Merge branch 'master' into dev

README.md

@@ -35,6 +35,8 @@ In the [nuScenes 3D detection challenge](https://www.nuscenes.org/object-detecti
 
 Code and models for the best vision-only method, [FCOS3D](https://arxiv.org/abs/2104.10956), have been released. Please stay tuned for [MoCa](https://arxiv.org/abs/2012.12741).
 
+MMDeploy has supported some MMDetection3d model deployment.
+
 Documentation: https://mmdetection3d.readthedocs.io/
 
 ## Introduction
@@ -228,7 +230,6 @@ Results and models are available in the [model zoo](docs/en/model_zoo.md).
 | PGD                  | ✓        | ☐        | ☐        | ✗         | ✗         | ☐     | ☐        | ☐     | ✗
 | MonoFlex             | ✗        | ✗        | ✗        | ✗         | ✗         | ✗     | ✗        | ✗     | ✓
 
-
 **Note:** All the about **300+ models, methods of 40+ papers** in 2D detection supported by [MMDetection](https://github.com/open-mmlab/mmdetection/blob/master/docs/en/model_zoo.md) can be trained or used in this codebase.
 
 ## Installation
@@ -241,6 +242,9 @@ Please see [getting_started.md](docs/en/getting_started.md) for the basic usage
 
 Please refer to [FAQ](docs/en/faq.md) for frequently asked questions. When updating the version of MMDetection3D, please also check the [compatibility doc](docs/en/compatibility.md) to be aware of the BC-breaking updates introduced in each version.
 
+## Model deployment
+
+Now MMDeploy has supported some MMDetection3D model deployment. Please refer to [model_deployment.md](docs/en/tutorials/model_deployment.md) for more details.
 ## Citation
 
 If you find this project useful in your research, please consider cite:

README_zh-CN.md

@@ -35,6 +35,8 @@
 
 最好的纯视觉方法 [FCOS3D](https://arxiv.org/abs/2104.10956) 的代码和模型已经发布。请继续关注我们的多模态检测器 [MoCa](https://arxiv.org/abs/2012.12741)。
 
+MMDeploy 已经支持了部分 MMDetection3D 模型的部署。
+
 文档: https://mmdetection3d.readthedocs.io/
 
 ## 简介
@@ -241,6 +243,10 @@ MMDetection3D 是一个基于 PyTorch 的目标检测开源工具箱, 下一代
 
 请参考 [FAQ](docs/zh_cn/faq.md) 查看一些常见的问题与解答。在升级 MMDetection3D 的版本时，请查看[兼容性文档](docs/zh_cn/compatibility.md)以知晓每个版本引入的不与之前版本兼容的更新。
 
+## 模型部署
+
+现在 MMDeploy 已经支持了一些 MMDetection3D 模型的部署。请参考 [model_deployment.md](docs/zh_cn/tutorials/model_deployment.md)了解更多细节。
+
 ## 引用
 
 如果你觉得本项目对你的研究工作有所帮助，请参考如下 bibtex 引用 MMdetection3D

docs/en/tutorials/model_deployment.md

+# Tutorial 8: MMDetection3D model deployment
+
+To meet the speed requirement of the model in practical use, usually, we deploy the trained model to inference backends. [MMDeploy](https://github.com/open-mmlab/mmdeploy) is OpenMMLab model deployment framework. Now MMDeploy has supported MMDetection3D model deployment, and you can deploy the trained model to inference backends by MMDeploy.
+
+## Prerequisite
+
+### Install MMDeploy
+
+```bash
+git clone -b master git@github.com:open-mmlab/mmdeploy.git
+cd mmdeploy
+git submodule update --init --recursive
+```
+
+### Install backend and build custom ops
+
+According to MMDeploy documentation, choose to install the inference backend and build custom ops. Now supported inference backends for MMDetection3D include [OnnxRuntime](https://mmdeploy.readthedocs.io/en/latest/backends/onnxruntime.html), [TensorRT](https://mmdeploy.readthedocs.io/en/latest/backends/tensorrt.html), [OpenVINO](https://mmdeploy.readthedocs.io/en/latest/backends/openvino.html).
+
+## Export model
+
+Export the Pytorch model of MMDetection3D to the ONNX model file and the model file required by the backend. You could refer to MMDeploy docs [how to convert model](https://mmdeploy.readthedocs.io/en/latest/tutorials/how_to_convert_model.html).
+
+```bash
+python ./tools/deploy.py \
+    ${DEPLOY_CFG_PATH} \
+    ${MODEL_CFG_PATH} \
+    ${MODEL_CHECKPOINT_PATH} \
+    ${INPUT_IMG} \
+    --test-img ${TEST_IMG} \
+    --work-dir ${WORK_DIR} \
+    --calib-dataset-cfg ${CALIB_DATA_CFG} \
+    --device ${DEVICE} \
+    --log-level INFO \
+    --show \
+    --dump-info
+```
+
+### Description of all arguments
+
+* `deploy_cfg` : The path of deploy config file in MMDeploy codebase.
+* `model_cfg` : The path of model config file in OpenMMLab codebase.
+* `checkpoint` : The path of model checkpoint file.
+* `img` : The path of point cloud file or image file that used to convert model.
+* `--test-img` : The path of image file that used to test model. If not specified, it will be set to `None`.
+* `--work-dir` : The path of work directory that used to save logs and models.
+* `--calib-dataset-cfg` : Only valid in int8 mode. Config used for calibration. If not specified, it will be set to `None` and  use "val" dataset in model config for calibration.
+* `--device` : The device used for conversion. If not specified, it will be set to `cpu`.
+* `--log-level` : To set log level which in `'CRITICAL', 'FATAL', 'ERROR', 'WARN', 'WARNING', 'INFO', 'DEBUG', 'NOTSET'`. If not specified, it will be set to `INFO`.
+* `--show` : Whether to show detection outputs.
+* `--dump-info` : Whether to output information for SDK.
+
+### Example
+
+```bash
+cd mmdeploy
+python tools/deploy.py \
+    configs/mmdet3d/voxel-detection/voxel-detection_tensorrt_dynamic-kitti.py \
+    ${$MMDET3D_DIR}/configs/pointpillars/hv_pointpillars_secfpn_6x8_160e_kitti-3d-3class.py \
+    ${$MMDET3D_DIR}/checkpoints/hv_pointpillars_secfpn_6x8_160e_kitti-3d-3class_20200620_230421-aa0f3adb.pth \
+    ${$MMDET3D_DIR}/demo/data/kitti/kitti_000008.bin \
+    --work-dir work-dir \
+    --device cuda:0 \
+    --show
+```
+
+## Inference Model
+
+Now you can do model inference with the APIs provided by the backend. But what if you want to test the model instantly? We have some backend wrappers for you.
+
+```python
+from mmdeploy.apis import inference_model
+
+result = inference_model(model_cfg, deploy_cfg, backend_files, img=img, device=device)
+```
+
+The `inference_model` will create a wrapper module and do the inference for you. The result has the same format as the original OpenMMLab repo.
+
+## Evaluate model (Optional)
+
+You can test the accuracy and speed of the model in the inference backend. You could refer to MMDeploy docs [how to measure performance of models](https://mmdeploy.readthedocs.io/en/latest/tutorials/how_to_measure_performance_of_models.html).
+
+```bash
+python tools/test.py \
+    ${DEPLOY_CFG} \
+    ${MODEL_CFG} \
+    --model ${BACKEND_MODEL_FILES} \
+    [--out ${OUTPUT_PKL_FILE}] \
+    [--format-only] \
+    [--metrics ${METRICS}] \
+    [--show] \
+    [--show-dir ${OUTPUT_IMAGE_DIR}] \
+    [--show-score-thr ${SHOW_SCORE_THR}] \
+    --device ${DEVICE} \
+    [--cfg-options ${CFG_OPTIONS}] \
+    [--metric-options ${METRIC_OPTIONS}] \
+    [--log2file work_dirs/output.txt]
+```
+
+### Example
+
+```bash
+cd mmdeploy
+python tools/test.py \
+    configs/mmdet3d/voxel-detection/voxel-detection_onnxruntime_dynamic.py \
+    ${MMDET3D_DIR}/configs/centerpoint/centerpoint_02pillar_second_secfpn_circlenms_4x8_cyclic_20e_nus.py \
+    --model work-dir/end2end.onnx \
+    --metrics bbox \
+    --device cpu
+```
+
+## Supported models
+
+| Model                | TorchScript | OnnxRuntime | TensorRT | NCNN  | PPLNN | OpenVINO | Model config                                                                           |
+| -------------------- | :---------: | :---------: | :------: | :---: | :---: | :------: | -------------------------------------------------------------------------------------- |
+| PointPillars         |      ?      |      Y      |    Y     |   N   |   N   |    Y     | [config](https://github.com/open-mmlab/mmdetection3d/blob/master/configs/pointpillars) |
+| CenterPoint (pillar) |      ?      |      Y      |    Y     |   N   |   N   |    Y     | [config](https://github.com/open-mmlab/mmdetection3d/blob/master/configs/centerpoint)  |
+
+## Note
+
+* MMDeploy version >= 0.4.0.
+* Currently, CenterPoint has only supported the pillar version.

docs/zh_cn/tutorials/model_deployment.md

+# 教程 8: MMDet3D 模型部署
+
+为了满足在实际使用过程中遇到的算法模型的速度需求，通常我们会将训练好的模型部署到各种推理后端上。 [MMDeploy](https://github.com/open-mmlab/mmdeploy) 是 OpenMMLab 系列算法库的部署框架，现在 MMDeploy 已经支持了 MMDetection3D，我们可以通过 MMDeploy 将训练好的模型部署到各种推理后端上。
+
+## 准备
+
+### 安装 MMDeploy
+
+```bash
+git clone -b master git@github.com:open-mmlab/mmdeploy.git
+cd mmdeploy
+git submodule update --init --recursive
+```
+
+### 安装推理后端编译自定义算子
+
+根据 MMDeploy 的文档选择安装推理后端并编译自定义算子，目前 MMDet3D 模型支持了的推理后端有 [OnnxRuntime](https://mmdeploy.readthedocs.io/en/latest/backends/onnxruntime.html)，[TensorRT](https://mmdeploy.readthedocs.io/en/latest/backends/tensorrt.html)，[OpenVINO](https://mmdeploy.readthedocs.io/en/latest/backends/openvino.html)。
+
+## 模型导出
+
+将 MMDet3D 训练好的 Pytorch 模型转换成 ONNX 模型文件和推理后端所需要的模型文件。你可以参考 MMDeploy 的文档 [how_to_convert_model.md](https://github.com/open-mmlab/mmdeploy/blob/master/docs/zh_cn/tutorials/how_to_convert_model.md)。
+
+```bash
+python ./tools/deploy.py \
+    ${DEPLOY_CFG_PATH} \
+    ${MODEL_CFG_PATH} \
+    ${MODEL_CHECKPOINT_PATH} \
+    ${INPUT_IMG} \
+    --test-img ${TEST_IMG} \
+    --work-dir ${WORK_DIR} \
+    --calib-dataset-cfg ${CALIB_DATA_CFG} \
+    --device ${DEVICE} \
+    --log-level INFO \
+    --show \
+    --dump-info
+```
+
+### 参数描述
+
+* `deploy_cfg` : MMDeploy 代码库中用于部署的配置文件路径。
+* `model_cfg` : OpenMMLab 系列代码库中使用的模型配置文件路径。
+* `checkpoint` : OpenMMLab 系列代码库的模型文件路径。
+* `img` : 用于模型转换时使用的点云文件或图像文件路径。
+* `--test-img` : 用于测试模型的图像文件路径。如果没有指定，将设置成 `None`。
+* `--work-dir` : 工作目录，用来保存日志和模型文件。
+* `--calib-dataset-cfg` : 此参数只在 int8 模式下生效，用于校准数据集配置文件。如果没有指定，将被设置成 `None`，并使用模型配置文件中的 'val' 数据集进行校准。
+* `--device` : 用于模型转换的设备。如果没有指定，将被设置成 cpu。
+* `--log-level` : 设置日记的等级，选项包括 `'CRITICAL'，'FATAL'，'ERROR'，'WARN'，'WARNING'，'INFO'，'DEBUG'，'NOTSET'`。如果没有指定，将被设置成 INFO。
+* `--show` : 是否显示检测的结果。
+* `--dump-info` : 是否输出 SDK 信息。
+
+### 示例
+
+```bash
+cd mmdeploy
+python tools/deploy.py \
+    configs/mmdet3d/voxel-detection/voxel-detection_tensorrt_dynamic-kitti.py \
+    ${$MMDET3D_DIR}/configs/pointpillars/hv_pointpillars_secfpn_6x8_160e_kitti-3d-3class.py \
+    ${$MMDET3D_DIR}/checkpoints/hv_pointpillars_secfpn_6x8_160e_kitti-3d-3class_20200620_230421-aa0f3adb.pth \
+    ${$MMDET3D_DIR}/demo/data/kitti/kitti_000008.bin \
+    --work-dir work-dir \
+    --device cuda:0 \
+    --show
+```
+
+## 模型推理
+
+现在你可以使用推理后端提供的 API 进行模型推理。但是，如果你想立即测试模型怎么办？我们为您准备了一些推理后端的封装。
+
+```python
+from mmdeploy.apis import inference_model
+
+result = inference_model(model_cfg, deploy_cfg, backend_files, img=img, device=device)
+```
+
+`inference_model` 将创建一个推理后端的模块并为你进行推理。推理结果与模型的 OpenMMLab 代码库具有相同的格式。
+
+## 测试模型（可选）
+
+可以测试部署在推理后端上的模型的精度和速度。你可以参考 [how to measure performance of models](https://mmdeploy.readthedocs.io/en/latest/tutorials/how_to_measure_performance_of_models.html)。
+
+```bash
+python tools/test.py \
+    ${DEPLOY_CFG} \
+    ${MODEL_CFG} \
+    --model ${BACKEND_MODEL_FILES} \
+    [--out ${OUTPUT_PKL_FILE}] \
+    [--format-only] \
+    [--metrics ${METRICS}] \
+    [--show] \
+    [--show-dir ${OUTPUT_IMAGE_DIR}] \
+    [--show-score-thr ${SHOW_SCORE_THR}] \
+    --device ${DEVICE} \
+    [--cfg-options ${CFG_OPTIONS}] \
+    [--metric-options ${METRIC_OPTIONS}] \
+    [--log2file work_dirs/output.txt]
+```
+
+### 示例
+
+```bash
+cd mmdeploy
+python tools/test.py \
+    configs/mmdet3d/voxel-detection/voxel-detection_onnxruntime_dynamic.py \
+    ${MMDET3D_DIR}/configs/centerpoint/centerpoint_02pillar_second_secfpn_circlenms_4x8_cyclic_20e_nus.py \
+    --model work-dir/end2end.onnx \
+    --metrics bbox \
+    --device cpu
+```
+
+## 支持模型列表
+
+| Model                | TorchScript | OnnxRuntime | TensorRT | NCNN  | PPLNN | OpenVINO | Model config                                                                           |
+| -------------------- | :---------: | :---------: | :------: | :---: | :---: | :------: | -------------------------------------------------------------------------------------- |
+| PointPillars         |      ?      |      Y      |    Y     |   N   |   N   |    Y     | [config](https://github.com/open-mmlab/mmdetection3d/blob/master/configs/pointpillars) |
+| CenterPoint (pillar) |      ?      |      Y      |    Y     |   N   |   N   |    Y     | [config](https://github.com/open-mmlab/mmdetection3d/blob/master/configs/centerpoint)  |
+
+## 注意
+
+* MMDeploy 的版本需要 >= 0.4.0。
+* 目前 CenterPoint 仅支持了 pillar 版本的。

mmdet3d/core/visualizer/open3d_vis.py

@@ -95,7 +95,7 @@ def _draw_bboxes(bbox3d,
         center = bbox3d[i, 0:3]
         dim = bbox3d[i, 3:6]
         yaw = np.zeros(3)
-        yaw[rot_axis] = -bbox3d[i, 6]
+        yaw[rot_axis] = bbox3d[i, 6]
         rot_mat = geometry.get_rotation_matrix_from_xyz(yaw)
 
         if center_mode == 'lidar_bottom':

mmdet3d/utils/__init__.py

@@ -2,10 +2,11 @@
 from mmcv.utils import Registry, build_from_cfg, print_log
 
 from .collect_env import collect_env
+from .compat_cfg import compat_cfg
 from .logger import get_root_logger
 from .setup_env import setup_multi_processes
 
 __all__ = [
     'Registry', 'build_from_cfg', 'get_root_logger', 'collect_env',
-    'print_log', 'setup_multi_processes'
+    'print_log', 'setup_multi_processes', 'compat_cfg'
 ]

mmdet3d/utils/compat_cfg.py

+# Copyright (c) OpenMMLab. All rights reserved.
+import copy
+import warnings
+
+from mmcv import ConfigDict
+
+
+def compat_cfg(cfg):
+    """This function would modify some filed to keep the compatibility of
+    config.
+
+    For example, it will move some args which will be deprecated to the correct
+    fields.
+    """
+    cfg = copy.deepcopy(cfg)
+    cfg = compat_imgs_per_gpu(cfg)
+    cfg = compat_loader_args(cfg)
+    cfg = compat_runner_args(cfg)
+    return cfg
+
+
+def compat_runner_args(cfg):
+    if 'runner' not in cfg:
+        cfg.runner = ConfigDict({
+            'type': 'EpochBasedRunner',
+            'max_epochs': cfg.total_epochs
+        })
+        warnings.warn(
+            'config is now expected to have a `runner` section, '
+            'please set `runner` in your config.', UserWarning)
+    else:
+        if 'total_epochs' in cfg:
+            assert cfg.total_epochs == cfg.runner.max_epochs
+    return cfg
+
+
+def compat_imgs_per_gpu(cfg):
+    cfg = copy.deepcopy(cfg)
+    if 'imgs_per_gpu' in cfg.data:
+        warnings.warn('"imgs_per_gpu" is deprecated in MMDet V2.0. '
+                      'Please use "samples_per_gpu" instead')
+        if 'samples_per_gpu' in cfg.data:
+            warnings.warn(
+                f'Got "imgs_per_gpu"={cfg.data.imgs_per_gpu} and '
+                f'"samples_per_gpu"={cfg.data.samples_per_gpu}, "imgs_per_gpu"'
+                f'={cfg.data.imgs_per_gpu} is used in this experiments')
+        else:
+            warnings.warn('Automatically set "samples_per_gpu"="imgs_per_gpu"='
+                          f'{cfg.data.imgs_per_gpu} in this experiments')
+        cfg.data.samples_per_gpu = cfg.data.imgs_per_gpu
+    return cfg
+
+
+def compat_loader_args(cfg):
+    """Deprecated sample_per_gpu in cfg.data."""
+
+    cfg = copy.deepcopy(cfg)
+    if 'train_dataloader' not in cfg.data:
+        cfg.data['train_dataloader'] = ConfigDict()
+    if 'val_dataloader' not in cfg.data:
+        cfg.data['val_dataloader'] = ConfigDict()
+    if 'test_dataloader' not in cfg.data:
+        cfg.data['test_dataloader'] = ConfigDict()
+
+    # special process for train_dataloader
+    if 'samples_per_gpu' in cfg.data:
+
+        samples_per_gpu = cfg.data.pop('samples_per_gpu')
+        assert 'samples_per_gpu' not in \
+               cfg.data.train_dataloader, ('`samples_per_gpu` are set '
+                                           'in `data` field and ` '
+                                           'data.train_dataloader` '
+                                           'at the same time. '
+                                           'Please only set it in '
+                                           '`data.train_dataloader`. ')
+        cfg.data.train_dataloader['samples_per_gpu'] = samples_per_gpu
+
+    if 'persistent_workers' in cfg.data:
+
+        persistent_workers = cfg.data.pop('persistent_workers')
+        assert 'persistent_workers' not in \
+               cfg.data.train_dataloader, ('`persistent_workers` are set '
+                                           'in `data` field and ` '
+                                           'data.train_dataloader` '
+                                           'at the same time. '
+                                           'Please only set it in '
+                                           '`data.train_dataloader`. ')
+        cfg.data.train_dataloader['persistent_workers'] = persistent_workers
+
+    if 'workers_per_gpu' in cfg.data:
+
+        workers_per_gpu = cfg.data.pop('workers_per_gpu')
+        cfg.data.train_dataloader['workers_per_gpu'] = workers_per_gpu
+        cfg.data.val_dataloader['workers_per_gpu'] = workers_per_gpu
+        cfg.data.test_dataloader['workers_per_gpu'] = workers_per_gpu
+
+    # special process for val_dataloader
+    if 'samples_per_gpu' in cfg.data.val:
+        # keep default value of `sample_per_gpu` is 1
+        assert 'samples_per_gpu' not in \
+               cfg.data.val_dataloader, ('`samples_per_gpu` are set '
+                                         'in `data.val` field and ` '
+                                         'data.val_dataloader` at '
+                                         'the same time. '
+                                         'Please only set it in '
+                                         '`data.val_dataloader`. ')
+        cfg.data.val_dataloader['samples_per_gpu'] = \
+            cfg.data.val.pop('samples_per_gpu')
+    # special process for val_dataloader
+
+    # in case the test dataset is concatenated
+    if isinstance(cfg.data.test, dict):
+        if 'samples_per_gpu' in cfg.data.test:
+            assert 'samples_per_gpu' not in \
+                   cfg.data.test_dataloader, ('`samples_per_gpu` are set '
+                                              'in `data.test` field and ` '
+                                              'data.test_dataloader` '
+                                              'at the same time. '
+                                              'Please only set it in '
+                                              '`data.test_dataloader`. ')
+
+            cfg.data.test_dataloader['samples_per_gpu'] = \
+                cfg.data.test.pop('samples_per_gpu')
+
+    elif isinstance(cfg.data.test, list):
+        for ds_cfg in cfg.data.test:
+            if 'samples_per_gpu' in ds_cfg:
+                assert 'samples_per_gpu' not in \
+                       cfg.data.test_dataloader, ('`samples_per_gpu` are set '
+                                                  'in `data.test` field and ` '
+                                                  'data.test_dataloader` at'
+                                                  ' the same time. '
+                                                  'Please only set it in '
+                                                  '`data.test_dataloader`. ')
+        samples_per_gpu = max(
+            [ds_cfg.pop('samples_per_gpu', 1) for ds_cfg in cfg.data.test])
+        cfg.data.test_dataloader['samples_per_gpu'] = samples_per_gpu
+
+    return cfg

mmdet3d/utils/setup_env.py

@@ -27,7 +27,13 @@ def setup_multi_processes(cfg):
 
     # setup OMP threads
     # This code is referred from https://github.com/pytorch/pytorch/blob/master/torch/distributed/run.py  # noqa
-    if 'OMP_NUM_THREADS' not in os.environ and cfg.data.workers_per_gpu > 1:
+    workers_per_gpu = cfg.data.get('workers_per_gpu', 1)
+    if 'train_dataloader' in cfg.data:
+        workers_per_gpu = \
+            max(cfg.data.train_dataloader.get('workers_per_gpu', 1),
+                workers_per_gpu)
+
+    if 'OMP_NUM_THREADS' not in os.environ and workers_per_gpu > 1:
         omp_num_threads = 1
         warnings.warn(
             f'Setting OMP_NUM_THREADS environment variable for each process '
@@ -37,7 +43,7 @@ def setup_multi_processes(cfg):
         os.environ['OMP_NUM_THREADS'] = str(omp_num_threads)
 
     # setup MKL threads
-    if 'MKL_NUM_THREADS' not in os.environ and cfg.data.workers_per_gpu > 1:
+    if 'MKL_NUM_THREADS' not in os.environ and workers_per_gpu > 1:
         mkl_num_threads = 1
         warnings.warn(
             f'Setting MKL_NUM_THREADS environment variable for each process '

tests/test_utils/test_compat_cfg.py

+import pytest
+from mmcv import ConfigDict
+
+from mmdet3d.utils.compat_cfg import (compat_imgs_per_gpu, compat_loader_args,
+                                      compat_runner_args)
+
+
+def test_compat_runner_args():
+    cfg = ConfigDict(dict(total_epochs=12))
+    with pytest.warns(None) as record:
+        cfg = compat_runner_args(cfg)
+    assert len(record) == 1
+    assert 'runner' in record.list[0].message.args[0]
+    assert 'runner' in cfg
+    assert cfg.runner.type == 'EpochBasedRunner'
+    assert cfg.runner.max_epochs == cfg.total_epochs
+
+
+def test_compat_loader_args():
+    cfg = ConfigDict(dict(data=dict(val=dict(), test=dict(), train=dict())))
+    cfg = compat_loader_args(cfg)
+    # auto fill loader args
+    assert 'val_dataloader' in cfg.data
+    assert 'train_dataloader' in cfg.data
+    assert 'test_dataloader' in cfg.data
+    cfg = ConfigDict(
+        dict(
+            data=dict(
+                samples_per_gpu=1,
+                persistent_workers=True,
+                workers_per_gpu=1,
+                val=dict(samples_per_gpu=3),
+                test=dict(samples_per_gpu=2),
+                train=dict())))
+    with pytest.warns(None) as record:
+        cfg = compat_loader_args(cfg)
+    # 5 warning
+    assert len(record) == 5
+    # assert the warning message
+    assert 'train_dataloader' in record.list[0].message.args[0]
+    assert 'samples_per_gpu' in record.list[0].message.args[0]
+    assert 'persistent_workers' in record.list[1].message.args[0]
+    assert 'train_dataloader' in record.list[1].message.args[0]
+    assert 'workers_per_gpu' in record.list[2].message.args[0]
+    assert 'train_dataloader' in record.list[2].message.args[0]
+    assert cfg.data.train_dataloader.workers_per_gpu == 1
+    assert cfg.data.train_dataloader.samples_per_gpu == 1
+    assert cfg.data.train_dataloader.persistent_workers
+    assert cfg.data.val_dataloader.workers_per_gpu == 1
+    assert cfg.data.val_dataloader.samples_per_gpu == 3
+    assert cfg.data.test_dataloader.workers_per_gpu == 1
+    assert cfg.data.test_dataloader.samples_per_gpu == 2
+
+    # test test is a list
+    cfg = ConfigDict(
+        dict(
+            data=dict(
+                samples_per_gpu=1,
+                persistent_workers=True,
+                workers_per_gpu=1,
+                val=dict(samples_per_gpu=3),
+                test=[dict(samples_per_gpu=2),
+                      dict(samples_per_gpu=3)],
+                train=dict())))
+
+    with pytest.warns(None) as record:
+        cfg = compat_loader_args(cfg)
+    # 6 warning
+    assert len(record) == 6
+    assert cfg.data.test_dataloader.samples_per_gpu == 3
+
+    # assert can not set args at the same time
+    cfg = ConfigDict(
+        dict(
+            data=dict(
+                samples_per_gpu=1,
+                persistent_workers=True,
+                workers_per_gpu=1,
+                val=dict(samples_per_gpu=3),
+                test=dict(samples_per_gpu=2),
+                train=dict(),
+                train_dataloader=dict(samples_per_gpu=2))))
+    # samples_per_gpu can not be set in `train_dataloader`
+    # and data field at the same time
+    with pytest.raises(AssertionError):
+        compat_loader_args(cfg)
+    cfg = ConfigDict(
+        dict(
+            data=dict(
+                samples_per_gpu=1,
+                persistent_workers=True,
+                workers_per_gpu=1,
+                val=dict(samples_per_gpu=3),
+                test=dict(samples_per_gpu=2),
+                train=dict(),
+                val_dataloader=dict(samples_per_gpu=2))))
+    # samples_per_gpu can not be set in `val_dataloader`
+    # and data field at the same time
+    with pytest.raises(AssertionError):
+        compat_loader_args(cfg)
+    cfg = ConfigDict(
+        dict(
+            data=dict(
+                samples_per_gpu=1,
+                persistent_workers=True,
+                workers_per_gpu=1,
+                val=dict(samples_per_gpu=3),
+                test=dict(samples_per_gpu=2),
+                test_dataloader=dict(samples_per_gpu=2))))
+    # samples_per_gpu can not be set in `test_dataloader`
+    # and data field at the same time
+    with pytest.raises(AssertionError):
+        compat_loader_args(cfg)
+
+
+def test_compat_imgs_per_gpu():
+    cfg = ConfigDict(
+        dict(
+            data=dict(
+                imgs_per_gpu=1,
+                samples_per_gpu=2,
+                val=dict(),
+                test=dict(),
+                train=dict())))
+    cfg = compat_imgs_per_gpu(cfg)
+    assert cfg.data.samples_per_gpu == cfg.data.imgs_per_gpu

tools/test.py

@@ -25,6 +25,13 @@ if mmdet.__version__ > '2.23.0':
 else:
     from mmdet3d.utils import setup_multi_processes
 
+try:
+    # If mmdet version > 2.23.0, compat_cfg would be imported and
+    # used from mmdet instead of mmdet3d.
+    from mmdet.utils import compat_cfg
+except ImportError:
+    from mmdet3d.utils import compat_cfg
+
 
 def parse_args():
     parser = argparse.ArgumentParser(
@@ -139,6 +146,8 @@ def main():
     if args.cfg_options is not None:
         cfg.merge_from_dict(args.cfg_options)
 
+    cfg = compat_cfg(cfg)
+
     # set multi-process settings
     setup_multi_processes(cfg)
 
@@ -147,23 +156,6 @@ def main():
         torch.backends.cudnn.benchmark = True
 
     cfg.model.pretrained = None
-    # in case the test dataset is concatenated
-    samples_per_gpu = 1
-    if isinstance(cfg.data.test, dict):
-        cfg.data.test.test_mode = True
-        samples_per_gpu = cfg.data.test.pop('samples_per_gpu', 1)
-        if samples_per_gpu > 1:
-            # Replace 'ImageToTensor' to 'DefaultFormatBundle'
-            cfg.data.test.pipeline = replace_ImageToTensor(
-                cfg.data.test.pipeline)
-    elif isinstance(cfg.data.test, list):
-        for ds_cfg in cfg.data.test:
-            ds_cfg.test_mode = True
-        samples_per_gpu = max(
-            [ds_cfg.pop('samples_per_gpu', 1) for ds_cfg in cfg.data.test])
-        if samples_per_gpu > 1:
-            for ds_cfg in cfg.data.test:
-                ds_cfg.pipeline = replace_ImageToTensor(ds_cfg.pipeline)
 
     if args.gpu_ids is not None:
         cfg.gpu_ids = args.gpu_ids[0:1]
@@ -181,18 +173,35 @@ def main():
         distributed = True
         init_dist(args.launcher, **cfg.dist_params)
 
+    test_dataloader_default_args = dict(
+        samples_per_gpu=1, workers_per_gpu=2, dist=distributed, shuffle=False)
+
+    # in case the test dataset is concatenated
+    if isinstance(cfg.data.test, dict):
+        cfg.data.test.test_mode = True
+        if cfg.data.test_dataloader.get('samples_per_gpu', 1) > 1:
+            # Replace 'ImageToTensor' to 'DefaultFormatBundle'
+            cfg.data.test.pipeline = replace_ImageToTensor(
+                cfg.data.test.pipeline)
+    elif isinstance(cfg.data.test, list):
+        for ds_cfg in cfg.data.test:
+            ds_cfg.test_mode = True
+        if cfg.data.test_dataloader.get('samples_per_gpu', 1) > 1:
+            for ds_cfg in cfg.data.test:
+                ds_cfg.pipeline = replace_ImageToTensor(ds_cfg.pipeline)
+
+    test_loader_cfg = {
+        **test_dataloader_default_args,
+        **cfg.data.get('test_dataloader', {})
+    }
+
     # set random seeds
     if args.seed is not None:
         set_random_seed(args.seed, deterministic=args.deterministic)
 
     # build the dataloader
     dataset = build_dataset(cfg.data.test)
-    data_loader = build_dataloader(
-        dataset,
-        samples_per_gpu=samples_per_gpu,
-        workers_per_gpu=cfg.data.workers_per_gpu,
-        dist=distributed,
-        shuffle=False)
+    data_loader = build_dataloader(dataset, **test_loader_cfg)
 
     # build the model and load checkpoint
     cfg.model.train_cfg = None