[Help]: Error Code: 9: Skipping tactic 0x0000000000000000 due to exception Assertion pluginUtils::isSuccess(status) failed. #172
Description
问题确认 Search before asking
Bug组件 Bug Component
Export
Bug描述 Describe the Bug
使用trtyolo export完成的onnx文件在trtexec的编译过程中报错。这个issue的平台和我一样，他的命令是错误的，我的命令是正确的，但是error相同，不过这个issue没有后续了。
command:
/usr/src/tensorrt/bin/trtexec --onnx=../yolo.onnx --saveEngine=./yolo11mpose.engine \
--fp16 --staticPlugins=/home/jetson/workspace/TensorRT-YOLO/lib/plugin/libcustom_plugins.so \
--setPluginsToSerialize=/home/jetson/workspace/TensorRT-YOLO/lib/plugin/libcustom_plugins.so \
--verbose --memPoolSize=workspace:8000 2>&1 | tee trtexec.log
Key error message:
[06/12/2025-18:11:21] [V] [TRT] =============== Computing costs for /model.23/EfficientIdxNMS_TRT
[06/12/2025-18:11:21] [V] [TRT] *************** Autotuning format combination: Float(57960,4,1), Float(43470,3,1) -> Int32(1,1), Float(400,4,1), Float(100,1), Int32(100,1), Int32(100,1) ***************
[06/12/2025-18:11:21] [V] [TRT] --------------- Timing Runner: /model.23/EfficientIdxNMS_TRT (PluginV2[0x80000019])
[06/12/2025-18:11:21] [E] Error[9]: Error Code: 9: Skipping tactic 0x0000000000000000 due to exception Assertion pluginUtils::isSuccess(status) failed.
[06/12/2025-18:11:21] [V] [TRT] /model.23/EfficientIdxNMS_TRT (PluginV2[0x80000019]) profiling completed in 0.0595078 seconds. Fastest Tactic: 0xd15ea5edd15ea5ed Time: inf
[06/12/2025-18:11:21] [V] [TRT] *************** Autotuning format combination: Half(57960,4,1), Half(43470,3,1) -> Int32(1,1), Half(400,4,1), Half(100,1), Int32(100,1), Int32(100,1) ***************
[06/12/2025-18:11:21] [V] [TRT] --------------- Timing Runner: /model.23/EfficientIdxNMS_TRT (PluginV2[0x80000019])
[06/12/2025-18:11:21] [E] Error[9]: Error Code: 9: Skipping tactic 0x0000000000000000 due to exception Assertion pluginUtils::isSuccess(status) failed.
[06/12/2025-18:11:21] [V] [TRT] /model.23/EfficientIdxNMS_TRT (PluginV2[0x80000019]) profiling completed in 0.00345909 seconds. Fastest Tactic: 0xd15ea5edd15ea5ed Time: inf
[06/12/2025-18:11:21] [E] Error[10]: IBuilder::buildSerializedNetwork: Error Code 10: Internal Error (Could not find any implementation for node /model.23/EfficientIdxNMS_TRT.)
[06/12/2025-18:11:21] [E] Engine could not be created from network
[06/12/2025-18:11:21] [E] Building engine failed
[06/12/2025-18:11:21] [E] Failed to create engine from model or file.
[06/12/2025-18:11:21] [E] Engine set up failed
&&&& FAILED TensorRT.trtexec [TensorRT v100300] # /usr/src/tensorrt/bin/trtexec --onnx=../yolo.onnx --saveEngine=./yolo11mpose.engine --fp16 --staticPlugins=/home/jetson/workspace/TensorRT-YOLO/lib/plugin/libcustom_plugins.so --setPluginsToSerialize=/home/jetson/workspace/TensorRT-YOLO/lib/plugin/libcustom_plugins.so --verbose --memPoolSize=workspace:8000
Important log message:
[06/12/2025-17:56:34] [I] === System Options ===
[06/12/2025-17:56:34] [I] Device: 0
[06/12/2025-17:56:34] [I] DLACore:
[06/12/2025-17:56:34] [I] Plugins: /home/jetson/workspace/TensorRT-YOLO/lib/plugin/libcustom_plugins.so
[06/12/2025-17:56:34] [I] setPluginsToSerialize: /home/jetson/workspace/TensorRT-YOLO/lib/plugin/libcustom_plugins.so
[06/12/2025-17:56:34] [I] dynamicPlugins:
[06/12/2025-17:56:34] [I] ignoreParsedPluginLibs: 0
[06/12/2025-17:56:34] [I]
可以看到我已经正确配置了许多issue中遇到的未配置staticPlugins和setPluginsToSerialize参数， 而且我确定我的plugins.so文件有效。但是问题仍然出现。
jetson@ubuntu:~/workspace/models/yolo_trt$ ldd /home/jetson/workspace/TensorRT-YOLO/lib/plugin/libcustom_plugins.so
linux-vdso.so.1 (0x0000ffffb00fe000)
libcudart.so.12 => /usr/local/cuda-12.6/lib64/libcudart.so.12 (0x0000ffffafd60000)
libnvinfer.so.10 => /lib/aarch64-linux-gnu/libnvinfer.so.10 (0x0000ffffa1ab0000)
libstdc++.so.6 => /lib/aarch64-linux-gnu/libstdc++.so.6 (0x0000ffffa1880000)
libm.so.6 => /lib/aarch64-linux-gnu/libm.so.6 (0x0000ffffa17e0000)
libgcc_s.so.1 => /lib/aarch64-linux-gnu/libgcc_s.so.1 (0x0000ffffa17b0000)
libc.so.6 => /lib/aarch64-linux-gnu/libc.so.6 (0x0000ffffa1600000)
/lib/ld-linux-aarch64.so.1 (0x0000ffffb00c5000)
libdl.so.2 => /lib/aarch64-linux-gnu/libdl.so.2 (0x0000ffffa15e0000)
libpthread.so.0 => /lib/aarch64-linux-gnu/libpthread.so.0 (0x0000ffffa15c0000)
librt.so.1 => /lib/aarch64-linux-gnu/librt.so.1 (0x0000ffffa15a0000)
libnvdla_compiler.so => /usr/lib/aarch64-linux-gnu/nvidia/libnvdla_compiler.so (0x0000ffffa0db0000)
libcudla.so.1 => /usr/local/cuda-12.6/lib64/libcudla.so.1 (0x0000ffffa0d80000)
libnvos.so => /usr/lib/aarch64-linux-gnu/nvidia/libnvos.so (0x0000ffffa0d50000)
libnvcudla.so => /lib/aarch64-linux-gnu/libnvcudla.so (0x0000ffffa0d20000)
libcuda.so => /lib/aarch64-linux-gnu/libcuda.so (0x0000ffff9e4b0000)
libnvdla_runtime.so => /usr/lib/aarch64-linux-gnu/nvidia/libnvdla_runtime.so (0x0000ffff9de50000)
libnvrm_gpu.so => /usr/lib/aarch64-linux-gnu/nvidia/libnvrm_gpu.so (0x0000ffff9ddd0000)
libnvrm_mem.so => /usr/lib/aarch64-linux-gnu/nvidia/libnvrm_mem.so (0x0000ffff9ddb0000)
libnvrm_host1x.so => /usr/lib/aarch64-linux-gnu/nvidia/libnvrm_host1x.so (0x0000ffff9dd80000)
libnvsocsys.so => /usr/lib/aarch64-linux-gnu/nvidia/libnvsocsys.so (0x0000ffff9dd60000)
libnvtegrahv.so => /usr/lib/aarch64-linux-gnu/nvidia/libnvtegrahv.so (0x0000ffff9dd40000)
libnvrm_sync.so => /usr/lib/aarch64-linux-gnu/nvidia/libnvrm_sync.so (0x0000ffff9dd20000)
libnvsciipc.so => /usr/lib/aarch64-linux-gnu/nvidia/libnvsciipc.so (0x0000ffff9dce0000)
libnvrm_chip.so => /usr/lib/aarch64-linux-gnu/nvidia/libnvrm_chip.so (0x0000ffff9dcc0000)
Full log file is attached
复现环境 Environment
- OS: Jetpack 6.1
- Platform: Jetson Orin NX 16G
- TensorRT: 10.3
Bug描述确认 Bug description confirmation
- 我确认已经提供了Bug复现步骤、代码改动说明、以及环境信息，确认问题是可以复现的。I confirm that the bug replication steps, code change instructions, and environment information have been provided, and the problem can be reproduced.