Hello, I am a beginner in backdoor attacks and am very interested in this paper. However, due to my limited understanding and abilities, I am having difficulty fully grasping the article and its related code. I would appreciate it if you could point out the core code, as I am trying to reproduce the SWARM method mentioned in the paper. Additionally, I noticed that the train.py file contains two training methods: train and backdoortrain. However, the paper mentions that there are two training methods only during the testing phase, which I find confusing. I hope to receive your clarification on this.
Hello, I am a beginner in backdoor attacks and am very interested in this paper. However, due to my limited understanding and abilities, I am having difficulty fully grasping the article and its related code. I would appreciate it if you could point out the core code, as I am trying to reproduce the SWARM method mentioned in the paper. Additionally, I noticed that the train.py file contains two training methods: train and backdoortrain. However, the paper mentions that there are two training methods only during the testing phase, which I find confusing. I hope to receive your clarification on this.