From 6ea3577e0e7bb7957a3929695ccf7baf1da341f6 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Kaan=20=C3=87olak?= <kaancolak95@gmail.com>
Date: Tue, 14 Nov 2023 00:03:30 +0300
Subject: [PATCH 01/12] refactor(lidar_centerpoint): add training docs
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Kaan Çolak <kaancolak95@gmail.com>
---
 perception/lidar_centerpoint/README.md | 173 +++++++++++++++++++++++++
 1 file changed, 173 insertions(+)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index ed71349d5bd7f..2402058732580 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -68,6 +68,173 @@ You can download the onnx format of trained models by clicking on the links belo
 `Centerpoint` was trained in `nuScenes` (~28k lidar frames) [8] and TIER IV's internal database (~11k lidar frames) for 60 epochs.
 `Centerpoint tiny` was trained in `Argoverse 2` (~110k lidar frames) [9] and TIER IV's internal database (~11k lidar frames) for 20 epochs.
 
+## Training CenterPoint Model and Deploying to the Autoware
+
+### Overview
+This guide provides instructions on training a CenterPoint model using the **mmdetection3d** repository
+and seamlessly deploying it within the Autoware.  
+
+
+### Installation
+
+#### Install prerequisites
+**Step 1.** Download and install Miniconda from the [official website](https://mmpretrain.readthedocs.io/en/latest/get_started.html).
+
+**Step 2.** Create a conda virtual environment and activate it
+
+```bash
+conda create --name train-centerpoint python=3.8 -y
+conda activate train-centerpoint
+```
+
+**Step 3.** Install PyTorch
+
+Please ensure you have PyTorch installed, compatible with CUDA 11.6, as it is a requirement for current Autoware.
+
+```bash
+conda install pytorch==1.13.1 torchvision==0.14.1 pytorch-cuda=11.6 -c pytorch -c nvidia
+```
+#### Install mmdetection3d
+**Step 1.** Install MMEngine, MMCV and MMDetection using MIM
+
+```bash
+pip install -U openmim
+mim install mmengine
+mim install 'mmcv>=2.0.0rc4'
+mim install 'mmdet>=3.0.0'
+```
+
+**Step 2.** Install mmdetection3d forked repository
+
+Introduced several valuable enhancements in our fork of the mmdetection3d repository.
+Notably, we've made the PointPillar z voxel feature input optional to maintain compatibility with the original paper. 
+In addition, we've integrated a PyTorch to ONNX converter and a Tier4 Dataset format reader for added functionality.
+
+
+```bash
+git clone https://github.com/autowarefoundation/mmdetection3d.git -b dev-1.x-autoware
+cd mmdetection3d
+pip install -v -e .
+```
+
+### Preparing NuScenes dataset for training
+
+**Step 1.** Download the NuScenes dataset from the [official website](https://www.nuscenes.org/download) and extract the dataset to a folder of your choice.
+
+**Step 2.** Create a symbolic link to the dataset folder
+
+```bash
+ln -s /path/to/nuscenes/dataset/ /path/to/mmdetection3d/data/nuscenes/
+```
+
+**Step 3.** Prepare the NuScenes data by running:
+
+```bash
+cd mmdetection3d
+python tools/create_data.py nuscenes --root-path ./data/nuscenes --out-dir ./data/nuscenes --extra-tag nuscenes
+```
+
+### Training CenterPoint with NuScenes Dataset
+
+#### Prepare the config file
+
+The configuration file that illustrates how to train the CenterPoint model with the NuScenes dataset is 
+located at mmdetection3d/configs/centerpoint/centerpoint_custom.py. This configuration file is a derived version of the 
+centerpoint_pillar02_second_secfpn_head-circlenms_8xb4-cyclic-20e_nus-3d.py configuration file from mmdetection3D.
+In this custom configuration, the **use_voxel_center_z parameter** is set to **False** to deactivate the z coordinate of the voxel center,
+aligning with the original paper's specifications and making the model compatible with Autoware. Additionally, the filter size is set as **[32, 32]**.
+
+The CenterPoint model can be tailored to your specific requirements by modifying various parameters within the configuration file.
+This includes adjustments related to preprocessing operations, training, testing, model architecture, dataset, optimizer, learning rate scheduler, and more.
+
+#### Start training
+
+```bash
+python tools/train.py configs/centerpoint/centerpoint_custom.py --work-dir ./work_dirs/centerpoint_custom
+```
+
+#### Evaluation of the trained model
+
+For evaluation purposes, we have included a sample dataset captured from vehicle which consists of the following LiDAR sensors: 
+1 x Velodyne VLS128, 4 x Velodyne VLP16, and 1 x Robosense RS Bpearl. This dataset comprises 600 LiDAR frames and encompasses 5 distinct classes, 6905 cars, 3951 pedestrians,
+75 cyclists, 162 buses, and 326 trucks 3D annotation. In the sample dataset, frames annotatated as a 2 frame, each second. You can employ this dataset for a wide range of purposes,
+including training, evaluation, and fine-tuning of models. It is organized in the Tier4Dataset format. 
+
+##### Download the sample dataset
+```bash 
+TODO(kaancolak): add the link to the sample dataset
+
+#Extract the dataset to a folder of your choice
+
+#Create a symbolic link to the dataset folder
+ln -s /PATH/TO/DATASET/ /PATH/TO/mmdetection3d/data/tier4_dataset/
+```
+
+##### Prepare dataset and evaluate trained model
+
+Create .pkl files for the purposes of training, evaluation, and testing.
+
+```bash
+
+python tools/create_data.py Tier4Dataset --root-path data/sample_dataset/ --out-dir data/sample_dataset/ --extra-tag Tier4Dataset --version sample_dataset --annotation-hz 2
+```
+
+Run evaluation
+
+```bash
+python tools/test.py ./configs/centerpoint/test-centerpoint.py /PATH/OF/THE/CHECKPOINT  --task lidar_det
+```
+
+Evaluation result could be relatively low because of the e to variations in sensor modalities between the sample dataset
+and the training dataset. The model's training parameters are originally tailored to the NuScenes dataset, which employs a single lidar 
+sensor positioned atop the vehicle. In contrast, the provided sample dataset comprises concatenated point clouds positioned at
+the base link location of the vehicle.
+
+### Deploying CenterPoint model to Autoware
+
+#### Convert CenterPoint PyTorch model to ONNX Format
+The lidar_centerpoint implementation requires two ONNX models as input the voxel encoder and the backbone-neck-head of the CenterPoint model, other aspects of the network,
+such as preprocessing operations, are implemented externally. Under the fork of the mmdetection3d repository, 
+we have included a script that converts the CenterPoint model to Autoware compitible ONNX format. 
+You can find it in `mmdetection3d/tools/centerpoint_onnx_converter.py` file.
+ 
+
+```bash
+python tools/centerpoint_onnx_converter.py --cfg configs/centerpoint/centerpoint_custom.py --ckpt work_dirs/centerpoint_custom/YOUR_BEST_MODEL.pth -work-dir ./work_dirs/onnx_models
+```
+
+#### Create the config file for custom model 
+
+Create a new config file named **centerpoint_custom.param.yaml** under the config file directory of the lidar_centerpoint node. Sets the parameters of the config file like
+point_cloud_range, point_feature_size, voxel_size, etc. according to the training config file. 
+
+```yaml
+/**:
+  ros__parameters:
+    class_names: ["CAR", "TRUCK", "BUS", "BICYCLE", "PEDESTRIAN"]
+    point_feature_size: 4
+    max_voxel_size: 40000
+    point_cloud_range: [-51.2, -51.2, -3.0, 51.2, 51.2, 5.0,]
+    voxel_size: [0.2, 0.2, 8.0]
+    downsample_factor: 1
+    encoder_in_feature_size: 9
+    # post-process params
+    circle_nms_dist_threshold: 0.5
+    iou_nms_target_class_names: ["CAR"]
+    iou_nms_search_distance_2d: 10.0
+    iou_nms_threshold: 0.1
+    yaw_norm_thresholds: [0.3, 0.3, 0.3, 0.3, 0.0]
+```
+
+#### Launch the lidar_centerpoint node
+
+```bash
+cd /YOUR/AUTOWARE/PATH/Autoware
+source install/setup.bash
+ros2 launch lidar_centerpoint lidar_centerpoint.launch.xml  model_name:=centerpoint_custom  model_path:=/PATH/TO/ONNX/FILE/
+```
+
+
 ### Changelog
 
 #### v1 (2022/07/06)
@@ -142,3 +309,9 @@ Example:
 [v1-head-centerpoint]: https://awf.ml.dev.web.auto/perception/models/centerpoint/v1/pts_backbone_neck_head_centerpoint.onnx
 [v1-encoder-centerpoint-tiny]: https://awf.ml.dev.web.auto/perception/models/centerpoint/v1/pts_voxel_encoder_centerpoint_tiny.onnx
 [v1-head-centerpoint-tiny]: https://awf.ml.dev.web.auto/perception/models/centerpoint/v1/pts_backbone_neck_head_centerpoint_tiny.onnx
+
+## Legal Notice
+*The nuScenes dataset is released publicly for non-commercial use under the Creative 
+Commons Attribution-NonCommercial-ShareAlike 4.0 International Public License. 
+Additional Terms of Use can be found at https://www.nuscenes.org/terms-of-use. 
+To inquire about a commercial license please contact nuscenes@motional.com.*

From 20ef77c2f464424073fe2b912142d3b1812dfb26 Mon Sep 17 00:00:00 2001
From: "pre-commit-ci[bot]"
 <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Date: Mon, 13 Nov 2023 21:06:48 +0000
Subject: [PATCH 02/12] style(pre-commit): autofix

---
 perception/lidar_centerpoint/README.md | 45 ++++++++++++++------------
 1 file changed, 24 insertions(+), 21 deletions(-)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index 2402058732580..2ae4335549092 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -71,13 +71,14 @@ You can download the onnx format of trained models by clicking on the links belo
 ## Training CenterPoint Model and Deploying to the Autoware
 
 ### Overview
-This guide provides instructions on training a CenterPoint model using the **mmdetection3d** repository
-and seamlessly deploying it within the Autoware.  
 
+This guide provides instructions on training a CenterPoint model using the **mmdetection3d** repository
+and seamlessly deploying it within the Autoware.
 
 ### Installation
 
 #### Install prerequisites
+
 **Step 1.** Download and install Miniconda from the [official website](https://mmpretrain.readthedocs.io/en/latest/get_started.html).
 
 **Step 2.** Create a conda virtual environment and activate it
@@ -94,7 +95,9 @@ Please ensure you have PyTorch installed, compatible with CUDA 11.6, as it is a
 ```bash
 conda install pytorch==1.13.1 torchvision==0.14.1 pytorch-cuda=11.6 -c pytorch -c nvidia
 ```
+
 #### Install mmdetection3d
+
 **Step 1.** Install MMEngine, MMCV and MMDetection using MIM
 
 ```bash
@@ -107,10 +110,9 @@ mim install 'mmdet>=3.0.0'
 **Step 2.** Install mmdetection3d forked repository
 
 Introduced several valuable enhancements in our fork of the mmdetection3d repository.
-Notably, we've made the PointPillar z voxel feature input optional to maintain compatibility with the original paper. 
+Notably, we've made the PointPillar z voxel feature input optional to maintain compatibility with the original paper.
 In addition, we've integrated a PyTorch to ONNX converter and a Tier4 Dataset format reader for added functionality.
 
-
 ```bash
 git clone https://github.com/autowarefoundation/mmdetection3d.git -b dev-1.x-autoware
 cd mmdetection3d
@@ -138,8 +140,8 @@ python tools/create_data.py nuscenes --root-path ./data/nuscenes --out-dir ./dat
 
 #### Prepare the config file
 
-The configuration file that illustrates how to train the CenterPoint model with the NuScenes dataset is 
-located at mmdetection3d/configs/centerpoint/centerpoint_custom.py. This configuration file is a derived version of the 
+The configuration file that illustrates how to train the CenterPoint model with the NuScenes dataset is
+located at mmdetection3d/configs/centerpoint/centerpoint_custom.py. This configuration file is a derived version of the
 centerpoint_pillar02_second_secfpn_head-circlenms_8xb4-cyclic-20e_nus-3d.py configuration file from mmdetection3D.
 In this custom configuration, the **use_voxel_center_z parameter** is set to **False** to deactivate the z coordinate of the voxel center,
 aligning with the original paper's specifications and making the model compatible with Autoware. Additionally, the filter size is set as **[32, 32]**.
@@ -155,13 +157,14 @@ python tools/train.py configs/centerpoint/centerpoint_custom.py --work-dir ./wor
 
 #### Evaluation of the trained model
 
-For evaluation purposes, we have included a sample dataset captured from vehicle which consists of the following LiDAR sensors: 
+For evaluation purposes, we have included a sample dataset captured from vehicle which consists of the following LiDAR sensors:
 1 x Velodyne VLS128, 4 x Velodyne VLP16, and 1 x Robosense RS Bpearl. This dataset comprises 600 LiDAR frames and encompasses 5 distinct classes, 6905 cars, 3951 pedestrians,
 75 cyclists, 162 buses, and 326 trucks 3D annotation. In the sample dataset, frames annotatated as a 2 frame, each second. You can employ this dataset for a wide range of purposes,
-including training, evaluation, and fine-tuning of models. It is organized in the Tier4Dataset format. 
+including training, evaluation, and fine-tuning of models. It is organized in the Tier4Dataset format.
 
 ##### Download the sample dataset
-```bash 
+
+```bash
 TODO(kaancolak): add the link to the sample dataset
 
 #Extract the dataset to a folder of your choice
@@ -186,27 +189,27 @@ python tools/test.py ./configs/centerpoint/test-centerpoint.py /PATH/OF/THE/CHEC
 ```
 
 Evaluation result could be relatively low because of the e to variations in sensor modalities between the sample dataset
-and the training dataset. The model's training parameters are originally tailored to the NuScenes dataset, which employs a single lidar 
+and the training dataset. The model's training parameters are originally tailored to the NuScenes dataset, which employs a single lidar
 sensor positioned atop the vehicle. In contrast, the provided sample dataset comprises concatenated point clouds positioned at
 the base link location of the vehicle.
 
 ### Deploying CenterPoint model to Autoware
 
 #### Convert CenterPoint PyTorch model to ONNX Format
+
 The lidar_centerpoint implementation requires two ONNX models as input the voxel encoder and the backbone-neck-head of the CenterPoint model, other aspects of the network,
-such as preprocessing operations, are implemented externally. Under the fork of the mmdetection3d repository, 
-we have included a script that converts the CenterPoint model to Autoware compitible ONNX format. 
+such as preprocessing operations, are implemented externally. Under the fork of the mmdetection3d repository,
+we have included a script that converts the CenterPoint model to Autoware compitible ONNX format.
 You can find it in `mmdetection3d/tools/centerpoint_onnx_converter.py` file.
- 
 
 ```bash
 python tools/centerpoint_onnx_converter.py --cfg configs/centerpoint/centerpoint_custom.py --ckpt work_dirs/centerpoint_custom/YOUR_BEST_MODEL.pth -work-dir ./work_dirs/onnx_models
 ```
 
-#### Create the config file for custom model 
+#### Create the config file for custom model
 
 Create a new config file named **centerpoint_custom.param.yaml** under the config file directory of the lidar_centerpoint node. Sets the parameters of the config file like
-point_cloud_range, point_feature_size, voxel_size, etc. according to the training config file. 
+point_cloud_range, point_feature_size, voxel_size, etc. according to the training config file.
 
 ```yaml
 /**:
@@ -214,7 +217,7 @@ point_cloud_range, point_feature_size, voxel_size, etc. according to the trainin
     class_names: ["CAR", "TRUCK", "BUS", "BICYCLE", "PEDESTRIAN"]
     point_feature_size: 4
     max_voxel_size: 40000
-    point_cloud_range: [-51.2, -51.2, -3.0, 51.2, 51.2, 5.0,]
+    point_cloud_range: [-51.2, -51.2, -3.0, 51.2, 51.2, 5.0]
     voxel_size: [0.2, 0.2, 8.0]
     downsample_factor: 1
     encoder_in_feature_size: 9
@@ -234,7 +237,6 @@ source install/setup.bash
 ros2 launch lidar_centerpoint lidar_centerpoint.launch.xml  model_name:=centerpoint_custom  model_path:=/PATH/TO/ONNX/FILE/
 ```
 
-
 ### Changelog
 
 #### v1 (2022/07/06)
@@ -311,7 +313,8 @@ Example:
 [v1-head-centerpoint-tiny]: https://awf.ml.dev.web.auto/perception/models/centerpoint/v1/pts_backbone_neck_head_centerpoint_tiny.onnx
 
 ## Legal Notice
-*The nuScenes dataset is released publicly for non-commercial use under the Creative 
-Commons Attribution-NonCommercial-ShareAlike 4.0 International Public License. 
-Additional Terms of Use can be found at https://www.nuscenes.org/terms-of-use. 
-To inquire about a commercial license please contact nuscenes@motional.com.*
+
+_The nuScenes dataset is released publicly for non-commercial use under the Creative
+Commons Attribution-NonCommercial-ShareAlike 4.0 International Public License.
+Additional Terms of Use can be found at <https://www.nuscenes.org/terms-of-use>.
+To inquire about a commercial license please contact nuscenes@motional.com._

From 48afb5c03b0564f7b22497ed4fcc7fbab1fd3be0 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Kaan=20=C3=87olak?= <kaancolak95@gmail.com>
Date: Thu, 7 Dec 2023 15:38:12 +0900
Subject: [PATCH 03/12] refactor(lidar_centerpoint): and link and small fix
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Kaan Çolak <kaancolak95@gmail.com>
---
 perception/lidar_centerpoint/README.md | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index 2ae4335549092..902bef0abdeac 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -73,7 +73,7 @@ You can download the onnx format of trained models by clicking on the links belo
 ### Overview
 
 This guide provides instructions on training a CenterPoint model using the **mmdetection3d** repository
-and seamlessly deploying it within the Autoware.
+and seamlessly deploying it within Autoware.
 
 ### Installation
 
@@ -90,7 +90,7 @@ conda activate train-centerpoint
 
 **Step 3.** Install PyTorch
 
-Please ensure you have PyTorch installed, compatible with CUDA 11.6, as it is a requirement for current Autoware.
+Please ensure you have PyTorch installed, and compatible with CUDA 11.6, as it is a requirement for current Autoware.
 
 ```bash
 conda install pytorch==1.13.1 torchvision==0.14.1 pytorch-cuda=11.6 -c pytorch -c nvidia
@@ -98,7 +98,7 @@ conda install pytorch==1.13.1 torchvision==0.14.1 pytorch-cuda=11.6 -c pytorch -
 
 #### Install mmdetection3d
 
-**Step 1.** Install MMEngine, MMCV and MMDetection using MIM
+**Step 1.** Install MMEngine, MMCV, and MMDetection using MIM
 
 ```bash
 pip install -U openmim
@@ -114,7 +114,7 @@ Notably, we've made the PointPillar z voxel feature input optional to maintain c
 In addition, we've integrated a PyTorch to ONNX converter and a Tier4 Dataset format reader for added functionality.
 
 ```bash
-git clone https://github.com/autowarefoundation/mmdetection3d.git -b dev-1.x-autoware
+git clone https://github.com/autowarefoundation/mmdetection3d.git
 cd mmdetection3d
 pip install -v -e .
 ```
@@ -157,25 +157,25 @@ python tools/train.py configs/centerpoint/centerpoint_custom.py --work-dir ./wor
 
 #### Evaluation of the trained model
 
-For evaluation purposes, we have included a sample dataset captured from vehicle which consists of the following LiDAR sensors:
+For evaluation purposes, we have included a sample dataset captured from the vehicle which consists of the following LiDAR sensors:
 1 x Velodyne VLS128, 4 x Velodyne VLP16, and 1 x Robosense RS Bpearl. This dataset comprises 600 LiDAR frames and encompasses 5 distinct classes, 6905 cars, 3951 pedestrians,
-75 cyclists, 162 buses, and 326 trucks 3D annotation. In the sample dataset, frames annotatated as a 2 frame, each second. You can employ this dataset for a wide range of purposes,
+75 cyclists, 162 buses, and 326 trucks 3D annotation. In the sample dataset, frames are annotated as 2 frames for each second. You can employ this dataset for a wide range of purposes,
 including training, evaluation, and fine-tuning of models. It is organized in the Tier4Dataset format.
 
 ##### Download the sample dataset
 
 ```bash
-TODO(kaancolak): add the link to the sample dataset
 
+wget https://autoware-files.s3.us-west-2.amazonaws.com/dataset/lidar_detection_sample_dataset.tar.gz
 #Extract the dataset to a folder of your choice
-
+tar -xvf lidar_detection_sample_dataset.tar.gz
 #Create a symbolic link to the dataset folder
 ln -s /PATH/TO/DATASET/ /PATH/TO/mmdetection3d/data/tier4_dataset/
 ```
 
 ##### Prepare dataset and evaluate trained model
 
-Create .pkl files for the purposes of training, evaluation, and testing.
+Create .pkl files for training, evaluation, and testing.
 
 ```bash
 
@@ -188,7 +188,7 @@ Run evaluation
 python tools/test.py ./configs/centerpoint/test-centerpoint.py /PATH/OF/THE/CHECKPOINT  --task lidar_det
 ```
 
-Evaluation result could be relatively low because of the e to variations in sensor modalities between the sample dataset
+Evaluation results could be relatively low because of the e to variations in sensor modalities between the sample dataset
 and the training dataset. The model's training parameters are originally tailored to the NuScenes dataset, which employs a single lidar
 sensor positioned atop the vehicle. In contrast, the provided sample dataset comprises concatenated point clouds positioned at
 the base link location of the vehicle.
@@ -199,14 +199,14 @@ the base link location of the vehicle.
 
 The lidar_centerpoint implementation requires two ONNX models as input the voxel encoder and the backbone-neck-head of the CenterPoint model, other aspects of the network,
 such as preprocessing operations, are implemented externally. Under the fork of the mmdetection3d repository,
-we have included a script that converts the CenterPoint model to Autoware compitible ONNX format.
+we have included a script that converts the CenterPoint model to Autoware compatible ONNX format.
 You can find it in `mmdetection3d/tools/centerpoint_onnx_converter.py` file.
 
 ```bash
 python tools/centerpoint_onnx_converter.py --cfg configs/centerpoint/centerpoint_custom.py --ckpt work_dirs/centerpoint_custom/YOUR_BEST_MODEL.pth -work-dir ./work_dirs/onnx_models
 ```
 
-#### Create the config file for custom model
+#### Create the config file for the custom model
 
 Create a new config file named **centerpoint_custom.param.yaml** under the config file directory of the lidar_centerpoint node. Sets the parameters of the config file like
 point_cloud_range, point_feature_size, voxel_size, etc. according to the training config file.

From 06ab87ba7865f56545965bd6718934f3e3210579 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Kaan=20=C3=87olak?= <kaancolak95@gmail.com>
Date: Mon, 22 Jan 2024 18:26:15 +0900
Subject: [PATCH 04/12] refactor(lidar_centerpoint): update docs.
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Kaan Çolak <kaancolak95@gmail.com>
---
 perception/lidar_centerpoint/README.md | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index 902bef0abdeac..f9803ce0af208 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -104,7 +104,7 @@ conda install pytorch==1.13.1 torchvision==0.14.1 pytorch-cuda=11.6 -c pytorch -
 pip install -U openmim
 mim install mmengine
 mim install 'mmcv>=2.0.0rc4'
-mim install 'mmdet>=3.0.0'
+mim install 'mmdet>=3.0.0rc5, <3.3.0'
 ```
 
 **Step 2.** Install mmdetection3d forked repository
@@ -141,7 +141,7 @@ python tools/create_data.py nuscenes --root-path ./data/nuscenes --out-dir ./dat
 #### Prepare the config file
 
 The configuration file that illustrates how to train the CenterPoint model with the NuScenes dataset is
-located at mmdetection3d/configs/centerpoint/centerpoint_custom.py. This configuration file is a derived version of the
+located at mmdetection3d/projects/AutowareCenterPoint/configs. This configuration file is a derived version of the
 centerpoint_pillar02_second_secfpn_head-circlenms_8xb4-cyclic-20e_nus-3d.py configuration file from mmdetection3D.
 In this custom configuration, the **use_voxel_center_z parameter** is set to **False** to deactivate the z coordinate of the voxel center,
 aligning with the original paper's specifications and making the model compatible with Autoware. Additionally, the filter size is set as **[32, 32]**.
@@ -152,7 +152,7 @@ This includes adjustments related to preprocessing operations, training, testing
 #### Start training
 
 ```bash
-python tools/train.py configs/centerpoint/centerpoint_custom.py --work-dir ./work_dirs/centerpoint_custom
+python tools/train.py projects/AutowareCenterPoint/configs/centerpoint_custom.py --work-dir ./work_dirs/centerpoint_custom
 ```
 
 #### Evaluation of the trained model
@@ -185,7 +185,7 @@ python tools/create_data.py Tier4Dataset --root-path data/sample_dataset/ --out-
 Run evaluation
 
 ```bash
-python tools/test.py ./configs/centerpoint/test-centerpoint.py /PATH/OF/THE/CHECKPOINT  --task lidar_det
+python tools/test.py projects/AutowareCenterPoint/configs/centerpoint_custom_test.py /PATH/OF/THE/CHECKPOINT  --task lidar_det
 ```
 
 Evaluation results could be relatively low because of the e to variations in sensor modalities between the sample dataset
@@ -200,10 +200,10 @@ the base link location of the vehicle.
 The lidar_centerpoint implementation requires two ONNX models as input the voxel encoder and the backbone-neck-head of the CenterPoint model, other aspects of the network,
 such as preprocessing operations, are implemented externally. Under the fork of the mmdetection3d repository,
 we have included a script that converts the CenterPoint model to Autoware compatible ONNX format.
-You can find it in `mmdetection3d/tools/centerpoint_onnx_converter.py` file.
+You can find it in `mmdetection3d/projects/AutowareCenterPoint` file.
 
 ```bash
-python tools/centerpoint_onnx_converter.py --cfg configs/centerpoint/centerpoint_custom.py --ckpt work_dirs/centerpoint_custom/YOUR_BEST_MODEL.pth -work-dir ./work_dirs/onnx_models
+python projects/AutowareCenterPoint/centerpoint_onnx_converter.py --cfg projects/AutowareCenterPoint/configs/centerpoint_custom.py --ckpt work_dirs/centerpoint_custom/YOUR_BEST_MODEL.pth --work-dir ./work_dirs/onnx_models
 ```
 
 #### Create the config file for the custom model

From b7d671a305ad04d50e46b68b08f2e0c9492580f5 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Kaan=20=C3=87olak?= <kaancolak95@gmail.com>
Date: Tue, 27 Feb 2024 18:50:53 +0300
Subject: [PATCH 05/12] fix(lidar_centerpoint): change dataset name
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Kaan Çolak <kaancolak95@gmail.com>
---
 perception/lidar_centerpoint/README.md | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index f9803ce0af208..e51c43c626991 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -111,7 +111,7 @@ mim install 'mmdet>=3.0.0rc5, <3.3.0'
 
 Introduced several valuable enhancements in our fork of the mmdetection3d repository.
 Notably, we've made the PointPillar z voxel feature input optional to maintain compatibility with the original paper.
-In addition, we've integrated a PyTorch to ONNX converter and a Tier4 Dataset format reader for added functionality.
+In addition, we've integrated a PyTorch to ONNX converter and a T4 format reader for added functionality.
 
 ```bash
 git clone https://github.com/autowarefoundation/mmdetection3d.git
@@ -160,7 +160,7 @@ python tools/train.py projects/AutowareCenterPoint/configs/centerpoint_custom.py
 For evaluation purposes, we have included a sample dataset captured from the vehicle which consists of the following LiDAR sensors:
 1 x Velodyne VLS128, 4 x Velodyne VLP16, and 1 x Robosense RS Bpearl. This dataset comprises 600 LiDAR frames and encompasses 5 distinct classes, 6905 cars, 3951 pedestrians,
 75 cyclists, 162 buses, and 326 trucks 3D annotation. In the sample dataset, frames are annotated as 2 frames for each second. You can employ this dataset for a wide range of purposes,
-including training, evaluation, and fine-tuning of models. It is organized in the Tier4Dataset format.
+including training, evaluation, and fine-tuning of models. It is organized in the T4 format.
 
 ##### Download the sample dataset
 
@@ -179,7 +179,7 @@ Create .pkl files for training, evaluation, and testing.
 
 ```bash
 
-python tools/create_data.py Tier4Dataset --root-path data/sample_dataset/ --out-dir data/sample_dataset/ --extra-tag Tier4Dataset --version sample_dataset --annotation-hz 2
+python tools/create_data.py T4Dataset --root-path data/sample_dataset/ --out-dir data/sample_dataset/ --extra-tag T4Dataset --version sample_dataset --annotation-hz 2
 ```
 
 Run evaluation

From 808a7c7cd7d2a604535820f578db4d512f57da6d Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Kaan=20=C3=87olak?= <kaancolak95@gmail.com>
Date: Wed, 28 Feb 2024 11:57:05 +0300
Subject: [PATCH 06/12] fix(lidar_centerpoint): add docker instruction
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Kaan Çolak <kaancolak95@gmail.com>
---
 perception/lidar_centerpoint/README.md | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index e51c43c626991..2fd56c6011d02 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -119,6 +119,25 @@ cd mmdetection3d
 pip install -v -e .
 ```
 
+#### Use Training Repository with Docker
+Alternatively, you can use Docker to run the mmdetection3d repository.We provide a Dockerfile to build a Docker image with the mmdetection3d repository and its dependencies.
+
+Clone fork of the mmdetection3d repository
+```bash
+git clone https://github.com/autowarefoundation/mmdetection3d.git
+```
+
+Build the Docker image by running the following command
+```bash
+cd mmdetection3d
+docker build -t mmdetection3d -f docker/Dockerfile .
+```
+
+Run the Docker container
+```bash
+docker run --gpus all --shm-size=8g -it -v {DATA_DIR}:/mmdetection3d/data mmdetection3d
+```
+
 ### Preparing NuScenes dataset for training
 
 **Step 1.** Download the NuScenes dataset from the [official website](https://www.nuscenes.org/download) and extract the dataset to a folder of your choice.

From 3345fd8fae61cc9baf532937c428ac01589c09c1 Mon Sep 17 00:00:00 2001
From: "pre-commit-ci[bot]"
 <66853113+pre-commit-ci[bot]@users.noreply.github.com>
Date: Wed, 28 Feb 2024 08:59:04 +0000
Subject: [PATCH 07/12] style(pre-commit): autofix

---
 perception/lidar_centerpoint/README.md | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index 2fd56c6011d02..c09d6ad5c721e 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -120,20 +120,24 @@ pip install -v -e .
 ```
 
 #### Use Training Repository with Docker
+
 Alternatively, you can use Docker to run the mmdetection3d repository.We provide a Dockerfile to build a Docker image with the mmdetection3d repository and its dependencies.
 
 Clone fork of the mmdetection3d repository
+
 ```bash
 git clone https://github.com/autowarefoundation/mmdetection3d.git
 ```
 
 Build the Docker image by running the following command
+
 ```bash
 cd mmdetection3d
 docker build -t mmdetection3d -f docker/Dockerfile .
 ```
 
 Run the Docker container
+
 ```bash
 docker run --gpus all --shm-size=8g -it -v {DATA_DIR}:/mmdetection3d/data mmdetection3d
 ```

From 2f780bd65a6f31d9cd8f89124dd109edba7124ef Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Kaan=20=C3=87olak?= <kaancolak95@gmail.com>
Date: Tue, 26 Mar 2024 15:30:27 +0300
Subject: [PATCH 08/12] fix(lidar_centerpoint): fix spell
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: Kaan Çolak <kaancolak95@gmail.com>
---
 perception/lidar_centerpoint/README.md | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index c09d6ad5c721e..692bf27d8059e 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -146,6 +146,8 @@ docker run --gpus all --shm-size=8g -it -v {DATA_DIR}:/mmdetection3d/data mmdete
 
 **Step 1.** Download the NuScenes dataset from the [official website](https://www.nuscenes.org/download) and extract the dataset to a folder of your choice.
 
+**Note:** The NuScenes dataset is large and requires significant disk space. Ensure you have enough storage available before proceeding.
+
 **Step 2.** Create a symbolic link to the dataset folder
 
 ```bash
@@ -165,7 +167,7 @@ python tools/create_data.py nuscenes --root-path ./data/nuscenes --out-dir ./dat
 
 The configuration file that illustrates how to train the CenterPoint model with the NuScenes dataset is
 located at mmdetection3d/projects/AutowareCenterPoint/configs. This configuration file is a derived version of the
-centerpoint_pillar02_second_secfpn_head-circlenms_8xb4-cyclic-20e_nus-3d.py configuration file from mmdetection3D.
+`centerpoint_pillar02_second_secfpn_head-circlenms_8xb4-cyclic-20e_nus-3d.py` configuration file from mmdetection3D.
 In this custom configuration, the **use_voxel_center_z parameter** is set to **False** to deactivate the z coordinate of the voxel center,
 aligning with the original paper's specifications and making the model compatible with Autoware. Additionally, the filter size is set as **[32, 32]**.
 

From ef15632554a4f58b5b479f56fb599d86c0eb3e08 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?M=2E=20Fatih=20C=C4=B1r=C4=B1t?= <mfc@leodrive.ai>
Date: Tue, 26 Mar 2024 15:50:59 +0300
Subject: [PATCH 09/12] small fixes and spellcheck
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Signed-off-by: M. Fatih Cırıt <mfc@leodrive.ai>
---
 perception/lidar_centerpoint/README.md | 19 +++++++++----------
 1 file changed, 9 insertions(+), 10 deletions(-)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index 692bf27d8059e..ff6bc4c0c0abb 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -62,7 +62,7 @@ ros2 launch lidar_centerpoint lidar_centerpoint.launch.xml model_name:=centerpoi
 
 You can download the onnx format of trained models by clicking on the links below.
 
-- Centerpoint : [pts_voxel_encoder_centerpoint.onnx](https://awf.ml.dev.web.auto/perception/models/centerpoint/v2/pts_voxel_encoder_centerpoint.onnx), [pts_backbone_neck_head_centerpoint.onnx](https://awf.ml.dev.web.auto/perception/models/centerpoint/v2/pts_backbone_neck_head_centerpoint.onnx)
+- Centerpoint: [pts_voxel_encoder_centerpoint.onnx](https://awf.ml.dev.web.auto/perception/models/centerpoint/v2/pts_voxel_encoder_centerpoint.onnx), [pts_backbone_neck_head_centerpoint.onnx](https://awf.ml.dev.web.auto/perception/models/centerpoint/v2/pts_backbone_neck_head_centerpoint.onnx)
 - Centerpoint tiny: [pts_voxel_encoder_centerpoint_tiny.onnx](https://awf.ml.dev.web.auto/perception/models/centerpoint/v2/pts_voxel_encoder_centerpoint_tiny.onnx), [pts_backbone_neck_head_centerpoint_tiny.onnx](https://awf.ml.dev.web.auto/perception/models/centerpoint/v2/pts_backbone_neck_head_centerpoint_tiny.onnx)
 
 `Centerpoint` was trained in `nuScenes` (~28k lidar frames) [8] and TIER IV's internal database (~11k lidar frames) for 60 epochs.
@@ -121,7 +121,7 @@ pip install -v -e .
 
 #### Use Training Repository with Docker
 
-Alternatively, you can use Docker to run the mmdetection3d repository.We provide a Dockerfile to build a Docker image with the mmdetection3d repository and its dependencies.
+Alternatively, you can use Docker to run the mmdetection3d repository. We provide a Dockerfile to build a Docker image with the mmdetection3d repository and its dependencies.
 
 Clone fork of the mmdetection3d repository
 
@@ -129,14 +129,14 @@ Clone fork of the mmdetection3d repository
 git clone https://github.com/autowarefoundation/mmdetection3d.git
 ```
 
-Build the Docker image by running the following command
+Build the Docker image by running the following command:
 
 ```bash
 cd mmdetection3d
 docker build -t mmdetection3d -f docker/Dockerfile .
 ```
 
-Run the Docker container
+Run the Docker container:
 
 ```bash
 docker run --gpus all --shm-size=8g -it -v {DATA_DIR}:/mmdetection3d/data mmdetection3d
@@ -166,9 +166,10 @@ python tools/create_data.py nuscenes --root-path ./data/nuscenes --out-dir ./dat
 #### Prepare the config file
 
 The configuration file that illustrates how to train the CenterPoint model with the NuScenes dataset is
-located at mmdetection3d/projects/AutowareCenterPoint/configs. This configuration file is a derived version of the
-`centerpoint_pillar02_second_secfpn_head-circlenms_8xb4-cyclic-20e_nus-3d.py` configuration file from mmdetection3D.
-In this custom configuration, the **use_voxel_center_z parameter** is set to **False** to deactivate the z coordinate of the voxel center,
+located at `mmdetection3d/projects/AutowareCenterPoint/configs`. This configuration file is a derived version of
+[this centerpoint configuration file](https://github.com/autowarefoundation/mmdetection3d/blob/5c0613be29bd2e51771ec5e046d89ba3089887c7/configs/centerpoint/centerpoint_pillar02_second_secfpn_head-circlenms_8xb4-cyclic-20e_nus-3d.py)
+from mmdetection3D.
+In this custom configuration, the **use_voxel_center_z parameter** is set as **False** to deactivate the z coordinate of the voxel center,
 aligning with the original paper's specifications and making the model compatible with Autoware. Additionally, the filter size is set as **[32, 32]**.
 
 The CenterPoint model can be tailored to your specific requirements by modifying various parameters within the configuration file.
@@ -190,7 +191,6 @@ including training, evaluation, and fine-tuning of models. It is organized in th
 ##### Download the sample dataset
 
 ```bash
-
 wget https://autoware-files.s3.us-west-2.amazonaws.com/dataset/lidar_detection_sample_dataset.tar.gz
 #Extract the dataset to a folder of your choice
 tar -xvf lidar_detection_sample_dataset.tar.gz
@@ -200,10 +200,9 @@ ln -s /PATH/TO/DATASET/ /PATH/TO/mmdetection3d/data/tier4_dataset/
 
 ##### Prepare dataset and evaluate trained model
 
-Create .pkl files for training, evaluation, and testing.
+Create `.pkl` files for training, evaluation, and testing.
 
 ```bash
-
 python tools/create_data.py T4Dataset --root-path data/sample_dataset/ --out-dir data/sample_dataset/ --extra-tag T4Dataset --version sample_dataset --annotation-hz 2
 ```
 

From b3b609aee4fb9f258732d3e88f9cefa4d11925e3 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Kaan=20=C3=87olak?= <kaancolak95@gmail.com>
Date: Fri, 24 May 2024 11:46:14 +0300
Subject: [PATCH 10/12] docs(lidar_centerpoint): add version

---
 perception/lidar_centerpoint/README.md | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index 5435b14c54034..ee212f59638da 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -204,6 +204,8 @@ ln -s /PATH/TO/DATASET/ /PATH/TO/mmdetection3d/data/tier4_dataset/
 
 Create `.pkl` files for training, evaluation, and testing.
 
+The dataset was formatted according to T4Dataset specifications, with 'sample_dataset' designated as one of its versions.
+
 ```bash
 python tools/create_data.py T4Dataset --root-path data/sample_dataset/ --out-dir data/sample_dataset/ --extra-tag T4Dataset --version sample_dataset --annotation-hz 2
 ```

From e3f719491841675770d70742ffba9ecb63a15cb8 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Kaan=20=C3=87olak?= <kaancolak95@gmail.com>
Date: Tue, 4 Jun 2024 11:07:26 +0300
Subject: [PATCH 11/12] Update README.md

docs(lidar_centerpoint): add thanks
---
 perception/lidar_centerpoint/README.md | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index ee212f59638da..39893d05cf0ea 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -340,6 +340,10 @@ Example:
 [v1-encoder-centerpoint-tiny]: https://awf.ml.dev.web.auto/perception/models/centerpoint/v1/pts_voxel_encoder_centerpoint_tiny.onnx
 [v1-head-centerpoint-tiny]: https://awf.ml.dev.web.auto/perception/models/centerpoint/v1/pts_backbone_neck_head_centerpoint_tiny.onnx
 
+## Acknowledgment: deepen.ai's 3D Annotation Tools Contribution
+
+Special thanks to deepen.ai for providing their 3D Annotation tools, which have been instrumental in creating our sample dataset.
+
 ## Legal Notice
 
 _The nuScenes dataset is released publicly for non-commercial use under the Creative

From 26f028ca04f246b3f58a5c105de5adba67c30e30 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Kaan=20=C3=87olak?= <kaancolak95@gmail.com>
Date: Tue, 4 Jun 2024 11:10:09 +0300
Subject: [PATCH 12/12] docs(lidar_centerpoint): add link

---
 perception/lidar_centerpoint/README.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/perception/lidar_centerpoint/README.md b/perception/lidar_centerpoint/README.md
index 39893d05cf0ea..ff672c79bd253 100644
--- a/perception/lidar_centerpoint/README.md
+++ b/perception/lidar_centerpoint/README.md
@@ -342,7 +342,7 @@ Example:
 
 ## Acknowledgment: deepen.ai's 3D Annotation Tools Contribution
 
-Special thanks to deepen.ai for providing their 3D Annotation tools, which have been instrumental in creating our sample dataset.
+Special thanks to [Deepen AI](https://www.deepen.ai/) for providing their 3D Annotation tools, which have been instrumental in creating our sample dataset.
 
 ## Legal Notice