Release v1.0.0rc1

open-mmlab · Apr 6, 2022 · 333536f · 333536f
2 parents 9c7270d + f747daa
commit 333536f
Show file tree

Hide file tree

Showing 219 changed files with 2,555 additions and 14,403 deletions.
diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
@@ -103,3 +103,44 @@ jobs:
           env_vars: OS,PYTHON
           name: codecov-umbrella
           fail_ci_if_error: false
+
+  build_windows:
+    runs-on: ${{ matrix.os }}
+    strategy:
+      matrix:
+        os: [windows-2022]
+        python: [3.8]
+        platform: [cpu]
+    steps:
+      - uses: actions/checkout@v2
+      - name: Set up Python ${{ matrix.python }}
+        uses: actions/setup-python@v2
+        with:
+          python-version: ${{ matrix.python }}
+      - name: Upgrade pip
+        run: pip install pip --upgrade --user
+      - name: Install PyTorch
+        # As a complement to Linux CI, we test on PyTorch LTS version
+        run: pip install torch==1.8.2+${{ matrix.platform }} torchvision==0.9.2+${{ matrix.platform }} -f https://download.pytorch.org/whl/lts/1.8/torch_lts.html
+      - name: Install mmdet3d dependencies
+        run: |
+          pip install mmcv-full -f https://download.openmmlab.com/mmcv/dist/cpu/torch1.8/index.html --only-binary mmcv-full
+          python -m pip install mmdet==2.19.0
+          python -m pip install mmsegmentation==0.20.0
+          python -m pip install -r requirements/build.txt -r requirements/runtime.txt -r requirements/tests.txt
+      - name: Build and install
+        run: pip install -e .
+      - name: Run unittests and generate coverage report
+        run: coverage run --branch --source mmdet3d -m pytest tests/
+      - name: Generate coverage report
+        run: |
+          coverage xml
+          coverage report -m
+      - name: Upload coverage to Codecov
+        uses: codecov/codecov-action@v2
+        with:
+          file: ./coverage.xml
+          flags: unittests
+          env_vars: OS,PYTHON
+          name: codecov-umbrella
+          fail_ci_if_error: false
diff --git a/README.md b/README.md
@@ -25,11 +25,11 @@
 [![license](https://img.shields.io/github/license/open-mmlab/mmdetection3d.svg)](https://github.com/open-mmlab/mmdetection3d/blob/master/LICENSE)
 
 
-**News**: We released the codebase v1.0.0rc0.
+**News**: We released the codebase v1.0.0rc1.
 
 Note: We are going through large refactoring to provide simpler and more unified usage of many modules.
 
-The compatibilities of models are broken due to the unification and simplification of coordinate systems. For now, most models are benchmarked with similar performance, though few models are still being benchmarked.
+The compatibilities of models are broken due to the unification and simplification of coordinate systems. For now, most models are benchmarked with similar performance, though few models are still being benchmarked. In this version, we update some of the model checkpoints after the refactor of coordinate systems. See more details in the [Changelog](docs/en/changelog.md).
 
 In the [nuScenes 3D detection challenge](https://www.nuscenes.org/object-detection?externalData=all&mapData=all&modalities=Any) of the 5th AI Driving Olympics in NeurIPS 2020, we obtained the best PKL award and the second runner-up by multi-modality entry, and the best vision-only results.
 
@@ -83,7 +83,7 @@ This project is released under the [Apache 2.0 license](LICENSE).
 
 ## Changelog
 
-v1.0.0rc0 was released in 18/2/2022.
+v1.0.0rc1 was released in 1/4/2022.
 Please refer to [changelog.md](docs/en/changelog.md) for details and release history.
 
 ## Benchmark and model zoo
@@ -266,20 +266,21 @@ We wish that the toolbox and benchmark could serve the growing research communit
 ## Projects in OpenMMLab
 
 - [MMCV](https://github.com/open-mmlab/mmcv): OpenMMLab foundational library for computer vision.
-- [MIM](https://github.com/open-mmlab/mim): MIM Installs OpenMMLab Packages.
+- [MIM](https://github.com/open-mmlab/mim): MIM installs OpenMMLab packages.
 - [MMClassification](https://github.com/open-mmlab/mmclassification): OpenMMLab image classification toolbox and benchmark.
 - [MMDetection](https://github.com/open-mmlab/mmdetection): OpenMMLab detection toolbox and benchmark.
-- [MMDetection3D](https://github.com/open-mmlab/mmdetection3d): OpenMMLab next-generation platform for general 3D object detection.
+- [MMDetection3D](https://github.com/open-mmlab/mmdetection3d): OpenMMLab's next-generation platform for general 3D object detection.
+- [MMRotate](https://github.com/open-mmlab/mmrotate): OpenMMLab rotated object detection toolbox and benchmark.
 - [MMSegmentation](https://github.com/open-mmlab/mmsegmentation): OpenMMLab semantic segmentation toolbox and benchmark.
+- [MMOCR](https://github.com/open-mmlab/mmocr): OpenMMLab text detection, recognition, and understanding toolbox.
+- [MMPose](https://github.com/open-mmlab/mmpose): OpenMMLab pose estimation toolbox and benchmark.
+- [MMHuman3D](https://github.com/open-mmlab/mmhuman3d): OpenMMLab 3D human parametric model toolbox and benchmark.
+- [MMSelfSup](https://github.com/open-mmlab/mmselfsup): OpenMMLab self-supervised learning toolbox and benchmark.
+- [MMRazor](https://github.com/open-mmlab/mmrazor): OpenMMLab model compression toolbox and benchmark.
+- [MMFewShot](https://github.com/open-mmlab/mmfewshot): OpenMMLab fewshot learning toolbox and benchmark.
 - [MMAction2](https://github.com/open-mmlab/mmaction2): OpenMMLab's next-generation action understanding toolbox and benchmark.
 - [MMTracking](https://github.com/open-mmlab/mmtracking): OpenMMLab video perception toolbox and benchmark.
-- [MMPose](https://github.com/open-mmlab/mmpose): OpenMMLab pose estimation toolbox and benchmark.
+- [MMFlow](https://github.com/open-mmlab/mmflow): OpenMMLab optical flow toolbox and benchmark.
 - [MMEditing](https://github.com/open-mmlab/mmediting): OpenMMLab image and video editing toolbox.
-- [MMOCR](https://github.com/open-mmlab/mmocr): OpenMMLab text detection, recognition and understanding toolbox.
 - [MMGeneration](https://github.com/open-mmlab/mmgeneration): OpenMMLab image and video generative models toolbox.
-- [MMFlow](https://github.com/open-mmlab/mmflow): OpenMMLab optical flow toolbox and benchmark.
-- [MMFewShot](https://github.com/open-mmlab/mmfewshot): OpenMMLab fewshot learning toolbox and benchmark.
-- [MMHuman3D](https://github.com/open-mmlab/mmhuman3d): OpenMMLab 3D human parametric model toolbox and benchmark.
-- [MMSelfSup](https://github.com/open-mmlab/mmselfsup): OpenMMLab self-supervised learning toolbox and benchmark.
-- [MMRazor](https://github.com/open-mmlab/mmrazor): OpenMMLab Model Compression Toolbox and Benchmark.
-- [MMDeploy](https://github.com/open-mmlab/mmdeploy): OpenMMLab Model Deployment Framework.
+- [MMDeploy](https://github.com/open-mmlab/mmdeploy): OpenMMLab model deployment framework.
diff --git a/README_zh-CN.md b/README_zh-CN.md
@@ -25,11 +25,11 @@
 [![license](https://img.shields.io/github/license/open-mmlab/mmdetection3d.svg)](https://github.com/open-mmlab/mmdetection3d/blob/master/LICENSE)
 
 
-**新闻**: 我们发布了版本 v1.0.0rc0.
+**新闻**: 我们发布了版本 v1.0.0rc1.
 
 说明：我们正在进行大规模的重构，以提供对许多模块更简单、更统一的使用。
 
-由于坐标系的统一和简化，模型的兼容性会受到影响。目前，大多数模型都以类似的性能对齐了精度，但仍有少数模型在进行基准测试。
+由于坐标系的统一和简化，模型的兼容性会受到影响。目前，大多数模型都以类似的性能对齐了精度，但仍有少数模型在进行基准测试。在这个版本中，我们更新了部分坐标系重构后的模型权重文件。您可以在 [变更日志](docs/en/changelog.md) 中查看更多详细信息。
 
 在第三届 [nuScenes 3D 检测挑战赛](https://www.nuscenes.org/object-detection?externalData=all&mapData=all&modalities=Any)（第五届 AI Driving Olympics, NeurIPS 2020）中，我们获得了最佳 PKL 奖、第三名和最好的纯视觉的结果，相关的代码和模型将会在不久后发布。
 
@@ -83,7 +83,7 @@ MMDetection3D 是一个基于 PyTorch 的目标检测开源工具箱, 下一代
 
 ## 更新日志
 
-最新的版本 v1.0.0rc0 在 2022.2.18 发布。
+最新的版本 v1.0.0rc1 在 2022.4.1 发布。
 如果想了解更多版本更新细节和历史信息，请阅读[更新日志](docs/zh_cn/changelog.md)。
 
 ## 基准测试和模型库
@@ -269,18 +269,19 @@ MMDetection3D 是一款由来自不同高校和企业的研发人员共同参与
 - [MMClassification](https://github.com/open-mmlab/mmclassification): OpenMMLab 图像分类工具箱
 - [MMDetection](https://github.com/open-mmlab/mmdetection): OpenMMLab 目标检测工具箱
 - [MMDetection3D](https://github.com/open-mmlab/mmdetection3d): OpenMMLab 新一代通用 3D 目标检测平台
+- [MMRotate](https://github.com/open-mmlab/mmrotate): OpenMMLab 旋转框检测工具箱与测试基准
 - [MMSegmentation](https://github.com/open-mmlab/mmsegmentation): OpenMMLab 语义分割工具箱
-- [MMAction2](https://github.com/open-mmlab/mmaction2): OpenMMLab 新一代视频理解工具箱
-- [MMTracking](https://github.com/open-mmlab/mmtracking): OpenMMLab 一体化视频目标感知平台
-- [MMPose](https://github.com/open-mmlab/mmpose): OpenMMLab 姿态估计工具箱
-- [MMEditing](https://github.com/open-mmlab/mmediting): OpenMMLab 图像视频编辑工具箱
 - [MMOCR](https://github.com/open-mmlab/mmocr): OpenMMLab 全流程文字检测识别理解工具包
-- [MMGeneration](https://github.com/open-mmlab/mmgeneration): OpenMMLab 图片视频生成模型工具箱
-- [MMFlow](https://github.com/open-mmlab/mmflow): OpenMMLab 光流估计工具箱与测试基准
-- [MMFewShot](https://github.com/open-mmlab/mmfewshot): OpenMMLab 少样本学习工具箱与测试基准
+- [MMPose](https://github.com/open-mmlab/mmpose): OpenMMLab 姿态估计工具箱
 - [MMHuman3D](https://github.com/open-mmlab/mmhuman3d): OpenMMLab 人体参数化模型工具箱与测试基准
 - [MMSelfSup](https://github.com/open-mmlab/mmselfsup): OpenMMLab 自监督学习工具箱与测试基准
 - [MMRazor](https://github.com/open-mmlab/mmrazor): OpenMMLab 模型压缩工具箱与测试基准
+- [MMFewShot](https://github.com/open-mmlab/mmfewshot): OpenMMLab 少样本学习工具箱与测试基准
+- [MMAction2](https://github.com/open-mmlab/mmaction2): OpenMMLab 新一代视频理解工具箱
+- [MMTracking](https://github.com/open-mmlab/mmtracking): OpenMMLab 一体化视频目标感知平台
+- [MMFlow](https://github.com/open-mmlab/mmflow): OpenMMLab 光流估计工具箱与测试基准
+- [MMEditing](https://github.com/open-mmlab/mmediting): OpenMMLab 图像视频编辑工具箱
+- [MMGeneration](https://github.com/open-mmlab/mmgeneration): OpenMMLab 图片视频生成模型工具箱
 - [MMDeploy](https://github.com/open-mmlab/mmdeploy): OpenMMLab 模型部署框架
 
 ## 欢迎加入 OpenMMLab 社区

diff --git a/configs/3dssd/README.md b/configs/3dssd/README.md
@@ -28,9 +28,9 @@ Some settings in our implementation are different from the [official implementat
 
 |  Backbone   |Class| Lr schd | Mem (GB) | Inf time (fps) | mAP |Download |
 | :---------: | :-----: | :------: | :------------: | :----: |:----: | :------: |
-|    [PointNet2SAMSG](./3dssd_4x4_kitti-3d-car.py)| Car |72e|4.7||78.69(81.27)<sup>1</sup>|[model](https://download.openmmlab.com/mmdetection3d/v0.1.0_models/3dssd/3dssd_kitti-3d-car_20210602_124438-b4276f56.pth) &#124; [log](https://download.openmmlab.com/mmdetection3d/v0.1.0_models/3dssd/3dssd_kitti-3d-car_20210602_124438.log.json)|
+|    [PointNet2SAMSG](./3dssd_4x4_kitti-3d-car.py)| Car |72e|4.7||78.58(81.27)<sup>1</sup>|[model](https://download.openmmlab.com/mmdetection3d/v1.0.0_models/3dssd/3dssd_4x4_kitti-3d-car/3dssd_4x4_kitti-3d-car_20210818_203828-b89c8fc4.pth) &#124; [log](https://download.openmmlab.com/mmdetection3d/v1.0.0_models/3dssd/3dssd_4x4_kitti-3d-car/3dssd_4x4_kitti-3d-car_20210818_203828.log.json)|
 
-[1]: We report two different 3D object detection performance here. 78.69mAP is evaluated by our evaluation code and 81.27mAP is evaluated by the official development kit （so as that used in the paper and official code of 3DSSD ）. We found that the commonly used Python implementation of [`rotate_iou`](https://github.com/traveller59/second.pytorch/blob/e42e4a0e17262ab7d180ee96a0a36427f2c20a44/second/core/non_max_suppression/nms_gpu.py#L605) which is used in our KITTI dataset evaluation, is different from the official implementation in [KITTI benchmark](http://www.cvlibs.net/datasets/kitti/eval_object.php?obj_benchmark=3d).
+[1]: We report two different 3D object detection performance here. 78.58mAP is evaluated by our evaluation code and 81.27mAP is evaluated by the official development kit （so as that used in the paper and official code of 3DSSD ）. We found that the commonly used Python implementation of [`rotate_iou`](https://github.com/traveller59/second.pytorch/blob/e42e4a0e17262ab7d180ee96a0a36427f2c20a44/second/core/non_max_suppression/nms_gpu.py#L605) which is used in our KITTI dataset evaluation, is different from the official implementation in [KITTI benchmark](http://www.cvlibs.net/datasets/kitti/eval_object.php?obj_benchmark=3d).
 
 ## Citation
 

diff --git a/configs/3dssd/metafile.yml b/configs/3dssd/metafile.yml
@@ -25,5 +25,5 @@ Models:
       - Task: 3D Object Detection
         Dataset: KITTI
         Metrics:
-          mAP: 78.69
-    Weights: https://download.openmmlab.com/mmdetection3d/v0.1.0_models/3dssd/3dssd_kitti-3d-car_20210602_124438-b4276f56.pth
+          mAP: 78.58
+    Weights: https://download.openmmlab.com/mmdetection3d/v1.0.0_models/3dssd/3dssd_4x4_kitti-3d-car/3dssd_4x4_kitti-3d-car_20210818_203828-b89c8fc4.pth
diff --git a/configs/_base_/datasets/waymoD5-3d-3class.py b/configs/_base_/datasets/waymoD5-3d-3class.py
@@ -25,7 +25,7 @@
     points_loader=dict(
         type='LoadPointsFromFile',
         coord_type='LIDAR',
-        load_dim=5,
+        load_dim=6,
         use_dim=[0, 1, 2, 3, 4],
         file_client_args=file_client_args))
 

diff --git a/configs/_base_/datasets/waymoD5-3d-car.py b/configs/_base_/datasets/waymoD5-3d-car.py
@@ -23,7 +23,7 @@
     points_loader=dict(
         type='LoadPointsFromFile',
         coord_type='LIDAR',
-        load_dim=5,
+        load_dim=6,
         use_dim=[0, 1, 2, 3, 4],
         file_client_args=file_client_args))
 

diff --git a/configs/_base_/models/cascade_mask_rcnn_r50_fpn.py b/configs/_base_/models/cascade_mask_rcnn_r50_fpn.py
@@ -126,11 +126,10 @@
             pos_weight=-1,
             debug=False),
         rpn_proposal=dict(
-            nms_across_levels=False,
             nms_pre=2000,
             nms_post=2000,
-            max_num=2000,
-            nms_thr=0.7,
+            max_per_img=2000,
+            nms=dict(type='nms', iou_threshold=0.7),
             min_bbox_size=0),
         rcnn=[
             dict(
@@ -187,11 +186,10 @@
         ]),
     test_cfg=dict(
         rpn=dict(
-            nms_across_levels=False,
             nms_pre=1000,
             nms_post=1000,
-            max_num=1000,
-            nms_thr=0.7,
+            max_per_img=1000,
+            nms=dict(type='nms', iou_threshold=0.7),
             min_bbox_size=0),
         rcnn=dict(
             score_thr=0.05,

diff --git a/configs/_base_/models/mask_rcnn_r50_fpn.py b/configs/_base_/models/mask_rcnn_r50_fpn.py
@@ -89,8 +89,8 @@
             nms_across_levels=False,
             nms_pre=2000,
             nms_post=1000,
-            max_num=1000,
-            nms_thr=0.7,
+            max_per_img=1000,
+            nms=dict(type='nms', iou_threshold=0.7),
             min_bbox_size=0),
         rcnn=dict(
             assigner=dict(
@@ -114,8 +114,8 @@
             nms_across_levels=False,
             nms_pre=1000,
             nms_post=1000,
-            max_num=1000,
-            nms_thr=0.7,
+            max_per_img=1000,
+            nms=dict(type='nms', iou_threshold=0.7),
             min_bbox_size=0),
         rcnn=dict(
             score_thr=0.05,

diff --git a/configs/dynamic_voxelization/README.md b/configs/dynamic_voxelization/README.md
@@ -23,7 +23,7 @@ We implement Dynamic Voxelization proposed in  and provide its results and model
 |  Model   |Class| Lr schd | Mem (GB) | Inf time (fps) | mAP | Download |
 | :---------: | :-----: |:-----: | :------: | :------------: | :----: | :------: |
 |[SECOND](./dv_second_secfpn_6x8_80e_kitti-3d-car.py)|Car    |cyclic 80e|5.5||78.83|[model](https://download.openmmlab.com/mmdetection3d/v0.1.0_models/dynamic_voxelization/dv_second_secfpn_6x8_80e_kitti-3d-car/dv_second_secfpn_6x8_80e_kitti-3d-car_20200620_235228-ac2c1c0c.pth) &#124; [log](https://download.openmmlab.com/mmdetection3d/v0.1.0_models/dynamic_voxelization/dv_second_secfpn_6x8_80e_kitti-3d-car/dv_second_secfpn_6x8_80e_kitti-3d-car_20200620_235228.log.json)|
-|[SECOND](./dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class.py)| 3 Class|cosine 80e|5.5||65.10|[model](https://download.openmmlab.com/mmdetection3d/v0.1.0_models/dynamic_voxelization/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class_20200620_231010-6aa607d3.pth) &#124; [log](https://download.openmmlab.com/mmdetection3d/v0.1.0_models/dynamic_voxelization/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class_20200620_231010.log.json)|
+|[SECOND](./dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class.py)| 3 Class|cosine 80e|5.5||65.27|[model](https://download.openmmlab.com/mmdetection3d/v1.0.0_models/dynamic_voxelization/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class_20210831_054106-e742d163.pth) &#124; [log](https://download.openmmlab.com/mmdetection3d/v1.0.0_models/dynamic_voxelization/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class_20210831_054106.log.json)|
 |[PointPillars](./dv_pointpillars_secfpn_6x8_160e_kitti-3d-car.py)| Car|cyclic 80e|4.7||77.76|[model](https://download.openmmlab.com/mmdetection3d/v0.1.0_models/dynamic_voxelization/dv_pointpillars_secfpn_6x8_160e_kitti-3d-car/dv_pointpillars_secfpn_6x8_160e_kitti-3d-car_20200620_230844-ee7b75c9.pth) &#124; [log](https://download.openmmlab.com/mmdetection3d/v0.1.0_models/dynamic_voxelization/dv_pointpillars_secfpn_6x8_160e_kitti-3d-car/dv_pointpillars_secfpn_6x8_160e_kitti-3d-car_20200620_230844.log.json)|
 
 ## Citation

diff --git a/configs/dynamic_voxelization/metafile.yml b/configs/dynamic_voxelization/metafile.yml
@@ -37,8 +37,8 @@ Models:
       - Task: 3D Object Detection
         Dataset: KITTI
         Metrics:
-          mAP: 65.10
-    Weights: https://download.openmmlab.com/mmdetection3d/v0.1.0_models/dynamic_voxelization/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class_20200620_231010-6aa607d3.pth
+          mAP: 65.27
+    Weights: https://download.openmmlab.com/mmdetection3d/v1.0.0_models/dynamic_voxelization/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class/dv_second_secfpn_2x8_cosine_80e_kitti-3d-3class_20210831_054106-e742d163.pth
 
   - Name: dv_pointpillars_secfpn_6x8_160e_kitti-3d-car
     In Collection: Dynamic Voxelization