[Feature] ImVoteNet complete model #337

yezhen17 · 2021-03-05T05:58:43Z

Added the ImVoteNet complete model, including initialization, multi-tower training and test (simple and aug).

* fix bug caused by mmcv upgrade; delete pdb breakpoint * fix typos

* Modify README * modify the mmdetection to mmdetection3d in CONTRIBUTING.md and add model zoo link before demo in getting_started.md * arxiv->Arxiv * Add model zoo link

* Move train_cfg/test_cfg to model * Move train_cfg/test_cfg to model * Move train_cfg/test_cfg to model * Move train_cfg/test_cfg to model * Move train_cfg/test_cfg to model * Move train_cfg/test_cfg to model * Move train_cfg/test_cfg to model * Move train_cfg and test_cfg into model * modify centerpoint configs * Modify docs * modify build_detector * modify test_config_build_detector * modify build_detector parameters * Adopt the same strategy in build_detector

* Add chinese version readme * Delete useless words * modify the link of dynamic voxelization * modify readme * remove it * remove it * modify chinese version readme * delete the wrong whitespace * Modify chinese readme * Modify chinese readme * modify chinese version readme

* fix bug caused by mmcv upgrade; delete pdb breakpoint * fix typos * use torch.cuda api * update unittest for points_in_bbox_gpu and points_in_boxes_batch * Added comments for explanation

* add h3d backbone * add h3d backbone * add h3dnet * modify scannet config * fix bugs for proposal refine * fix bugs for test backbone * add primitive head test * modify h3dhead * modify h3d head * update loss weight config * fix bugs for h3d head loss * modify h3d head get targets function * update h3dnet base config * modify weighted loss * Revert "Merge branch 'h3d_u2' into 'master'" This reverts merge request !5 * add open3d visual * modify docstring * use bbox and coord convert * modify docstring * fix bugs for unittest * fix bugs for unittest * modify doc * add visualize script and modify docs * modify docs * fix bugs

…en-mmlab#318) * a faster & more memory-efficient implementation of DynamicScatter * fix format issues and add pytest skip code for tests on machines without cuda support * some trivial changes: decrease the number of kernel threads per block to 512, to enable inference on GPUs with computing capability lower than 2.0 change the backpropagation behavior of max-reduction. when there are multiple points shares the same maximum feature value, only the first point (with lowest row index) among them is chosen to propagate the output gradient back. before this change, all points with the same maximum feature value can propagate the output gradient back. this change makes the max-reduction behaves in consistence with torch.max. this change may cause gradcheck failure in test_dynamic_scatter.py. please do not worry about it because torch.max fails the gradcheck too. * fix typo Co-authored-by: zhanggefan <1152009@tongji.edu.cn>

* Reduce the open3d dependency * Reduce the open3d dependency * Reduce dependency of open3d if open3d is not used * remove the warning in show_result

* fix nonzero warning with as_tuple * fix nonzero warning with as_tuple open-mmlab#320

…ation when using GPUs with compute capability lower than 6.x (open-mmlab#326) * fix a bug that may cause compilation failure of dynamic voxelization when using gpus with compute capability lower than 6.x fix imperfection kernel code that may unintentionally discard valid points when input points count is larger than 50000 * 512 (nearly impossible though). * Modified scatter_points_cuda.cu to ensure backward compatibility with PyTorch1.5 on CUDA9.0 * fix the issue of DynamicScatter gradient check failure by explicit mark non-floating-point tensor as non-differentiable.

* Update README.md * Update README_zh-CN.md * Update getting_started.md * Update version.py * Update changelog.md * Update changelog.md * Update changelog.md

codecov · 2021-03-05T10:00:57Z

Codecov Report

Merging #337 (8fa5b7f) into imvotenet (8ed6ed9) will decrease coverage by 1.21%.
The diff coverage is 17.10%.

@@              Coverage Diff              @@
##           imvotenet     #337      +/-   ##
=============================================
- Coverage      50.81%   49.59%   -1.22%     
=============================================
  Files            177      178       +1     
  Lines          11950    12319     +369     
  Branches        1867     1934      +67     
=============================================
+ Hits            6072     6110      +38     
- Misses          5493     5823     +330     
- Partials         385      386       +1

Flag	Coverage Δ
unittests	`49.59% <17.10%> (-1.22%)`	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
mmdet3d/models/dense_heads/anchor3d_head.py	`61.85% <ø> (ø)`
mmdet3d/models/dense_heads/ssd_3d_head.py	`13.93% <0.00%> (ø)`
mmdet3d/models/dense_heads/train_mixins.py	`6.47% <ø> (ø)`
mmdet3d/models/detectors/centerpoint.py	`15.66% <ø> (+0.36%)`	⬆️
mmdet3d/models/fusion_layers/point_fusion.py	`67.02% <ø> (ø)`
...det3d/models/roi_heads/bbox_heads/h3d_bbox_head.py	`23.23% <0.00%> (ø)`
...et3d/models/roi_heads/mask_heads/primitive_head.py	`18.91% <ø> (ø)`
mmdet3d/ops/roiaware_pool3d/points_in_boxes.py	`32.25% <0.00%> (-11.23%)`	⬇️
mmdet3d/core/visualizer/open3d_vis.py	`10.58% <10.58%> (ø)`
mmdet3d/models/detectors/imvotenet.py	`18.80% <12.18%> (-16.83%)`	⬇️
... and 18 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 8ed6ed9...c8749c9. Read the comment docs.

mmdet3d/models/detectors/imvotenet.py

mmdet3d/models/model_utils/img_feat_mlp_module.py

Tai-Wang

Just need some minor adjustments. There should be another two PRs about benchmark/configs and unit test I guess for complete the overall model.

yezhen17 · 2021-03-07T12:52:00Z

Just fixed some bugs after I tested the whole model. The next PR should be ready soon.

* add faq * faq revised * faq revised * faq revised * faq revised * faq revised --- add title & add line to index.rst

* refactor * refactor train/test * refactor train/test * refactor train.py/test.py * refactor train.py/test.py --- specify default of seed to 0 & save mmdet version instead of mmdet3d git hash * refactor train.py/test.py --- specify default of seed to 0 & save mmdet version instead of mmdet3d git hash

mmdet3d/models/detectors/imvotenet.py

Tai-Wang

Please finally check all the docstrings one more time and elaborate the incomplete ones. Also fix the failing of CI.

yezhen17 · 2021-03-10T08:46:26Z

For CI fix, refer to #350 @Tai-Wang

yezhen17 and others added 30 commits February 3, 2021 17:27

integrated vote fusion

b373499

coords transform and unit test

3be3690

Update docstring

cce6534

Merge branch 'imvotenet_fork' into imvotenet

b88085b

refactor and add unit test

747cfd9

fix bug caused by mmcv upgrade; delete pdb breakpoint (open-mmlab#304)

ac9a3e8

add point fusion unittest

1e4f12c

remove unused file

b9cf3ca

[Fix] Fix typos and error in comments (open-mmlab#308)

5592a50

* fix bug caused by mmcv upgrade; delete pdb breakpoint * fix typos

updates

4415f49

[Enhance] Modify readme (open-mmlab#310)

a347ac7

* Modify README * modify the mmdetection to mmdetection3d in CONTRIBUTING.md and add model zoo link before demo in getting_started.md * arxiv->Arxiv * Add model zoo link

add assertion info

a96d949

update

c419f9b

[Fix] Fix error when tensors are not in the same device (open-mmlab#317)

bdeacec

* fix bug caused by mmcv upgrade; delete pdb breakpoint * fix typos * use torch.cuda api * update unittest for points_in_bbox_gpu and points_in_boxes_batch * Added comments for explanation

add unittest

c8c466f

add vote fusion unittest

97552b5

add vote fusion unittest

770ed19

[Enhance]: Hide the import of open3d (open-mmlab#323)

ccd3047

* Reduce the open3d dependency * Reduce the open3d dependency * Reduce dependency of open3d if open3d is not used * remove the warning in show_result

Fix nonzero warning (open-mmlab#330)

df505fe

* fix nonzero warning with as_tuple * fix nonzero warning with as_tuple open-mmlab#320

Bump to v0.11.0 (open-mmlab#324)

e366272

* Update README.md * Update README_zh-CN.md * Update getting_started.md * Update version.py * Update changelog.md * Update changelog.md * Update changelog.md

minor update

f289bc5

docstring

abf54e8

Merge branch 'imvotenet_fork' into _imvotenet_backbone

a47a6ff

Merge branch '_imvotenet_fusion' into _imvotenet_backbone

a80769e

initial update of imvotenet

68f07fc

Tai-Wang reviewed Mar 7, 2021

View reviewed changes

mmdet3d/models/detectors/imvotenet.py Show resolved Hide resolved

Tai-Wang reviewed Mar 7, 2021

View reviewed changes

mmdet3d/models/detectors/imvotenet.py Show resolved Hide resolved

Tai-Wang reviewed Mar 7, 2021

View reviewed changes

mmdet3d/models/model_utils/img_feat_mlp_module.py Outdated Show resolved Hide resolved

Tai-Wang reviewed Mar 7, 2021

View reviewed changes

yezhen17 added 4 commits March 7, 2021 14:13

add docstring

ef04ac5

Merge branch '_imvotenet_backbone' into imvotenet

0da8c3f

fix bugs

ba361de

Merge branch '_imvotenet_backbone' into imvotenet

122d829

yezhen17 and others added 10 commits March 8, 2021 14:04

fix bug

a542dde

fix bug

b955263

Merge branch '_imvotenet_backbone' into imvotenet

4464dd9

[Enhance] Add FAQ for common problems (open-mmlab#333)

43dde9d

* add faq * faq revised * faq revised * faq revised * faq revised * faq revised --- add title & add line to index.rst

Fix missing coord_type in db-sampler config. (open-mmlab#345)

b65f9b0

fix bug

cce3e38

Merge branch '_imvotenet_backbone' into imvotenet

5bb4136

fix bug

5baaa1f

Merge branch '_imvotenet_backbone' into imvotenet

69c8d9b

Tai-Wang reviewed Mar 10, 2021

View reviewed changes

mmdet3d/models/detectors/imvotenet.py Outdated Show resolved Hide resolved

Tai-Wang approved these changes Mar 10, 2021

View reviewed changes

yezhen17 added 2 commits March 10, 2021 15:20

add docstring

7714432

Merge branch '_imvotenet_backbone' into imvotenet

27c478e

yezhen17 and others added 3 commits March 10, 2021 17:33

using ubuntu 18.04 in github workflow (open-mmlab#350)

e5f55bd

merge conflicts

453a952

move train/test cfg

c8749c9

ZwwWayne closed this Mar 17, 2021

tpoisonooo pushed a commit to tpoisonooo/mmdetection3d that referenced this pull request Sep 5, 2022

fix intro (open-mmlab#337)

1a35cf8

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[Feature] ImVoteNet complete model #337

[Feature] ImVoteNet complete model #337

yezhen17 commented Mar 5, 2021

codecov bot commented Mar 5, 2021 •

edited

Loading

Tai-Wang left a comment •

edited

Loading

yezhen17 commented Mar 7, 2021

Tai-Wang left a comment •

edited

Loading

yezhen17 commented Mar 10, 2021

[Feature] ImVoteNet complete model #337

[Feature] ImVoteNet complete model #337

Conversation

yezhen17 commented Mar 5, 2021

codecov bot commented Mar 5, 2021 • edited Loading

Codecov Report

Tai-Wang left a comment • edited Loading

Choose a reason for hiding this comment

yezhen17 commented Mar 7, 2021

Tai-Wang left a comment • edited Loading

Choose a reason for hiding this comment

yezhen17 commented Mar 10, 2021

codecov bot commented Mar 5, 2021 •

edited

Loading

Tai-Wang left a comment •

edited

Loading

Tai-Wang left a comment •

edited

Loading