代码问题咨询

[shu616048151]

代码问题咨询

[helion-du-mas-des-bourboux-thales]

Google clasifies your text as Chinese, language I do not speek, however, here is the google traduction:

By pulling the code, you can find the algorithm proposed by you. There is no outstanding performance on the RML2016 data set. But it performs well on the AugMod dataset.
Question 1: I want to ask that this algorithm only performs well on AugMod?
Question 2: Is AugMod a data set made by myself? ? ? ? Is there any way to generate it by myself? ? ?

Here are my answers:

There is no outstanding performance on the RML2016 data set. But it performs well on the AugMod dataset.

Our conclusion, which you can find in the paper there: https://arxiv.org/abs/2003.12260, in table 3, is that it performs as well as, or better than, other state of the art architectures on RadioML2016.04C, RadioML2016.10A or RadioML2016.10B.
We think the reason is that the problem starts then to be dataset limited instead of model limited.
For example the RadioML2016.04C, RadioML2016.10A or RadioML2016.10B datasets have inner problems, that were investigated by others.

Question 1: I want to ask that this algorithm only performs well on AugMod?

No it performs well, either as well as, or better than, on all the five datasets tested in the paper.
However this difference starts to be significant for RadioML2018.01A and is very significant for AugMod.

Question 2: Is AugMod a data set made by myself? ? ? ? Is there any way to generate it by myself? ? ?

The Augmod dataset was made by our team, and is publicly available on kaggle.
However, we do not have currently plans to release the code to produce it.

Does all of these answer your questions?

[helion-du-mas-des-bourboux-thales]

Hello @shu616048151, any remarks on this issue?

[shu616048151]

在公开数据集的流程图中无法训练，并且希望能够公开数据集制作代码

[helion-du-mas-des-bourboux-thales]

Hello @shu616048151, could you try using English?
Here is the translation of what ou just send:

Unable to train in the flow chart of the public data set, and hope to make the code of the public data set

Why are you unable to train? Did you follow the jupyter-notebook?
What are the errors you get?

[helion-du-mas-des-bourboux-thales]

Any news @shu616048151

[helion-du-mas-des-bourboux-thales]

Close, for not clear what is asked.