Timer

Hi, I have the same question, does the current version of mmrazor support the instance segmentation algorithm?

How to modify your configuration file, can you teach me?

Thank you very much for your prompt help. I haven't tried it yet, I want to ask if all classification algorithms support multi-class, or only co-training algorithm？

I re-tried the latest package. But it seems like only the first class is predicted. Result/Co_Training_BreastCancer.txt： accuracy 0.4148936170212766 precision 0.13829787234042554 Recall 0.3333333333333333 F1 0.1954887218045113 AUC 0.6487264250900614 Confusion_matrix [[1. 0. 0.]...

[breast_cancer.csv](https://github.com/YGZWQZD/LAMDA-SSL/files/10073766/breast_cancer.csv) Here is my raw data. It contains 625 data, with 343 features, corresponding to 3 classes I believe the data format is correct.

thanks for your help. I re-tried the lastest version (https://github.com/YGZWQZD/LAMDASSL/tree/master/LAMDA_SSL/Algorithm/Classification/Co_Training.py). But it seems like only the first class is predicted. as the confusion matrix is [[1. 0. 0.] [1. 0....