autovision-cloud / sa-det3d Goto Github PK

View Code? Open in Web Editor NEW

165.0 165.0 32.0 1.74 MB

[ICCVW-2021] SA-Det3D: Self-attention based Context-Aware 3D Object Detection

License: MIT License

Python 99.63% Shell 0.37%

3d-deep-learning 3d-detection autonomous-driving object-detection point-cloud pytorch

sa-det3d's Introduction

Hi there 👋

🌱 I'm Prarthana Bhattacharyya, currently a Ph.D. Candidate at University of Waterloo.
🔭 I'm broadly interested in deep learning and its applications to the field of computer vision.
🎯 I focus on the following areas, particularly in the context of autonomous robots:
- 3D geometric understanding and perception
- temporal motion forecasting
- self-supervised representation learning

sa-det3d's People

Contributors

Stargazers

Watchers

sa-det3d's Issues

How can I get all the result of KITTI?

Thanks for your great work!

I want to get all the result of train, val and test set for my tracking work. How can I get these via script like

python test.py --cfg_file cfgs/kitti_models/pointpillar_fsa.yaml --batch_size 4 --ckpt ${SAVED_CKPT_PATH}/pointpillar_fsa.pth

Look forward to your reply！

`dsa` model training problem

Hello,

I am using the newest OpenPCDet, and copy your codes to it.

I met a problem while training DSA model of second/pointpillar/pvrcnn, while not meeting this problem in pointrcnn_dsa, errors is as follows:

Expected to have finished reduction in the prior iteration before starting a new one. This error indicates that your module has parameters that were not used in producing loss. You can enable unused parameter detection by passing the keyword argument `find_unused_parameters=True` to `torch.nn.parallel.DistributedDataParallel`, and by 
making sure all `forward` function outputs participate in calculating loss. 
If you already have done the above, then the distributed data parallel module wasn't able to locate the output tensors in the return value of your module's `forward` function. Please include the loss function and the structure of the return value of `forward` of your module when reporting this issue (e.g. list, dict, iterable).
Parameter indices which did not receive grad for rank 0: 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87
 In addition, you can set the environment variable TORCH_DISTRIBUTED_DEBUG to either INFO or DETAIL to print out information about which particular parameters did not receive gradient on this rank as part of this error

so I follow the instructions and set find_unused_parameters=True in torch.nn.parallel.DistributedDataParallel, it can training normal, did you meet this problem while not setting find_unused_parameters=True in torch.nn.parallel.DistributedDataParallel?

here env is:
pytorch 1.10.1

Name: torch
Version: 1.10.1+cu111
Summary: Tensors and Dynamic neural networks in Python with strong GPU acceleration
Home-page: https://pytorch.org/
Author: PyTorch Team
Author-email: [email protected]
License: BSD-3
Location: /home/deze/.conda/envs/torch110_py38/lib/python3.8/site-packages
Requires: typing-extensions
Required-by: thop, torchvision

cuda 11.1

spconv_cu111

Name: spconv-cu111
Version: 2.1.25
Summary: spatial sparse convolution
Home-page: https://github.com/traveller59/spconv
Author: Yan Yan
Author-email: [email protected]
License: Apache License 2.0
Location: /home/deze/.conda/envs/torch110_py38/lib/python3.8/site-packages
Requires: ccimport, cumm-cu111, fire, numpy, pccm, pybind11
Required-by:

Regards,
Deze

About the cyclist on KITTI test

Dear author, i train your pvrcnn_fas.yaml，only cyclist，and split the training set to 80% and 20% for training and validating, then choose the best ckpt to run the test set, but only get: 75.59 % | 58.31 % | 52.01 % for easy, mod, hard , which are far away from your results in your paper. Can you tell me how to reproduce your results on cyclist?? Thank you!

Improving speed of Pointpillars

Hello @AutoVision-cloud, thanks for the release of your nice work.

I have a question regarding the speed of Pointpillars. Since you determined that Pointpillars-FSA (or -DSA) runs with nearly half the amount of G-FLOPs, I assume this means your model should have a faster inference speed than the baseline...
if my assumption is correct, do you know how much faster can it run?
I tested Pointpillars-FSA with custom data and compared the speed with base Pointpillars, but didn't see any difference in speed....
To me it's not clear how Pointpillars-FSA would perform less G-FLOPs if you are adding extra layers to the baseline, con you please explain? Or for faster speed should I pass only the context features to the BEVEncoder, and ignore the pillar features, thus avoiding concatenation of pillar and context features?

[WARN] Cannot find rule for <class 'pcdet.models.detectors.pv_rcnn.PVRCNN'>. Treat it as zero Macs and zero Params.

I trained the code on the local sever(TITAN RTX), the result is worse than the paper reported. I got some WARNING, that may make the performance decrease. How could I solve this?

2021-09-18 18:13:00,168   INFO  ==> Loading parameters from checkpoint ../output/kitti_models/pvrcnn_fsa/default/ckpt/checkpoint_epoch_80.pth to GPU
2021-09-18 18:13:00,259   INFO  ==> Checkpoint trained from version: pcdet+0.3.0+a7cf536
2021-09-18 18:13:00,764   INFO  ==> Done (loaded 382/382)
2021-09-18 18:13:00,778   INFO  *************** EPOCH 80 EVALUATION *****************
eval:   0%|                                             | 0/943 [00:00<?, ?it/s][WARN] Cannot find rule for <class 'pcdet.models.backbones_3d.vfe.mean_vfe.MeanVFE'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'spconv.conv.SubMConv3d'>. Treat it as zero Macs and zero Params.
[INFO] Register count_bn() for <class 'torch.nn.modules.batchnorm.BatchNorm1d'>.
[INFO] Register zero_ops() for <class 'torch.nn.modules.activation.ReLU'>.
[WARN] Cannot find rule for <class 'spconv.modules.SparseSequential'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'spconv.conv.SparseConv3d'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.models.backbones_3d.spconv_backbone.SlimVoxelBackBone8x'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.models.backbones_2d.map_to_bev.height_compression.HeightCompression'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.ops.pointnet2.pointnet2_stack.pointnet2_utils.QueryAndGroup'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'torch.nn.modules.container.ModuleList'>. Treat it as zero Macs and zero Params.
[INFO] Register count_convNd() for <class 'torch.nn.modules.conv.Conv2d'>.
[INFO] Register count_bn() for <class 'torch.nn.modules.batchnorm.BatchNorm2d'>.
[WARN] Cannot find rule for <class 'torch.nn.modules.container.Sequential'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.ops.pointnet2.pointnet2_stack.pointnet2_modules.StackSAModuleMSG'>. Treat it as zero Macs and zero Params.
[INFO] Register count_linear() for <class 'torch.nn.modules.linear.Linear'>.
[INFO] Register count_convNd() for <class 'torch.nn.modules.conv.Conv1d'>.
[WARN] Cannot find rule for <class 'torch.nn.modules.normalization.GroupNorm'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'torch.nn.modules.activation.Softmax'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.models.backbones_3d.sa_block.SA_block'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.models.backbones_3d.pfe.sa_voxel_set_abstraction.SAVoxelSetAbstraction'>. Treat it as zero Macs and zero Params.
[INFO] Register zero_ops() for <class 'torch.nn.modules.padding.ZeroPad2d'>.
[WARN] Cannot find rule for <class 'pcdet.models.backbones_2d.encoder_2d.bev_encoder.BaseBEVEncoder'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.models.backbones_3d.cfe.voxel_fsa.PositionalEncoding'>. Treat it as zero Macs and zero Params.
[INFO] Register zero_ops() for <class 'torch.nn.modules.dropout.Dropout'>.
[WARN] Cannot find rule for <class 'torch.nn.modules.normalization.LayerNorm'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.models.backbones_3d.cfe.voxel_fsa.VoxelContext3D_fsa'>. Treat it as zero Macs and zero Params.
[INFO] Register count_convNd() for <class 'torch.nn.modules.conv.ConvTranspose2d'>.
[WARN] Cannot find rule for <class 'pcdet.models.backbones_2d.decoder_2d.bev_decoder.ConcatVoxelDecoder'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.utils.loss_utils.SigmoidFocalClassificationLoss'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.utils.loss_utils.WeightedSmoothL1Loss'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.utils.loss_utils.WeightedCrossEntropyLoss'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.models.dense_heads.anchor_head_single.AnchorHeadSingle'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.models.dense_heads.point_head_simple.PointHeadSimple'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.models.roi_heads.target_assigner.proposal_target_layer.ProposalTargetLayer'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.models.roi_heads.pvrcnn_head.PVRCNNHead'>. Treat it as zero Macs and zero Params.
[WARN] Cannot find rule for <class 'pcdet.models.detectors.pv_rcnn.PVRCNN'>. Treat it as zero Macs and zero Params.
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/torch/nn/functional.py:718: UserWarning: Named tensors and all their associated APIs are an experimental feature and subject to change. Please do not use them for anything important until they are released as stable. (Triggered internally at  /pytorch/c10/core/TensorImpl.h:1156.)
  return torch.max_pool2d(input, kernel_size, stride, padding, dilation, ceil_mode)
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/torch/nn/functional.py:3613: UserWarning: Default upsampling behavior when mode=bilinear is changed to align_corners=False since 0.4.0. Please specify align_corners=True if the old behavior is desired. See the documentation of nn.Upsample for details.
  "See the documentation of nn.Upsample for details.".format(mode)
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/torch/nn/functional.py:3658: UserWarning: The default behavior for interpolate/upsample with float scale_factor changed in 1.6.0 to align with other frameworks/libraries, and now uses scale_factor directly, instead of relying on the computed output size. If you wish to restore the old behavior, please set recompute_scale_factor=True. See the documentation of nn.Upsample for details. 
  "The default behavior for interpolate/upsample with float scale_factor changed "
Model PARAMS:  , Model G-FLOPS:  9959985.0 	 64276345568.0
eval: 100%|█| 943/943 [05:37<00:00,  2.79it/s, recall_0.3=(16989, 17000) / 17558
2021-09-18 18:18:38,355   INFO  *************** Performance of EPOCH 80 *****************
2021-09-18 18:18:38,355   INFO  Generate label finished(sec_per_example: 0.0896 second).
2021-09-18 18:18:38,355   INFO  recall_roi_0.3: 0.967593
2021-09-18 18:18:38,355   INFO  recall_rcnn_0.3: 0.968220
2021-09-18 18:18:38,355   INFO  recall_roi_0.5: 0.923112
2021-09-18 18:18:38,355   INFO  recall_rcnn_0.5: 0.930003
2021-09-18 18:18:38,355   INFO  recall_roi_0.7: 0.676558
2021-09-18 18:18:38,355   INFO  recall_rcnn_0.7: 0.743365
2021-09-18 18:18:38,357   INFO  Average predicted number of objects(3769 samples): 9.800
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
  warn(NumbaDeprecationWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
  warn(NumbaDeprecationWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
  warn(NumbaDeprecationWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
  warn(NumbaDeprecationWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
  warn(NumbaDeprecationWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
  warn(NumbaDeprecationWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
  warn(NumbaDeprecationWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
  warn(NumbaDeprecationWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
  warn(NumbaDeprecationWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/decorators.py:110: NumbaDeprecationWarning: Eager compilation of device functions is deprecated (this occurs when a signature is provided)
  warn(NumbaDeprecationWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (24) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (30) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (35) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (20) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (40) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (28) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (25) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (96) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (24) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (30) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (30) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (24) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (35) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (20) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (40) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (28) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (25) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
/home/ds/anaconda3/envs/pcdet/lib/python3.7/site-packages/numba/cuda/compiler.py:865: NumbaPerformanceWarning: Grid size (96) < 2 * SM count (144) will likely result in GPU under utilization due to low occupancy.
  warn(NumbaPerformanceWarning(msg))
2021-09-18 18:18:58,305   INFO  Car [email protected], 0.70, 0.70:
bbox AP:95.6558, 89.2137, 88.9727
bev  AP:90.0087, 87.4705, 86.9144
3d   AP:89.0232, 78.7346, 78.2206
aos  AP:95.61, 89.10, 88.80
Car [email protected], 0.70, 0.70:
bbox AP:97.8422, 93.7267, 91.9560
bev  AP:92.8356, 88.4083, 88.2203
3d   AP:91.2983, 82.0211, 79.9662
aos  AP:97.81, 93.58, 91.75
Car [email protected], 0.50, 0.50:
bbox AP:95.6558, 89.2137, 88.9727
bev  AP:95.7055, 89.2747, 89.0953
3d   AP:95.6466, 89.2571, 89.0537
aos  AP:95.61, 89.10, 88.80
Car [email protected], 0.50, 0.50:
bbox AP:97.8422, 93.7267, 91.9560
bev  AP:97.8848, 94.1767, 94.1172
3d   AP:97.8545, 94.0966, 93.9559
aos  AP:97.81, 93.58, 91.75
Pedestrian [email protected], 0.50, 0.50:
bbox AP:71.9394, 66.0419, 63.8174
bev  AP:67.1435, 60.5519, 56.4955
3d   AP:65.0513, 57.2925, 53.5970
aos  AP:64.38, 59.06, 56.54
Pedestrian [email protected], 0.50, 0.50:
bbox AP:72.2893, 67.0041, 64.0620
bev  AP:67.4582, 59.9565, 55.6927
3d   AP:65.0522, 56.9932, 52.4130
aos  AP:63.73, 58.78, 55.69
Pedestrian [email protected], 0.25, 0.25:
bbox AP:71.9394, 66.0419, 63.8174
bev  AP:76.9071, 73.0327, 70.4967
3d   AP:76.8903, 73.0165, 70.3640
aos  AP:64.38, 59.06, 56.54
Pedestrian [email protected], 0.25, 0.25:
bbox AP:72.2893, 67.0041, 64.0620
bev  AP:78.6493, 74.0522, 71.0445
3d   AP:78.6324, 74.0364, 70.9557
aos  AP:63.73, 58.78, 55.69
Cyclist [email protected], 0.50, 0.50:
bbox AP:89.5657, 76.8325, 75.1482
bev  AP:87.4041, 72.6511, 69.3511
3d   AP:85.7404, 70.5434, 64.8741
aos  AP:89.38, 76.23, 74.44
Cyclist [email protected], 0.50, 0.50:
bbox AP:94.6328, 79.5327, 75.5670
bev  AP:90.2565, 73.6993, 69.4862
3d   AP:88.3381, 70.5201, 66.3709
aos  AP:94.37, 78.81, 74.86
Cyclist [email protected], 0.25, 0.25:
bbox AP:89.5657, 76.8325, 75.1482
bev  AP:88.4720, 73.7940, 72.0709
3d   AP:88.4720, 73.7940, 72.0709
aos  AP:89.38, 76.23, 74.44
Cyclist [email protected], 0.25, 0.25:
bbox AP:94.6328, 79.5327, 75.5670
bev  AP:93.1332, 76.1067, 73.0396
3d   AP:93.1332, 76.1067, 73.0396
aos  AP:94.37, 78.81, 74.86

2021-09-18 18:18:58,309   INFO  Result is save to /home/ds/anaconda3/envs/pcdet/SA-Det3D/OpenPCDet/output/kitti_models/pvrcnn_fsa/default/eval/epoch_80/val/default
2021-09-18 18:18:58,309   INFO  ****************Evaluation done.*****************

How to compute the GFLOPs of one PV-RCNN?

Dear authors,

In your nice paper, you mention the GFLOPs and I want to know how you compute them?

Thanks so much.

Point cloud cropping error

I use the pre-trained model as well as my own point cloud data and then run demo.py. The result is the rear point cloud of the vehicle rather than the front Point Cloud.

Small gap between the reproduced and the reported results

Dear authors,

Thanks for your good work. I have reproduced the FSA results on KITTI val split with 40 recall positions across PointPillars, SECOND, Point-RCNN, and PV-RCNN. I use 4 GPUs for training and use the training cfg files provided. Below is the results.

---	3D	BEV
PointPillars (Report)	79.04	88.47
PointPillars (Reproduce)	78.64 (-0.4)	88.11
SECOND (Report)	81.86	90.01
SECOND (Reproduce)	81.48 (-0.38)	90.20
Point R-CNN (Report)	82.10	88.37
Point R-CNN (Reproduce)	81.86 (-0.24)	88.29
PV-RCNN (Report)	84.95	90.92
PV-RCNN (Reproduce)	84.68 (-0.27)	91.01

Would you please provide some suggestions to remove the gap? Thank you very much in advance.

Regards,
Yukang Chen

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.