EfficientNMS_TRT plugin compiled by myself can not work

[pango99]

Description

For some reaon, I need to compile the EfficientNMS_TRT plugin by myself, I clone the code from Github, and add the files in "efficientNMSPlugin" and "common" folders to my vs project, compile it success, then I add the plugin I compiled to network, but build engine failed, and tensorRT report below infos:

00000018 4.94450378 [21768] TrtZLog-testYOLOv8(INFO): start to build trt engine...
00000019 6.44722557 [21768] TRT-testYOLOv8(INFO): [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +934, GPU +316, now: CPU 10810, GPU 1662 (MiB)
00000020 6.45917130 [21768] TRT-testYOLOv8(INFO): [MemUsageChange] Init cuDNN: CPU +3, GPU +8, now: CPU 10813, GPU 1670 (MiB)
00000021 6.45966244 [21768] TRT-testYOLOv8(INFO): Local timing cache in use. Profiling results in this builder pass will not be stored.
00000098 389.70767212 [21768] TRT-testYOLOv8(INFO): Some tactics do not have sufficient workspace memory to run. Increasing workspace size will enable more tactics, please check verbose output for requested sizes.
00000099 619.68878174 [21768] TRT-testYOLOv8(WARNING): Skipping tactic 0x0000000000000000 due to exception Assertion status == kSTATUS_SUCCESS failed.
00000100 619.68914795 [21768] TRT-testYOLOv8(WARNING): Skipping tactic 0x0000000000000000 due to exception Assertion status == kSTATUS_SUCCESS failed.
00000101 619.73187256 [21768] TRT-testYOLOv8(ERROR): 4: [optimizer.cpp::nvinfer1::builder::cgraph::LeafCNode::computeCosts::3726] Error Code 4: Internal Error (Could not find any implementation for node (Unnamed Layer 707) [PluginV2DynamicExt] due to insufficient workspace. See verbose log for requested sizes.)
00000102 620.01159668 [21768] TRT-testYOLOv8(ERROR): 2: [builder.cpp::nvinfer1::builder::Builder::buildSerializedNetwork::751] Error Code 2: Internal Error (Assertion engine != nullptr failed. )*

If I use the EfficientNMS plugin shipped with trt sdk, there is no error encounted during engine build, so what's wrong with the plugin I compiled?

Environment

TensorRT Version: 8.5.3.1

NVIDIA GPU: RTX 4070 Super

NVIDIA Driver Version: 551.61

CUDA Version: 11.6.0

CUDNN Version: 8.6.0.163

Operating System: Windows 10 Pro

Python Version (if applicable):

Tensorflow Version (if applicable):

PyTorch Version (if applicable):

Baremetal or Container (if so, version):

[lix19937]

Internal Error (Could not find any implementation for node (Unnamed Layer 707) [PluginV2DynamicExt] due to insufficient workspace. See verbose log for requested sizes.)

If insufficient workspace is provided, it is possible that TensorRT will not be able to find an implementation for a layer. By default, the workspace is set to the total global memory size of the given device. TRy to increase the workspace size as suggested.