Is it possible to see the benchmarks on jetson devices?(my goal is nano)

[aiLover2]

No description provided.

[Broad-sky]

This project only provides an example. If you want to use it on Nano, you can quantify the model to see the effect.

[Wulingtian]

This project only provides an example. If you want to use it on Nano, you can quantify the model to see thQINGWEN

This project only provides an example. If you want to use it on Nano, you can quantify the model to see the effect.
请问如何生成onnx模型，我这边基于自己自己的数据训练了一个模型，成功生成onnx，onnx转换为trt模型报错了，如下：
WARNING: Skipping tactic 0 due to Myelin error: myelinTargetSetPropertyMemorySize called with invalid memory size (0).
ERROR: 10: [optimizer.cpp::nvinfer1::builder::`anonymous-namespace'::LeafCNode::computeCosts::2011] Error Code 10: Internal Error (Could not find any implementation for node {ForeignNode[Slice_362...Concat_599]}.)

[Broad-sky]

This project only provides an example. If you want to use it on Nano, you can quantify the model to see thQINGWEN

This project only provides an example. If you want to use it on Nano, you can quantify the model to see the effect.
请问如何生成onnx模型，我这边基于自己自己的数据训练了一个模型，成功生成onnx，onnx转换为trt模型报错了，如下：
WARNING: Skipping tactic 0 due to Myelin error: myelinTargetSetPropertyMemorySize called with invalid memory size (0).
ERROR: 10: [optimizer.cpp::nvinfer1::builder::`anonymous-namespace'::LeafCNode::computeCosts::2011] Error Code 10: Internal Error (Could not find any implementation for node {ForeignNode[Slice_362...Concat_599]}.)

对于自己训练的模型，将网络的输出作为onnx的输出，此时导出onnx。
如果使用官方模型，直接用readme文档提供的模型即可，不复杂！

[Wulingtian]

This project only provides an example. If you want to use it on Nano, you can quantify the model to see thQINGWEN

This project only provides an example. If you want to use it on Nano, you can quantify the model to see the effect.
请问如何生成onnx模型，我这边基于自己自己的数据训练了一个模型，成功生成onnx，onnx转换为trt模型报错了，如下：
WARNING: Skipping tactic 0 due to Myelin error: myelinTargetSetPropertyMemorySize called with invalid memory size (0).
ERROR: 10: [optimizer.cpp::nvinfer1::builder::`anonymous-namespace'::LeafCNode::computeCosts::2011] Error Code 10: Internal Error (Could not find any implementation for node {ForeignNode[Slice_362...Concat_599]}.)

对于自己训练的模型，将网络的输出作为onnx的输出，此时导出onnx。 如果使用官方模型，直接用readme文档提供的模型即可，不复杂！

现在r50模型可以正常推理了，但是我试了下r18 backbone的模型，发现推理时间比r50 还慢，只是把segmentation_trt.cpp里面的256替换为128