search

lyuwenyu / RT-DETR

[CVPR 2024] Official RT-DETR (RTDETR paddle pytorch), Real-Time DEtection TRansformer, DETRs Beat YOLOs on Real-time Object Detection. 🔥 🔥 🔥
Apache License 2.0
2.61k stars 303 forks source link

求助:ppdet导出的模型转ONNX之后,预测报错 #428

Closed David-dotcom666 closed 2 months ago

David-dotcom666 commented 2 months ago

Star RTDETR 请先在RTDETR主页点击star以支持本项目 Star RTDETR to help more people discover this project. 环境: paddle2.5.2 ppdet_dev和2.7都试过 onnx==1.13.0 paddle2onnx==1.0.5

Describe the bug 报错信息 onnxruntime.capi.onnxruntime_pybind11_state.InvalidArgument: [ONNXRuntimeError] : 2 : INVALID_ARGUMENT : Non-zero status code returned while running Tile node. Name:'p2o.Tile.3' Status Message: the tensor to be tiled using Tile OP must be atleast 1 dimensional 4

To Reproduce Steps to reproduce the behavior.

lyuwenyu commented 2 months ago

这 自己训练的模型？用torch的也试一下

David-dotcom666 commented 2 months ago

试了下直接下载预训练模型然后按照流程导出再转成onnx，也是报错，下面这个是因为我把onnxruntime换成了1.17的. 有没有转成onnx的模型文件啊，我想看看跟我转出来的有啥区别不 9

David-dotcom666 commented 2 months ago

onnxok paddle换成2.4.2后可以正常,但是不知道为什么2.4.2infer原生的paddle模型速度比2.5.2/paddle3.0b的慢了100ms(100ms->200ms)