[Paddle2ONNX] Oops, there are some operators not supported yet, including uniform_random

[MaciejPykoszAigo]

Please fill in the information below so that we can solve the problem quickly, Thanks !

Describe the bug
I tried to export model to onnx, and the terminal shows "there are some operator not supported yet..."

Before that, I re-trained the SegNeXt model (https://github.com/PaddlePaddle/PaddleSeg/tree/release/2.8/configs/segnext), then converted the .pdparams as instructed in https://github.com/PaddlePaddle/PaddleSeg/blob/release/2.8/docs/model_export.md, and then tried to convert to onnx as instructed in https://github.com/PaddlePaddle/Paddle2ONNX/blob/develop/README_en.md

Informations (please complete the following information):

• Inference engine for deployment: onnx runtime/tensorRT
• Why convert to onnx：Performance
• Paddle2ONNX Version: 1.3.1
• Email/Wechat/Phone:

Screenshots

Additional context

• Paddle2ONNX: 1.3.1
• PaddlePaddle: 2.6.0
• PaddleSeg: 2.8.0
• Python: 3.10.13

[Zheng-Bicheng]

这个算子暂时 ONNX 中没有对应的算子

[MaciejPykoszAigo]

Thanks for your answer.
In ONNX documentation I found RandomUniform https://onnx.ai/onnx/operators/onnx__RandomUniform.html. Probably this is what we need. Is it possible to implement it in Paddle2ONNX, the same as it was resolved here: #1299 or #1088.
Could you add it?

[MaciejPykoszAigo]

Hi, @Zheng-Bicheng
Do you think this is doable?

[4399123]

我也遇到了这个问题，模型训练好了，没办法部署，这个算法存在有什么意义，paddleseg中一堆没办法转化的算法，pidnet，rtformer等，都没人去解决，一个自适应平均池化通过普通的池化转写都不会修改。

[MaciejPykoszAigo]

@Zheng-Bicheng Could I have an answer please?

[Zheng-Bicheng]

@MaciejPykoszAigo There is currently no such operator in ONNX, so we cannot map this operator from Paddle model to ONNX.

[Zheng-Bicheng]

You can see the list of ONNX operators here:

• https://onnx.ai/onnx/operators/

[MaciejPykoszAigo]

@Zheng-Bicheng I found RandomUniform in your list (https://onnx.ai/onnx/operators/onnx__RandomUniform.html#l-onnx-doc-randomuniform). Do you think it's the same operator as uniform_random in the error in the screenshot above? Could you please verify this?