fix: ml detector buffer capacity and parameterization (autowarefounda…

…tion#7936) * fix: increased the buffer capacity in ML-based detectors (1M -> 2M due to new models) but parameterized to users can adapt it to their use cases Signed-off-by: Kenzo Lobos-Tsunekawa <[email protected]> * chore: added schema parts Signed-off-by: Kenzo Lobos-Tsunekawa <[email protected]> * chore: points are countable ! Signed-off-by: Kenzo Lobos-Tsunekawa <[email protected]> --------- Signed-off-by: Kenzo Lobos-Tsunekawa <[email protected]> Signed-off-by: chtseng <[email protected]>
chihungtzeng · Jul 23, 2024 · 35b0c47 · 35b0c47
1 parent 089dbb3
commit 35b0c47
Show file tree

Hide file tree

Showing 15 changed files with 48 additions and 25 deletions.
diff --git a/perception/image_projection_based_fusion/config/pointpainting.param.yaml b/perception/image_projection_based_fusion/config/pointpainting.param.yaml
@@ -5,6 +5,7 @@
     head_onnx_path: "$(var model_path)/pts_backbone_neck_head_$(var model_name).onnx"
     head_engine_path: "$(var model_path)/pts_backbone_neck_head_$(var model_name).engine"
     trt_precision: fp16
+    cloud_capacity: 2000000
     post_process_params:
       # post-process params
       circle_nms_dist_threshold: 0.3

diff --git a/...d_fusion/include/image_projection_based_fusion/pointpainting_fusion/pointpainting_trt.hpp b/...d_fusion/include/image_projection_based_fusion/pointpainting_fusion/pointpainting_trt.hpp
@@ -23,7 +23,6 @@
 
 namespace image_projection_based_fusion
 {
-static constexpr size_t CAPACITY_POINT = 1000000;
 class PointPaintingTRT : public centerpoint::CenterPointTRT
 {
 public:

diff --git a/perception/image_projection_based_fusion/src/pointpainting_fusion/node.cpp b/perception/image_projection_based_fusion/src/pointpainting_fusion/node.cpp
@@ -103,6 +103,7 @@ PointPaintingFusionNode::PointPaintingFusionNode(const rclcpp::NodeOptions & opt
     this->declare_parameter<int>("densification_params.num_past_frames");
   // network param
   const std::string trt_precision = this->declare_parameter<std::string>("trt_precision");
+  const std::size_t cloud_capacity = this->declare_parameter<std::int64_t>("cloud_capacity");
   const std::string encoder_onnx_path = this->declare_parameter<std::string>("encoder_onnx_path");
   const std::string encoder_engine_path =
     this->declare_parameter<std::string>("encoder_engine_path");
@@ -181,9 +182,9 @@ PointPaintingFusionNode::PointPaintingFusionNode(const rclcpp::NodeOptions & opt
   centerpoint::DensificationParam densification_param(
     densification_world_frame_id, densification_num_past_frames);
   centerpoint::CenterPointConfig config(
-    class_names_.size(), point_feature_size, max_voxel_size, pointcloud_range, voxel_size,
-    downsample_factor, encoder_in_feature_size, score_threshold, circle_nms_dist_threshold,
-    yaw_norm_thresholds, has_variance_);
+    class_names_.size(), point_feature_size, cloud_capacity, max_voxel_size, pointcloud_range,
+    voxel_size, downsample_factor, encoder_in_feature_size, score_threshold,
+    circle_nms_dist_threshold, yaw_norm_thresholds, has_variance_);
 
   // create detector
   detector_ptr_ = std::make_unique<image_projection_based_fusion::PointPaintingTRT>(

diff --git a/perception/lidar_centerpoint/config/centerpoint.param.yaml b/perception/lidar_centerpoint/config/centerpoint.param.yaml
@@ -6,6 +6,7 @@
     head_onnx_path: "$(var model_path)/pts_backbone_neck_head_$(var model_name).onnx"
     head_engine_path: "$(var model_path)/pts_backbone_neck_head_$(var model_name).engine"
     trt_precision: fp16
+    cloud_capacity: 2000000
     post_process_params:
       # post-process params
       circle_nms_dist_threshold: 0.5

diff --git a/perception/lidar_centerpoint/config/centerpoint_tiny.param.yaml b/perception/lidar_centerpoint/config/centerpoint_tiny.param.yaml
@@ -6,6 +6,7 @@
     head_onnx_path: "$(var model_path)/pts_backbone_neck_head_$(var model_name).onnx"
     head_engine_path: "$(var model_path)/pts_backbone_neck_head_$(var model_name).engine"
     trt_precision: fp16
+    cloud_capacity: 2000000
     post_process_params:
       # post-process params
       circle_nms_dist_threshold: 0.5

diff --git a/perception/lidar_centerpoint/include/lidar_centerpoint/centerpoint_config.hpp b/perception/lidar_centerpoint/include/lidar_centerpoint/centerpoint_config.hpp
@@ -24,14 +24,16 @@ class CenterPointConfig
 {
 public:
   explicit CenterPointConfig(
-    const std::size_t class_size, const float point_feature_size, const std::size_t max_voxel_size,
-    const std::vector<double> & point_cloud_range, const std::vector<double> & voxel_size,
-    const std::size_t downsample_factor, const std::size_t encoder_in_feature_size,
-    const float score_threshold, const float circle_nms_dist_threshold,
-    const std::vector<double> yaw_norm_thresholds, const bool has_variance)
+    const std::size_t class_size, const float point_feature_size, const std::size_t cloud_capacity,
+    const std::size_t max_voxel_size, const std::vector<double> & point_cloud_range,
+    const std::vector<double> & voxel_size, const std::size_t downsample_factor,
+    const std::size_t encoder_in_feature_size, const float score_threshold,
+    const float circle_nms_dist_threshold, const std::vector<double> yaw_norm_thresholds,
+    const bool has_variance)
   {
     class_size_ = class_size;
     point_feature_size_ = point_feature_size;
+    cloud_capacity_ = cloud_capacity;
     max_voxel_size_ = max_voxel_size;
     if (point_cloud_range.size() == 6) {
       range_min_x_ = static_cast<float>(point_cloud_range[0]);
@@ -85,6 +87,7 @@ class CenterPointConfig
   };
 
   // input params
+  std::size_t cloud_capacity_{};
   std::size_t class_size_{3};
   const std::size_t point_dim_size_{3};  // x, y and z
   std::size_t point_feature_size_{4};    // x, y, z and time-lag

diff --git a/perception/lidar_centerpoint/include/lidar_centerpoint/centerpoint_trt.hpp b/perception/lidar_centerpoint/include/lidar_centerpoint/centerpoint_trt.hpp
@@ -31,8 +31,6 @@
 
 namespace centerpoint
 {
-static constexpr size_t CAPACITY_POINT = 1000000;
-
 class NetworkParam
 {
 public:

diff --git a/perception/lidar_centerpoint/lib/centerpoint_trt.cpp b/perception/lidar_centerpoint/lib/centerpoint_trt.cpp
@@ -85,7 +85,7 @@ void CenterPointTRT::initPtr()
   mask_size_ = config_.grid_size_x_ * config_.grid_size_y_;
 
   // host
-  points_.resize(CAPACITY_POINT * config_.point_feature_size_);
+  points_.resize(config_.cloud_capacity_ * config_.point_feature_size_);
 
   // device
   voxels_d_ = cuda::make_unique<float[]>(voxels_size_);
@@ -100,7 +100,7 @@ void CenterPointTRT::initPtr()
   head_out_dim_d_ = cuda::make_unique<float[]>(grid_xy_size * config_.head_out_dim_size_);
   head_out_rot_d_ = cuda::make_unique<float[]>(grid_xy_size * config_.head_out_rot_size_);
   head_out_vel_d_ = cuda::make_unique<float[]>(grid_xy_size * config_.head_out_vel_size_);
-  points_d_ = cuda::make_unique<float[]>(CAPACITY_POINT * config_.point_feature_size_);
+  points_d_ = cuda::make_unique<float[]>(config_.cloud_capacity_ * config_.point_feature_size_);
   voxels_buffer_d_ = cuda::make_unique<float[]>(voxels_buffer_size_);
   mask_d_ = cuda::make_unique<unsigned int[]>(mask_size_);
   num_voxels_d_ = cuda::make_unique<unsigned int[]>(1);

diff --git a/perception/lidar_centerpoint/lib/preprocess/voxel_generator.cpp b/perception/lidar_centerpoint/lib/preprocess/voxel_generator.cpp
@@ -61,7 +61,7 @@ std::size_t VoxelGenerator::generateSweepPoints(float * points_d, cudaStream_t s
     float time_lag = static_cast<float>(
       pd_ptr_->getCurrentTimestamp() - rclcpp::Time(pc_cache_iter->header.stamp).seconds());
 
-    if (point_counter + sweep_num_points > CAPACITY_POINT) {
+    if (point_counter + sweep_num_points > config_.cloud_capacity_) {
       RCLCPP_WARN_STREAM(
         rclcpp::get_logger("lidar_centerpoint"),
         "Requested number of points exceeds the maximum capacity. Current points = "

diff --git a/perception/lidar_centerpoint/schema/centerpoint.schemal.json b/perception/lidar_centerpoint/schema/centerpoint.schemal.json
@@ -12,6 +12,12 @@
           "default": "fp16",
           "enum": ["fp32", "fp16"]
         },
+        "cloud_capacity": {
+          "type": "integer",
+          "description": "Capacity of the point cloud buffer (should be set to at least the maximum theoretical number of points).",
+          "default": 2000000,
+          "minimum": 1
+        },
         "post_process_params": {
           "type": "object",
           "properties": {

diff --git a/perception/lidar_centerpoint/src/node.cpp b/perception/lidar_centerpoint/src/node.cpp
@@ -50,6 +50,7 @@ LidarCenterPointNode::LidarCenterPointNode(const rclcpp::NodeOptions & node_opti
   const int densification_num_past_frames =
     this->declare_parameter<int>("densification_params.num_past_frames");
   const std::string trt_precision = this->declare_parameter<std::string>("trt_precision");
+  const std::size_t cloud_capacity = this->declare_parameter<std::int64_t>("cloud_capacity");
   const std::string encoder_onnx_path = this->declare_parameter<std::string>("encoder_onnx_path");
   const std::string encoder_engine_path =
     this->declare_parameter<std::string>("encoder_engine_path");
@@ -104,9 +105,9 @@ LidarCenterPointNode::LidarCenterPointNode(const rclcpp::NodeOptions & node_opti
       "The size of voxel_size != 3: use the default parameters.");
   }
   CenterPointConfig config(
-    class_names_.size(), point_feature_size, max_voxel_size, point_cloud_range, voxel_size,
-    downsample_factor, encoder_in_feature_size, score_threshold, circle_nms_dist_threshold,
-    yaw_norm_thresholds, has_variance_);
+    class_names_.size(), point_feature_size, cloud_capacity, max_voxel_size, point_cloud_range,
+    voxel_size, downsample_factor, encoder_in_feature_size, score_threshold,
+    circle_nms_dist_threshold, yaw_norm_thresholds, has_variance_);
   detector_ptr_ =
     std::make_unique<CenterPointTRT>(encoder_param, head_param, densification_param, config);
 

diff --git a/perception/lidar_transfusion/config/transfusion.param.yaml b/perception/lidar_transfusion/config/transfusion.param.yaml
@@ -3,6 +3,7 @@
     # network
     class_names: ["CAR", "TRUCK", "BUS", "BICYCLE", "PEDESTRIAN"]
     trt_precision: fp16
+    cloud_capacity: 2000000
     voxels_num: [5000, 30000, 60000]  # [min, opt, max]
     point_cloud_range: [-92.16, -92.16, -3.0, 92.16, 92.16, 7.0]  # [x_min, y_min, z_min, x_max, y_max, z_max]
     voxel_size: [0.24, 0.24, 10.0]  # [x, y, z]

diff --git a/perception/lidar_transfusion/include/lidar_transfusion/transfusion_config.hpp b/perception/lidar_transfusion/include/lidar_transfusion/transfusion_config.hpp
@@ -25,11 +25,13 @@ class TransfusionConfig
 {
 public:
   TransfusionConfig(
-    const std::vector<int64_t> & voxels_num, const std::vector<double> & point_cloud_range,
-    const std::vector<double> & voxel_size, const int num_proposals,
-    const float circle_nms_dist_threshold, const std::vector<double> & yaw_norm_thresholds,
-    const float score_threshold)
+    const std::size_t cloud_capacity, const std::vector<int64_t> & voxels_num,
+    const std::vector<double> & point_cloud_range, const std::vector<double> & voxel_size,
+    const std::size_t num_proposals, const float circle_nms_dist_threshold,
+    const std::vector<double> & yaw_norm_thresholds, const float score_threshold)
   {
+    cloud_capacity_ = cloud_capacity;
+
     if (voxels_num.size() == 3) {
       max_voxels_ = voxels_num[2];
 
@@ -86,7 +88,7 @@ class TransfusionConfig
   }
 
   ///// INPUT PARAMETERS /////
-  const std::size_t cloud_capacity_{1000000};
+  std::size_t cloud_capacity_{};
   ///// KERNEL PARAMETERS /////
   const std::size_t threads_for_voxel_{256};  // threads number for a block
   const std::size_t points_per_voxel_{20};

diff --git a/perception/lidar_transfusion/schema/transfusion.schema.json b/perception/lidar_transfusion/schema/transfusion.schema.json
@@ -21,6 +21,12 @@
           "default": "fp16",
           "enum": ["fp16", "fp32"]
         },
+        "cloud_capacity": {
+          "type": "integer",
+          "description": "Capacity of the point cloud buffer (should be set to at least the maximum theoretical number of points).",
+          "default": 2000000,
+          "minimum": 1
+        },
         "voxels_num": {
           "type": "array",
           "items": {

diff --git a/perception/lidar_transfusion/src/lidar_transfusion_node.cpp b/perception/lidar_transfusion/src/lidar_transfusion_node.cpp
@@ -27,11 +27,14 @@ LidarTransfusionNode::LidarTransfusionNode(const rclcpp::NodeOptions & options)
   class_names_ = this->declare_parameter<std::vector<std::string>>("class_names", descriptor);
   const std::string trt_precision =
     this->declare_parameter<std::string>("trt_precision", descriptor);
+  const std::size_t cloud_capacity =
+    this->declare_parameter<std::int64_t>("cloud_capacity", descriptor);
   const auto voxels_num = this->declare_parameter<std::vector<int64_t>>("voxels_num", descriptor);
   const auto point_cloud_range =
     this->declare_parameter<std::vector<double>>("point_cloud_range", descriptor);
   const auto voxel_size = this->declare_parameter<std::vector<double>>("voxel_size", descriptor);
-  const int num_proposals = (this->declare_parameter<int>("num_proposals", descriptor));
+  const std::size_t num_proposals =
+    this->declare_parameter<std::int64_t>("num_proposals", descriptor);
   const std::string onnx_path = this->declare_parameter<std::string>("onnx_path", descriptor);
   const std::string engine_path = this->declare_parameter<std::string>("engine_path", descriptor);
 
@@ -74,8 +77,8 @@ LidarTransfusionNode::LidarTransfusionNode(const rclcpp::NodeOptions & options)
   DensificationParam densification_param(
     densification_world_frame_id, densification_num_past_frames);
   TransfusionConfig config(
-    voxels_num, point_cloud_range, voxel_size, num_proposals, circle_nms_dist_threshold,
-    yaw_norm_thresholds, score_threshold);
+    cloud_capacity, voxels_num, point_cloud_range, voxel_size, num_proposals,
+    circle_nms_dist_threshold, yaw_norm_thresholds, score_threshold);
 
   const auto allow_remapping_by_area_matrix =
     this->declare_parameter<std::vector<int64_t>>("allow_remapping_by_area_matrix", descriptor);