Jetson nano yolov8部署 onnx ->> trt

[icesheeters]

Jetpack:4.6
CUDA:10.2.3
cuDNN:8.2.1.32
TensorRt: 8.2.3.0
python:3.8

执行：python export.py -o robot.onnx -e jetson.trt --end2end --v8 -v -w 4
[03/19/2024-16:10:14] [TRT] [V] --------------- Timing Runner: (Unnamed Layer* 302) [PluginV2DynamicExt] (PluginV2)
[03/19/2024-16:10:14] [TRT] [V] Deleting timing cache: 1439 entries, 2433 hits
[03/19/2024-16:10:14] [TRT] [I] [MemUsageChange] Init cuBLAS/cuBLASLt: CPU +0, GPU +0, now: CPU 908, GPU 2849 (MiB)
[03/19/2024-16:10:14] [TRT] [E] 2: [pluginV2DynamicExtRunner.cpp::execute::115] Error Code 2: Internal Error (Assertion status == kSTATUS_SUCCESS failed.)
[03/19/2024-16:10:14] [TRT] [E] 2: [builder.cpp::buildSerializedNetwork::417] Error Code 2: Internal Error (Assertion enginePtr != nullptr failed.)
Traceback (most recent call last):
File "export.py", line 316, in
main(args)
File "export.py", line 276, in main
builder.create_engine(args.engine, args.precision, args.calib_input, args.calib_cache, args.calib_num_images,
File "export.py", line 269, in create_engine
with self.builder.build_serialized_network(self.network, self.config) as engine, open(engine_path, "wb") as f:
AttributeError: enter

大佬，这个是因为cuda或者TensorRt的版本太低了导致的吗？

[icesheeters]

解决了，不用端到端优化就好了

[Linaom1214]

@icesheeters EFNMS plugin 在8.4才引入