How are the results of the one vs rest experiment and the cross-dataset experiment calculated?

[labixiaochao]

Is the result of the experiment in ONE vs REST averaged after training and testing each class, what about the results across datasets, and what are its training commands? Please ask the specific commands and training methods and how the results are calculated. Can I find a pre-trained model for all classes in the code?

[zqhang]

We use the same checkpoints to report one-vs-rest and cross-dataset settings

[labixiaochao]

We use the same checkpoints to report one-vs-rest and cross-dataset settings
Is it enough to average the results of the three categories directly?

[zqhang]

In our experiments, this could perform well. With more data, it may achieve better performance. Please refer to this

[labixiaochao]

In our experiments, this could perform well. With more data, it may achieve better performance. Please refer to this
Thank you for your reply, I mean the results of the entire dataset reported in the final paper are the average results of the three classes above

[zqhang]

Yes, u are right.

[labixiaochao]

Yes, u are right.

thank u very much, can you send me the code of the other models compared in the paper