11-19 21:16:39,047 INFO cfg.OPTIMIZATION.LR_CLIP: 1e-07 2022-11-19 21:16:39,047 INFO cfg.OPTIMIZATION.LR_WARMUP: False 2022-11-19 21:16:39,047 INFO cfg.OPTIMIZATION.WARMUP_EPOCH: 1 2022-11-19 21:16:39,047 INFO cfg.OPTIMIZATION.GRAD_NORM_CLIP: 10 2022-11-19 21:16:39,048 INFO cfg.TAG: CaDDN 2022-11-19 21:16:39,048 INFO cfg.EXP_GROUP_PATH: content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/cfgs/kitti_models 2022-11-19 21:16:39,048 INFO Loading KITTI dataset 2022-11-19 21:16:39,048 INFO Total samples for KITTI dataset: 0 /content/gdrive/MyDrive/IITH_VCA/CaDDN/checkpoints/deeplabv3_resnet101_coco-586e9e4e.pth 2022-11-19 21:16:42,793 INFO ==> Loading parameters from checkpoint /content/gdrive/MyDrive/IITH_VCA/CaDDN/checkpoints/deeplabv3_resnet101_coco-586e9e4e.pth to GPU Traceback (most recent call last): File "/content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/test.py", line 200, in main() File "/content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/test.py", line 196, in main eval_single_ckpt(model, test_loader, args, eval_output_dir, logger, epoch_id, dist_test=dist_test) File "/content/gdrive/MyDrive/IITH_VCA/CaDDN/tools/test.py", line 58, in eval_single_ckpt model.load_params_from_file(filename=args.ckpt, logger=logger, to_cpu=dist_test) File "/content/gdrive/MyDrive/IITH_VCA/CaDDN/pcdet/models/detectors/detector3d_template.py", line 359, in load_params_from_file model_state_disk = checkpoint['model_state'] KeyError: 'model_state'

    opened by prabhata07 1
  • the result is not reproduced

    the result is not reproduced

    I trained the model with 2 Tesla V100 GPUs in the default configuration and got Easy@R40, Moderate@R40, Hard@R40 21.203, 15.1487,12.7407 respectively， which is lower than the 23.77,16.07, 13.61 you declared. But when I use the pre-trained weight, I can get 23.77,16.07, 13.61. So, are there any training tricks that have not been published?

    opened by leo038 0
Releases(v0.3.1)
Owner
Toronto Robotics and AI Laboratory
Toronto Robotics and AI Laboratory
GitHub
Monocular Depth Estimation Using Laplacian Pyramid-Based Depth Residuals

LapDepth-release This repository is a Pytorch implementation of the paper "Monocular Depth Estimation Using Laplacian Pyramid-Based Depth Residuals" M

Minsoo Song 205 Dec 30, 2022
Official implementation of the network presented in the paper "M4Depth: A motion-based approach for monocular depth estimation on video sequences"

M4Depth This is the reference TensorFlow implementation for training and testing depth estimation models using the method described in M4Depth: A moti

Michaël Fonder 76 Jan 3, 2023
Apply our monocular depth boosting to your own network!

MergeNet - Boost Your Own Depth Boost custom or edited monocular depth maps using MergeNet Input Original result After manual editing of base You can

Computational Photography Lab @ SFU 142 Dec 17, 2022
Light-weight network, depth estimation, knowledge distillation, real-time depth estimation, auxiliary data.

light-weight-depth-estimation Boosting Light-Weight Depth Estimation Via Knowledge Distillation, https://arxiv.org/abs/2105.06143 Junjie Hu, Chenyou F

Junjie Hu 13 Dec 10, 2022
Code for Transformers Solve Limited Receptive Field for Monocular Depth Prediction

Official PyTorch code for Transformers Solve Limited Receptive Field for Monocular Depth Prediction. Guanglei Yang, Hao Tang, Mingli Ding, Nicu Sebe,

stanley 152 Dec 16, 2022
AdelaiDepth is an open source toolbox for monocular depth prediction.

AdelaiDepth is an open source toolbox for monocular depth prediction.

Adelaide Intelligent Machines (AIM) Group 743 Jan 1, 2023
Code and datasets for the paper "Combining Events and Frames using Recurrent Asynchronous Multimodal Networks for Monocular Depth Prediction" (RA-L, 2021)

Combining Events and Frames using Recurrent Asynchronous Multimodal Networks for Monocular Depth Prediction This is the code for the paper Combining E

Robotics and Perception Group 69 Dec 26, 2022
Boosting Monocular Depth Estimation Models to High-Resolution via Content-Adaptive Multi-Resolution Merging

Boosting Monocular Depth Estimation Models to High-Resolution via Content-Adaptive Multi-Resolution Merging This repository contains an implementation

Computational Photography Lab @ SFU 1.1k Jan 2, 2023
[CVPR 2021] Monocular depth estimation using wavelets for efficiency

Single Image Depth Prediction with Wavelet Decomposition Michaël Ramamonjisoa, Michael Firman, Jamie Watson, Vincent Lepetit and Daniyar Turmukhambeto

Niantic Labs 205 Jan 2, 2023
[ICCV 2021] Excavating the Potential Capacity of Self-Supervised Monocular Depth Estimation

EPCDepth EPCDepth is a self-supervised monocular depth estimation model, whose supervision is coming from the other image in a stereo pair. Details ar

Rui Peng 110 Dec 23, 2022
Advancing Self-supervised Monocular Depth Learning with Sparse LiDAR

Official implementation for paper "Advancing Self-supervised Monocular Depth Learning with Sparse LiDAR"

Ziyue Feng 72 Dec 9, 2022
This repo is for Self-Supervised Monocular Depth Estimation with Internal Feature Fusion(arXiv), BMVC2021

DIFFNet This repo is for Self-Supervised Monocular Depth Estimation with Internal Feature Fusion(arXiv), BMVC2021 A new backbone for self-supervised d

Hang 3 Oct 22, 2021
the official code for ICRA 2021 Paper: "Multimodal Scale Consistency and Awareness for Monocular Self-Supervised Depth Estimation"

G2S This is the official code for ICRA 2021 Paper: Multimodal Scale Consistency and Awareness for Monocular Self-Supervised Depth Estimation by Hemang

NeurAI 4 Jul 27, 2022
SimpleDepthEstimation - An unified codebase for NN-based monocular depth estimation methods

SimpleDepthEstimation Introduction This is an unified codebase for NN-based monocular depth estimation methods, the framework is based on detectron2 (

null 8 Dec 13, 2022
ONNX-GLPDepth - Python scripts for performing monocular depth estimation using the GLPDepth model in ONNX

ONNX-GLPDepth - Python scripts for performing monocular depth estimation using the GLPDepth model in ONNX

Ibai Gorordo 18 Nov 6, 2022
ONNX-PackNet-SfM: Python scripts for performing monocular depth estimation using the PackNet-SfM model in ONNX

Python scripts for performing monocular depth estimation using the PackNet-SfM model in ONNX

Ibai Gorordo 14 Dec 9, 2022
Repository for "Toward Practical Monocular Indoor Depth Estimation" (CVPR 2022)

Toward Practical Monocular Indoor Depth Estimation Cho-Ying Wu, Jialiang Wang, Michael Hall, Ulrich Neumann, Shuochen Su [arXiv] [project site] DistDe

Meta Research 122 Dec 13, 2022
A working implementation of the Categorical DQN (Distributional RL).

Categorical DQN. Implementation of the Categorical DQN as described in A distributional Perspective on Reinforcement Learning. Thanks to @tudor-berari

Florin Gogianu 98 Sep 20, 2022
The implemention of Video Depth Estimation by Fusing Flow-to-Depth Proposals

Flow-to-depth (FDNet) video-depth-estimation This is the implementation of paper Video Depth Estimation by Fusing Flow-to-Depth Proposals Jiaxin Xie,

null 32 Jun 14, 2022